DoE
EPA
unnea aiates
Department
of Energy
Fossil Energy Division
Fossil Fuel Utilization
Washington DC 20540
ANL/CEN/FE-78-10
United States Environmental Industrial Environmental Research EPA-600/7-79-203
Protection Agency Laboratory August 1979
Office of Research and Development Research Triangle Park NC 27711
Support Studies in
Fluidized-bed
Combustion
1978 Annual Rep&rt
Interagency
Energy/Environment
R&D Program Report
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy-Environment Research and Development
8. "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the INTERAGENCY ENERGY-ENVIRONMENT
RESEARCH AND DEVELOPMENT series. Reports in this series result from the
effort funded under the 17-agency Federal Energy/Environment Research and
Development Program. These studies relate to EPA's mission to protect the public
health and welfare from adverse effects of pollutants associated with energy sys-
tems. The goal of the Program is to assure the rapid development of domestic
energy supplies in an environmentally-compatible manner by providing the nec-
essary environmental data and control technology. Investigations include analy-
ses of the transport of energy-related pollutants and their health and ecological
effects; assessments of, and development of, control technologies for energy
systems; and integrated assessments of a wide range of energy-related environ-
mental issues.
EPA REVIEW NOTICE
This report has been reviewed by the participating Federal Agencies, and approved
for publication. Approval does not signify that the contents necessarily reflect
the views and policies of the Government, nor does mention of trade names or
commercial products constitute endorsement or recommendation for use.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/7-79-203
August 1979
Support Studies in Fluidized-bed
Combustion
1978 Annual Report
by
I. Johnson, G.J. Vogel, S.H.D. Lee, D.S. Moulton,
F.F. Nunes, J.A. Shearer, G.W. Smith, E.B. Smyk,
R.B. Snyder, W.M. Swift, F.G. Teats, C.B. Turner,
W.I. Wilson, and A.A. Jonke
Argonne National Laboratory
9700 South Cass Avenue
Argonne, Ilinois 60439
EPA No. IAG-D5-E681
DoE No. W-31-109-Eng-38
Program Element No. INE825
Project Officers:
David A. Kirchgessner John F. Geffken
EPA/Industrial Environmental DoE/Fossil Fuel Utilization
Research Laboratory Washington, DC 20540
Research Triangle Park, NC 27711
Prepared for
U.S. ENVIRONMENTAL PROTECTION AGENCY U.S. DEPARTMENT OF ENERGY
Office of Research and Development Fossil Fuel Utilization
Washington, DC 20460 Washington, DC 20540
-------
TABLE OF CONTENTS
Page
ABSTRACT 1
SUMMARY 1
TASK A. GASEOUS POLLUTANT EMISSION CONTROL IN FBCs 10
1. Enhancement of Limestone Sulfation 10
a. Introduction 10
b. Effects of CaCl2 Additive on
Limestone Sulfation 11
c. Effects of MgCl2 and Other Salts on
Limestone Sulfation 17
2. Investigation of Sulfation Reactions during
Fluidized-Bed Combustion of Coal 23
3. Evaluation of Coal Pyrolysis Char as a
Feedstock for FBC 30
a. Material 30
b. PDU Combustion System 30
c. Test Plan and Experimental Procedure .... 32
d. Results and Conclusions 33
4. The Use of Oil Shale for S02 Emission Control
in Atmospheric Pressure Fluidized-Bed Coal
Combustors 36
a. Materials 37
b. Experimental 37
c. Sulfation Results 37
d. Prediction of S02 Retention for
Oil Shales 40
e. Evaluation of Oil Shale Sorbent for AFBC
Sulfur-Removal Systems 40
f. Evaluation of Oil Shale Use in a Fluidized-
Bed Combustion Plant Employing a Carbon. . .
Burnup Cell 40
g. Attrition Results 41
h. Conclusions 42
5. Comparison of Limestone Calcium Utilization
in an AFBC with TGA Projections 42
6. Prediction of Limestone Requirements for
an FBC-CBC Combustor 46
iii
-------
TABLE OF CONTENTS (Cont'd.)
Page
7. Estimation of Limestone Requirements for AFBC . . 48
8. Petrographic Examination of Limestones 53
9. Effect of Water on S02 Reactivity 53
10. Limestone Attrition in a Fluidized Bed 53
TASK B. TURBINE CORRODENT STUDIES 55
1. Emission of Alkali Metals during the
Combustion of Coal 55
2. Effect of Additives on the Retention of Alkali
Metals in the Bed during the Combustion of Coal . . 56
3. Removal of Alkali Metal Compounds from the Hot
Combustion Gas of Coal 58
a. Screening Tests 59
b. Parametric Tests 64
TASK C. TRACE POLLUTANT CONTROL IN FBC 72
1. Trace Element Behavior in Sorbent during
Cyclic Utilization 72
a. Analytical 72
b. Experimental 73
c. Results 73
TASK D. PARTICULATE CONTROL STUDIES
1. Evaluation of On-Line Light-Scattering
Particle. Analyzers 79
a. Introduction 79
b. System and Procedure for Flue-Gas Particle
Measurements 80
c. Experimental Evaluation of the Multi-
particle Analyzer 80
d. Conclusions 83
2. Particle Removal from Flue Gas 84
a. Granular-Bed Filters 84
b. Acoustic Dust Conditioning 88
c. High Efficiency Cyclone 91
iv
-------
TABLE OF CONTENTS (Cont'd.)
Page
MISCELLANEOUS STUDIES 91
1. Pulsed L-Valve Tests 91
a. Introduction 91
b. Equipment and Procedure 92
c. Test Results 93
APPENDIX A. Compositions (wt %) of Limestones and Dolomites . . 95
APPENDIX B. Limestone Designations and Suppliers 97
REFERENCES 99
-------
LIST OF FIGURES
No. Title Page
1. Effect of CaCl2 Addition on Limestone Sulfation at 850°C
for 6 h in 0.3% S02, 20% C02 , 5% 02, Balance N2 13
2. Percent Conversion to Sulfate vs. Average Pore Diameter
for Limestones Treated with Ca"cT2 at 850°C and Sulfated
6 h in 0.3% S02, 5% 02, 20% C02, and the Balance N2 16
3. Changes in Porosity Curves with 0.5 mol % NaCl Addition
to Limestones Calcined at 850°C for 1 h in 5% 02, 20% C02,
and the Balance N2 19
4. Effect of 0.5 mol % CaCl2 Addition on Porosity of Limestones
Calcined at 850°C for 1 h in 5% 02, 20% C02, and the
Balance N2 20
5. Changes in Porosity Curves with 0.5 mol % MgCl2 Addition
to Limestones Calcined at 850°C for 1 h in 5% 02, 20% C02,
and the Balance N2 21
6. Porosity Curves of ANL-9701 Comparing Combustor-Sulfated
Samples (C) and Samples Sulfated in a Laboratory Tube
Furnace (A and B) 24
7. Porosity Curves Showing the Effect of Calcining on Lime-
stone ANL-9701 in a Laboratory Tube Furnace 26
8. Porosity Curves of ANL-9501 Comparing Combustor-Sulfated
Samples (C) with Samples Sulfated in a Laboratory Tube
Furnace (A and B) 27
9. Porosity Curves for (A) Heat-Treated ANL-9501 and (B)
CaCl2-Treated ANL-9501 Sulfated in a Laboratory Tube Furnace . . 29
10. Simplified Equipment Flowsheet of PDU Fluidized-Bed
Combustor and Associated Equipment 32
11. Combustion Efficiency of Coal Char as a Function of Bed
Temperature, Combustor Pressure, and Fluidizing-Gas Velocity . . 35
12. Conversion (Measured with a TGA) of CaO to CaSC>4 in
Precalcined Spent Green River Oil Shale at 700 to
1050 °C 38
VI
-------
LIST OF FIGURES (Cont'd.)
No. Title Page
13. Conversion (Measured with a TGA) of CaO to CaSO^ in
Spent Oil Shale, Tymochtee Dolomite, Greer Limestone,
and Germany Valley Limestone, Using 0.3% S02~5% 02 in
N2 at 900 °C 39
14. Calcium Utilization in Seven Precalcined Limestones at
900°C as a Function of Time 44
15. Experimental and Calculated Calcium Utilizations 47
16. TGA Reactivity Curves for ANL-9901 and ANL-9902
Limestones 51
17. Calcium Utilization of Limestone versus their
CaC03 Concentrations 52
18. Calcium Utilization as a Function of Pore Size for
Calcined Stones 52
19. Effect of Sorbent Bed Temperature on Sorption Capacity
as a Function of Experiment Duration 64
20. Effect of Superficial Gas Velocity on Sorption Capacity
at 800 °C 68
21. Effect of the Contact Time of Flue Gas with Sorbent on
NaCl Vapor Capture 71
22. Typical Pulse Jet Tested in the Pulse-Jet Development
Program 89
23. Schematic of the Pulse Jet/Resonant Manifold System 90
vii
-------
LIST OF TABLES
No- Title Page
1. Effects of CaCl2 and NaCl upon Six-Hour Sulfation
of Limestones at 850°C in 0.3% S02, 5% 02 , 20% C02,
and the Balance N2 12
2. Average Pore Diameter and Percent Conversion to Sulfated
State at 850°C for Limestones Treated with CaCl2 15
3. Percent Conversion to Sulfate after 6 h with 0.3% S02
at 850°C for Limestones Precalcined with MgCl2 Additive
at 850 °C 1 h in 5% 02, 20% C02, Balance N2 18
4. Percent Conversion to Sulfate after 6 h with 0.3% S02 at
850°C for Limestones Precalcined with Inorganic Additives
at 850°C 1 h in 5% 02) 20% C02, and the Balance N2 22
5. Sulfation Conversions and Average Pore Diameters for
ANL-9701 Stone Under Various Reaction Conditions 25
6. Percent Conversion to Sulfate and Average Pore Diameters
for ANL-9501 Stone Under Various Reaction Conditions 28
7. Average Particle Size and Chemical Properties of Occidental
Research Corporation's Flash-Pyrolysis Coal Char 31
8. Combustion Efficiency of Flash-Pyrolysis Coal Char in
Fluidized-Bed Combustion Process Development Unit 34
9. Concentrations (in wt %) of Major Constituents of
Calcareous Materials 37
10. Requirements for Green River Oil Shale, Germany Valley
Limestone, Greer Limestone, and Tymochtee Dolomite
to Meet S02-Emission Standard 41
11. Comparison of Predicted and Pilot-Plant Calcium
Utilization 45
12. Projected Limestone Requirements, kg Stone/kg of Coal, to
Meet 0.5 g S02/MJ Standard for AFBC 49
13. Material Balances of Sodium and Potassium from Combustion
of 0.5 wt % NaCl-Impregnated Activated Coconut Charcoal
Mixed with 5 wt % Additive 57
vi±±
-------
LIST OF TABLES (Cont'd.)
No. Title Page
14. Experimental Conditions for Testing Diatomaceous Earth
and Activated Bauxite Sorbents for K/jSC^ Vapor Capture .... 61
15. Material Balances of K/jSC^ from Tests of K.2S04 Vapor
Capture by Diatomaceous Earth and Activated Bauxite ..... 62
16. Distribution of Potassium Ion in the Sorbents ........ 63
17. Sodium Chloride Distributions from Tests of NaCl-Vapor
Capture by Diatomaceous Earth as a Function of Superficial
Gas Velocity of Flue Gas ................... 66
18. Sodium Chloride Distributions from Tests of NaCl-Vapor
Capture by Activated Bauxite as a Function of Superficial
Gas Velocity of Flue Gas .................. 67
19. Sodium Chloride Distributions from Tests of NaCl-Vapor
Capture by Diatomaceous Earth as a Function of Gas
Hourly Space Velocity .................... 70
20. Sodium Chloride Distributions from Tests of NaCl-Vapor
Capture by Activated Bauxite as a Function of Gas
Hourly Space Velocity .................... 71
21. Samples Analyzed in the Trace Element Study ......... 74
22. Trace-Element Analyses for Tymochtee Dolomite Samples from
the First and Tenth Combustion Cycles ............ 75
23. Concentrations and Material Balances for Trace Elements in
Steady-State Samples of Solids Entering and Leaving the
Regeneration Reactor during the Tenth Regeneration Cycle ... 78
24. Particle Sizes and Loadings Obtained with the Multi-
particle Analyzer, the Andersen Impactor^ and the
Membrane Filter in Coal Combustion and Cold Elutriation
Experiments ......................... 81
ix
-------
SUPPORTIVE STUDIES IN
FLUIDIZED BED COMBUSTION
Annual Report
July 1977 - September 1978
by
Irving Johnson, G. J. Vogel, S. H. D. Lee, D. S. Moulton, F. F. Nunes,
J. A. Shearer, G. W. Smith, E. B. Smyk, R. B. Snyder, W. M. Swift,
F. G. Teats, C. B. Turner, W. I. Wilson, and A. A. Jonke
ABSTRACT
These laboratory and process development-scale studies
support the Fossil Energy development program for atmospheric
and pressurized fluidized-bed coal combustion. The specific
objectives of the current program are (1) to establish the
basic understanding needed to optimize the use of limestones
for the control of S02 emission from FBCs and (2) to
develop the technical basis for processes for the treatment
of high-temperature, high-pressure gases from PFBCs so that
these gases can be used to operate gas turbines.
This report presents information on the enhancement of
limestone sulfation by the use of chemical additives, the
evaluation of coal char as a fuel for FBCs, the use of
oil shale in place of limestone for S02 emission control,
development of a model for the prediction of the performance
of limestones in FBCs from laboratory-data, studies of the
emission of alkali metal compounds from coal combustion
systems, development of sorbents for the removal of
gaseous KC1 and NaCl from hot gas streams, studies of the
fate of trace elements in a FBC combustion-regeneration
system, evaluation of two laser instruments for the in
situ measurement of particle size and concentration in a
hot gas stream, evaluation of a high efficiency cyclone,
the development of a granular bed filter for particulate
removal, the development of a high-intensity sound source
to enhance particulate agglomeration and the development
of a pulsed L valve for the metering of the flow of
solids.
SUMMARY
Task A. Gaseous Pollutant Emission Control in Fluidized-Bed
Combustors (FBCs)
Enhancement of Limestone Sulfation. The use of various salts (NaCl,
CaCl2, etc.) to increase the rate and the extent of sulfation of limestones
-------
is being investigated on a laboratory scale, using a horizontal tube furnace.
The results of studies using NaCl, CaCl2, MgCl2, and other inorganic salts
are reported.
Detailed studies have been made of the mechanism by which NaCl additive
changes the sulfation characteristics of limestones. Earlier work showed that
the addition of NaCl increases the size of the pores in the calcined limestone.
Therefore the gaseous S02/02 mixture can more readily penetrate the limestone
particle and react with the CaO to form CaSC>4. As stated previously, the NaCl
is believed to form a liquid film on the surface of the CaO grains which facili-
tates the growing together of these grains and their recrystallization. This
process leads to larger pores. The large pores permit a greater extent of
reaction before they are closed off by CaSCy,., which has a larger volume than
does CaO. The presence of the NaCl in the system appears to lead to the
recrystallization of CaO and CaS04 which allows the gaseous S02/02 mixture
to more easily diffuse through the layer of CaS04 that forms on the CaO grains.
The pore size of the calcined limestone was earlier found to increase as
the amount of NaCl added was increased. In current work, the extent of
sulfation was observed to pass through a maximum value when the average pore
diameter was in the range, 0.3 to 0.4 ym. The decrease in the extent of
sulfation for average pore diameters larger than this maximum is believed
due to a decrease in internal surface area. For some limestones, an increase
in the extent of limestone sulfation similar to that obtained by NaCl treatment
can be achieved by heating the calcined limestone at elevated temperatures or
by calcination in an atmosphere of C02 at high pressure. Both of these
treatments also increase the average pore diameter.
The addition of CaCl2 is more effective than the addition of NaCl in
increasing the extent of sulfation of limestones. The average pore diameter of
the calcined limestone is increased by the addition of _< 0.5 mol % CaCl2 to
give a maximum sulfation and hence the mechanism of the process is the same as
for NaCl. Large (>1 mol %) additions of CaCl2 produce a second maximum in
the plot of extent of sulfation _y_£. average pore diameter. This second
maximum (which has not been observed for NaCl) is believed due to the formation
of a liquid CaCl2-CaO phase (CaO is soluble in liquid CaCl2) which facili-
tates the S02-02-CaO reaction. The effect of 0.1 and 0.5 mol % CaCl2 additions
on approximately thirty different limestones was determined. In 60% of the
cases a larger extent of sulfation was obtained for the 0.1 mol % addition.
The amount of CaCl2 needed for optimum extent of sulfation of a specific
limestone depends on the amount and type of impurities present in the limestone
The effect of MgCl2 additive on (1) the pore size of the calcined
limestone and (2) the extent of sulfation was very similar to that observed
for CaCl2. Also, the effects of NaOH, Na2C03, Na2S04, Na2Si03, Na2Si03'
9H20, Ca(OH)2, CaF2, CaS04, and 113603 on the extent of sulfation of four
different limestones were tested. Most of these substances had a positive
effective on the extent of sulfation. The extent of sulfation was strongly
influenced by the impurity content of the stone.
Investigation of Sulfation Reaction during Fluidized-Bed Coal Combustion.
Studies of the sulfation process which occurs in fluidized bed coal combustion
are under way. In these studies, samples of limestones which had undergone
-------
sulfation in an FBC have been compared with samples which had been sulfated in
a small laboratory tube furnace using similar sulfation conditions. In a set
of comparisons for one limestone type, the extents of sulfation of the limestone
in the FBC and in the laboratory-scale apparatus were very similar; furthermore,
the porosities of the samples were nearly identical. However, for a second
limestone type and a different FBC, the extent of sulfation of samples from
the FBC was significantly greater than for a sample of the same limestone
sulfated in a laboratory-scale tube furnace. In the latter comparison, the
average pore diameter of the sulfated material from the FBC was larger than
that obtained in the laboratory-scale sulfator. The greater extent of sulfation
obtained in the FBC is due to the average pore diameter being closer to
optimum. The cause of this additional pore growth in the combustor has not
been determined.
Evaluation of Coal-Pyrolysis Char as a Feedstock for FBC. As part of a
program to evaluate flash pyrolysis coal char for power generation, a series
of combustion tests was performed to measure the effects of bed temperature,
combustor pressure, and fluidizing-gas velocity on the combustion efficiency
of the char and to determine whether the char would be an acceptable feedstock
for FBC units. The char used in the study was the product of the flash
pyrolysis of a Wyoming subbituminous coal at a pyrolysis temperature of •x/650°C.
The combustion tests were performed in a 15.2-cm-ID fluidized-bed combustion
process development unit at the following nominal conditions of the three
independent variables: bed temperatures of 800 and 900°C, combustor pressures
of 405 and 810 kPa, and fluidizing-gas velocities of 0.76 and 1.1 m/s. In all
experiments, excess air was held to 3% oxygen (a.17% excess air) in the dry
flue gas.
Feeding of the extremely fine particle size char proved to be a significant
problem in testing. Attempts to blend the char with sorbent and to feed both
materials from the same hopper were not successful in eliminating the problem.
However, the data obtained are considered to be sufficiently accurate to
assess the combustibility of the char in an FBC and to observe the effects of
the changes made in operating conditions.
Combustion efficiency of the char, based on the rates of unburned carbon
leaving the combustor and of fresh carbon fed to the combustor, was measured
and found to range from >\,94 to 99%. Combustion efficiencies of 99% have been
measured for coal combustion in FBCs, but only at much higher excess air
levels, i.e., ^95%.
Of the three independent variables, only temperature was observed to
affect the combustion efficiency of the char. Combustion efficiency increased
from an average of -\,95 at 800°C to 98% at 900°C.
The potential of the char as a fuel for FBCs has been demonstrated. The
two factors most likely to be responsible for the high combustion efficiencies
measured were (1) the extremely fine particle size of the char, and (2) the
fact that the char originated from a subbituminous coal. It should not be
concluded that all coal chars would be equally attractive for FBC applications.
Use of Oil Shales for SOg-Emission Control in AFBCs. The use of oil shales
for the control of SC>2 emission in atmospheric-pressure fluidized-bed coal
combustors has been studied by measuring the rate and extent of sulfation of oil
-------
shale, using a thermogravimetric analyzer (TGA). From these TGA data and
similar data on various limestones, it was estimated that to meet the Federal
S02~emission standards for a coal containing 3% sulfur, 0.6 and 1.4 kg of
spent oil shale and virgin oil shale, respectively, per kilogram of coal would
be needed. For the same coal, 0.4 kg of a limestone or 0.2 kg of a dolomite
per kilogram of coal would be needed. Greater quantities of oil shale are
required due to its lower CaO content. The attrition rate of Green River oil
shale was found to be similar to the attrition rates for typical limestones.
Comparison of Calcium Utilization in an AFBC with Predictions from TGA
J)at_a. A method has been developed to predict the utilizaton of calcium in
limestones used for SC^-emission control in AFBCs. Rate constants for the
reaction of bed materials with S02 were measured, using a thermogravimetric
analyzer. These constants, as well as specific values of the Ca/S feed ratio,
superfical gas velocity, fluidized bed height, and bed voidage are used in the
Keairns equation to predict calcium utilization. Predicted values of calcium
utilization differed from observed PDU values by about +4%. This method can
be used to determine what Ca/S in the feed will give the SC>2 retention
required by Federal S02 emission standards.
Prediction of Limestone Requirements for AFBC-CBC Systems. To improve
the combustion efficiency of AFBCs, the carbon-rich material removed from the
flue-gas stream by cyclones may be burned in a high-temperature (1100°C),
low-gas-velocity (1.8 m/s) fluidized-bed carbon burnup cell (CBC). The method
previously developed to predict the calcium utilizations in AFBCs from TGA
data (see above section) has been used to predict the overall limestone
requirements of an AFBC having a CBC. These computations indicate that about
60% of the sulfur in the elutriated material can be retained in the CBC if
fresh limestone or partially sulfated stone is added to the cell. The actual
SC>2 retentions obtained in a CBC depend on the particular limestone used and
the CBC temperature.
Estimation of Limestone Requirements in AFBCs. Sixty-one different lime-
stones have been tested for their reactivity with SC>2 in a TGA, and the
performance of each stone in an AFBC predicted. The predictions were made for
0.9-m deep beds at 2.4 and 3.6 m/s linear gas velocity. The coal contained
4.3% sulfur and required an 83% sulfur retention. For the dolomitic limestones
tested, between 0.27 and 1.2 kg of stone per kg of coal was required at 2.4 m/s
(8 ft/s) gas velocity and between 0.36 and 2.4 kg stone/kg coal at 3.6 m/s (12
ft/s) gas velocity. For the calcitic limestones tested, the estimates were
0.37 to 1.5 kg stone/kg coal and 0.37 to 2.0 kg stone/kg coal at 2.4 and 3.6
m/s gas velocities, respectively. These studies point out the large variations
in limestone requirements which can exist, depending on the nature of the
limestone. The average pore size of about 40 of the calcined stones was
measured. Although there was a general trend showing the extent of sulfation
increasing with increasing average pore diameter; there was a large scatter in
the data which indicated that one or more additional factors are also important.
Petrographic Examination of Limestone. Previous studies revealed that lime-
stones from various quarries vary greatly in SC^-scavenging capacity. Five
stones have been examined using petrographic techniques. Results have been
-------
obtained on too few stones to allow the morphology of the stone to be corre-
lated with the SC>2 reactivity. However, in the case of two dolomitic stones
examined, the less reactive stone had very coarse grains. The grains from one
stone which is a "popper" (i.e. , explodes upon calcination) contain a large
number of inclusions; however, another stone (not one of the five studied
here) which is not a popper also has a large number of inclusions. This
suggests that the type of inclusions as well as their number may be important
in determining whether popping occurs.
Task B. Turbine-Corrodent Studies
The objective of these studies is to understand the factors which influence
the emission of alkali-metal compounds from coal combustion systems and to
develop methods for reducing the alkali-metal compound content of hot combustion
gases. These studies are in support of the development of a pressurized
fluidized-bed coal combustion system utilizing a gas turbine. The transport
of alkali-metal compounds from a combustor to a gas turbine is expected to be
the major cause of hot corrosion of turbine blades.
Emission of Alkali Metals during Coal Combustion. The emission of alkali-
metal compounds during the combustion of various coals, charcoal, and lignite
has been previously studied. Additionial studies have now been made on
emissions during the combustion of lignite. These studies show that during
the combustion of lignite, potassium is more volatile than sodium. When NaCl
was added to a lignite sample prior to combustion, emission of potassium (as
KC1) was enhanced.
Effect of Additives on the Retention of Alkali Metals in the Bed during
Coal Combustion. Studies of the emission of alkali metals from coal during
combustion indicate that the fraction of the alkali in the coal that is
emitted is inversely proportional to the ash content of the coal. A series of
different additives was added to charcoal which had been impregnated with 0.5%
NaCl, and the quantity of sodium and potassium retained in the residue after
combustion was measured. (Charcoal contains a large amount of potassium.) The
retention of sodium ranged from 76 to 10%, and that of potassium from 71 to
5%. The order of effectiveness of these mineral additives in retaining sodium
is as follows: silica gel, montmorillonite, kaolinite, bauxite, Greer lime-
stone, Tymochtee dolomite, alumina powder, Dolowhite (a dolomite), no additive,
and pure alumina. The order for effectiveness of potassium retention was
about the same. The clay minerals were very effective in retaining alkali
metals.
Removal of Gaseous Alkali Metal Compounds from the Hot Combustion Gases
of Coal. Solid granular sorbents are under investigation for the removal of
gaseous alkali-metal compounds (viz., NaCl or KC1) from hot combustion gases.
In these studies, a hot (850°C) simulated flue gas is loaded with gaseous NaCl
or KCl and then passed through a hot packed bed of the granular substance being
tested to determine its effectiveness for removing the alkali-metal compound.
After screening tests with gaseous NaCl using alundum, diatomaceous earth,
silica gel, kaolinite, attapulgus clay and activated bauxite, systematic
studies with NaCl, KCl, and K^SO^ continued. Tests with diatomaceous
earth and activated bauxite are continuing and have thus far shown that both
of these substances are effective in removing gaseous NaCl, KCl, and
from hot gases .
-------
Tests have been made to determine the effect of bed temperature, linear
gas velocity, and gas contact time on the retention of gaseous NaCl by diatoma-
ceous earth and activated bauxite. Increasing the temperature from 800 to 880°C
increased the amount of gaseous NaCl sorbed by diatomaceous earth and decreased
the amount sorbed by activated bauxite. This difference is believed to be due
to an endothermic chemical reaction occurring between the diatomaceous earth
and the gaseous NaCl contrasted with physical adsorption of gaseous NaCl
on the activted bauxite.
For linear gas velocities of 25, 66, and 155 cm/s and constant gas
contact time, the amount of gaseous NaCl sorbed was found to be the same for
both sorbents. Thus, at the test conditions, the rate of the sorption
reaction is not controlled by the mass transfer of gas from the bulk to the
external surface of the granular sorbent. In tests with activated bauxite,
increasing the gas contact time from 0.05 to 0.19 s increased the removal of
gaseous NaCl from the gas stream from about 80% to 98%. These results suggest
that gas contact times of 0.2 s will be practical.
Task C. Trace-Pollutant Control in FBCs
Trace-Element Behavior in Sorbent During Cyclic (Sulfation-Regeneration)
Utilization. The objectives of the work reported here were (1) to measure
the level of selected trace elements in the sorbent over several cycles of
sulfation and regeneration and to determine the tendency of trace elements to
be enriched or depleted in the sorbent in several utilization cycles and (2)
to make a material balance (of particulate samples only) for a single regenera-
tion experiment, assessing changes in the trace-element concentrations in the
sorbent for evidence of trace-element losses by volatilization.
Samples of (1) particulate sulfated Tymochtee dolomite from the first and
tenth combustion cycles of a 10-cycle combustion-regeneration experiment
and (2) particulate sulfated and regenerated sorbent, coal, and recovered
fly ash from the tenth regeneration cycle of the same cyclic experiment
have been analyzed for trace elements by spark-source mass spectrometry.
Based on the analyses of sulfated dolomite sorbent from the first and
tenth combustion cycles, the concentrations of at least 9 of 31 elements
appear to have increased. This enrichment was probably due to the buildup of
coal ash on the surface of the sorbent particles during the ten combustion-
regeneration cycles. During the ten cycles, none of the 31 elements appear to
have been depleted.
Trace element concentrations in sulfated sorbent feed and in regenerated
sorbent product from the tenth regeneration cycle were compared. Of 31
elements, only nickel showed any evidence of being depleted during regeneration.
For the tenth regeneration half-cycle, material balances were taken
around the regeneration reactor for 27 of the 31 elements studied. There was
a large scatter in the material balances of the various elements, making the
behavior of the elements during regeneration difficult to assess.
-------
Task D. Particulate Control Studies
Evaluation of On-Line Light-Scattering Particle Analyzers. Two light-
scattering particle-size analyzers have been tested at ANL in the process
development unit (PDU) fluidized-bed combustion system. The analyzers are (1)
a single-particle analyzer developed by Spectron Development Laboratory and
(2) a multiparticle analyzer developed by Leeds and Northrup. Particle-size
distributions and mass loadings determined at various flue-gas duct locations
with the Spectron and the Leeds and Northrup instruments were compared with
those obtained with (1) an Andersen cascade impactor, (2) a Coulter counter,
and (3) positive filters. These comparisons were used to evaluate the two
instruments at their present stage of development.
Based on the experimental work performed, the multiparticle analyzer
developed by Leeds and Northrup is at a stage of development allowing it to be
used with little operator training. Operation of the single-particle analyzer
developed by Spectron Development Laboratory requires much more care, and data
reduction is presently very time-consuming. The single-particle analyzer is a
first-generation instrument and has much potential. On the other hand, the
multiparticle analyzer is a more advanced instrument and with some refinement
can be useful now in the development of PFBC technology.
Particulate Removal From Flue Gas. An experimental program is under way
at ANL to test and evaluate promising high temperature/pressure flue-gas
cleaning methods in the off-gas system of the 15.2-cm-dia (i.e., process
development unit or PDU) fluidized-bed combustor. Techniques identified for
investigation are granular-bed filtration (using beds of limestone sorbent
from the FBC), high-efficiency controlled-vortex cyclones, and acoustic
agglomeration.
Filtration experiments have been performed using small (7.8 or 15.4-cm-ID),
horizontal fixed beds of limestone material and a downward flow of flue gas
through the granular-bed filter.
The effects of bed depth, bed-particle size, and filtration velocity on
filtration efficiency have been studied. For a -6 +14 mesh bed, filtration
efficiency increased from ^91% at a 5.1-cm bed depth to ^98% at a 40.6-cm bed
depth. Over the range of bed depths studied, using a -14 +30 mesh bed material
rather than the coarse material decreased the penetration of particles through
the filter by 35 to 40%.
In experiments using a shallow (5.1-cm) bed of -6 +14 mesh virgin
dolomite, no effect of gas velocity on filtration efficiency was observed over
the range of 15 to 60 cm/s. However, at a bed depth of 20.4 cm with the -6
+14 mesh virgin dolomite and at a bed depth of 5.1 cm with a sulfated
limestone (mean particle diameter of 'WOO pm), filtration efficiency was
observed to increase with decreasing gas velocity, increasing from -\,94% at 60
cm/s to -v99% at 15 cm/s for the sulfated limestone.
Particle loadings in the flue gas leaving the granular-bed filter generally
ranged from 0.01 to 0.07 g/m^. There was little change in mass mean diameter
of the dust as a result of passing through the granular-bed filter, and
differences in filtration efficiency were not significant over the range of
dust-particle sizes entering the filter.
-------
For a 200-MWe PFBC, the rates of limestone usage and flue-gas volumetric
flow were used to assess the permissible range of parameters such as bed
depth, face velocity, and bed-replacement rate for the granular bed filter.
The results indicated that at the filtration conditions of the PDU tests, the
limestone requirements of the granular-bed filter would exceed the quantity
available from a PFBC. The PDU tests were performed, however, with very high
particulate loadings to the granular-bed filter, resulting in very short
filtration times and high limestone requirements. Further testing of the
granular bed filter with more typical inlet dust loadings is planned.
Acoustic dust conditioning is a technique to enhance the natural tendency
of polydispersed particulates to impact upon each other and to agglomerate.
Thus, acoustic dust conditioning is designed to increase the efficiency of
downstream dust collectors.
Work is currently being carried out under a subcontract with the University
of Toronto to develop and fabricate a pulse-jet sound generator, a resonant
manifold, and acoustic treatment sections for subsequent installation and
testing in the flue-gas system of the ANL PDU combustor.
Prototype pulse jets have been fabricated and tested at ambient conditions
to finalize the design of the pulse jet for operation at elevated pressure. A
sound intensity of 155 to 160 dB at a frequency of ^280 Hz has been achieved.
Design of the resonant manifold has been completed, and testing of the
unit at ambient pressure is being performed. Installation of the pneumatic
control system for testing both the pulse jet and the resonant manifold at
elevated pressures is nearly complete.
A high-efficiency, controlled-vortex cyclone.(TAN-JET) has been obtained
from the Donaldson Co. and installed in the flue gas system of the ANL combustor
as a secondary cyclone. Experiments to establish the performance of the unit
as a particulate-removal device are being initiated. The cyclone's performance
will be tested, and its use will also permit testing of the granular-bed
filter at lower loadings than those used in earlier granular-bed experiments.
In addition, the cyclone will be used in evaluating the ability of the pulse-jet
acoustic agglomeration system to increase the particle-removal efficiency of
mechanical collectors.
Miscellaneous Studies
Pulsed L-Valve Tests. The L-valve, a type of nonmechanical valve, is a
device which can be used to control the flow of solids into either dense-phase
or lean-phase media, such as fluidized beds and pneumatic transport lines.
A small test program was initiated to develop an intermittently aerated
L-valve feeder capable of feeding solids at very low rates. The L-valve
tested was a standard 9.52-mm (3/8-in.) stainless steel tube cross fitting.
Solids flowed by gravity from a hopper into the top of the L-valve. By use of
a timer and a solenoid valve, pulsed aeration gas was supplied at right
angles to the solids downward flow. The gas transported the solids horizontally
in dense-phase flow to a point where the solids were discharged into a pneumatic
transport line.
-------
Test results demonstrated the ability of the valve to feed both coal and
coal-limestone mixtures at low feed rates and within rigid specifications for
feed-rate variance. With refinement of the timer used in the tests, the
L-valve could be easily adapted for on-line control of solids feed rates.
-------
10
TASK A. GASEOUS POLLUTANT EMISSION CONTROL IN FBCs
1. Enhancement of Limestone Sulfation
(J. Shearer and C. Turner)
a. Introduction
We have studied the effects of adding NaCl to a fluidized-bed coal
combustion system in order to enhance S02 capture by a limestone sorbent to
form CaS04. The proposed mechanism of interaction between NaCl and limestones
is based on structural changes in the stone induced by the presence of trace
amounts of liquid phase. Eutectic mixtures of salt and CaO form locally at
temperatures greater than 700°C, and the increased ionic mobility enhances
structural rearrangement and recrystallization which, in turn, affects the
porosity of the lime produced. The resulting increase in average pore diameter
causes increased permeability to 502/02 mixtures. The larger pore sizes
also permit further reaction to form CaS04 which has a larger molar volume
than the original CaC03. The buildup of an impervious layer of CaSO^. is
prevented by these dynamic structural changes occurring throughout the reaction.
As pore size increases, total surface area decreases and, at some point, loss
of surface area available for reaction causes a decrease in sulfation despite
the increasing permeability. Thus, in a plot of percent conversion of CaO to
CaS04 versus average pore diameter, the curve at first rises to a maximum
conversion at an optimum pore diameter and then decreases as the average pore
size increases further. Laboratory data show that for natural limestones
average pore diameters for calcines formed at 850°C in 20% O>2,* 5% 02
and the balance N2 are generally below this optimum value of 0.3-0.4 ym.
Thus, slight increases in average pore diameter will maximize the
potential for capture of S02. Data given below for 29 limestones show that
in most cases, 1 mol % NaCl or less is needed to achieve this increase.
Sulfations were carried out in a synthetic flue gas with the intention of
applying the results to actual coal combustion systems.
A similar situation occurs for dolomites. However, upon calcination,
of large amounts of MgC03 species, which are unreactive to S02/02 mixtures,
extensive porosity is produced. Thus, in natural dolomites a greater percentage
of the CaO component is converted to CaS04 since the inert MgO prevents the
CaS04 from forming a shell that blocks access to internal pores. For this
reason, maximum conversion occurs at a somewhat smaller average pore diameter
than for limestones and smaller amounts of salt additive are required to
increase the average to this optimum.
Changes produced by the addition of NaCl then lead to enhanced
sulfation in both limestones and dolomites. Laboratory studies indicate that
any phenomenon that affects porosity can be used to enhance sulfation. Thus,
high-temperature sintering, high C02 pressure, and the addition of other
inorganic salts will all have a beneficial effect on limestone sulfation if
the extent of pore growth can be controlled and optimized.
A more detailed presentation of the experimental results and conclu-
sions can be found in Ref. 1.
*A11 gas concentrations are given in vol %,
-------
11
b. Effects of CaCl2 Additive on Limestone Sulfation
Earlier work in this area (ANL/CEN/FE-77-3) suggests that CaCl2
enhances S(>2 capture by limestones. (It has since been pointed out that
substantial amounts of CaCl2 are routinely added to coal storage piles to
prevent freezing during the winter in colder areas.) A more complete study
was initiated to demonstrate the effectiveness of CaCl2 as a sulfation
enhancer and to compare the mode of interaction with that of NaCl. Experiments
were performed in a horizontal tube furnace capable of achieving 1100°C and
with a gas-mixing system capable of producing appropriate combinations of
02, CC>2, SC>2, and N2. Limestone samples (18-20 mesh) were contained in
quartz boats in the furnace. The source of salt on the treated limestones was
evaporated aqueous slurry.
Table 1 presents data on sulfation of calcite spar, ANL-9501,
ANL-8903, and ANL-8001 limestones in the CaC^-treated and NaCl-treated
states. These four stones represent a fairly broad range of impurity levels,
as noted in the table. (Data for NaCl-treated stones were obtained earlier.)
The results are given as percent conversion of CaO to CaS04. Figure 1 is a
bar graph showing the effect of CaCl2.
The effectiveness of CaCl2 is apparent from Table 1 and Figure 1
and is comparable to the effectiveness of NaCl. However, in contrast to
NaCl, there are apparently two concentration regions of CaCl2 where sulfation
of limestones is enhanced (Fig. 1). The first is at very low concentrations
of CaCl2 «0.5 mol %) .
The low melting point of CaCl2 and its capability of forming
low-melting eutectics with CaCK and CaSCy,. support the mechanism proposed
for NaCl effectiveness (reported earlier in ANL/CEN/FE-77-3) which results in
changing the pore diameters and increasing the stone permeability. Porosity
measurements given below show that with small additions of salt, "average"
pore diameters of CaCl2-treated stones shift to a much greater extent than
does the pore diameter of stones treated with the same percent NaCl. Therefore,
an optimum porosity structure will be reached with only trace amounts of
CaCl2 additive, as shown in Fig. 1. The stone with the greatest impurity
content, ANL-8001 (Greer limestone), achieves maximum sulfation with a slightly
higher concentration of CaCl2 than gives maximum sulfation in the other
stones. The Ca^+ in the additive can react with silica in the Greer
limestone, effectively reducing the interaction of CaCl2 with CaO and
CaS04 so that maximum sulfation occurs at the higher salt concentrations.
The second concentration region where sulfation is enhanced (Fig. 1)
is at higher CaCl2 concentrations (>1 mol %). For NaCl in the range of
concentrations investigated, this phenomenon has not been observed. Apparently,
for all CaCl2~treated stones studied, sufficiently large amounts of CaCl2
would cause the entire CaO content of the limestone to be sulfated. It is
believed that complete sulfation would be due to the formation in this system
of a substantial amount of liquid containing a large amount of dissolved CaO.
The phase diagram for CaO-CaCl2 shows that a liquid containing up to 20%
CaO can be formed at these temperatures with locally high salt concentrations.^a
The presence of CaS04 may even enhance this effect. With so much CaO in a
"liquid" form, sulfation can proceed rapidly by dissolution and crystallization
-------
12
Table 1. Effects of CaCl2 and NaCl upon Six-Hour
Sulfation of Limestones at 850°C in 0.3%
S02, 5% 02, 20% C02, and the Balance N2
Quantity of
Salt Added,
Limestone mol %
Calcite Spar
(no impurities)
ANL-9501 (Grove)
(^3% impurities)
ANL-8903
(^9% impurities)
ANL-8001 (Greer)
(^20% impurities)
0
0.1
0.2
0.5
1.0
2.0
3.0
4.0
5.0
0
0.1
0.2
0.5
1.0
2.0
3.0
4.0
5.0
0
0.1
0.5
1.0
2.0
3.0
4.0
5.0
0
0.1
0.5
1.0
2.0
3.0
4.0
5.0
% Conversion
of CaO to CaS04
with CaCl2
Addition
5.2
35.0
33.8
30.1
19.2
24.4
33.0
46.0
73.9
13.0
50.0
45.0
28.5
19.9
27.1
40.6
53.8
71.1
43.0
55.4
30.0
31.1
48.2
53.0
65.2
69.0
38.0
49.3
52.8
50.6
40.2
43.6
52.9
55.0
% Conversion
of CaO to CaSC>4
with NaCl
Addition
5.2
21.0
27.0
39.8
34.0
23.0
—
20.7
—
13.0
—
—
52.2
44.6
31.6
—
19.8
—
43.0
—
47.5
—
—
—
27.9
—
38.0
—
43.9
45.5
55.0
—
55.5
—
-------
13
en
o
o
o
o
O
z
o
en
CE
LJ
>
z
o
o
\J\J
90
80
70
60
50
40
30
20
10
n
—
.
—
,
—
—
—
—
0
r
1
nf
>
.5
i
5
4
2
3
—
c
4
0
.1
.2
3
I2
1
o
5
rT
h 3
0
.?
I
2
.5. 4n
0
.1
I
3
2T
—
—
—
—
CALCITE 1359(GROVE) CV GREER
(0.0% IMPURITIES) (~ 3% IMPURITIES) (~9 % IMPURITIES) (~20 % IMPURITIES)
Fig. 1. Effect of CaCl-2 Addition on Limestone Sulfation
at 850°C for 6 h in 0.3% S02, 20% C02, 5% 02,
Balance N2. (All stones precalcined without S02
present.) Numbers above bars each refer to mol /
CaCl2 added.
-------
14
and can progress through entire particles as long as enough liquid is present.
In terms of practical application in a fluidized bed, the large amounts of
CaCl2 required plus the possibility of agglomeration would seem to prohibit
the application of this liquid formation for sulfation enhancement. However,
as shown, the use of even small amounts of CaCl2 has major effects and
deserves further attention.
In order to clearly define the effect of CaCl2, a larger number of
limestones (with added CaCl2) representing large variations in composition
and morphology were reacted in a simulated flue gas. The flue gas consisted
of 0.3% S02, 5% 02, 20% CC>2, and the balance N2. Reaction was carried out at
850°C for 6 h using both untreated stones and stones treated by evaporating
CaCl2 from an aqueous solution. Chemical analyses for these stones are
listed in Appendix A. Results of sulfation experiments carried out with these
limestones are presented in Table 2 as percent conversions of CaO to €3804.
Data are presented for untreated stones, limestones treated with a 0.1 mol %
CaCl2, and stones treated wtih 0.5 mol % CaCl2. As suggested earlier in
this report, one range of CaCl2 additions at which maximum conversions to
sulfate are achieved is <0.5 mol %; larger quantities of salt are generally
required for stones with the greatest impurity content. In 60% of the
cases a larger extent of sulfation was obtained for the 0.1 mol % addition.
In most cases, except for some very pure limestones, the maximum conversion to
sulfate approaches 50-60%. Even in these pure stones, the increase in reactivity
relative to the reactivity of untreated stones is substantial.
For an FBC in which introduction of a major amount of corrosive
species is to be avoided the effectiveness of such small amounts of CaCl2
for sulfation enhancement makes this salt appear more promising then NaCl
(which required M.O mol %) . Moreover, when CaCl2 is used routinely as a
freeze-proofing material for coal storage piles, the quantities used are
comparable to those used in this work, with no apparent major corrosive
effects. In work in a PDU-scale unit paralleling this sulfation-enhancement
work,-5 a series of metal corrosion specimens is being exposed to flue gas in
the presence of limestone containing CaCl2 and in the freeboard above the
limestone bed. Metallographic analysis will be performed to determine
corrosive effects.
The effects of CaCl2 on the calcined limestones are similar to the
effects of NaCl addition. When either salt is added to the limestone during
calcination, the average pore diameters of the calcines increase. Representative
curves are presented in a later portion of this report. Calcium chloride is
slightly more effective than is NaCl in increasing the pore diameters, smaller
amounts of CaCl2 being required to produce the same magnitude of change as
NaCl. The porosity reaches an optimum pore-size distribution with respect to
sulfation with small amounts of CaCl2. This distribution has a maximum
sulfation reactivity occurring near 0.3 ^m, as shown in Fig 2, which plots
average pore diameter versus percent conversion for stones treated with
various amounts of CaCl2- These average pore diameters were calculated from
surface areas and volumes for pores >0.01 wm (since pores smaller than 0.01
pm are ineffective in sulfation) and corrected for mercury compression. The
range of pore-size distributions which results in optimum sulfation reactivity
-------
15
Table 2. Average Pore Diameter and Percent Conversion to Sulfated State at
850°C for Limestones Treated with CaCl2. Sulfated 6 h in 0.3%
S02) 5% 02, 20% C02, and the Balance N2.
Untreated Limestone
Stone
ANL-5304
ANL-6702
ANL-7401
ANL-8001
ANL-8101
ANL-8301
ANL-8701
ANL-8902
ANL-8903
ANL-9201
ANL-9401
ANL-9402
ANL-9501
ANL-9502
ANL-9503
ANL-9504
ANL-9505
ANL-9601
ANL-9602
ANL-9603
ANL-9701
ANL-9702
ANL-9703
ANL-9704
ANL-9801
ANL-9802
ANL-9901
ANL-9902
Calcite
Avg.
Pore
Dia,
ym
0.140
0.187
0.171
0.224
0.144
0.096
0.232
0.134
0.159
0.094
0.229
0.089
0.100
0.205
0.170
0.124
0.163
0.090
0.118
0.127
0.117
0.074
0.182
0.186
0.146
0.198
0.141
0.180
0.039
Conver . , %
42.6
38.4
41.5
38.0
33.5
29.3
33.2
29.9
43.0
28.1
51.7
17.5
13.0
34.9
42.8
25.9
31.2
30.0
38.8
16.5
7.0
35.8
53.2
33.3
42.8
48.4
36.4
43.6
5.2
0.1 mol % CaCl2
Added
Avg.
Pore
Dia,
ym
0.376
0.549
0.217
0.244
0.166
0.122
0.306
0.237
0.288
0.143
0.366
0.372
0.144
0.280
0.379
0.328
0.483
0.232
0.183
0.228
0.186
0.105
0.280
0.586
0.536
0.524
0.216
0.672
0.475
Conver . , %
58.4
60.9
48.6
49.3
49.9
42.7
50.3
49.0
55.4
43.9
56.3
36.0
50.0
38.6
45.2
44.2
40.1
36.7
43.5
43.8
37.7
41.0
41.6
37.7
41.2
47.2
39.7
50.8
35.0
0.5 mol
Avg.
Pore
Dia,
ym
0.574
1.84
0.305
0.382
0.292
0.253
0.577
0.304
0.542
0.521
0.812
1.18
1.32
0.753
1.85
0.778
1.68
0.884
0.962
0.620
0.896
0.966
1.37
1.29
1.24
1.64
0.781
1.53
0.803
% CaCl2
Added
Conver. , %
55.5
56.6
65.2
52.8
44.6
59.9
59.4
47.4
30.0
30.5
30.0
34.3
28.5
39.2
58.1
35.4
38.6
46.0
32.8
36.5
36.9
37.6
47.0
36.9
42.7
63.9
24.8
51.7
30.1
-------
16
o
CO
o
o
o
o
O
g
CO
o:
UJ
o
o
100
90
80
70
60
50
40
30
20
10
0
1 1 1 1—
• UNTREATED STONE
A O.lmol% CaCI2 ADDED
1 1 1 1
0.5mol % CoCI2 ADDED
> 1.0mol % CaCI2 ADDED
I
1
1
1
1
0 0.2 0.4 0.6 0.8 1.0 1.2 1.4
, AVERAGE PORE DIAMETER,
1.6
1.8 2.0
Fig. 2. Percent Conversion to Sulfate versus Average Pore
Diameter for Limestones Treated with CaCl2 at 850°C
and Sulfated for 6 h in 0.3% S02, 5% 02, 20% C02, and
the Balance N
-------
17
when the additive is CaCl2 is also the optimum range observed for NaCl
addition. This correlation supports the concept that pore size and distri-
bution are the most important factors determining the extent of sulfation of
limestones in 862/02 mixtures for limestone particles averaging 18-20 mesh size,
The second increase in SC>2 capture with increasing CaCl2 concen-
trations (see Fig. 2) occurs at average pore diameters greater than 1.0 ym.
As discussed earlier, this phenomenon is believed to be a result of the
formation of large amounts of a liquid phase containing dissolved CaO. Higher
concentrations of CaCl2 were also tried that resulted in pore diameters much
greater than 2.0 pm but are not presented in this graph. The presence of a
liquid phase leading to major structural rearrangement is supported by preli-
minary scanning electron micrographs .
c. Effects of MgCl2 and Other Salts on Limestone Sulfation
In a continuing study of the effects of various salts, MgCl2 at
very low concentrations was found to be an effective enhancer of limestone
sulfation. Table 3 lists the percent conversions of available CaO to CaSO^
for four limestones treated with MgCl2 (introduced via evaporation from an
aqueous slurry). The same two-maximum effect observed for CaCl2 is present
for MgCl2- The low-MgC^-concentration region (<0.5 mol %) results in
porosity changes favorable to sulfation via trace amounts of a liquid phase;
intermediate amounts (1-2 mol %) cause large amounts of a liquid phase to form
with substantial amounts of dissolved CaO accelerating the formation of
The magnitude of the MgCl2 effect may be compared with that of NaCl
and CaCl2 by examining Figs. 3, 4, and 5, which show porosity curves measured
on four limestones treated with (l) NaCl, (2) CaCl2 , and (3) MgCl2 ,
respectively. The horizontal bars with arrowheads mark the extent of shift in
pore diameters of the majority of pores from the pore distribution of untreated
stone to the pore distribution of salt-treated (0.5 mol %) stone. In all
three figures, the effect is greatest for the purest limestone and decreases
as the impurity content increases until finally there are very small effects
for ANL-8001 which contains ^20% impurities. The relative effectiveness of
the three salts can be represented by MgCl2 >_ CaCl2 > NaCl for the same
mole percent addition of salt. In terms of weight percent addition, MgCl2
and CaCl2 have comparable effects; by weight, more of each is required than
of NaCl. At this salt concentration (0.5 mol %) , most treated limestones have
greater total porosity as well as larger average pore diameters than do the
same stones when not treated.
The effects of other inorganic salts on limestone sulfation were
examined concurrently with the NaCl and CaCl2 work. Table 4 lists percent
conversions to CaS04 for four limestones treated with a variety of salts at
concentrations of 1 mol %. The results for NaCl, CaCl2, and MgCl2 are
also included to allow comparison. The data in Table 4 do show that most
salts tested have some measurable positive effects and hence may be sulfation
enhancers at the proper salt concentrations.
-------
18
Table 3. Percent Conversion to Sulfate after 6 h with 0.3% SC>2
at 850°C for Limestones Precalcined with MgCl2 Additive
at 850°C 1 h in 5% 02, 20% C02; Balance N2
MgCl2 Added, % Conversion of
Limestone mol % CaO to CaSC>4
Calcite Spar
(^0% impurities)
ANL-9501
(^3% impurities)
ANL-8903
(^9% impurities)
ANL-8001 (Greer)
(^20% impurities)
0.0
0.1
0.5
1.0
4.0
0.0
0.1
0.5
1.0
4.0
0.0
0.1
0.5
1.0
4.0
0.0
0.1
0.5
1.0
4.0
5.0
5.2
24.0
25.7
21.0
36.8
13.0
28.4
36.6
20.9
32.9
43.0
57.9
45.6
33.9
52.4
38.0
49.2
54.7
57.0
41.3
48.5
-------
19
I111 I ' ' '—\ 1" ' I ' ' '—T
ANL-8001 (Greer) (20% Impurities)
(a) No salt; (a1) 0.5 mol % NaCI
ANL-8903 (9% Impurities)
(b) No salt; (b1) 0.5 mol % NaCI
ANL-9501 (3% Impurities)
(c) No salt; (c1) 0.5 mol % NaCI
- 0.45 h Calcite Spar (0 Impurities)
0.65
0.60
0.55
0.50
UJ
=> 0.40
o
> 0.35
LU
p 0.30
=1 0.25
o 0.20
0.15
0.10
0.05
0.00
_ (d) No salt; (d1) 0.5 mol % NaCI
TT—r
-rm-r
10
I 0.6 0.2 O.I 0.06
PORE DIAMETER,
0.02 0.01 0.006
Fig. 3. Changes in Porosity Curves with 0.5 mol % NaCI Addition
to Limestones Calcined at 850 °C 1 h in 5% QI, 20% CC>2,
and the Balance N2
-------
20
o> 0.50
0.40
g 0.35
^ 0.25
S
o 0.20
—11 i i r i—i—i 1 111 i I i—i—i r
•ANL-8001 (Greer) (20% Impurities)
(a) No salt; (a1) 0.5 mol % CaCI2
'ANL-8903 (9% Impurities)
-(b) No salt; (b1) 0.5 mol % CaCI2
ANL-9501 (3% Impurities)
(c) No salt; (c1) 0.5 mol % CaCI2
-Calcite Spar (0 Impurities)
_(d) No salt; (d1) 0.5 mol % CaCI2
I 0.6 0.2 O.I 0.06
PORE DIAMETER, fj.m
0.02 0.01 0.006
Fig 4. Effect of 0.5 mol % CaCl2 Addition on Porosity
of Limestones Calcined at 850°C for 1 h in 5% 02,
20% C02, and the Balance N2
-------
21
0.65
0.60
0.55
^ 0.50
E
o
- 0.45
UJ
I 0.40
o
> 0.35
UJ
> 0.30
h-
<
=| 0.25
o 0.20
0.15
0.10
0.05
0.00
ANL-8001 (Greer) (20% Impurities)
(a) No salt; (a1) 0.5 mol % MgCI2
ANL-8903 (9% Impurities)
(b) No salt; (b1) 0.5 mol % MgCI2
ANL-9501 (3% Impurities)
(c) No salt; (c1) 0.5 mol % MgCI2
Calcite Spar (0 Impurities)
- (d) No salt; (d1) 0.5 mol % MgCI2
10
I 0.6 0.2 O.I 0.06
PORE DIAMETER, fan
0.02 0.01 0.006
Fig. 5. Changes in Porosity Curves with 0.5 mol % MgCl2 Addition
to Limestones Calcined at 850°C 1 h in 5% 02, 20% C02,
and the Balance N2
-------
22
Table 4. Percent Conversion to Sulfate of Limestones after 6 h with
0.3% S02 at 850°C for Limestones Precalcined with Inorganic
Additives at 850°C 1 h in 5% 02, 20% C02 , and the Balance N2
Salt Added
(1 mol %)
Untreated
Stones
NaCl
NaOH
Na2C03
Na2SC>4
Na2SiC>3
Na2Si03-9H20
CaCl2
MgCl2
Ca(OH)2
CaF2
CaS04
H3B03
Calcite
Spar (o,0%
Impurities)
5.2
33.3
20.2
23.6
23.8
5.2
21.4
19.2
21.0
10.4
20.6
8 0
5.1
ANL-9501
(Grove) U3%
Impurities)
13.0
44.4
27.6
43.6
41.5
12.3
36.8
19.9
20.9
12.4
26.3
13.6
17.3
ANL-8903 (^9%
Impurities)
43.0
47.5
43.5
41.0
43.2
48.4
39.8
34.7
33.0
47.5
53.6
44.4
43.9
ANL-8001
(Greer) (^20%
Impurities)
38.0
43.9
41.0
40.0
46.8
40.3
44.9
50.6
54.7
49.0
45.8
46.1
46.8
-------
23
We know from our understanding of the effects of NaCl and CaCl.2
that maximum sulfation enhancement cannot be found by using only one concentra-
tion of salt in a single sulfation experiment. Porosity curves at several
salt concentrations should be used to locate the salt concentration most
likely to give maximum reactivity with S02/C>2, assuming that an optimum
pore distribution for limestone sulfation can be reached despite differing
limestone compositions. Once this salt concentration is found, a series of
bracketing sulfation experiments should serve to indicate maximum sulfations
achievable for the particular stone and salt used.
2. Investigation of Sulfation Reactions during Fluidized-Bed Combustion of
Coal
(J. Shearer, R. Snyder, J. Lenc, and C. Turner)
To understand the effects of salt addition in a real combustor system,
the basic reaction of limestone with 502/02 must be understood first.
Comparisons were made between sulfations carried out in horizontal tube
furnaces and in actual small-scale-combustor runs under similar conditions.
Limestone ANL-9701 (Germany Valley) was reacted for ^6 h in a PDU
fluidized-bed coal combustor at 850°C with steady-state SC>2 levels in the
exhaust gas. Bed material was continually added to and removed from the
combustor. For comparison, 6-h runs were performed in the tube furnace
assemblies, using a synthetic flue gas of 0.3% S02, 5% 02, 20% C02, and
the balance N2. Gas concentratons in the combustor (except (>2 which was
12%) were estimated to be similar. Figure 6 (A, B, and C) shows porosity
curves for time fractions of various sulfation systems. Plot A shows curves
for tube furnace samples of limestone ANL-9701 precalcined and sulfated for
various intervals as shown. Plot B shows a similar set of porosity curves for
ANL-9701 stone simultaneously calcined and sulfated in tube furnaces. Plot C
is a series of porosity curves measured on an over-flow sample from the coal
combustor. The sample was separated by color into fractions representing
increased length of exposure to the combustor gases from (a) to (e).
A basic similarity of all three graphs is the shifting of the average
pore diameter to larger pore sizes during the course of reaction. The extents
of shift for both of the tube-furnace samples and the combustor samples are
similar. Table 5 lists conversions to sulfate and average pore diameters for
the three cases, as well as average pore diameters for a sample that was
calcined (without SC>2 present) at 850°C in 5% 02, 20% C02, and the balance
N2 (for which porosity curves are shown in Fig. 7). From these data
(Table 5), one can see that similar average pore diameters correspond to
similar extents of sulfation in all three systems; the last experiment (with
no 802) shows that the apparent gradual increase in pore sizes can be
accounted for by long-term (6-h) heating under these conditions. On the basis
of the above data, the results obtained in (1) the tube furnace assembly and
(2) this particular combustor system appeared to be in excellent agreement.
In contrast to the above results, the data from a study done on another
limestone (ANL-9501) in another PDU fluidized-bed coal combustor and in our
tube furnaces were not in such good agreement. Figure 8 (A, B, C) shows
-------
24
D>
\
K)
0.35
0.30
0.25
0.20
0.15
0.10
0.05
0
0.35
- 0.30
LJ
3
0.25
O 0.20
0.10
1 T
Ih
A. Precalcined ANL-9701 sulfated in a tube
furnace at 850°C in 0.3% S02, 5% 02,
20% C02, and the balance N2.
2.0 1.0 0.5 0.2 O.I 0.05 0.020.01 0.005
T
T
1
I I
1
^/
C^zr-^-'eh
0.05
°
0.35
0.30
0.25
0.20
0.15
0.10 -
0.05 -
0
/
lj I B. ANL-9701 simultaneously calcined and
// / sulfated in a tube furnace at 850°C
// / in 0.3% S02l 5% 02, 20% C02, and
and the balance N2.
2.0 1.0 0.5
0.2 O.I 0.05 0.020.01 0.005
~~l 1 1 1 1 1
d C. ANL-9701 sulfated in
in a small fluidized-bed
b coal combustor at 850°C
in 0.3% S02t 12% 02,
Q 16% C02, and the
balance N2.
0.2 O.I 0.05 0.02 0.01 ^0.005
PORE DIAMETER,
2.0 1.0 0.5
Fig. 6. Porosity Curves of ANL-9701 Comparing Combustor-Sulfated
Samples (O and Samples Sulfated in a Laboratory Tube
Furnace (A and B)
-------
25
Table 5. Sulfation Conversions and Average Pore Diameters
for ANL-9701 Stone Under Various Reaction Conditions
Average Conversion
Pore Diameter CaO to CaSC>4,
Reaction Conditions at 850°C ym %
Precalcined 1 h without 862
Present and then Sulfated in
0.3% S02, 5% 02, 20% C02 and
the balance N2 for:
1 h 0.144 8.0
3 h 0.162 10.0
6 h 0.199 11.0
Simultaneously Calcined and
Sulfated in 0.3% S02, 5% 02,
20% C02, and the balance N2 for:
1 h 0.124 6.4
3 h 0.153 8.0
6 h 0.163 10.8
Sulfated Samples from
Fluidized-Bed Coal Combustor:a
a partially calcined undefined
b 0.065 4.6
c 0.121 7.5
d 0.173 9.1
e 0.203 11.5
Calcined in 5% 02, 20% C02,
and the balance N2 for:
0.5 h
1 h
2 h
4 h
6 h
0.090
0.110
0.120
0.175
0.190
—
—
—
—
aFlue gas: ^0.3% S02, 12% 02, 16% C02, and the balance N2. Entries are
in order of increasing length of time in the combustor.
-------
26
to
E
o
LJ
2
O
>
I-
<
o
2.0 1.0 0.5 0.2 O.I 0.05 0.02 0.01 0.005
PORE DIAMETER, /im
Fig. 7. Porosity Curves Showing the Effect of Calcining
on Limestone ANL-9701 in a Laboratory Tube
Furnace. Limestone Calcined at 850°C in 5% C>2 ,
20% O>2 and the balance N2 .
porosity curves for (A) precalcined ANL-9501 sulfated in a tube furnace, (B)
ANL-9501 simulatenously calcined and sulfated in a tube furnace, and (C)
separated fractions of a partially sulfated limestone ANL-9501 sample from the
fluidized-bed combustor test. Notable differences between results for the
tube-furnace samples and the combustor samples are apparent. In particular,
the extent of sulfation is much greater for the combustor samples. Table 6
lists percent conversions to sulfate and average pore diameters for the
various samples.
Whereas for limestone ANL-9701 sulfated in the earlier mentioned combustor,
there was excellent correlation of average pore diameter and a low extent of
sulfation. In this system the much greater sulfation fills the pores to such
an extent that the assumed increase in pore diameter is masked.
Figure 9A shows porosity curves for 1-h and 6-h calcined ANL-9501 limestone
in a laboratory tube furnace. The average pore diameter shifted, due to ionic
migration and diffusion. Figure 9B shows porosity curves for a 0.2 mol%
CaCl2~treated sample of ANL-9501 undergoing calcination and then sulfation.
The average pore diameter of the calcined limestone was much larger than in
the absence of salt and, as sulfation progressed, the apparent average pore
diameter decreased due to infilling of pore space by newly formed CaSO^. In
this particular case, the average pore diameter apparently changed from >1.0 ym
to <0.7 ym due to CaSCfy formation. The extent of conversion after 5 h of
sulfation is >40%. The difficulty in determining the average pore diameter of
the limestone when only porosity curves for the end product are available is
apparent.
-------
27
0.35
0.30
0.25
0.20
0.15
0.10
0.05
o> 0
^ 0.40
o 0.35
uJ 0.30
O
o
0.25
0.20
0.15
0.10
0.05
0
Ih
-<• 3h
- 6h
A. ANL-9501 precolcined at 850°C in 5% 02,
20% C02, and the balance N2 and then
sulfated with 0.3% S02 added.
2.0 1.0 0.5
0.2 O.I 0.05 0.02 001 0.005
~~1 I 1 I I I
Ih
3h
6h
B. ANL-9501 simultaneously calcined
sulfated at 850°C in 0.3% S02,
20% C02, and the balance N2.
and
5% 02,
1.0 0.5 0.2 O.I 0,05 0.020.01 0.005
T
T
T
C. ANL-9501 sulfated in
fluidized-bed combustor
at 850°C.
i
i
1.0 0.5 0.2 O.I 0.05 0.02 0.0\ 0.005
PORE DIAMETER,
Fig. 8. Porosity Curves of ANL-9501 Comparing Combustor-
Sulfated Samples (C) with Samples Sulfated in
a Laboratory Tube Furnace (A and B)
-------
28
Table 6. Percent Conversion to Sulfate and Average Pore Diameters
for ANL-9501 Stone Under Various Reaction Conditions
Average Pore Conversion
Diameter, CaO to CaS04,
Reaction Conditions at 850°C ym %
A. Precalcined 1 h without S02
and then Sulfated in 0.3% S02,
5% 02, 20% C02> and the balance
N2 in a Tube Furnace for:
1 h 0.155 6.5
3 h 0.177 7.0
6 h 0.222 10.1
B. Simultaneously Calcined and
Sulfated in a Tube Furnace in
0.3% S02, 5% 02, 20% C02) and
the balance N2 for:
1 h 0.128 6.3
3 h 0.195 7.5
6 h 0.232 13.0
C. Sulfated Samples from
Fluidized-Bed Coal Combustor3
a 0.140 17.0
b 0.211 23.5
c 0.232 27.0
d 0.269 38 5
e 0.340 50.0
f (0.300) (8.7% as CaS04)
D. Calcined in 5% 02, 20% C02,
and the balance N2 for:
1 h 0.089
6 h 0.200
aFlue gas: ^0.3% S02, 3% 02, 16% C02, and the balance N2.
-------
29
rO
0.45
0.40 -
0.35 -
0.30
0.25
0.20
0.15
0.10
0 0.05
III 1
1 III
1 h CALCINE
^^^r^Z"-^ 6 n CALCINE
/
/
/
;
/
/ \
I I
/
—
A. Calcined at 850°C in 5% 02, ~
- // 20% C02, and the balance N2. -
III 1
1 III
0.02 0.01 0.005
h SULFATION
3h SULFATION
5 h SULFATION
B.ANL-9501+0.2 mol% CaCI2 precalcined
and sulfated at 850°C in 0.3% S02,
5% 02, 20% C02, and the balance N2.
Ill ill
1.0 0.5
0.2 O.I 0.05 0.02 0.01
PORE DIAMETER, pm
0.005
Fig. 9. Porosity Curves for (A) Heat-Treated ANL-9501 and
(B) CaCl2-Treated ANL-9501 Sulfated in a Laboratory
Tube Furnace
In the case of the combustor samples, growth of pores due to long-term
heating and possibly other factors is masked by the apparent decrease in pore
size due to infilling with CaS04. When the extent of sulfation is low as
for the aforementioned ANL-9701, the average pore diameter is not significantly
affected and still correlates with the extent of sulfation. When sulfation is
rapid and extensive infilling of pore space occurs, the calculated average
pore diameter is lower than the extent of sulfation would suggest.
-------
30
From previous work with NaCl and CaCl.2 on many limestones, maximum
sulfation occurs when average pore diameter is in the range 0.3-0.5 ym. At
larger pore sizes, reactivity decreases. The combustor data presented here
for ANL-9501 suggest that, because the extent of sulfation of some stone in
the system is as high as 50%, the average pore diameters of those stones are in
the optimum range of 0.3-0.4 ym or larger. The reasons for the greater
increase in average pore diameter for the combustor samples than in the
tube-furnace samples are not yet fully understood. One reason may be increased
reaction time due to longer turnover times in the fluidized-bed for this
particular combustor. Also, lower oxygen concentrations in the combustor may
increase the likelihood of reducing zones forming, which have been found to
enhance sulfation in laboratory studies. Possible causes are being investigated
in laboratory-scale furnaces and a small quartz fluidized-bed unit.
3. Evaluation of Coal-Pyrolysis Char as a Feedstock for FBC
(W. M. Swift)
In many coal-conversion processes, char is produced as a by-product. One
such process is Occidental Research Corporation's flash-pyrolysis process for
converting coal to liquids and gases. As part of a program to evaluate flash-
pyrolysis coal char for power generation via direct combustion, fluid-bed
combustion, and magnetohydrodynamics, a series of combustion tests was performed
in the ANL 15.2-cm-ID fluidized-bed combustion process development unit
(PDU). The objectives of the tests were to measure the effects of bed temper-
ature, combustor pressure, and fluidizing-gas velocity on the combustion
efficiency of the char and to determine whether the char would be an acceptable
feedstock for FBC units.
a. Material
The char was produced at Occidental Research Corporation's flash-
pyrolysis process development unit located in LaVerne, CA. The particular
char tested in the ANL combustion experiments was produced from Wyoming
subbituminous coal (Big Horn) at a pyrolysis temperature of ^650°C. A size
analysis and a chemical analysis (supplied by Occidental Research Corporation)
for the char used in the tests are presented in Table 7.
As indicated by the size-analysis data given in Table 7, the char
was a very finely divided material. Extrapolation of the size-analysis data
on log-probability graph paper indicates a mass mean particle diameter of
only 30 ym for the char. According to Occidental Research Corporation, the
material tested was somewhat finer than that usually produced by the flash-
pyrolysis process; typically, the char is only 70% through 200 mesh. The
finely divided char had poor flow characteristics, and there were severe
feeding problems in the combustion experiments.
»
b. PDU Combustion System
The experimental equipment and instrumentation of the PDU at Argonne
consist of a 15.2-cm-ID fluidized-bed combustor that can be operated at
pressures up to 1014 kPa; a compressor to provide fluidizing-combustion air;
-------
31
Table 7- Average Particle Size and Chemical Properties
of Occidental Research Corporation's Flash-
Pyrolysis Coal Char3
Size Analysis
Cummulative wt "/<
Screen Mesh Particle Size, ym through Screen
400
270
200
150
100
60
38
53
75
106
147
246
63.0
80.5
89.7
96.3
98.8
99.8
Chemical Analysis
Proximate Analysis Ultimate Analysis
Component wt % Component wt %
Moisture
Ash
Volatile Matter
Fixed Carbon
Negligible
14.7
6.9
78.4
Carbon
Hydrogen
Oxygen
Nitrogen
Sulfur
Ash
79.6
1.8
2.1
1.1
0.7
14.7
aData provided by Occidental Research Corporation.
a preheater for the fluidizing-combustion air; peripherally sealed rotary
feeders for metering solids into an air stream fed into the combustor; two
cyclone separators and a filter in series for solids removal from the flue
gas; associated heating and cooling arrangements and controls; and devices for
sensing and displaying temperature and pressure. A simplified schematic
flowsheet of the combustion equipment is presented in Fig. 10.
The flue gas (off-gas) is sampled continuously and is analyzed for
the constituents of primary importance. Nitrogen oxide and total NOX are
analyzed using a chemiluminescent analyzer; sulfur dioxide, methane, carbon
monoxide, and carbon dioxide determinations are made using infrared analyzers;
oxygen is monitored using a paramagnetic analyzer; and total hydrocarbons are
analyzed by flame ionization.
-------
32
TO GAS ANALYSIS
SYSTEM
TEST FILTER
STAINLESS STEEL
AIR
TJ
STEEL
FILTER
PRESSURE
CONTROL
VALVE
VENTILATION
EXHAUST
SECONDARY
CYCLONE
Fig. 10. Simplified Equipment Flowsheet of PDU Fluidized-Bed Combustor
and Associated Equipment. The "additive feeder" is actually
a "sorbent feeder."
c. Test Plan and Experimental Procedure
To measure the effects of bed temperature, combustor pressure, and
fluidizing gas velocity on combustion efficiency of the char in the ANL PDU, a
complete factorial series of experiments was planned at the following nominal
conditions: bed temperatures of 800 and 900°C; combustor pressures of 405 and
810 kPa; and gas velocities of 0.76 and 1.1 m/s. In all experiments, com-
bustion was performed at the condition of 3% oxygen in the dry flue gas.
When there is complete combustion, 3% oxygen in the dry flue gas corresponds
to ^16.5% excess air
In an experiment, partially sulfated limestone from a previous
experiment was charged to the reactor to provide an initial bed of material
for start-up. The limestone used in most of the char experiments was Greer
limestone (44.8 wt % CaO, 1.9 wt % MgO). In a few of the experiments, Grove
limestone (53.4 wt % CaO, 0.6 wt % MgO) was used.
-------
33
The starting bed temperature was raised to about 430°C by passing
preheated fluidizing air through the combustor and simultaneously employing
resistance heaters on the combustor wall. At this point, the system was
raised to the operating pressure, and coal was pneumatically injected into the
bed. Coal, rather than char, was used during start-up for safety since the
combustion characteristics of the char at low temperatures were unknown.
To prevent carbon accumulation in the fluidized bed at the initially
low bed temperatures, the coal was first injected intermittently until a
rapidly increasing bed temperature and a changing flue-gas composition confirmed
sustained combustion. Coal was then injected continuously, and the bed
temperature was raised to the combustion temperature selected for the experiment
The switchover in fuel from coal to char was accomplished by initially
charging the coal feed hopper (Fig. 10) with just enough coal for startup.
Sufficient char was added to the hopper on top of the coal to provide adequate
operating time to acquire the necessary data for determining combustion
efficiency. Once the combustor was operating on char and conditions were
established, the primary and secondary cyclone hoppers in the flue-gas system
were emptied, and steady-state samples of solids carried over in the flue gas
were collected.
During the combustion experiments, limestone was also fed to the
combustor. Sulfated sorbent was removed from the combustor by overflow into a
standpipe to maintain a constant fluidized-bed level of 0.9 m.
d. Results and Conclusions
Char Feed Problems. The only real problem encountered in performing
the char combustion experiments was in maintaining a steady feed rate of char
to the combustor. As indicated above, the char was extremely fine (estimated
mass mean particle diameter of '^30 ym) . The flow of char from the feed hopper
via a rotary valve into the pneumatic-transport feed line was very poor. As a
result, maintaining the combustion conditions at exactly the design nominal
operating conditions was virtually impossible. The duration of many experiments
was considerably shorter than planned due to complete stoppage of char feed.
In several experiments, the char and sorbent were premixed and fed
from a single hopper in an attempt to improve the flow characteristics of the
char. This operation gave only a slight improvement; there was the added
complication that the two materials segregated during feeding, which made
accurate determination of the char feed rate considerably more difficult.
In spite of the feeding difficulties, however, the data obtained
are considered to be sufficiently accurate to assess the combustibility of the
char in the PDU and to observe the effects of the gross changes in the operating
conditions on combustion. Since the carbon holdup in the FBC is very low,
the system can, for practical purposes, be assumed to be at steady state
relative to combustion efficiency at all times. Thus, even relatively short
combustion experiments of ^1 h were sufficient for this particular assessment.
-------
34
In those cases where it was necessary to make assumptions regarding
char feed rate, the assumptions were based on results of experiments in which
char was metered separately from the sorbent. It is also noteworthy that
combustion efficiency is not particularly sensitive to the char feed rate.
Errors of a factor of two in the char feed rate would alter the calculated
combustion efficiencies by only about 5% of the calculated value.
Combustion Efficiency. Combustion efficiency for each experiment
was calculated according to Eq. 1:
„,.,.. ,nnv (Rate of Unburned-Carbon Carryover) (1)
Combustion Efficiency = 100%. - r-r— 7—-—r -—-r\
•r (Rate of Carbon Feed)
Unburned carbon carryover included carbon in the limestone overflow, as well
as carbon recovered in the primary and secondary cyclones. Concentrations of
carbon monoxide, methane, and other gases representing incomplete combustion
were sufficiently low that they did not enter into the combustion efficiency
calculations.
The calculated results of the experiments are presented in Table 8.
The combustion efficiencies are also plotted in Fig. 11 as a function of the
three independent variables: bed temperature, combustor pressure, and gas
velocity.
Table 8. Combustion Efficiency of Flash-Pyrolysis Coal Char
in Fluidized-Bed Combustion Process Development
Unit (Combustion experiments were performed with
3% oxygen in the dry flue gas.)
Bed
Temperature,
°C
800
800
800
800
900
900
900
900
Combustor
Pressure,
kPa
405
405
810
810
405
405
810
810
Fluid izing-Gas
Velocity
m/s
0.76
1.10
0.76
1.10
0.76
1.10
0.76
1.10
Combustion
Efficiency,3
%
95, 94b
96
94 ,b 95b
no data
97
97
99, 98b
98b
aMore than one value indicates replicate experiments.
bBased on estimate of char feed rate when feeding char/sorbent
mixture.
-------
35
100
95
90
I
o
z
UJ
o
750 800 850 900 950
BED TEMPERATURE, °C
100
UJ 95
C/5
13
00
S
O
O
90
I
200 400 600 800 1000
COMBUSTOR PRESSURE, kPa
100
95
90
I
0.50 0.70 0.90 1.10 1.30
FLUIDIZING-GAS VELOCITY, m/s
Fig. 11. Combustion Efficiency of Coal Char as a Function of Bed
Temperature, Combustor Pressure, and Fluidizing-Gas Velocity
As mentioned above and as indicated in Table 8, it was necessary to
estimate the char feed rate in experiments where the char and sorbent were
premixed and fed to the combustor from a common feed hopper. The estimate was
required because the char and sorbent tended to segregate during feeding,
(evidenced by the feed rate needed to maintain 3% oxygen in the dry flue gas
constantly decreasing with experiment time). For replicate experiments in
which an estimate of the char feed rate was required for one of the replicates
(lines 1 and 7, Table 8) there was excellent agreement of combustion efficiencies,
The fact that in each case, the combustion efficiency based on the estimated
char feed rate was 1% lower than the other is a reflection of an intentional
bias used in estimating the char feed rate. For the five experiments in which
the char feed rate was directly measured, the char/air feed ratios were
determined. The highest ratio (rather than the average) was then used—the
-------
36
char feed rate in those experiments using char/sorbent feed mixtures was
estimated from the air feed rate. Thus, the combustion efficiencies calculated
from the estimated feed rates are considered conservative, with the reported
values equal to or less than the actual combustion efficiencies.
Calculated combustion efficiencies for the char varied from a low of
%94% to a high of ^99% in these experiments. These are considerably higher
than the combustion efficiencies obtained from burning coal in the same PDU
and under similar combustion conditions. Previous studies^ have indicated
combustion efficiencies of 90% or less for coal at 900°C and M.7% excess
combustion air; combustion efficiencies of ^98% were obtained with coal at
900°C, but a75% excess air was required.
The plots of the data in Fig. 11 show that temperature is the only
variable which appears to have a significant effect on the combustion efficiency;
combustion efficiency increasing from an average of ^95% at 800°C to a/98% at
900°C. A thorough statistical analysis of the results could not be performed
due to missing data at one set of operating conditions (Table 8). Efforts to
perform this experiment were repeatedly unsuccessful due to char feeding
difficulties.
Conclusions. The potential of this char as fuel for FBC has been
demonstrated. The two factors most likely to be responsible for the high
combustion efficiency of the char are (1) its extremely fine particle size and
(2) the fact that the char originated from a subbituminous coal. It cannot be
concluded from these results that all coal chars would be equally suitable for
FBC applications.
4. The Use of Oil Shale for SC^-Emission Control in Atmospheric Pressure
Fluidized-Bed Coal Combustors (AFBC)
(W. I. Wilson and R. B. Snyder)
The feasibility of using oil shale, rather than limestone or dolomite,
for S02 emission control in atmospheric fluidized-bed combustors (AFBCs) has
been investigated. The oil shale used in this study was from the Green River
formation in northwest Colorado. The Green River deposits are reported to
contain total reserves equivalent to 644x10° m^ (4050 billion barrels) of
oil; of this total, the equivalent of 99x10^ m^ (620 billion barrels) of
oil shale at concentrations ranging from 0.1 to 0.27 m^/Mg (25 to 65 gal/ton)
is considered to be economically recoverable.^ The processing of this vast
energy resource by retorting will yield a substantial quantity of spent oil
shale.
In our experiments, the SC>2 reactivity of both virgin shale and spent
shale (kerogen removed) was determined by thermogravimetric analysis, and the
results were used to estimate the quantities of shale required to reduce the
SC>2 concentration in the effluent gas sufficiently for compliance with
federal standards. These quantities were then compared with the required
quantities of Germany Valley limestone, Greer limestone, and Tymochtee dolomite.
The attrition rate of the virgin shale was also compared with those of the
limestones and dolomite.
-------
37
a. Materials
The compositions of the virgin and spent oil shales studied are given
in Table 9. Spent oil shale was prepared by heating virgin oil shale at 400°C
in air for 3 h to remove the organic material (kerogen). The chemical composi-
tions of Germany Valley limestone, Greer limestone, and Tymochtee dolomite are
also given in Table 9.
Table 9. Concentrations (in wt %) of Major Constituents of
Calcareous Materials
MgCC-3 Fe203 A1203 Si02 Na20 K20
Virgin Oil
Shale 27.7 13.8 2.60 4.14 30.1 2.19 1.21 0.89
Spent Oil
Shale 33.5 16.7 3.14 5.0 36.4 2.64 1.46 0.66
Tymochtee
(ANL-5101) 51.8 43.3 0.41 1.46 3.61 0.07
Greer
(ANL-8001) 80.4 3.50 1.24 3.18 10.34 2.23
Germany
Valley
(ANL-9701) 97.75 0.6 0.1 1.8 0.2 0.25
b. Experimental
A thermogravimetric analyzer (TGA) was used for the kinetic studies.
The TGA apparatus has been discussed in detail in a previous paper." The
oil shale reactions with S02 were performed at temperatures of 700 to 1050°C.
The attrition tests were performed for 2 h at 870°C, using a fluidizing
gas composed of 0.3% S02-5% C02 and the balance N2 at a superficial gas velocity
of 1.45 m/s. The apparatus (described in detail in ANL/CEN/FE-78-4) was a
5-cm-ID externally heated fluidized-bed reactor.
c. Sulfation Results
Spent oil shale was precalcined at 900°C in 20% C02-N2; sulfation was
for 3 h with the same gas except that 0.3% S02 and 5% 02 were added. In Fig. 12,
the conversion of CaO to CaS04 in spent, calcined oil shale at various temperatures
-------
38
REACTION TEMPERATURE
700° C SPENT OIL
750° C SPENT OIL
800° C SPENT OIL
C SPENT OIL
C SPENT OIL
SHALE
SHALE
SHALE
SHALE
SHALE
SHALE
SHALE
1050° C SPENT OIL
900°C VIRGIN OIL
Fig. 12. Conversion (Measured with a TGA) of CaO to CaS04 in
Precalcined Spent Green River Oil Shale at 700 to
1050°C. Reaction conditions: -50 +70 mesh material
precalcined in 20% C02-balance N2; sulfation gas, 0.3%
S02-5% 02-20% C02 in N2.
(700-1050°C) is given as a function of time. During the first hour of exper-
iments performed at 700-900°C, the reaction rates were the same; however,
during the remaining two hours, the reaction rates for all experiments except
those at 700 and 750°C varied significantly with temperature. The reaction
rates at 1000 and 1050°C were much lower than those at the lower temperatures.
This inverse variation of reaction rate with temperature is believed to be due
to the competing reaction of CaSCfy with Si02 and other impurities at the
higher temperatures. The latter reactions release S02 and form the products,
(Ca,Fe,Mg)Si03 and Ca(Fe,Mg)Si02. However, the shale sulfated between 700 and
800°C contained CaSffy, the double salt 3MgS04 -CaSC>4, silica, and silicates.
-------
39
At the lower temperatures, 700 to 800°C, the weight gain determined
with the TGA was larger than would be obtained if the sole reaction was 100%
of the CaO being converted to CaSC>4 (Fig. 12). Results from X-ray diffraction
and wet chemical analyses confirmed that at these temperatures, SMgSO^-CaSCfy
forms, as mentioned above. The formation of this compound would account for
the excessive weight gain.
Also, virgin oil shale was sulfated under these conditons. Conversion
of calcium to CaSO/^ for the virgin oil shale was slightly lower (73%) than for
the spent oil shale, (86%). Both reactions were essentially complete in 1 h
(Fig. 12). Spent oil shale was simultaneously calcined and sulfated at 900°C,
using a synthetic combustion gas of 0.3% S02~5% 02-20% (X^-balance N2. The
reaction rate was the same as for the precalcined material.
Fig. 13 compares the reaction rate of spent oil shale with S02 with
those of Tymochtee dolomite, Greer limestone, and Germany Valley limestone.^
During the first hour, the percentage of calcium converted to CaS04 for the
oil shale is higher than that for Tymochtee Dolomite (which is a highly
reactive dolomite). Nevertheless, by the end of the 3-h run, 98% of the CaO in
the Tymochtee dolomite was utilized to capture S02, compared with 83.5% for
the spent oil shale. The reaction rates of Greer limestone and Germany Valley
limestone are much lower than those of spent shale and Tymochtee dolomite.
100
o
03
<
-1
<
z
o
O
O
80
60
40 —,
20
1 I ' I
SPENT OIL SHALE
TYMOCHTEE DOLOMITE
GREER LIMESTONE
GERMANY VALLEY
2
TIME.hr
Fig. 13. Conversion (Measured with a TGA) of CaO to CaS04
in Spent Oil Shale, Tymochtee Dolomite, Greer
Limestone, and Germany Valley Limestone, Using
0.3% S02-5% 02 in N2 at 900°C
-------
40
d. Prediction of S02 Retention for Oil Shales
The reactivities of oil shale with SC>2 as determined with the TGA
were used to predict calcium utilization by this sorbent and its SC>2 retention
in an FBC pilot plant. The experimental kinetic data at 900°C were used to
estimate the quantity of oil shale necessary to meet the SC^-emission standard
of 0.5 g SC>2/MJ (1.2 Ib SC>2 emission/10^ Btu) released in a combustor.
The S(>2 reactivity curves (Fig. 13) must be converted to curves for
SC>2 retention vs. Ca/S ratio before the quantity of calcareous material
necessary per unit of coal to meet the SC^-emission standard can be predicted.
The predicted S02 retention for spent and virgin oil shale for a fluidized bed
operated at 3.9 m/s with a bed height of 1.1 m is described in ANL/CEN/FE-78-3.
Because the virgin oil shale has a lower SC>2 reactivity than spent shale at
900°C (Fig. 12) the virgin material requires a higher Ca/S ratio.
e. Evaluation of Oil Shale Sorbent for AFBC Sulfur-Removal System
The results, presented in ANL/CEN/FE-78-3 described above, were used
to estimate the number of kilograms of oil shale necessary per kilogram of
coal to meet the SC^-emission standard and were compared with previous results
for limestone. The analysis was performed for 3% and 4.3% sulfur coals, both
of which had a heating value of 28,319 kJ/kg (12,183 Btu/lb). The spent oil
shale contained 33.5 wt % CaCC>3 and 0.66 wt % sulfur. The SC^-emission
standard allows only 0.73 kg of sulfur to be emitted per 100 kg of such a
coal.
The estimates of oil shale requirements were made for 3% and 4.3%
sulfur coals combusted in a fluidized bed operated at 3.8 m/s with a bed
height of 0.83 m; these estimates were then compared with similar requirements
for Tymochtee dolomite, Greer limestone, and Germany Valley limestone. The
results are presented in Table 10. The Ca/S ratio of the virgin oil shale is
similar to that of Greer limestone, a reactive stone used by Pope, Evans, and
Robbins at their Rivesville FBC pilot plant. However, since the oil shale has
a low CaC03 content in comparison with the limestones and dolomites (^30% vs.
50-98%) for both 3% and 4% sulfur coals, the requirements for shale are
projected to be considerably larger than for the other stones.
It should be noted that these estimates are only for atmospheric
FBCs operated at a high superficial gas velocity, not for pressurized FBCs.
Decreasing the gas velocity and/or increasing the boiler pressure should
decrease shale requirements substantially. Also, if Western low-sulfur coal
(1-2% sulfur) is used in an FBC, the shale requirement will be reduced by a
factor of approximately four.
f.. Evaluation of Oil Shale Use in a Fluidized-Bed Combustion Plant
Employing a Carbon Burnup Cell
For an FBC system that employs a carbon burnup cell (CBC) , CaSC>4 in
the partially sulfated oil shale would decompose at the high operating temper-
ature of the CBC (1000-1100°C) due to the reaction of CaSC>4 with SiC>2 to
release SC>2 (as previously discussed). Thus, if oil shale is used in an FBC-CBC
-------
41
Table 10. Requirements for Green River Oil Shale,
Germany Valley Limestone, Greer Limestone, and
Tymochtee Dolomite to Meet SC^-Emission Standard3
Calcium-Based
Stone
Ca/S Ratio
3% S Coal
4.3% S Coal
Germany Valley
Limestone
(ANL-9701) 3.8
Greer Limestone
(ANL-8001) 3.1
Tymochtee Dolomite
(ANL-5101) L.O
7.5
3.8
1.5
kg of Stone/kg of Coal
I S Coal 4.3% S Coal
Spent Oil Shaleb
Virgin Oil Shalec
1.9
3.1
not possible
4.3
0.58
1.4
not possible
3.3
0.36
0.36
0.18
1.0
0.6
0.4
aBasis: 28,319 kJ/kg coal; FBC operated at 3.8 m/s gas velocity and 0.83-m
bed depth.
^Spent shale contains 0.66% S.
cVirgin shale has a heating value of 7020 kJ/kg and contains 0.9% sulfur.
system, the S02 in elutriated, partially sulfated oil shale might be released in
the CBC, and, to meet the S02 emission standard, more S02 would have to be
retained in the combustor. Assuming that only 5% of the bed material elutriates
from the bed, it is estimated that S02 retention in the FBC would have to be
increased to 95% to meet the S02~emission standard for 4.3% sulfur coal. The
results of this evaluation indicate that oil shales similar to Green River
shale should not be used in a FBC-CBC system.
g.
Attrition Results
The attrition tendencies of virgin Green River oil shale, Tymochtee
dolomite, Greer limestone, and Germany Valley limestone as bed materials were
determined. In the limestone and dolomite tests, simultaneous calcination and
sulfation was conducted. From weight changes and chemical analyses of the
original material, the final overhead, and final bed material, the percentage
of the original material that remained in the bed was determined. The
amount of attrition in the 2-h tests was taken to be the percentage of original
material collected overhead.
In the tests of virgin oil shale, simultaneous kerogen combustion,
calcination, and sulfation occurred. Again, chemical analyses and weight-change
information were used to determine the percentage of original material collected
overhead.
-------
42
The extent of attrition of virgin oil shale was 22 wt %, whereas
those for Tymochtee dolomite, Greer limestone, and Germany Valley limestone
were 3%, 11%, and 38% respectively. These results suggest that the attrition
rate of virgin oil shale is not unlike that of limestones and dolomites.
h. Conclusions
Laboratory analyses suggest that the calcium in oil shales has a
high reactivity with SC>2 and can be used in fluidized-bed coal combustors to
reduce SC>2 emissions. It is predicted that to meet the EPA S02~emission
standard, more oil shale than limestone or dolomite would be required since
the calcium content of shales is relatively low. Also, the Green River shale
used in the tests contains approximately 1% sulfur, which may be released as
SC>2 in an FBC. The use of shales may be desirable if the FBC is operated at
low superficial gas velocities (less than 3.2 m/s) or with low-sulfur coals
(containing less than 3% sulfur). The oil shales should not be used in an
FBC-CBC unless the shale elutriation rate is minimal, since at the temperature
at which a CBC operates (1000-1100°C) , the CaSC>4 and the silicates in the
elutriated shale would react to release SC>2.
The attrition rate of Green River oil shale was similar to that
of limestones and a dolomite.
Only one oil shale, Green River oil shale, was used in this eval-
uation. Since the S(>2 reactivity and attrition rates of limestones vary
widely, a large variation is also expected for oil shales. Thus, these
results are not necessarily applicable to all oil shales. The Green River oil
shale was not tested in an experimental FBC, and its performance in such a
unit might differ from the results reported. However, since all oil shales
contain much less calcium than limestones contain, further investigation was
not considered.
5. Comparison of Limestone Calcium Utilization in an AFBC with TGA Projections
(R. B. Snyder and W. I. Wilson)
Various investigators^>o~13 have studied the reactivity of limestones
with S02; however, none has reported a method of accurately predicting the
limestone reactivity with SC>2 at specified fluidized-bed coal combustion
conditions. When a high-sulfur (4% S) coal is used, approximately 85% of the
S(>2 generated must be retained in the bed material in order to meet EPA
S02-emission standards. However, the percentage of sulfur that must be
removed varies, depending on the sulfur content and heat content of a coal.
Therefore, process R&D engineers, in reporting results of experimental com-
bustion tests, usually give SC>2 retention as a function of the molar ratio
of calcium (in the limestone)/sulfur (in the coal) fed. The calcium utili-
zation of the limestone can then be expressed as follows:
(2)
-------
43
where
U = the fraction of the calcium converted to calcium sulfate
R = fraction of the sulfur dioxide retained in the combustor bed
material
Ca/S = molar Ca/S feed ratio
The designer of a plant knows what value of R is required to meet the EPA
S02~emission standard for a designated coal but needs to know the Ca/S ratio
required for the designated limestone(s) to meet this EPA standard. If the
calcium utilization, U, for a given limestone can be predicted, the Ca/S ratio
required to meet the EPA SC^-emission standard (R) can be determined from Eq. 2,
However, the limestone calcium utilization, U, is also a function of the
Ca/S ratio—for example, when a Ca/S ratio of 4 is used, the maximum calcium
utilization is 0.25. Therefore, for a highly reactive limestone, the calcium
utilization will be limited by the availability of SC>2 or the Ca/S ratio,
whereas for a limestone having a low reactivity with SC>2, the calcium utili-
zation will be limited by the reactivity of the limestone. The value of R
that is just large enough to meet the EPA S02 standard gives the minimum
required Ca/S ratio.
This report presents an expression for U in terms of Ca/S which includes
kinetic data obtainable from thermogravimetric analysis of the limestone in
question, as well as values of specific fluidized-bed combustion operating
parameters. A comparison is made between U calculated using this function and
actual pilot-plant calcium utilization. The predictions presented below are
for atmospheric fluidized-bed combustion (AFBC) only and cannot be directly
used to predict pressurized fluidized-bed operating conditions.
Seven limestones which had been previously tested in pilot-scale FBCs*
were tested for reactivity with SC*2 and 02 in the thermogravimetric analyzer
(TGA). However, in these TGA studies, the particle size was -18 +30 mesh, and
the results did not correlate well with actual pilot-plant calcium utilizations.
In the results reported below, however, the particle size was smaller, -50 +70
mesh, and the correlation of TGA results and pilot-scale TGA results was
good. All limestones, sized to -50 +70 mesh (250 ym, median diameter), were
precalcined in 20% CC-2 prior to reaction with the synthetic combustion gas.
All sulfation reactions were performed at 900°C with a 0.3% S02~5% C^-balance
N2 synthetic combustion gas.
The percent conversion of the calcium in the limestone to calcium sulfate
was determined on the TGA and is given in Fig. 14 for the seven limestones.
Reaction was rapid in the first hour, then became limited either by the
gas-CaSC>4 solid diffusion process or by the decrease in the pore diffusion
process or by the decrease in the pore diffusion coefficient of S02 due to
pores becoming smaller.1^
Predictions of calcium utilization in an AFBC were made, using the
kinetic information obtained from the ANL TGA experiments (Fig. 14) and were
Pilot-scale FBCs are referred to in the report literature as process
development units, test units, and fluidized-bed boilers.
-------
44
o
CO
o
O
o
o
o
CO
oc
UJ
o
CJ
ANL-8701
ANL-8001
_. -
ANL-8IOI
0.5
1.0
1.5 2.0
TIME, h
Fig. 14. Calcium Utilization in Seven Precalcined Limestones at
900°C as a Function of Time. Calcined at 900°C, 20% C02
(-50 +70 mesh)
compared with various pilot-plant experimental values of calcium utilization
(Table 11). The pilot-plant data were from the following sources: Pope,
Evans, and Robbins (PER)° determined the desulfurization behavior of Greer
(ANL-8001), Germany Valley, (ANL-9701), and Chaney (ANL-8701), limestones;
Argonne National Laboratory (ANL) studied limestones 1359 (ANL-9501), 1360
(ANL-8101), and 1337 (ANL-5301), and Tymochtee dolomite (ANL-5101); Consoli-
dation Coal Co. (CCC) tested Tymochtee dolomite (ANL-5101); and Morgantown
Energy Research Center (MERC) tested Greer limestone (ANL-8001).
The predicted U values shown in Table 11 were found using the TGA kinetic
information along with a fluid-bed desulfurization equation (Eq. 3) developed
by Keairns et al.^ for the calculation of calcium utilization.
-kHe
-
kHe
1 - e
(3)
where
U = calcium utilization, fraction
Ca/S = calcium to sulfur mole ratio in the feed
-------
Table 11. Comparison of Predicted and Pilot-Plant Calcium Utilization (U)
Pilot Plant
Organization
PER
PER
PER
PER
PER
PER
PER
MERC
MERC
ANL
ANL
ANL
ANL
ANL
ANL
ANL
ANL
ANL
ANL
CCC
CCC
CCC
CCC
CCC
Limestone/
Dolomite
Germany Valley
Germany Valley
Chaney
Greer
Greer
Greer
Greer
Greer
Greer
1359
1359
1359
1359
1359
1359
1360
1337
Tymochtee
Tymochtee
Tymochtee
Tymochtee
Tymochtee
Tymochtee
Tymochtee
Ca/S
4.9
2.8
3.7
3.4
3.1
2.9
2.4
5.76
2.17
2.6
2.4
2.6
2.3
2.5
2.5
2.3
2.2
1.6
1.5
0.8
0.4
1.9
1.5
0.96
Particle
Residence
Time,
h
2.6
2.6
2.6
3
2.6
2.6
2.6
-
-
15
15
15
15
15
15
15
15
15
15
2.5
5
11
13
21
Calcium
Utili-
zation,3
%
16.0
25.0
21.6
26.3
26.6
27.9
28.0
13.6
40.6
20
22
20
36
32
34
32
37
54
53
12.3
23.3
47.4
62
82
TGA-Predicted
Calcium
Utili-
zation,3
%
13.3
16.0
24.3
24.2
26.2
27.6
31.7
16.9
40.9
29.0
29.6
29.0
29.7
29.4
29.4
34.7
42.8
56.0
59.2
12.0
23.7
47.4
57.0
71.0
-p-
(Jl
3The relative standard deviation between the observed and predicted calcium
utilizations is +4.1%.
-------
46
V = superficial gas velocity, m/s
H = fluidized-bed height, m
e = bed voidage, assumed to be 0.5
k = average particle reaction rate constant, s~^
This fluid-bed desulfurization equation gives the calcium utilization as a
function of the "average" reaction rate constant k of the particles in the
bed, the Ca/S feed ratio, and the specified AFBC operating parameters.
The slopes of the curves in Fig. 14 give the instantaneous rate constants,
k1 , as a function of calcium utilization. The average limestone calcium
utilization, U, is that value at which k and k' are equal (see Keairns
et al.^ for calculational details).
It should be noted that the pilot plant results in Table 11 represent a
wide range of operating conditions. The PER combustion runs were done at 815
to 832°C. The ANL and MERC pilot plant runs were all performed at 871°C. The
Consolidation Coal Co. runs with Tymochtee dolomite were performed at 982°C.
In contrast, all TGA runs were performed at 900°C.
The superficial gas velocities ranged from 0.61 m/s in the MERC runs to
0.51 m/s used by PER. Pope, Evans, and Robbins used <0.6-cm «1.4-in) limestone;
50% by weight of the material was smaller than 30 mesh (600-ym dia)- It
would be expected that very little of the material smaller than 30 mesh
remained in the bed for an appreciable time. In the ANL experimental runs,
the average particle size was 540-630 i_nn; Consolidation Coal Co. used -16 +28
mesh (1190-1000 ym) material. The TGA-determined k's and the actual Ca/S
ratios and operating parameters of the pilot-plant runs were substituted into
Eq. 3 to calculate the predicted U's shown in Table 11.
The predicted calcium utilizations are plotted against actual pilot plant
experimental values in Fig. 15. Agreement is within the experimental error
for the pilot plants and thus is satisfactory. The relative standard deviation
is ±4%.
6. Predicition of Limestone Requirements for an FBC-CBC Combustor
(R. B. Snyder and W. I. Wilson)
To reduce the size and cost of an AFBC, a high superficial gas velocity
(3-4.6 m/s) is required. However, high gas velocities cause high loadings of
dust (limestone and unburned coal) in the effluent gas stream. High gas
velocities, therefore, result in low combustion efficiencies. Pope, Evans, and
Robbins (PER)l-> reported combustion efficiencies of ^85% at a fluidizing-gas
velocity of 3.8 m/s in a process development unit (PDU) combustor. To increase
overall combustion efficiency, PER incorporated into their process a carbon-
burnup cell (CBC) which operates at a higher temperature than the combustor
and at a lower fluidizing-gas velocity (^1100°C, 1.8 m/s). Unburned coal
dust removed from the combustor effluent gas stream by cyclones was injected
into a fluidized bed in a CBC where the unburned fuel was combusted to recover
the heat values. Pope, Evans,and Robbins estimates that the overall combustion
efficiency can be increased to approximately 99%.^
-------
47
o
o
Q
Q
UJ
(T
Q.
— VARIATION IN
EXPERIMENTAL RESULTS
0 10 20 30 40 50 60 70 80 90 100 110
EXPERIMENTAL PILOT
PLANT CALCIUM UTILIZATION, %
Fig. 15. Experimental and Calculated Calcium Utilizations.
-50 +70 mesh stone. 20% CC>2 in N2 (precalcination)
0.3% S02-5% 02 in N2 (sulfation) 900°C.
Due to the high operating temperature (1100°C) in the CBC, the S02
released in the CBC might not be captured by either partially sulfated lime-
stones or fresh limestones injected into the CBC. As a result, 7% of the
sulfur would bypass the sulfur-removal system of the combustor and be released
as S02 from the CBC. Consequently, a greater percentage of the sulfur (released
as S02) would have to be captured in the combustor to meet the EPA S02-emission
standard (0.5 g S02/MJ produced by the entire system). Because of this
increased requirement for sulfur retention in the combustor, a greater overall
Ca/S feed ratio would be required (at a higher sulfur retention, calcium
utilization is lower). Thus, if it is assumed that no sulfur is retained in
the CBC, much larger amounts of limestone would be required to meet the EPA
S02-emission standard.
Babcock and Wilcox^" estimate that the weight ratio of unburned sulfur
to carbon in the coal dust elutriated from the combustor is one-half that in
the original coal. Thus, at a combustion efficiency of 85% in the combustor,
approximately 93% of the sulfur in the coal is oxidized to S02 in the combustor;
the other 7% is unburned and is elutriated from the combustor in the coal dust
along with the partially sulfated limestone.
Conflicting results!7,18 have been reported by investigators as to the
fate of the elutriated sulfur released as S02 in a CBC. The National Coal
Boardl? and Argonne National Laboratory^** found that above 870° C, the S02
reactivity of several limestones dramatically decreased. In contrast, PER*-'
found that the addition of limestone 1359 (ANL-9501) to their CBC caused a
two-thirds reduction in S02 emissions.
-------
48
The TGA was utilized to determine the reactivity of three stones at the
operating conditions of an FBC and a CBC. The results (reported previously in
ANL/CEN/FE-77-11) showed that feeding fresh limestone or limestone that had
been partially sulfated in the combustor to the CBC should reduce SC-2 emissions
from the CBC by more than 60%. The quantity of limestone necessary to capture
60% of the SC>2 in a CBC depends largely upon the type of limestone tested and
the CBC operating temperature. A typical range of limestone requirements for
different stones is 0.36-1.2 kg of stone per kg of coal.
7. Estimation of Limestone Requirements for AFBC
(R. B. Snyder and W. I. Wilson)
Using the TGA, projections of limestone requirements to meet the EPA
S02~emission standard for AFBCs were made. Sixty-one limestones* were tested
on the TGA for reactivity with SC>2 at 900°C using 0.3% SC>2 . Some results were
previously reported (ANL/CEN/FE-78-4), and additional results are now available.
There is a large variation in limestone-S02 reactivity and in the extent of
conversion of calcium carbonate to calcium sulfate. For the high-calcium
(>90% CaC03) limestones tested, the conversion of CaCC>3 to CaS04 ranged from
19 to 66%, while for the dolomites (40-60% CaCC^) the range was 21 to 100%.
By use of limestone-SC>2 reactivity curves (not shown), estimates were
made of the quantity of each limestone per unit weight of coal necessary to
meet the Federal SC^-emission standards. These estimates were made for a
fluidized-bed coal combustor operated with a 0.9-m bed depth and superficial
gas velocities of 2.4 m/s (8 ft/s) and 3.6 m/s (12 ft/s) and using a coal
containing 4.3% S and having a heating value of 28,300 kJ/kg (12,200 Btu/lb).
In Table 12, the limestone/kg coal required for the reactor at the two
different velocities are given for 57 of the 61 stones tested.
For a reactor operated at 2.4 m/s (8 ft/s), 54% of the total stones
tested required less than 0.5 kg/kg of coal to meet the SC^-emission standard.
This included 54% of the dolomites (with CaCC>3 concentrations of 48 to 60%)
and 55% of the limestones (with CaCC>3 concentrations larger than 90%). All
stones can meet the SC^-emission standard when the AFBC is operated at a
superficial gas velocity of 2.4 m/s. At 3.6 m/s, only one stone is projected
not to be able to reduce SC>2 emission sufficiently to meet the EPA standard.
These results indicate that a large percentage of the calcareous stones
are highly reactive with SC>2. Since limestones and dolomites are extremely
plentiful throughout the United States, these results suggest that it is likely
that there would be a high SC^-reactivity stone (which would minimize limestone
requirements and limestone disposal cost) near a chosen combustor site. The
highly reactive stones (54% of the stones tested) would require between 0.25
and 0.5 kg stone/kg coal to meet the present S02~emission standard (combustor
operated at 2.4 m/s).
It can be seen from Table 12 that increasing the superficial gas velocity
from 2.4 m/s (8 ft/s) to 3.6 m/s increases the limestone requirements by
approximately 90%. From the fluid-bed desulfurization equation developed by
*The compositions and sources of the limestones used in these studies
are given in Appendices A and B.
-------
49
Table 12. Projected Limestone Requirements, kg Stone/kg of Coal,
to Meet 0.5 g S02/MJ Standard for AFBC
Limestone
ANL-4801
ANL-4901
ANL-4902
ANL-4903
ANL-5001
ANL-5101
(Tymochtee)
ANL-5102
ANL-5201
ANL-5202
ANL-5203
ANL-5204
ANL-5205
ANL-5206
ANL-5207
ANL-5301 (1337)
ANL-5302
ANL-5303
ANL-5304
ANL-5401
ANL-5402
ANL-5403
ANL-5501
ANL-5601
(Dolowhite)
ANL-5602
ANL-5603
ANL-6101 (1351)
ANL-6301
ANL-6401
ANL-6501
ANL-6701
ANL-6702
ANL-7401
ANL-8001
(Greer)
ANL-8101 (1360)
ANL-8301
ANL-8701
(Chaney)
ANL-8901 (1343)
ANL-8902
ANL-8903
ANL-9201 (1336)
ANL-9401
Reactor Aa
0.30
0.73
0.30
0.46
0.40
0.27
0.60
0.48
0.46
0.45
0.74
0.59
0.39
1.2
0.34
0.98
0.83
1.0
0.33
0.32
0.39
2.4
-
0.89
0.95
-
0.37
0.46
0.49
0.28
-
0.62
0.48
0.47
0.79
0.30
-
0.52
0.46
-
0.35
Reactor B^
0.40
1.1
0.43
0.78
0.48
0.37
0.86
0.92
0.82
0.82
1.0
0.9
0.45
2.4
0.53
2.0
1.3
1.1
0.36
0.40
0.59
Std. not met
-
1.2
1.2
-
0.49
0.57
0.93
0.37
-
0.79
0.73
0.51
1.0
0.34
-
0.8
0.57
-
0.44
Total Ca Utili-
zation in TGA,C %
100.0
88.3
86.2
94.8
100.0
99.2
91.3
93.7
94.8
93.1
56.8
85.0
100.0
75.8
96.8
65.1
68.9
42.4
97.7
92.3
97.2
21.3
-
44.8
43.6
-
79.7
62.6
77.0
81.3
13.1
65.8
53.2
53.6
-
60.8
-
44.0
44.6
-
52.5
(Cont'd.)
-------
50
Table 12. (Cont'd.)
Total Ca Utili-
Limestone Reactor Aa Reactor B^ zation in TGA,C /
ANL-9402
ANL-9501
(Grove)
ANL-9502
ANL-9503
ANL-9504
ANL-9505
ANL-9601 (2203)
ANL-9602
ANL-9603
ANL-9701
(Germany Valley)
ANL-9702
ANL-9703
ANL-9704
ANL-9705
ANL-9706
ANL-9801
ANL-9802
ANL-9803
ANL-9901
ANL-9902
ANL-9903
Range
1.5
0.56
0.43
0.37
0.44
0.51
1.1
0.45
0.55
0.96
0.52
0.27
0.36
0.41
0.34
0.52
0.32
0.67
0.39
0.31
0.51
0.37-2.4
2.0
0.62
0.71
0.56
0.51
0.78
1.5
0.61
0.68
1.3
0.61
0.28
0.42
0.57
0.41
0.67
0.52
0.91
0.43
0.51
0.80
0.28-00
26.5
37.5
55.6
56.6
43.9
44.4
32
51.7
32.1
18.7
31.0
66.2
41.7
46.2
49.2
35.3
61.8
33.3
39.6
56.7
56.7
aReactor A is assumed to operate with a 0.9 m bed depth at 2.4 m/s.
Calcium-S02 rate constant is 31.5 s"-*-.
"Reactor B is assumed to operate with a 0.9 m bed depth at 3.6 m/s.
Calcium-S02 rate constant is 47 s~l
cThis column lists the percentages of calcium in the limestones (-50
+70 mesh) converted to CaSC>4 after the stone is first precalcined in
20% C02 at 900°C, then reacted at 900°C with a 0.3% S02-5% 02-
balance N2 gas mixture for 3 h.
Westinghouse,!^ the rate constant for the calcium-S02 reaction can be
calculated for the two sets of conditions studied. These reaction rate
constant values are 31.5 and 47 s~l.
It should be noted that total calcium utilization in the TGA (Table 12) is
not necessarily a good indicator of limestone performance in the combustor.
Two examples are as follows: (1) In a comparison of Greer (ANL-8001) with
ANL-7401, ANL-7401 has a greater calcium utilization on the TGA, but requires
a greater amount of stone/kg coal than does Greer. (2) In a comparison of
ANL-9902 with ANL-9901, 56.7% of the CaC03 in ANL-9902 is converted to €3804
-------
51
compared with 39.6% for ANL-9901. However, ANL-9901 requires less limestone
than ANL-9902 at a superficial gas velocity of 3.6 m/s. Figure 16 shows the
reactivity curves for the two stones. Sulfation of ANL-9901 is much faster
than the sulfation of ANL-9902. Since the maximum SC>2 residence time in a
fluid bed with 0.9 m bed depth and 3.6 m/s gas velocity is approximately 0.25 s,
the important factor in obtaining high SC>2 retention is a high calcium-SC>2
reactivity, not total conversion of calcium to calcium sulfate. Thus, ANL-9901
is superior to ANL-9902.
20 40 60 80 100 120
Fig. 16 TGA Reactivity Curves for ANL-9901
and ANL-9902 Limestones
The projections of limestone requirements given in Table 12 are sensitive
to the rate of the calcium-S02 reaction or, in other words, to the slopes of
the TGA curves of calcium conversion. Thus small changes in the rate of
calcium conversion to CaSO^ can produce large changes in projected limestone
requirements. The results given in Table 12 are more useful for comparing
limestones than are absolute values. Finally, Table 12 results are given to
two significant figures to allow comparison of limestones with each other; it
is not suggested that the projected limestone requirements are accurate to two
significant figures. Because the accuracy of the absolute numbers is not
known, they should be used with caution.
It would be useful to be able to correlate limestone physical properties
with limestone-S02 reactivity results. Since in general, the total calcium
utilization of a limestone does reflect the reactivity of a stone with SC>2,
investigators have attempted to correlate calcium utilization with limestone
properties. From Fig. 17, we see that the chemical composition (CaC03 concen-
tration) of the stone does not correlate well with calcium utilization. It
has been shown, however, that the average pore diameter of limestones that
have been treated with agents that enhance reactivity with S02 generally does
correlate with calcim utilization. Therefore, the pore-size distributions of
23 untreated limestones and 21 untreated dolomites (after calcination at 900°C
using a 20% CQ% in N2 gas mixture) were measured. The pore-size distributions
were thus identical to those of calcined limestones that were reacted with
SC>2 • Figure 18 shows the calcium utilization as a function of average pore
diameter. Although the correlation is poor, there is a general trend
-------
52
0
^-
o
CO
o
0
0
0
O
z~
o
CO
Q;
LJ
O
O
IUU
90
80
70
60
50
d
40
30
20
10
n
1 Q-^ncX
— o -
00
0 0
o -
0° o ° 0 °
°G9G 00
>Sb <5>
-------
53
suggesting that other physical properties such as grain size may be important.
Therefore, petrographic examinations of limestones were initiated to help
understand limestone-S02 reactivity. This work is reported below.
8. Petrographic Examination of Limestones
TR. B. Snyder, H. L. Fuchs, and W. I. Wilson)
As stated above, there is a large variation in calcium utilization for
various limestones which have essentially the same chemical composition. It
is believed that differences in SC>2 capacity of stones of similar composition
are due to variations in the stone such as porosity, grain size, and grain
defects. To investigate this, five stones (ANL-5402, ANL-5501, ANL-9802,
ANL-9701, and ANL-5001) were selected for an initial petrographic study. The
results were reported previously (ANL/CEN/FE-78-4) and are summarized here.
Two dolomites were studied, one with high S(>2 reactivity and one with low
SC>2 reactivity. The low SC>2 reactivity of one dolomite was due to its
exceptionally large grain size (produced by metamorphism). Two limestones
were also studied which had large differences in reactivity with SC>2. Their
reactivity differences could not be explained by petrographic properties.
The fifth stone is a "popper"* that is, it explodes upon calcination. Petro-
graphic examination indicated that this popping might be due to inclusions of
liquids, gases, or solid organics which form gases that rapidly expand upon
heating. Another dolomite (ANL-5301) which also has a large number of
inclusions is not a popper. This suggests that the type of inclusions, as
well as their numbers, may affect whether popping occurs.
9. Effect of Water on SC>2 Reactivity
(R. B. Snyder and W. I. Wilson)
The reactivity of limestones with SC>2 in the presence of excess oxygen
has been investigated, using a TGA. In all of these experiments, a synthetic
combustion gas composed of 0.3% S02~5% C^-balance N2 has been used. However,
in a fluidized-bed combustor, the combustion gas also contains approximately
7% H20. To determine any effect of water, three limestones were reacted with
0.3% S02~5% C>2 gas having various concentrations of I^O. The presence of 1^0
in the reactant gas had essentially no effect on limestone-S02 reactivity.
The results are also presented in a previous quarterly report (ANL/CEN/FE-78-4)
10. Limestone Attrition in a Fluidized-Bed
(R. B. Snyder and W. I. Wilson)
The effects of certain variables on limestone attrition have been
reported previously (ANL/CEN/FE-78-4) for ANL-8001, ANL-9601, and ANL-5101
stones. Several conclusions were drawn from these results and are summarized
here:
*"Poppers" are stones which, as the particles rupture or break up upon
calcination, make a popping sound. Some stones actually explode into very
fine particles. Of about 60 stones tested to date, approximately 20% are
"poppers."
-------
54
(1) An increase in temperature (20 to 870°C) increased the attrition
rates of both limestones ANL-9601 and ANL-8001 but not that of ANL-5101.
(These were 8-h batch tests with the stones fluidized by air.) More exper-
iments with additional stones are needed to determine the effect of temperature
on attrition.
(2) Calcination has no effect on attrition rate upon 8-h attrition at 870°C.
The attrition rate of precalcined and virgin material was identical for
all three stones. This conclusion does not apply to limestones or dolomites
that are "poppers."
(3) Sulfation of a limestone may decrease its attrition rate. When
sulfated, ANL-5101 and ANL-9601 stones had 30% and 40% as much attrition as
when tested with all the conditions the same except that SC>2 was absent.
ANL-8001 stone showed no attrition loss since this limestone in its natural
form is very attrition-resistant.
(4) The extent of sulfation or rate of sulfation is important. Even
though AN1-5101 stone is a softer stone than ANL-8001, simultaneous calcination
and sulfation of ANL-5101 produced a stone that was more attrition-resistant
than ANL-8001 stone. This is in agreement with results (ANL/CEN/FE-77-3) of
previous attrition studies of ANL-8001 and ANL-5101 in the PDU pressurized
combustor in which ANL-8001 and ANL-5101 losses due to attrition and elutriation
were 20% and 16%, respectively.
(5) Limestones are not more attrition-resistant than are dolomites. For
example, ANL-9601 (limestone 2203) which is very soft and has a high calicum
content (96% CaCX^) is less attrition resistant than ANL-5101 (Tymochtee
dolomite).
(6) Composition may be important (ANL/CEN/FE-77-3). Greater quantities
of Al, Si, and Fe in a stone may produce more attrition-resistant stones.
This can only be determined by examining a large number of limestones and
dolomites.
(7) Other limestone physical properties such as grain size and grain
defects are important. As shown previously (ANL/CEN/FE-77-3, p. 110), lime-
stones with similar compositions can have widely different attrition rates.
-------
55
TASK B. TURBINE-CORRODENT STUDIES
(S. H. D. Lee)
In the prospective application of pressurized fluidized-bed combustion of
coal to power generation, corrosion of turbine hardware due to attack by
alkali-metal compounds present in the hot flue gas is a potential problem.
This problem can be eliminated or alleviated by lowering the concentration of
alkali-metal compounds in the hot flue gas to a level tolerable to the turbine.
A direct method to accomplish this would be control of combustion so as to
minimize evolution of alkali metal compounds. The mechanism of evolution of
alkali-metal compounds during the combustion of coal at fluidized-bed combustion
temperatures is not well understood. An alternative method of lowering the
concentration of alkali metal compounds in the combustion gas is with a hot
granular-bed filter upstream from the turbine.
Research work has been undertaken (1) to obtain a better understanding of
the emission of alkali-metal compounds during the combustion of coal at the
temperature range of fluidized-bed combustion, (2) to investigate the effec-
tiveness of additives in controlling the emission of alkali-metal compounds
during the combustion of coal, and (3) to develop an effective granular-bed
filter sorbent material for the removal of alkali-metal compounds from hot
combustion gas.
1. Emission of Alkali Metals during the Combustion of Coal
Work on this subject has continued. A simulated high-chlorine, high-sodium
coal was investigated. This type of coal is of interest because of past
observations that fouling and corrosion of the fireside of a boiler furnace is
closely related to chlorine and sodium contents.19-21 Simulated high-chlorine,
high-sodium coals were made by impregnating coals with NaCl in a water
solution.
This study was carried out using a laboratory-scale horizontal fixed-bed
batch combustor system consisting of a preheating section which is preheated
at startup, a combustion section where the sample is combusted, a filtration
section containing a tubular alumina filter, and a cold-trap section containing
a water-cooled metal cold trap. This combustor system was described in an
earlier annual report in this series (ANL/ES-CEN-1016). Experimental results
obtained from the combustion of Illinois Herrin No. 6 coal and activated
coconut charcoal and part of results from the combustion of Glen-harold
lignite (North Dakota) were reported in ANL/CEN/FE-77-3. Presented here are
additional experimental results obtained from the combustion of Glen-harold
lignite.
In the study of transport of alkali metals from the combustion of coal,
it is important to understand the chemical form of the alkali metals present in
the coal. In order to obtain information on this, 20-g samples of lignite
were leached with distilled water at 95°C for 5 h and filtered, and the filter
cake was leached with 5% HC1 solution at 95°C for another 5 h. The water
leachant was analyzed for Na+, K+, Cl~, and S042~, and the acid leachant was
analyzed for Na+ and K+. Results showed that 44% of the sodium is present in
compound forms that are soluble in water and 40% of the sodium is in a form
soluble in 5% HCl solution. In contrast, potassium in lignite principally
occurs in water-insoluble, acid-insoluble forms. A low concentration of Cl~
-------
56
was observed in the water leachant indicating that the water-soluble sodium in
lignite is apparently not present as the chloride. The sodium in lignite has
been reported22 to be present primarily as salts of humic acids.
The material balances for sodium in three series of combustion experiments
with Glen-harold lignite were reported and discussed in the preceding annual
report (ANL/CEN/FE-77-3) and for potassium in a quarterly report (Table 7,
ANL/CEN/FE-77-8). Of the three series of experiments, one was with plain
lignite, and the other two with 0.5 wt % NaCl-impregnated lignite. In all of
these experiments, lignite was combusted in air at a flow rate of 3.5 L/min.
Each series of experiments consisted of five batch combustion runs; in each
run, 50 g of dry lignite was burned. Results indicated that the potassium in
lignite behaves as if it is much more volatile than sodium. At 850°C, 38.1%
of the potassium volatilized, whereas only 11.5% of the sodium volatilized; at
900°C, volatilization of potassium was 47.8% in comparison with 13.5% for
sodium. Results also showed that the addition of NaCl to lignite significantly
enhances the evolution of potassium during combustion. The potassium in the
condensate collected from the cold-trap finger was identified by X-ray dif-
fraction* as KC1.
As shown by a leaching study potassium occurs in lignite as compound
forms that mostly are not soluble in water. However, results obtained from
the combustion of plain lignite indicated that significant amounts of potassium
are converted into water-soluble compounds during combustion.
This phenomenon was previously observed for sodium. Results from the
leaching of the combustion residue showed that during the combustion, not only
is there physical vaporization of NaCl, but also NaCl undergoes the following
chemical reaction to release chlorine and to form sodium sulfate in the residue:
2 NaCl + S02 + 1/2 02 + H20 J Na2SC>4 + 2 HC1 t
2. Effect of Additives on the Retention of Alkali Metals in the Bed
During the Combustion of Coal
The previous annual report (ANL/CEN/FE-77-3) showed that during the
combustion of coals impregnated with 0.5 wt % NaCl, the greater the amount of
mineral matter in the coal, the more the added NaCl is tied up in the ash and
therefore the less NaCl is vaporized. Qualitatively, clay minerals and silica
have been reported to be effective "getters" of alkali metals during the
combustion of coal.2-* In order to quantitatively study the effectiveness of
some additives (such as clay minerals, silica, and limestones, etc.) on the
retention of alkali metals during combustion of coal, a series of experiments
has been completed in which 0.5 wt % NaCl-impregnated activated coconut
charcoal (in a mixture with additive) was burned. The same combustor as
described in Section 1 above was used. In each experiment, 5 wt % additive
was physically mixed with 20 g charcoal and the mixture was combusted at an
air flow rate of 3.5 L/min, 900°C, and atmospheric pressure.
Experimental results obtained from this series of experiments were
reported and discussed in detail in an earlier report (ANL/CEN/FE-77-8).
Table 13 is a summary of the material balances of sodium and potassium. Fairly
*work done by B. Tani
-------
57
Table 13. Material Balances of Sodium and Potassium
from Combustion of 0.5 wt % NaCl-Impregnated
Activated Coconut Charcoal Mixed with 5 wt %
Additive
Experiment
No. SL-10- Additive
Total Amount, mg
Input Output
Na K Na K
% Alkali Fixed as
Water Insoluble
Alkali
Na K
1
2
4
5
6
None 50 188 49
Kaolinite
No. 9 50 189 53
Montmorillonite
No. 27
Greer Limestone
Dolowhite
66 192 62
51 194 53
50 188 47
Tymochtee
Dolomite 50 194 48
125
154
174
141
154
150
12
64
65
57
22
46
5
62
71
25
10
22
7
8
9
10
Bauxite
Pure Alumina
Silica Gel
A1203 Filter
Powder
50
52
51
51
188
188
188
190
51
45
49
46
165
157
147
147
64
10
76
39
61
14
52
37
good material balances for sodium were obtained, except for the experiments in
which pure alumina and A1203 filter powder were the additives. For potassium,
the total "Output" was more than 10% lower than the total "Input" in each
experiment. Nonhomogeneity of the charcoal sample used, which could cause a
great variation in the potassium content of the charcoal, is a possible source
of the error since the material balance is very sensitive to the potassium
content used in calculating the "Input" term. In contrast, error due to the
variation in sodium content was not reflected in the sodium material balance
because of the low sodium content of the charcoal. The large variation in the
potassium content in the charcoal was previously observed (Table 30, ANL/CEN/FE-
77-3).
-------
58
In the two right-hand columns of Table 13 are the percentages of alkali
fixed as water-insoluble alkali in the ash. This is a suitable basis for
evaluating the effectiveness of an additive in retaining alkali metals because
the alkali metals tied up in these forms are not easily vaporized into the
combustion gas. As may be seen in Table 13 the retention effect of these
additives, in general, was similar for sodium and potassium. Clay minerals
(kaolinite and montmorillonite) tied up sodium in the ash very effectively.
The SC>2 sorbents tested (Greer limestone, Tymochtee dolomite, and Dolowhite)
were compared—the effectiveness of a sorbent for retaining sodium was found
to be related to the amount of impurity (such as A1203 and Si02) in the
sorbent. Greer limestone is an impure limestone (10% SiC>2 and 3% Al2C>3) and
showed some effectiveness in retaining sodium. In contrast, Dolowhite is a
very pure dolomite and only a poor "getter." Bauxite is an aluminum ore and
generally contains clay minerals. Its effectiveness was similar to that of
kaolinite and montmorillonite. Pure alumina was inert. On the other hand,
silica gel showed a substantial ability to tie up sodium. This is probably
related to its large surface area. The results of this series of experiments
clearly support the thought that clay minerals in the mineral matter of coal
are responsible for the retention of alkali in the coal ash during the com-
bustion of coal.
3. Removal of Gaseous Alkali Metal Compounds from the Hot Combustion Gas
of Coal
Work is under way with a laboratory-scale, batch-type, fixed-bed combustor
system (described in Section 1 above) to develop bed material for the granular-
bed filter that will remove alkali metal compounds from the hot combustion gas
of coal. In an experiment, a bed of candidate sorbent is placed between two
perforated steel plates in a stainless steel tube. The tube, in the filtration
section of the combustor, is fastened upstream from existing cold traps. A
weighed pure alkali metal compound* (such as NaCl or KC1) is heated and
vaporized in the combustion section of the combustor, which is heated by
induction heating. The vapor is carried by preheated flowing gas through the
sorbent bed and finally is condensed on the cold traps and a backup filter
(i.e. , a glass wool filter). The candidate sorbent is tested at temperatures
between 800 and 900°C and is evaluated for its effectiveness in removing the
alkali metal compound from the alkali-metal-compound-bearing hot flowing
gas.
The particle size range of each sorbent was carefully controlled by
screening; all sorbents tested were -8 +10 mesh. Before a sorbent was tested,
it was heat-treated in a muffle furnace at 900°C for 15 h in flowing air. The
purpose of this treatment was to remove any possible alkali metal compounds
present in the sorbent before the experiment which might be vaporized during
the experiment.
It was found that a small fraction of the alkali metal compound vapor in
the flowing gas was lost on the combustor wall upstream from the sorbent bed.
Since substantial amounts of steel scale peel off the combustor wall at the
*A.C.S. reagent
-------
59
end of each test, it is believed that the loss of alkali compound was mainly
due to reactions of alkali metal compound vapor with the hot stainless steel
wall of the combustor. To obtain the amount of alkali metal compound vapor
transported through the sorbent bed, the amount of alkali metal compound vapor
vaporized duirng a test needs to be corrected for this loss. For the screening
test experiments using the combustor system, the amount of loss was obtained
using data from blank runs with the experimental conditions the same as in a
regular experimental run except that the sorbent bed was absent. The candidate
sorbents tested in screening tests were evaluated based on the material
balance of alkali metal compound (obtained on a batch basis). In the rest of
the experiments (parametric tests), each sorbent was evaluated based on the
direct analysis of an alkali element in the sorbent that had been tested.
During this reporting period, several commercial products were screened
(as described below) by measuring their effectiveness for removing NaCl, KC1,
and K2SO^. vapors. From the screening test results, the two most promising
products, diatomaceous earth and activated bauxite were chosen for further
parametric tests. Results from these screening and parametric tests were
reported and discussed in detail in earlier quarterly reports (ANL/CEN/FE-77-11,
-77-3, and -78-4). Presented here is a summary of these results, as well as
additional results obtained since.
a. Screening Tests
NaCl Vapor Tests. Six commercial products (alundum, Celatom MP-91
diatomaceous earth, silica gel., Burgess No. 10 pigment, attapulgus clay, and
activated bauxite) were first screened by measuring their effectiveness in
removing NaCl vapor from hot NaCl-vapor-bearing flowing air. In each test, a
7.6-cm-thick sorbent bed was packed, and the bed was maintained at an average
temperature of 870°C under an air flow of 3.5 L/min. Under these conditions,
the loading of NaCl vpor in the air was about 60 ppm, and the linear velocity
of the air passing through the sorbent bed was 7.6 cm/s.
The NaCl material balances for these sorbent tests have already been
reported (ANL/CEN/FE-77-11). The alundum, a high-purity a-Al203, was inert.
A good material balance obtained from the test with inert a~Al203 demonstrated
that there was no physical condensation of NaCl vapor on the sorbent.
Celatom MP-91 diatomaceous earth, Burgess No. 10 pigment, and activated
bauxite effectively captured 74, 78, and 92% of the NaCl vapor, respectively.
In other work in which silica gel was an additive to activated coconut charcoal
which was combusted (Table 13), silica gel was effective in tying up sodium in
the ash bed; however, when silica gel was used as a sorbent, it was only
moderately effective in tying up NaCl vapor. This was due to the loss of
porosity of the silica gel, as will be seen in the porosity-measurement
study. Attapulgus clay showed fair effectiveness in capturing NaCl vapor;
however, this clay, as well as Burgess No. 10 pigment, became fragile after
the test. From a viewpoint of practical application, this behavior is not
acceptable for a candidate sorbent because fine particles would be produced,
increasing the load on the downstream particulate-removal facilities. In
general, the activities of sorbents for NaCl-vapor capture were found to be
related to their internal surface areas.
-------
60
The results of an earlier water-leaching study of these sorbents
indicated that for diatomaceous earth, Burgess No. 10 pigment, and attapulgus
clay, the NaCl vapor was primarily retained via chemical reactions with the
sorbents; for silica gel and activated bauxite, the NaCl vapor was retained by
a physical adsorption process.
To investigate the capability of these sorbents at high gas velocity,
Celatom MP-91 diatomaceous earth, Burgess No. 10 pigment, and activated
bauxite were further tested at a flow velocity of 25 cm/s, using either air or
a simulated flue gas. Average bed temperature was 880 C. The composition of
the simulated flue gas was 3% 02, 16% C02, 180 ppm H20, 300 ppm S02, 80 to 110
ppm NaCl vapor, and the balance N2. The material balances of NaCl for all of
these tests showed that the capability of all three sorbents for NaCl vapor
retention was substantially greater than at low gas velocities; 96, 85, and
98% NaCl captures were obtained by diatomaceous earth, Burgess No. 10 pigment,
and activated bauxite, respectively. A possible explanation for this increase
is that a higher linear velocity of the flue gas increases the mass transfer
of NaCl vapor from the bulk of flue gas to the surface of the sorbent, thereby
increasing sorption of NaCl vapor by the sorbent. Results obtained from the
leaching of sorbents tested in this series of experiments again showed that
the sodium removed by diatomaceous earth and Burgess No. 10 pigment was
essentially present in chemical forms not soluble in water. In contrast, most
of the sodium sorbed by activated bauxite was present in water-soluble form.
The importance of this is that sodium adsorbed on activated bauxite can be
easily removed by a simple leaching process, making it possible to regenerate
activated bauxite for reuse. The regenerability of activated bauxite and the
effect of regeneration on its sorption capability will be studied.
KCl-Vapor Tests. Except for low-rank Western coals, the potassium
content of typical U.S. coal is higher than the sodium content—approximately
two moles of potassium are present for each mole of sodium.24- Potassium
compounds and sodium compounds exist in coal in similar chemical forms.
Potassium chloride and sodium chloride are present in saline groundwater
permeating the rock and filling pores and cracks in the coal bed. Both
potassium and sodium are also constituents of clay minerals in coal, such as
illite and montmorillonite. Potassium chloride is known to have an appreciably
higher vapor pressure at the fluidized-bed combustion temperature range than
does sodium chloride (0.7 kPa at 900°C for KC1, as compared with 0.4 kPa for
NaCl); therefore, its vapor is expected to be present in the flue gas of
pressurized fluidized-bed combustors (PFBCs)-
Thertnodynamic calculations to identify the major species present at
the operating conditions for a fluidized-bed combustor have been made by
researchers at Westinghouse Research Laboratory.2^ Results of the calculations
indicate that gaseous potassium chloride is the major potassium carrier.
Based on these results, Celatom MP-91 diatomaceous earth, activated bauxite,
and Burgess No. 10 pigment were further tested for their sorption capability
for KC1 vapor, under the same conditions used to test them with NaCl vapor at
25 cm/s linear gas velocity.
Material balances of KC1 from tests of these sorbents for KCl-vapor
capture indicated that diatomaceous earth and activated bauxite captured KC1
vapor as effectively as they captured NaCl vapor. Ninety-eight and ninety-five
-------
61
percent KC1 vapor captures were obtained with diatoraaceous earth and activated
bauxite, respectively. In comparison with those two sorbents, Burgess No. 10
pigment was less effective for KCl vapor capture (74%), just as for NaCl vapor
capture (shown above). The similar sorption activity of these sorbents toward
both NaCl and KCl vapor is expected because of the similar chemical properties
of NaCl and KCl.
K2SC-4-Vapor Tests. In the flue gas from coal combustion, sodium and
potassium are expected to be present as sulfates also. Sodium sulfate and
potassium sulfate in liquid form are generally believed to be the precursors
for metal sulfidation, the most common mode of hot corrosion occurring in gas
turbines;26 therefore, the ability of candidate sorbents to retain alkali
metal sulfates must be demonstrated. Both sodium and potassium as sulfates
have substantially lower vapor pressures than as chlorides. Because the
maximum allowable operating temperature of the laboratory-scale fixed-bed
combustor is 900°C, the sorption capability of the sorbents for alkali metal
sulfates could not be determined using this combustor; therefore, a small-scale
sorption-test rig that can be operated up to 1250°C was assembled. The
apparatus is described in a previous report in this series (ANL/CEN/FE-78-4).
Two 24-h experimental runs were conducted, one each to test activated
bauxite and diatomaceous earth for their abilities to retain potassium sulfate
vapor in a simulated PFBC flue gas. Table 14 shows the experimental test
conditions, and Table 15 shows the material balances of ^804 from these two
tests.
Table 14. Experimental Conditions for Testing
Diatomaceous Earth and Activated Bauxite
Sorbents for ^SO^ Vapor Capture
Avg. Sorbent-Bed Temperature 850° C
System Pressure 100 kPa
Particle Size of Each Sorbent 8-10 mesh Tyler
Flow Gas Composition 2.3% C>2
13.0% C02
120 ppm H20
110 ppm S02
3 ppm K2S04 vapor
Balance N2
Gas Flow Rate 8.85 L/min at
room temperature
Linear Gas Velocity 27 cm/s at 850°C
Gas Hourly Space Velocity (GHSV) 62,000 IT1 at 850° C
Duration of experiment 24 h
-------
62
Table 15. Material Balances of K2S04 from Tests of K2S04
Vapor Capture by Diatomaceous Earth and Activated
Bauxite
mg
Expt. KSO-4, Expt. KSO-5,
Activated Diatomaceous
Bauxite Earth
Input
(1) K2S04 Vaporized 346 264
Output
(2) K2S04 Collected from
(a) Combustion Tube 8 23
(b) Downstream Line and
Condenser 2 3
(3) K2SC>4 Captured by Sorbent3 322 191
Total 332 217
(4) K2SC>4 Loss (by difference) 14 47
(5) % K2SC>4 Capture
[(3)/(l) x 100] 93.1 72.3
aPotassium concentration of the sorbent was obtained by dissolving
two representative samples each of activated bauxite and diatomaceous
earth in NH4HF2 and a mixture H2SC>4, HF, and HN03, respectively, and
then analyzing the resulting solutions by atomic absorption. AA
analyses were done by R. Bane.
As shown in Table 15, K2SC>4 vapor captures by activated bauxite and
diatomaceous earth were 93.1 and 72.3%, respectively, during the 24-h exper-
iments. The sorbents were tested at a fairly high gas hourly space velocity
of 62,000 h~l (or a contact time of about 0.06 s) and at a fairly low K2S04
concentration in the flue gas as shown in Table 14. In a granular-bed filter
operated with a PFBC, an increase in the percent K2SC>4 vapor capture by the
sorbent can be expected to occur at longer contact times of K2S04 vapor with
the sorbent bed and at the higher K2SC>4 vapor concentrations. The capabilities
of these sorbents for removing K2SC-4 vaPor from hot flue gas is comparable to
their capabilities for removing alkali metal chloride vapors (demonstrated
earlier). Since K2SC>4 and Na2S04 have similar chemical properties, one would
expect that retention of Na2SC>4 vapor by these two sorbents would be similar
to K2S04 vapor retention.
-------
63
When activated bauxite and diatomaceous earth were tested with NaCl
vapor, results obtained from water-leaching the sorbent indicated that NaCl
vapor is adsorbed by activated bauxite in the form of NaCl and that the NaCl
vapor reacts with diatomaceous earth to form products that are not soluble in
water.
To gain an understanding of the reactions of K^SO^. vapor with the
two sorbents at the experimental conditions, after the K2S04 experiments, the
two sorbents were each leached with distilled water using magnetic stirring at
95°C for 1 h. Table 16 shows the distribution of water-soluble and water-
insoluble potassium species in both activated bauxite and diatomaceous earth
sorbents.
Table 16. Distribution of Potassiuum Ion in the Sorbents
K+, g-mol x 103
Activated Diatomaceous
Bauxite Earth
Water-Soluble 0.61 (16.5%) 0.06 (2.7%)
Water-Insoluble3 3.09 (83.5%) 2.14 (97.3%)
Totalb 3.70 2.20
aObtained by the difference of total and water-soluble.
^Obtained by the analysis of K+ in the sorbent.
Table 16 indicates that in contrast to the water-soluble NaCl
deposited on activated bauxite, products formed upon the reaction of
vapor with activated bauxite are only slightly soluble in water.
The products of the K2S04 vapor reactions with diatomaceous earth
are essentially water-insoluble. At the end of the experiment, significant
amounts of diatomaceous earth particles were observed to be coated with a
layer of transparent, vitreous material, which was not observed on the activated
bauxite sorbent. The X-ray diffraction analysis* of the vitreous material
showed that it was an amorphous substance and was not melted I^SCy,.. This
indicates that during the experiments, the K2S04 melt in the sample did not
creep and that apparently, K2S04 vapor reacted with diatomaceous earth to
form glassy potassium silicates. The silicates so formed may be in chemical
forms that dissolve very slowly in water or they may further combine with
impurities such as oxides of aluminum, iron, calcium, and magnesium to form
water-insoluble products.
Done by B. Tani
-------
64
b. Parametric Tests
The experimental results presented in the previous screening tests
clearly indicate the effectiveness of both activated bauxite and diatomaceous
earth sorbents in removing vapors of NaCl, KC1, and K2SC>4 from simulated PFBC
flue gas. Consequently, systematic studies were initiated to investigate the
effects of some operating variables in the sorption performance of these two
materials. In this reporting period, the effect of sorbent-bed temperature,
superficial gas velocity, and gas hourly space velocity (GHSV) of the flue gas
were studied. Results of these studies are presented below.
Effect of Sorbent-Bed Temperature. Sorption capacities of diato-
maceous earth and activated bauxite were measured at average bed temperatures
of 800 and 880°C as a function of experiment duration. Tests were conducted
at atmospheric pressure, using a simulated PFBC flue gas (3% 02, 16% CC>2,
180 ppm H20, 300 ppm SC>2 and the balance N2) containing 69 to 98 ppm
NaCl vapor. The superficial gas velocity was 25 cm/s and the gas hourly space
velocity was 67,000 h~ - The quantities of NaCl captured by sorbents and
condensed on the cold traps were reported earlier (ANL/CEN/FE-78-4). Calcu-
lated sorption capacities (in mg NaCl/g sorbent) are plotted in Fig. 19.
50
m
g 40
CO
-A- 1073 K
-O 1153 K
DIATOMACEOUS
EARTH
30
en
s
UJ
cc
CL
<
o
20
10
1234
EXPERIMENT DURATION, h
Fig. 19. Effect of Sorbent Bed Temperature on Sorption
Capacity as a Function of Experiment Duration
-------
65
As can be seen in Fig. 19, the sorption capacity of diatomaceous
earth is substantially greater than that of activated bauxite on a weight
basis (i.e., mg NaCl/g sorbent). This is due to the activated bauxite being
denser than diatomaceous earth. For all of these experiments, a 1.3-cm-thick
bed was packed; therefore, on a volume basis (i.e., mg NaCl/mL sorbent) the
sorption capacity of activated bauxite is greater than that of diatomaceous
earth.
Figure 19 also shows that the sorption capacity of diatomaceous
earth increases with sorbent-bed temperature. This indicates that the reaction
between NaCl vapor and diatomaceous earth is endothermic. In contrast to
diatomaceous earth, the sorption capacities for activated bauxite decrease
with increasing sorbent-bed temperature. Since a physical adsorption process
is always exothermic, the amount of adsorbate sorbed on an adsorbent must
always decrease with increasing temperature according to the principle of
Le Chatelier; therefore, the present observation substantiates the conclusion
(made above) that NaCl vapor is essentially captured by activated bauxite
through a physical adsorption process. Figure 19 also shows that the sorption
capacity of both diatomaceous earth and activated bauxite increase nonlinearly
with time. This suggests that under the experimental conditions, neither the
reaction rate between NaCl and diatomaceous earth nor the rate of adsorption
of NaCl on activated bauxite is controlled by the rate of mass transfer of NaCl
vapor from the bulk flue gas to the external surface of the sorbent.27
Effect of Superficial Gas Velocity of Flue Gas at Constant GHSV. In
gas-solid reactions, as in the system considered in the present study, the
effect of superficial gas velocity of flue gas passing through the sorbent bed
on the sorption rate of NaCl vapor provides another means of measuring the
role played by the mass transfer of NaCl vapor from the bulk flue gas to the
external surface of the sorbent. For mass-transfer-controlled reactions, an
increase in superficial gas velocity increases the mass transfer coefficient,
thereby increasing the extent of reaction. The superficial gas velocity is
also an important parameter for designing a sorber vessel because it determines
the bed cross-sectional area for a given volumetric flow rate of flue gas to
be treated.
Based on these reasons, a series of tests has been conducted to
study the effect of superficial gas velocity on the sorption capacity of both
diatomaceous earth and activated bauxite. In this series of experiments, -8 +10
mesh sorbent was tested at 800°C and atmospheric pressure in a simulated
PFBC flue gas (3% 02, 16% C02, 180 ppm H20, 300 ppm S02, and the balance N2) at
superficial gas velocities of 25, 66, and 155 cm/s, and at a constant GHSV of
67,000 h~l. The NaCl vapor concentration in the flue gas ranged from 69 to 98
ppm. Increases in the superficial gas velocity of the flue gas at the sorbent
bed were accomplished by reducing the cross section of the sorbent bed, but
the amount of sorbent packed was kept constant to maintain a constant GHSV-
For each superficial gas velocity, the sorbent was tested at durations of both
2 and 3 h as a double-check.
Tables 17 and 18 show NaCl vapor capture by diatomaceous earth and
activated bauxite, respectively, as a function of superficial gas velocity.
Calculated average rate of NaCl capture (in mg NaCl/g sorbent) [row (4) in both
Tables 17 and 18 divided by experiment duration] are plotted as a function of
superficial gas velocity in Fig. 20.
-------
Table 17. Sodium Chloride Distributions from Tests of NaCl-Vapor Capture by
Diatomaceous Earth as a Function of Superficial Gas Velocity of Flue Gas
Sorbent (15.0 g) was tested at 800°C and atmospheric pressure in a
simulated PFBC dry flue gas at GHSV = 67,000 h"1
Expt.
HGC-23
Expt.
HGC-24
Expt.
HGC-37
Expt.
HGC-38
Expt .
HGC-42
Expt.
HGC-41R
Duration of
Experiment, h
Superficial Gas
Velocity, cm/s
(1) NaCl Collected by
(a) Cold Traps
(b) Glass-wool Filter
(2) NaCl Captured by
Sorbent3
(3) Total
(4) Sorption Capacity,
mg NaCl/g sorbent
I(2)/15]
25
42
9
251
302
16.7
25
56
31
343
430
22.9
66
NaCl,
29
14
328
371
21.9
(17.8)b
66
mg
52
28
422
502
28.1
(24.1)b
155
29
18
332
379
22.1
(17.6)b
155
82
52
397
531
26.5
(21.4)b
aSodium concentration in the sorbent was obtained by dissolving representative samples of the
sorbent in a mixture of H^SCv,., HF, and HNC>3 and then analyzing the solution by flame emission
spectrometry (FE). FE analyses were done by R. Bane.
bTo allow comparison, these are values which would be obtained if the NaCl vapor were
transported as in the 25 cm/s-series of experiments. For example, 17.8 was obtained from
302/
328
V371
-------
Table 18. Sodium Chloride Distributions from Tests of NaCl-Vapor Capture by
Activated Bauxite as a Function of Superficial Gas Velocity of Flue Gas
Sorbent (30.0 g) was tested at 800°C and atmospheric pressure in a
simulated PFBC dry flue gas at GHSV = 67,000 h"1
Expt. Expt. Expt. Expt. Expt. Expt.
HGC-26 HGC-27 HGC-40 HGC-39 HGC-43R HGC-44
Duration of
Experiment, h
Superficial Gas
Velocity, cm/s
25
25
66 66
NaCl, mg
155
155
(1)
(2)
(3)
(4)
NaCl Collected by
(a) Cold Traps
(b) Glass-wool Filter
NaCl Captured by
Sorbent3
Total
Sorption Capacity
mg NaCl/g-sorbent
t(2)/30]
22
10
343
375
11.4
40
22
444
506
14.8
42
25
281
348
9.4
(10.1)b
62
33
430
525
14.3
(13.8)b
33
20
351
404
11.7
(10.9)b
47
27
498
582
16
(14
.6
.4)b
aSodium concentration in the sorbent was obtained by fusing representative samples of the sorbent
with NH4HF2, dissolving the samples in distilled water, and then analyzing the solutions by flame
emission spectrometry (FE). FE analyses were done by R. Bane.
bTo allow comparison, these are the values which would be obtained if the NaCl vapor were transported
as in the 25 cm/s-series of experiments. For example, 10.1 was obtained from - 281
-------
68
UJ
CO
tr
o. l5
o
o
z
o>
e. 10
u
H-
0.
o
5
o
o
Z
u.
o
H 0
1 1
-O-
-A-
1
1
2- HOUR RUN
3-HOUR RUN
—
— DIATOMACEOUS EARTH —
o r A
A —
O
A
..ACTIVATED BAUXITE
O Jr
— A ^
~
u
A
^^- DIATOMACEOUS EARTH*
^^^1 — ACTIVATED BAUXITE*
/^
1 1
< 0 40
tr
1
1
80 120 160
SUPERFICIAL VELOCITY, cm/s
Fig. 20. Effect of Superficial Gas Velocity on Sorption
Capacity at 800°C. *Data were calculated from
the results of a previous set of experiments
(HGC-4 and HGC-8 in ANL/CEN/FE-77-11). These
two values were revised to conform to the amounts
of NaCl that would be transported in the 25 cm/s-
series of experiments. In these two experiments,
-8 +10 mesh sorbent was tested at 870°C and
atmospheric pressure and in an air flow at GHSV =
3500 h"1-
Within the limits of experimental and analytical error, the sorption
capacity is not affected by superficial gas velocity in any 2- or 3-h run, as
shown in Fig. 20. These results verify that under the experimental conditions,
the reaction between diatomaceous earth and NaCl vapor and the adsorption of
NaCl vapor on activated bauxite are not controlled by the mass transfer of
NaCl vapor from the bulk flue gas to the external surface of the sorbent. The
sorption observed for these sorbents is controlled by the rate of diffusion of
NaCl vapor through internal pores, the rate of adsorption of NaCl vapor on the
active sites of the sorbent, or the chemical kinetics (in the case of diatoma-
ceous earth).
Also included in Fig. 20 are the rates of NaCl capture for both of
the sorbents, calculated from the results of a previous set of experiments
(HGC-4 and HGC-8 in ANL/CEN/FE-77-11) revised to correspond to the amounts of
NaCl vapor that would be transported in a 25 cm/s series of experiments.
-------
69
In these two experiments, the two -8 +10 mesh sorbents were tested at 870°C and
atmospheric pressure and with air flowing at a superficial gas velocity of 7.5
cm/s and GHSV = 3500. Because the sorbents were not tested under experimental
conditions quantitatively comparable to those of high superficial gas velocity,
the results are included in the figure for qualitative comparison only. It is
seen that the rate is substantially smaller at a superficial gas velocity of
7.5 cm/s than at high superficial gas velocities. This indicates that at such
a low superficial gas velocity, sorption of NaCl by these sorbents was probably
mass-transfer controlled.
Effect of GHSV of Flue Gas at Constant Superficial Gas Velocity.
The GHSV is defined as the volumetric flow rate of flue gas in units of
sorbent volumes per hour. The reciprocal of the GHSV, which is space time, is
related to the contact time of flue gas with the sorbent bed. Thus, the GHSV
determines the sorbent volume required for a given volumetric flow rate of
flue gas to be treated.
A series of tests was conducted in which -8 +10 mesh sorbent was
tested at 800°C and atmospheric pressure using a simulated PFBC flue gas (3%
02, 16% C02, 180 ppm ^0, 300 ppm SC>2 , and balance N2) at a superficial gas
velocity of 66 cm/s. The average concentration of the NaCl vapor in the flue
gas was 80 ppm. The tests were each run for 2 h. All experiments were carried
out at a constant flow rate of flue gas; therefore, GHSV was varied by
varying the volume of the sorbent bed. Tables 19 and 20 show NaCl vapor
capture by diatomaceous earth and activated bauxite, respectively, as a
function of GHSV- Figure 21 is a plot of results given in Tables 19 and
20.
Figure 21 shows that, as expected, increased contact time of the
flue gas with the sorbent increases the NaCl-vapor capture for both diatomaceous
earth and activated bauxite sorbents. For Experiment HGC-37, the percent NaCl
captured appears to be high. This experiment should be repeated to check this
result. The results of this series of experiments indicate that at fairly
short contact times of flue gas with the sorbent bed (about 0.2 s) and at 2-h
times of experiments, 98.5% and 97% removal of NaCl vapor from hot flue gas
can be achieved by activated bauxite and diatomaceous earth, respectively.
In addition to the sorbent-bed temperature, the superficial gas
velocity, and the gas hourly space velocity of the flue gas, the sorption
ability of a sorbent can be affected by several other variables. Among them
are pressure, alkali metal compound vapor concentration, moisture content of
the flue gas, and particle size of the sorbent. Studies on the effects of
these variables on the sorption capability of activated bauxite and diatomaceous
earth will be continued. The regenerability of activated bauxite by water
leaching and desorption processes will also be explored. Finally, an economic
evaluation will be done using these materials as granular-bed sorbents for
removing alkali-metal compounds from PFBC hot combustion gas.
-------
70
Table 19. Sodium Chloride Distributions from Tests of NaCl-
Vapor Capture by Diatomaceous Earth as a Function
of Gas Hourly Space Velocity (GHSV)
Sorbent was tested at 800°C and atmospheric pressure
in a simulated PFBC dry flue gas at a superficial gas
velocity of 66 cm/s. Duration of each experiment was
2 h.
GHSV, h"1
( Contact Time, s)
Expt .
HGC-37
67,000
(0.05)
Expt .
HGC-47
33,500
(0.11)
Expt.
HGC-52R
18,600
(0.19)
(1) NaCl Collected by
(a) Cold Traps
(b) Glass-Wool Filter
(2) NaCl Captured by Sorbent3
(3) Total
(4) % NaCl Capture by Sorbent,
[(2)/(3)] x 100
29
14
328
371
88.4
NaCl, mg
29
18
292
339
86.1
6
4
318
328
97.0
aSodium concentration in the sorbent was obtained by dissolving repre-
sentative samples of the sorbent in a mixture of ^804, HF, and HNC>3, and
then analyzing the solution by flame emission spectrometry (FE). FE
analyses were done by R. Bane.
-------
71
Table 20. Sodium Chloride Distributions from Tests of
NaCl-Vapor Capture by Activated Bauxite as
a Function of Gas Hourly Space Velocity (GHSV)
Sorbent was tested at 800°C and atmospheric
pressure in a simulated PFBC dry flue gas at
a superficial gas velocity of 66 cm/s. Duration
of each experiment was 2 h.
GHSV, h"1
( Contact Time, s)
(1)
(2)
(3)
(4)
NaCl Collected by
(a) Cold Traps
(b) Glass-Wool Filter
NaCl Captured by Sorbent3
Total
% NaCl Capture by Sorbent,
[(2)/(3)] x 100
Expt .
HGC-40
67,000
(0.05)
42
25
281
348
80.7
Expt.
HGC-50
33,500
(0.11)
NaCl, mg
19
11
348
378
92.1
Expt.
HGC-51
18,600
(0.19)
3
2
330
335
98.5
aSodium concentration in the sorbent was obtained by fusing representative
samples of the sorbent with NH4HF2, dissolving the samples in slightly
acidified distilled water, and then analyzing the solutions by flame
emission spectrometry (FE). FE analyses were done by R. Bane.
100
- 90
o
u
§ 80
Q.
o 70
o
o
60
50
O ACTIVATED BAUXITE
A DIATOMACEOUS EARTH
I.I.I.
0 0.04 0.08 O.I 3
CONTACT TIME, s
0.16
0.20
Fig. 21. Effect of the Contact Time of Flue Gas
with Sorbent on NaCl Vapor Capture
-------
72
TASK C. TRACE-POLLUTANT CONTROL IN FBC
1. Trace-Element Behavior in Sorbent During Cyclic Utilization
(W. M. Swift, G. W. Smith, and F. G. Teats)
Results obtained at Argonne National Laboratory and reported pre-
viously^ >28.29 indicated that smaller quantities of trace elements (such as
Hg, F, and Br) would be emitted to the environment from fluidized-bed combustion
than would be expected from a conventional coal-fired boiler. This conclusion
was based on a comparison of (l) open mass balances made around the ANL
15.2-cm-dia pressurized fluidized-bed combustor with (2) similar mass balances
reported30~32 for conventional coal-fired boilers. The mass balances were
open in that trace element concentrations were determined only for solid feed
and effluent (and in the case of conventional boilers, liquid effluents) from
the combustion processes. Trace-element emissions to the environment were
assumed to be the percentages of the trace elements which had entered the
combustion process but were not accounted for in the analyzed effluents.
The previous ANL study failed to consider, however, the potential influence
of regeneration in a regenerative fluidized-bed combustion process on the
release of trace elements to the environment. The regeneration reactor operates
at temperatures as high as 'vLlOO °C, as compared with the ^900 °C operating
temperature of the fluid-bed combustor. Thus, concern is justified that trace
elements contained in the sulfated sorbent from the combustor might be released
by volatilizaton during regeneration. A study was made, therefore, (1) to
measure the levels of selected trace elements in the sorbent over several
cycles of sulfation and regeneration, and determine the tendency for those
elements to be enriched or depleted in the sorbent under these conditions, and
(2) to make a material balance (based on solid samples only) for a single
regeneration experiment, assessing changes in the trace element concentrations
in the sorbent for evidence of trace element losses by volatilization. A
summary of the experimental findings is presented here. A more detailed
account of the work has been presented in a quarterly report (ANL/CEN/FE-78-3)
issued during the past year.
a. Analytical
In a recently published proceedings of a workshop on the effects of
trace contaminants from coal combustion, selected trace elements present in
coal were classified in three broad categories of toxicity—i.e., high,
medium, and low potential, for both terrestrial and aquatic life. The classi-
fication represents a "best estimate" based on the experience, knowledge, and
intuition of the panel members.
In the previous trace-element investigation at ANL,^>28,29 mass
balances were made around the ANL 15.2-cm-dia PFBC for the following trace and
minor elements: As, Be, Br, Co, Cr, F, Fe, Hg, K, La, Mn, Na, Pb, and Sc. In
that investigation, the concentrations of Hg, Pb, Be, and F in particulate
samples were determined by wet chemical techniques; concentrations of the
remaining ten elements were determined by neutron activation analysis. Using
the classification referenced above, of the fourteen elements studied previously,
five (Be, Co, Cr, F, and Hg) are suspected of having a potentially high toxicity,
-------
73
one (Pb) a potentially medium toxicity, two (As and Mn) a potentially low
toxicity, and six (Br, Fe, K, La, Na, and Sc) are not listed.33 Of the last
six, the behavior of two elements (sodium and potassium) during PFBC is of
interest in relation to the corrosion of hardware downstream from the combustor.
In the selection of the analytical approach for the current study,
major consideration was given to selecting a method which would provide
informaton on most elements assessed as having potentially high-toxicity and
medium-toxicity effects on the environment. For this reason, spark-source
mass spectroscopy (SSMS) was selected as the most suitable instrumental method
of analysis.
The samples selected for analysis were sent to Oak Ridge National
Laboratory, where they were analyzed for trace element impurities by SSMS.
The results, except for twelve metals analyzed by stable isotope dilution,
were obtained by a visual inspection and plate sensitivity technique. For the
latter method, the reported results are considered accurate within a factor of
two; isotope dilution results are considered accurate within ±20% of the
reported value. A duplicate of a sample analyzed by the Oak Ridge Analytical
Chemistry Division was also sent to an independent commercial laboratory for
SSMS analysis to allow comparison of the analyses.
b. Experimental
To assess the fate of trace elements during a regenerative FBC
process, particulate samples of coal, ash, sorbent, and sulfated sorbent from
the 10-cycle combustion-regeneration experiment with Tymochtee dolomite
sorbent (ANL/CEN/FE-77-3, an earlier report in this series) were analyzed.
This experiment was performed with no fresh makeup sorbent added after the
first combustion half-cycle experiment. The nominal conditions for the combustion
half-cycles, performed with Tymochtee dolomite and Arkwright coal, were a 900°C
bed temperature, an 810-kPa system pressure, a 1.5 CaO/S mol ratio, ^17%
excess combustion air, a 0.9 m/s fluidizing-gas velocity, and a 0.9-m bed
height. The regeneration half-cycles were performed using Triangle coal at a
bed temperature of 1100°C, a system pressure of 158 kPa, and a fluidized-bed
height of a/46 cm. Table 21 lists and describes the samples analyzed.
c. Results
Verification of Analytical Results. Sample J-11270, sulfated
Tymochtee dolomite, was submitted to both Oak Ridge National Laboratory (ORNL)
and an independent commercial laboratory for comparative SSMS analyses of
trace elements. Of the analyses for 43 elements by ORNL, 33 results were in
agreement within stated analytical accuracies with the results reported by the
commercial laboratory. Although not a conclusive demonstration of the quanti-
tative accuracy of the ORNL results, this agreement is.a good indication that
the ORNL results are sufficiently reliable to make qualitative observations
regarding the behavior of trace elements in a regenerative PFBC system. The
ORNL results were used.
-------
74
Table 21. Samples Analyzed in the Trace Element Study
Sample
No.
Exp.
No.
Sample
Description
J-11270 REC-1K Steady-state overflow sample of sulfated
Tymochtee dolomite from first combustion
cycle of 10-cycle combustion-regeneration
experiment
J-11177 REC-1K Same as for J-11270
J-11178 REC-10 Same as for J-11270 except that the dolomite
is from the tenth combustion cycle
J-11179 CCS-10 Sulfated dolomite feed to the tenth regen-
eration cycle
J-11180 CCS-10 Steady-state sample of regenerated dolomite
from the tenth regeneration cycle
J-11181 CCS-10 Steady-state sample of cyclone ash from the
tenth regeneration cycle
J-11182 CCS-10 Sample of Triangle coal used in the tenth
regeneration cycle
J-11183 CCS-10 Steady-state sample of ash that escaped from
the cyclone and collected on a membrane
filter during the tenth regeneration cycle
Trace Elements in Sulfated Tymochtee Dolomite Samples from the First
and Tenth Combustion Cycles. Table 22 presents trace-element SSMS analyses
for (1) samples of Tymochtee dolomite which had been sulfated only once and
(2) samples which had been through nine sulfation-regeneration cycles and a
tenth sulfation. Each of the samples in a pair (J-11270 and J-11177; J-11178
and J-11179) is expected to have similar compositions. The members of a pair
are not, however, fractions of the same sample. The results for samples
J-11178 and J-11179 are in excellent agreement. The agreement in results for
samples J-11270 and J-11177 is less impressive, but is still considered quite
good for nonduplicate samples.
The stability of trace elements in the sorbent during repeated
utilization cycles was assessed by comparing the average concentrations of the
elements in the samples of sulfated dolomite from the tenth combustion cycle
with their average concentrations in the samples of sulfated dolomite from the
first combustion cycle. In the comparisons, the average concentrations of the
elements in the tenth-cycle-sulfated dolomite (Table 22) have each been
-------
75
Table 22. Trace-Element Analyses for Tymochtee Dolomite Samples from the
First and Tenth Combustion Cycles. (Concentrations are in ppm
except where indicated otherwise.)3
Sulfated Dolomite from
First Combustion Cycle
Element
High-Toxicity
Be
B
Cd
Co
Cr
Cu
F
Ni
Tl
V
Med ium-Toxic i ty
Bi
Mo
Pb
Se
Ti
Zn
Low-Toxicity
As
Mn
Toxic ity Not
Assessed
Al
Ba
Br
Ca
Cs
Fe
La
Ce
In
Pr
K
Li
Sample
J-11270
<;5
70
2 ± 0.4
<5
20 ± 4
30 ± 6
100
100
<1
3
<1
<3
4 ± 0.8
<10
200
25+5
10
60
2%
50 ± 10
10
20%
<1
5000 ± 1000
10
5
0.6 ± 0.1
10
0.5%
2
Sample
J-11177
<20
100
4.5 ± 0.9
<10
17 ± 3
25+5
240 ± 50
0.1 ± 0.1
20
<1
30 ± 6
10 ± 2
<1
1000
50 ± 10
2
70
1%
52 ± 10
20%
3600 ± 700
0.5 ± 0.3
0.6%
Sulfated Dolomite From
Tenth Combustion Cycle
Sample
J-11178
<20
100
2.6 ± 0.6
20
100 ± 20
33 ± 7
250 ± 50
<1
70
<1
40+8
10 ± 2
<1
1500
110 ± 20
7
70
2%
180 ± 40
20%
1.5 ± 0.3%
2 ± 0.4
1%
Sample
J-11179
<20
200
1.8+ 0.4
20
92 ± 18
30 ± 6
350 ± 70
1.3 ± 0.2
100
<1
45 ± 9
9+2
<1
1500
100 ± 20
10
100
2%
140 ± 30
20%
2 ± 0.4%
1.5 ± 0.3
0.6%
(Cont'd.)
-------
76
Table 22. (Cont'd.)
Element
Sulfated Dolomite from
First Combustion Cycle
Sample Sample
J-11270 J-11177
Sulfated Dolomite from
Tenth Combustion Cycle
Sample Sample
J-11178 J-11179
Toxicity Not
Assessed
(Cont'd.)
Mg
Na
Nb
P
Nd
Sm
Rb
S
Si
Sr
Th
U
Y
10%
0.2%
s 3
10
10
<5
20
10%
2%
210 ± 40
<5
^5
2
10%
0.5%
5
40
30
6%
^2%
190 ± 40
___
10%
1%
6
100
20
7%
^7%
450 ± 90
15%
1%
5
100
25
7%
^7%
450 ± 90
aFactor-of-two accuracy except where indicated otherwise.
multiplied by 1.25 to adjust the concentrations of the ash-encrusted stone to
the equivalent amount of virgin limestone in the once-sulfated stone. The
sulfated dolomite from the tenth combustion cycle was encrusted with approxi-
mately 13 g of coal ash per 100 g of virgin dolomite sorbent.
Based on the analyses, it appears that the following elements were
enriched in the ten combustion cycles: Cr, Tl, V, Mo, Zn, Ba, Fe, In, and
Sr. The term "enriched" is applied here to an element which has a higher
concentration in the sulfated dolomite from the tenth combustion cycle than
in the sulfated dolomite from the first combustion cycle (consideration is
given to the estimated accuracy of the analytical method for each element).
The enrichment of these elements is undoubtedly related to the buildup of
coal ash on the surface of the dolomite particles during repeated utilization
cycles. Of the 31 elements for which samples from the first and tenth
cycles were analyzed, none appeared to have been depleted during the ten
cycles. Elements such as As, Cd, Cr, Pb, Ni, Se, Tl, and Zn, which have all
exhibited a tendency to be volatile during coal combustion in conventional
coal fired systems,-^ were either enriched or remained in the dolomite in
approximately the same concentration during the ten combustion cycles.
-------
77
Open Mass Balances Around the Regeneration Reactor During the
Tenth Cycle Regeneration Step. Samples of the coal feed, the sulfated
limestone feed to the regenerator, the regenerated sorbent, and the cyclone
fly ash plus the fly ash that passed through the cyclones in the tenth cycle
regeneration step were analyzed for trace elements to make mass balances.
The material balances were open (the flue gas was not analyzed for volatile
species) but can be used to indicate the distribution of trace elements in
the various solids streams leaving the regenerator, as well as the potential
loss of trace elements by volatilization.
Table 23 presents (1) the results reported by Oak Ridge for
analysis of the samples by SSMS and (2) material balances based on those
analyses and on the mass flow rates of the various solids streams.
Comparison of the concentration values in columns 3 and 4 of Table
23 (for the sulfated and regenerated sorbent) shows the stability of the 31
elements in the dolomite. Only the change in the concentration of nickel is
greater than the estimated limits of analytical accuracy for the two samples;
depletion of the nickel in the sorbent during regeneration is indicated.
This result is not readily reconciled with the previous observation (based
on Table 22 results), that concentrations of nickel in sorbent from the
first and tenth combustion cycles were not significantly different, and may
simply reflect analytical errors.
The open material balances for the 31 elements (Table 23) range
from a low of ^39% for sulfur (does not include the S02 liberated during
regeneration) to a high of ^130% for cadmium and manganese. The apparent
depletion of nickel in the sorbent during regeneration resulted in a material
balance for this element of only 48%. The low material balance for titanium
(^57%) results from an unexplained loss of titanium in the system, as
indicated by the low level of titanium in the ash from the cyclones and
filter. The remaining material balances generally ranged from 70 to 130%.
Due to the scatter, quantitative assessments of the behavior of the elements
during regeneration are difficult. However, from the stability of the
elements in the dolomite, it seems safe to conclude that the major influence
of regeneration on trace-element emissions from an FBC system would result
from the additional coal used as reductant in the regeneration step. Since
this represents only a small fraction of the total coal requirement (^3%) of
the FBC, the impact of regeneration on FBC emissions should be minor.
In conclusion, for the trace elements studied in the investigation,
regeneration does not appear to result in the release of trace elements from
the sorbent. A comparison of sulfated samples from the first and tenth
combustion half-cycles actually indicated an increased concentration (in the
dolomite) for 9 of the 31 elements studied. This is attributed to the
buildup of coal ash on the sorbent over repeated utilization cycles. During
a regeneration experiment, nickel was the only element which gave evidence
of being depleted somewhat in the sorbent. However, analytical uncertainties
prevent placing too much emphasis on this observation.
-------
78
Table 23. Concentrations and Material Balances for Trace Elements in Steady-
State Samples of Solids Entering and Leaving the Regeneration
Reactor during the Tenth Regeneration Cycle
Concentration3 (in ppm except where
Coal Sulfated
Element Feed Sorbent
noted otherwise)
Regenerated Cyclone Fly Ash
Sorbent + Filter Fly Ash
Material
Balance, %,
(Out/In)xlOO
High-Toxicity
Be
B
Cdb
Co
Crb
Cub
Nib
Tlb
V
<10
10
0.5
7
17
80
30
0.2
30
<20
200
1.8
20
92
30
350
1.3
100
<20
200
2.6
20
89
30
180
1.7
120
<20
200
5.5
20
220
110
300
1.4
40
—
91
130
87
92
73
48
120
100
Medium- Toxici ty
Bi
Mob
Pbb
Seb
Ti
Znb
Low- Toxic ity
As
Mn
Toxicity Not
Al
Ba
Ca
Feb
Inb
K
Mg
Na
Nb
P
Rb
S
Si
Srb
<1
6
10
<1
500
35
0.5
5
Assessed
2%
94
0.2%
0.32%
0.2
1%
0.05%
0.05%
2
500
10
0.1%
>2%
150
<1
45
9
<1
1500
100
10
100
2%
140
20%
2%
1.5
0.6%
15%
1%
5
100
25
7%
>7%
450
<1
40
6
<1
1000
130
10
150
1.5%
130
22%
2.2%
1
1%
15%
1%
5
200
30
3%
>4%
400
<1
28
130
<1
800
170
5
30
>2%
350
10%
1.6%
1.1
1.5%
3%
1%
3
1000
50
3%
>5%
450
78
110
—
57
110
88
130
>60
83
98
97
60
120
88
91
84
120
110
39
_ —
81
aAccuracy estimated to be a factor of 2, except for elements for which other
accuracy is indicated.
bAccuracy estimated to be +20% of stated value.
-------
79
TASK D. PARTICULATE CONTROL STUDIES
1 . Evaluation of On-Line Light-Scattering Particle Analyzers
(J. Montagna, G. W. Smith, and F. G. Teats)
a. Introduction
In the development of pressurized f luidized-bed combustion (PFBC)
systems, an on-line light-scattering particle analyzer for the flue gas could
provide continuous particle size and loading analysis without disturbing the
off-gas stream. These measurements would be used (l) in measuring the effi-
ciency of upstream particulate-removing devices (cyclones and filters),
(2) for establishing gas turbine performance at different particulate loadings,
and (3) for protecting turbines or a test cascade in the event of a sudden
system upset.
In a PFBC system, the flue gas will be at ^900°C and ^1000 kPa
atm) between the boiler and the turbine. In the absence of an on-line particle
analyzer, routine batch sampling of the hot off-gas (using inertial impactors)
will be necessary. Batch sampling from a pressurized hot off-gas environment
is difficult; another disadvantage is the long time lag between sampling and
analysis of samples for this technique.
Two types of optical analyzers which have not yet been evaluated for
f luidized-bed combustion applications were evaluated for characterizing the
particles in a flue-gas stream that is representative of a stream that would
be fed to a gas turbine. Both instruments use laser light beams. One is a
single-particle analyzer (a particle morphokinetometer , developed by Spectron
Development Laboratory) which characterizes the scattered light from each
particle (laser interferometer method), and the other is a Microtrac multi-
particle analyzer developed by Leeds and Northrup, which characterizes the
entire particle distribution in its optical path.
Single-Particle Analyzer. The principles of the Spectron single-
particle analyzer are discussed by Farmer. 35 The experimental results obtained
with this instrument and an evaluation of its performance were presented in
the previous annual report (ANL/CEN/FE-77-3) and Reference 36.
Multiparticle Analyzer. The experimental results obtained using the
multiparticle analyzer are reported here in summary form. Details of the
study are available in another report in this series (ANL/CEN/FE-77-8) and in
a topical report-^ published during the past year.
The Leeds and Northrup multiparticle analyzer used in this work
utilizes a new measurement technique in which three measurements of the
scattered light from a single laser beam directed into the off-gas duct are
made. These measurements are used to determine the mean diameter and variance
of the distribution for particles in the 1-50 ym range. For particles with a
unimodal size distribution, the size distribution can be obtained by examining
the total scattering intensity as a function of scattering angle and laser
beam polarization and by comparing the experimental data with theoretical
calculations for assumed particle-size distributions .3' One measured signal
-------
80
is also proportional to the volume of the particles illuminated, making
possible the calculation of the concentration of particles (loading) in the
fluid stream.
b. System and Procedure for Flue-Gas Particle Measurements
The particle-size measurements were made in the flue-gas system of
the ANL fluidized-bed combustion process development unit (PDU). Pairs of
optical windows installed upstream from the primary cyclone and near the
system outlet made it possible to size (1) the coarse entrained particles in
the flue gas from the combustor, (2) the particles in the flue gas leaving the
secondary cyclone, and (3) the very small particles in the flue gas leaving
the sintered-metal filter assembly. The coarse particles leaving the combustor,
previously determined to have a multimodal distribution, were not sized with
the multiparticle analyzer, which is only suitable for unimodal distributions.
Particle-size measurements were obtained (1) using the on-line
optical particle analyzer, (2) using an Andersen cascade impactor on gas
samples extracted through sampling ports downstream from the optical window
locations, and (3) using sieve and Coulter counter analysis of particulate
samples obtained from cyclones and test filters. Particle-loading measurements
were made by passing through glass fiber membrane filters the gas samples
extracted through sampling ports.
The Coulter counter was calibrated with standard particles, and its
measurements of particles in the flue gas agreed well with cascade-impactor
measurements. Thus, it was assumed that these two measurements were represent-
ative of the true particle size distributions. The multiparticle analyzer was
evaluated only by comparing its measurements with cascade impactor measurements.
c. Experimental Evaluation of the Multiparticle Analyzer
Measurement of Mean Particle Size. Table 24 lists the measured
particle diameters and loadings for several experiments. Because less than
10% of the particles were smaller than 1 ym (mass basis), the impactor distri-
butions used for calculating the aerodynamic log-mean diameter,(DAE) and
loadings were not truncated to accommodate the range of the multiparticle
analyzer (1-20 ym dia). When the loading (gravimetrically obtained) was
relatively high (>0.1 g/m3 or >0.044 grain/scf) in coal combustion experiments,
the (DAI?) obtained with the impactor ranged from 2 to 3.5 ym and the
projected area mean diameter obtained with the multiparticle analyzer (DA)
ranged from 2.2 to 3.4 ym for all suspended particle measurements. The volume
mean diameters (Dy), also obtained with the multiparticle analyzer, ranged
from 2.9 to 5.5 ym for the same experiments.
In "cold" experiment LN-9-5, only elutriated virgin limestone
particles were measured. The loading obtained with the Andersen impactor was
also relatively high (>0.1 g/m3); DA was 2.7 ym, in comparison with 1.5 ym for
DAE- It ha^ been expected that the measurements made with the various instru-
ments on the suspended virgin limestone would agree better than those made on
the mixture of particles during combustion experiments (LN-4, LN-5, LN-6).
However, the agreement of DAE with DA for suspended particles from the com-
bustion experiments was slightly better than for the measurements in the
cold experiments under high-loading (>0.1 g/m3) conditions.
-------
81
Table 24. Particle Sizes and Loadings Obtained with the
Multiparticle Analyzer, the Andersen Impactor,
and the Membrane Filter in Coal Combustion and
Cold Elutriation Experiments.
FBC Combus tor Conditions
Pressure 308 kPa (3 atm)
Temperature 855°C
Fluidizing Gas Velocity 1 m/s
Sorbent Greer limestone
Coal Sewickley
Multiparticle Analyzer
Exp.
LN-4-16
LN-4-2 f
LN-5-16
LN-5-2f
LN-6-16
LN-6-lf
LN-10-3f
LN-10-4f
LN-11-16
LN-11-26
LN-11-36
LN-9-3e»h
LN-9-26 >h
LN-9-5f'h
LN-9-4f»h
DV,a
ym
5.1
2.9
5.5
3.9
4.4
3.6
3.8
-
11.9±0.3
-
10.5±0.2
2.2410.36
-
3.77±0.23
—
5A,b
Vim
2.8
2.2
3.4
2.3
2.8
2.2
2.2
-
5.510.11
-
4.9710.12
2.0110.06
-
2.7210.07
—
Loading,
g/m->
0.43
0.83
0.18
1.0
0.46
0.64
1.5
0.37
0.12+0.01
0.13+0.01
0.1110.02
0.21+0.04
0.10+0.02
1.2010.06
0.52+0.17
Andersen
Impactor
DAE>d
ym
2.2
3.5
6.08
3.2
3.0
3.5
2.0
-
1.9
-
1.8
0.63
—
1.5
~
Loading,
g/m->
0.33
1.1
0.21
0.98
0.12
1.1
0.56
-
0.01
-
0.01
0.001
—
0.115
^
Membrane
Filter
Loading,
g/m-*
0.64
1.3
0.41
0.40
0.54
1.0
_
0.63
0.009
0.003
0.110
aVolume mean diameter
"Projected particle area mean diameter
cl g/m^ = 0.437 grain/scf
^Aerodynamic mean diameter
eDownstream from sintered-metal filters
^Between secondary cyclone and filter
SSuspect. Possibily, unexpected particle reentrainment in the flue gas occurred,
elutriation experiments—only virgin limestone particles were fed
-------
82
For particles leaving the cyclones (loadings X).l g/m^) , there was better
agreement between the mass log-mean diameter measurements made with the
multiparticle analyzer and the impactor than between those made with the
single-particle analyzer (only particles between 1.5 and 23.9 ym were compared)
and the Coulter counter. A loading of 0.05 g/m^ for particle diameters of
1-10 ym is presently proposed as the approximate maximum tolerable loading for
turbines. Because the particulate loading tolerances for turbines remain
uncertain, it is desirable to have optical instruments capable of measuring
particle loadings smaller than 0.05 g/m^.
The agreement of DAE with D^ was poorer for measurements made at
lower loading conditions «0.01 g/m^) in experiments LN-9 and LN-11 than_in
experiments at higher loadings. In these measurements at low loadings, D^ was
approximately three times as large as_D^E. The aerodynamic mean diameter,
I>AE> ranged from 0.63 to 1.9 ym, and D^ ranged from 2.0 to 5.5 ym. Since it
was difficult to obtain accurate impactor measurements for the comparison,
this agreement can be considered fair at this stage. Although, as previously
reported-^ particle size could be measured at low loadings with the multi-
particle analyzer, no particle measurements could be made with the single
particle analyzer at these low loadings because not enough particles passed
through the sample space of that instrument.
Measurement of Particulate Loadings. In these experiments (Table
24), particulate loadings (as well as mean particle sizes) were determined
with the multiparticle analyzer. Values obtained with the multiparticle
analyzer were compared with values obtained by three backup methods. The
amounts of particulate collected on a flue gas filter located downstream from
the multiparticle analyzer windows and the sampling port provided one backup
measurement. The other backup measurements utilized grab samples which
consisted of (1) the particulate collected in the impactor, including a backup
filter to the impactor, and (2) the material collected on a membrane filter.
For the measurements made during the combustion experiments (Table
24) in which the loadings were high (>0.1 g/m-*, gravimetrically obtained from
impactor or membrane filter samples), the most disparate gravimetric loading
value was one-fourth the loading measured with the multiparticle analyzer
(Table 24). (The loadings from the multiparticle analyzer are proportional to
the particulate density assumed, which in this work was 1.0 g/cm^.) For many
high loading measurements such as for LN-4-2, the gravimetric and multiparticle
analyzer loadings were very similar.
Under low loading conditions «0.01 g/m^), gravimetric loadings were
one order of magnitude lower than the optical multiparticle analyzer loadings.
The semicontinuous background loading signal from the multiparticle analyzer
was compared with the background plus sample loading signal and was found to
be only 4% greater than the background signal; this can account for the one-
order-of-magnitude difference between the multiparticle analyzer loadings and
the gravimetric loadings. These measurements at low loadings were made on
suspended particles that had passed through sintered metal filters in the
flue-gas system.
For the high-loading cold elutriation experiment (LN-9-4 and LN-
9-5, Table 24), loadings determined gravimetrically were one order of magnitude
smaller than those obtained in the high-loading hot experiments; for the
low-loading cold elutriation experiments (LN-9-2 and LN-9-3), the measured
-------
83
loadings were about two orders of magnitude smaller than in the low-loading
hot experiments. The large difference between the gravimetric and optical data
for experiment LN-9 may be due to an increase (over a period of one hour) in
the multiparticle analyzer background signal during this cold experiment.
The average gas linear velocity in the flue-gas duct near the
optical windows was ^30 m/s during the cold experiments; during the hot
combustion experiments, the linear gas velocity was only ^3 m/s. A greater
gas velocity creates more turbulence and a greater chance for particle buildup
in the free space between the windows and the stream of flue gas and thus a
higher background signal. The background signal was observed to decrease to
normal after the flow of limestone-laden air was discontinued.
d. Conclusions
The multiparticle analyzer can characterize a distribution by
providing the mean and a nongeometric variance of the distribution. However,
only if the distribution is known and unimodal can the instrument describe the
distribution of the particles. The distribution downstream from the cyclones
of a PFBC system is expected to be log-normal, based on measurements in this
and other studies. Therefore, particle distributions in this stream can be
obtained from multiparticle analyzer measurements.
In measurements with the multiparticle analyzer at high loadings
(>0.1 g/m^), the projected particle area mean diameters were generally less
than 25% smaller than the aerodynamic mass log-mean particle diameters determined
with the impactor. At low loading conditions «0.01 g/m^), the mean diameters
obtained with the multiparticle analyzer were approximately three times the
log-mean diameters obtained with the impactor.
At high loadings, the most disparate gravimetric loading was one-
fourth the loading measurement made with the multiparticle analyzer. However,
many measurements made by these two techniques at these relatively high
loadings were very similar. At low loadings (£0.01 g/m-') , the gravimetric
loadings were one order of magnitude lower than the values obtained with the
multiparticle analyzer. At these low loadings, the loading measurement signal
of the multiparticle analyzer is only 4% greater than the background; this is a
potential source of error. If the background signal can be reduced electron-
ically or via cleaner windows and/or optical path, on-line measurements with
this instrument to protect the gas turbines in PFBC systems will be very
promising.
The multiparticle analyzer is at a stage of development allowing it
to be used with little operator training. Compared to the single-particle
analyzer previously tested, the multiparticle analyzer is a more advanced
instrument and with some refinement can be useful now in the development of
PFBC technology.
-------
84
2. Particle Removal From Flue Gas
(W. Swift, G. Teats, A. R. Pumphrey, S. Smith, and J. Stockbar)
In pressurized fluidized-bed combustion, the hot flue gas from the
combustor must be expanded through a gas turbine to recover energy and
make the process economic. To prevent erosion of the turbine blades by
particulate matter entrained in the flue gas, the particulate loading must
be reduced to acceptably low levels.
Although the air-quality requirements for a gas turbine have yet to be
experimentally demonstrated, estimates of what constitute "acceptably low
levels" of particulate mass loading for a gas turbine generally range (depending
on the particle size distribution from ^0.05 to ^0.005 g/m^.^o If it is
assumed that the loading in the flue gas as it leaves the combustor is
of the order of 30 g/m^, the total particulate-removal efficiency required
to meet the estimated turbine air quality requirements would be between 99.8
and 99.98%.
Existing devices readily adaptable to high-temperature, high-pressure
particulate removal (e.g., conventional cyclones) are not very efficient in
removing particulate matter with diameters smaller than ^10 ym. Achieving the
"acceptable loading" necessary for PFBC requires, therefore, that highly
efficient methods be developed for removing from flue gas the particulate
solids having diameters between 2 and 10 ym.
An experimental program is continuing at ANL to test and evaluate promising
flue-gas cleaning methods in the off-gas system of the 15.2-cm-dia fluidized
bed combustor. Techniques under investigation are granular-bed filtration,
high-efficiency controlled-vortex cyclones, and acoustic agglomeration.
a. Granular-Bed Filters
Granular-bed filters currently under development can be generally
classified by the condition of the granular bed during filtration, i.e.,
fixed-bed, moving-bed, or fluidized-bed collectors. The concept of collection
being investigated at ANL is the use of fresh or sulfated limestone or dolomite
as granular-bed material in a fixed-bed collector with periodic bed replacement.
Use of sorbent from the combustion process as the granular-bed material
eliminates the need for "backflush" cleaning of the filter as in other fixed-bed
concepts or for external cleaning and recycle of the granular bed as in
moving-bed concepts. As is done with fixed granular beds that employ backflush
cleaning, several granular-bed modules would be operated in parallel. Peri-
odically, each module would be taken off line, and sorbent plus trapped
particulate matter would be replaced with virgin bed material. The sorbent
containing the trapped particulate matter could either be transferred to a
regenerator (if sulfated sorbent from the combustor had been used as the
filter medium) or disposed of.
To test the concept of using the sorbent material in a granular-
bed filter, a small test-filter was assembled and installed in the flue-gas
system of the ANL, 15.2-cm-dia fluidized-bed combustion PDU. The results of
tests made using the test filter have been reported, as well as calculated
-------
85
working ranges of parameters (ANL/CEN/FE-77-11 and ANL/CEN/FE-78-4). The
following represents a summary of the previous reports.
Equipment and Experimental Procedure. The granular-bed test chamber
is circular in cross section and can be modified to have an inside diameter of
either 7.8 or 15.4 cm. The fixed, horizontal granular-bed filter is supported
by a wire mesh screen. After the gas passes through the granular bed, the
gas, now cleaned, exits through a port in the side of the granular-bed filter
housing. Pressure taps located just above the top surface of the granular bed
and below the support screen measure the pressure drop across the bed. A
thermocouple just below the bed-support screen monitors the gas temperature
during each test.
In a test of the granular-bed filter, a preweighed amount of bed
material is charged into the granular-bed filter. While the fluidized-bed
combustion system is brought to steady-state operating conditions, the flue
gas (which has already passed through the primary and secondary cyclones),
bypasses the granular-bed filter test loop. Once steady operating conditions
are reached, the flow of flue gas is diverted through the granular-bed test
loop. In the test loop, the flow of gas is either passed through the granular-
bed filter or bypassed around the filter to a sample port where the flue gas
is sampled using an Andersen cascade impactor and/or a membrane filter to
determine particulate loading and/or size distribution.
During a granular-bed filtration cycle, the following quantities are
recorded: (1) the initial pressure drop, Ap^, across the granular-bed, (2) the
pressure drop across the filter as a function of time, (3) the time for the
pressure drop across the filter to equal AP^ plus 6.9 kPa (1.0 psi), (4) the
weight change of the granular bed during the filtration cycle, (5) the weight
of material recovered on a sintered-metal filter downstream from the granular
bed, and (6) the particulate sampler results.
Results. Experiments were performed to evaluate the effects of bed
particle size, gas velocity, and bed depth on filtration efficiency of the
granular bed. Virgin Tymochtee dolomite in two size ranges, -6 +14 mesh and
-14 +30 mesh, were used in the experiments reported here. The mean particle
diameters of the -6 +14 and -14 +30 mesh bed materials are ^2120 and 'W40 ym,
respectively.
In experiments to measure the effect of bed depth on filtration
efficiency for the -6 +14 mesh dolomite at a filtration velocity of 30 cm/s,
filtration efficiency increased from ^91% at a 5.1-cm bed depth to ^98% at a
40.6-cm bed depth. In terms of filtration per unit depth, however, the
filtraton efficiency decreased from 'x-18%/(cm of bed) to ^2.4%/(cm of bed) when
the bed depth was increased from 5.1 cm to 40.6 cm.
The effect of bed depth on filtration efficiency was also determined
for the -14 +30 mesh dolomite. As with the -6 +14 mesh bed, a slight increase
in filtration efficiency was indicated for a greater bed depth—from ^95% at a
7.6-cm bed depth to <\,97% at a 30-cm bed depth. The small effect of bed depth
on filtration efficiency for both the -14 +30 and -6 +14 mesh beds indicates
that in both cases, filtration occurs at or near the bed surface.
-------
86
A comparison of the results obtained for the two different mesh-size
materials shows that at a filtration velocity of 'x-SO cm/s, decreasing the mean
bed-particle size from ^120 ym to ^740 ym resulted in increased filtration
efficiency over the range of bed depths used. Expressed in terms of percent
penetration (100% - filtration efficiency), decreasing the particle size at the
bed depths used reduced penetration by 35 to 40% (e.g., from 9% to 5%).
Tests designed to measure the effect of gas velocity on filtration
efficiency have been inconclusive. Filtration tests were performed at gas
velocities of 15, 30, and 60 cm/s. In addition to -6 +14 mesh and -14 +30
mesh beds of dolomite, sulfated limestone bed material obtained from the Pope,
Evans, and Robbins atmospheric-bed test facility in Alexandria, West Virginia,
was tested. Although this sulfated material had a mass mean particle diameter
(WOO ym) approximately equal to that (o,740 ym) of the -14 +30 mesh dolomite
bed material, the sulfated limestone had a much broader particle size distri-
bution (2 wt % on 6 mesh and 19 wt % smaller than 170 mesh).
For -6 +14 mesh virgin dolomite at a bed depth of 5.1 cm, no change
in filtration efficiency was apparent over the range of velocities tested.
However, for the -6 +14 mesh bed material at a bed depth of 20.4 cm and for the
sulfated limestone at a bed depth of 5.1 cm, filtration efficiency apparently
improved with decreasing gas velocity. The filtration efficiency for the
sulfated bed material increased from 94% to 99% when the gas velocity was
decreased from 60 cm/s to 15 cm/s, an 80% decrease in particle penetration
through the bed. If this effect of velocity is real, it suggests that
factors other than inertial impaction of the flue-gas particulate on the bed
granules are important in governing particulate removal in the granular
bed.
It is both interesting and encouraging to note that the best filtration
results obtained in the tests reported here were obtained with the sulfated-
limestone bed. This is considered to be the material most likely to be used
in a practical application of the concept. The probable reason for its
performance being superior to that of the sized bed materials is its broader
particle-size distribution, and thus its lower bed porosity.
Particle loadings in the flue gas leaving the granular-bed filter
generally ranged from 0.01 to 0.07 g/m3. These loadings fall in the higher
end of the estimated range of acceptable particle loadings for gas-turbine
operation, 0.05 to 0.005 g/m3.38
Because the particle loadings in the flue gas entering the granular-
bed filter (which ranged from 0.3 to as high as 1.4 g/m3) could not be indepen-
dently controlled, no attempt was made to correlate the particle loadings in
the clean flue gas leaving the granular-bed filter with operating conditions.
Particle-size distributions of dust in the flue gas entering and
leaving the granular-bed filter were also measured with the inertial impactor.
Mass mean diameters of the dust, both entering and leaving the granular bed,
were generally of the order of 2 to 4 ym. Also, a comparison of inlet and
outlet particle-size distributions revealed that filtration efficiency was
constant over the range of dust particle sizes entering the filter, i.e., a
1-ym dust particle entering the granular-bed filter was captured as efficiently
-------
87
as was a 10-ym dust particle entering the filter. This result was unexpected
and may simply indicate that the sensitivity of the tests was insufficient to
measure any effect.
It should be noted that dust leaving the filter generally contained
between 10 and 25 wt % dust having particle diameters larger than 10 ym.
Thus, although the mass loadings in the flue gas leaving the granular-bed
filter were within the range of the estimated requirements, the amount of
>10-ym particles in the effluent gas may prove to be too great to be acceptable.
It should also be emphasized that the cascade impactor provides no information
about particle sizes larger than 10 ym. The distribution of large particles
will have to be determined by other techniques.
The initial pressure drop across the clean bed at the beginning of a
filtration cycle for gas velocities below 30 cm/s ranged from a low of 'vl kPa
(4 in. of H20) for a 5.0-cm bed of the -6 +14 mesh bed material to a high of
15 kPa (^60 in. of ^0) for a 30-cm bed of -14 +30 mesh bed material. Although
these pressure drops would be considered excessive in conventional particle-
removal devices for atmospheric combustors, they are not unreasonable for PFBC
applications.
The rate at which the pressure drop across the granular bed increased
during a filtration experiment varied, depending on the efficiency of filtration;
i.e., the more efficiently the filter performed, the more rapidly the pressure
drop across the filter increased. In one experiment in which the filtration
efficiency was ^99%, the initial pressure drop across the bed was 'W.7 kPa (31
in. water gauge) and increased at a rate of ^17.6 Pa/s (4.2 in. water gauge/min).
At these conditions, the granular-bed filter would have to be regenerated quite
frequently, perhaps every 0.5 h, to maintain an acceptable average pressure
drop across the filter.
Assessment of PFBC Limestone Requirements and Granular-Bed Filter
Design. A simple mass balance analysis was also performed to evaluate the
compatibility of sorbent usage in a PFBC process with limestone requirements
of the granular-bed filter. In the analysis, the rates of limestone usage and
the gas volumetric flow for a 200-MWe electric PFBC were estimated and the
permissible ranges of parameters such as bed thickness, gas velocity, and bed
replacement rate for a granular-bed filter which would be compatible with the PFBC
flows were assessed. The mass balance for the combustor (200 MWe) indicates
that ^17.3 Mg/h of solids overflow from the combustor would be available for
use as filtration medium in the granular-bed filter. Based on a bulk density
of 1370 kg/m3 for the sulfated sorbent, ^12.6 m3/h of the sulfated sorbent
would be available for use in the granular-bed filter. Irrespective of the
granular-bed filter design, a mass balance made on the filter indicates the
maximum filter surface area as a function of granular-bed-filter thickness and
frequency of filter regeneration, i.e., the number of times the granular
bed is changed per hour of operating time. The results of this analysis show
that for a given bed thickness, the maximum total filter area which could be
incorporated into the filter design decreases rapidly as the frequency of bed
regeneration (cleaning) increases. At a bed thickness of 10 cm, for example,
increasing the number of bed changes per hour from 0.5 to 3.0 decreases the
maximum filter area possible from ^255 m^ to ^42 m^.
-------
88
The material balance around the combustor also provides the volumetric
flow rate of the flue gas which would have to be filtered by the granular-bed
filter. Thus, the linear velocity of the flue gas through the granular-bed
filter can be calculated as a function of the filter surface area and the
operating temperature and pressure.
The significant observation made from this analysis is that to
maintain a relatively low gas velocity through the filter, i.e., 0.1 to
0.2 m/s, a total filter surface area between 0,160 m? and 630 m7 is required,
depending on the system operating pressure. A granular-bed filter with an
area of 160 m^ (the surface area necessary to handle 227,000 nP of gas at
815°C, 2.02 MPa, and a gas velocity of 0.2 m/s) and with a filter thickness
of 10 cm, would require bed replacement at intervals greater than one hour.
As indicated above, the pressure drop across the granular bed
increased quite rapidly during those experiments in which high filtration
efficiencies were obtained. In some experiments, the rate of pressure drop
increase was as high as a/14.3 Pa/s (1 psi every 8 min) . Thus, with bed
replacements at the rate of one per hour, the average pressure drop across the
filter would be 30-35 kPa (4.5-5.0 psi). In order to reduce this very high
value, the filter bed would have to be replaced more frequently, whereupon
the limestone required for the filter would exceed that available from the
PFBC process.
It should be emphasized, however, that the filtration experiment
results presented above were obtained under conditions of high inlet particle
loading, due to inefficient cyclones upstream from the granular-bed filter.
Additional tests are planned (following the addition of a high-efficiency
cyclone to the flue-gas system upstream from the granular-bed filter) to
evaluate the granular bed with more realistic particle loadings and size
distributions in the flue gas entering the granular-bed filter. Such additional
tests will be necessary before filter performance can be evaluated along with
the compatibility of sorbent usage in the PFBC process with material require-
ments of the granular-bed filter.
b. Acoustic Dust Conditioning
Acoustic dust conditioning is a technique used to enhance the
natural tendency of polydispersed particulates to impact upon each other and
agglomerate. Thus, the use of acoustics in controlling fine particle emissions
is a process whereby the mean size of the effluent particles is significantly
increased (and correspondingly their number is decreased) by exposure to
high-intensity finite-amplitude acoustic fields. As described, acoustic dust
conditioning is designed to increase the collection efficiency of downstream
dust collectors.
Work is currently being carried out under subcontract at the University
of Toronto to develop and fabricate a pulse-jet acoustic dust-conditioning
system consisting of a pulse-jet sound generator, a resonant manifold, and an
acoustic treatment section. A functional description of the various components
in the system was given previously (ANL/CEN/FE-77-3). Reported here is the
progress made at the University of Toronto toward completion of the fabricated
pulse jet, acoustic dust-conditioning system which is to be installed in the
flue-gas system of the ANL, 15.2-cm-dia combustor for testing and evaluation.
-------
89
Pulse-Jet Development. Figure 22 is a photograph of a typical pulse
jet used in the pulse-jet testing and development program. Pulse-jet development
consisted of the fabrication of a number of pulse jets to test the effects of
various geometrical configurations on the acoustic power and frequency range
of the device. The objective, of course, was to achieve a high level of sound
intensity, 155-160 dB, over a wide range of frequency. The geometrical
parameters examined included:
1. The intake and exhaust-pipe geometry
2. Combustion-chamber length and diameter
3. Fuel-inlet nozzle geometry.
Fig. 22. Typical Pulse Jet Tested in the
Pulse-Jet Development Program
Propane gas was used as fuel for the pulse jet. The unit was tested
using six different combustion chambers of similar geometry but differing in
length and diameter. Each combustion chamber was also tested with a variety
of intake- and exhaust-pipe geometries. The following summarizes the state of
development of the pulse jet:
1. A sound intensity of 155-160 dB has been obtained
at a frequency of Aj280 Hz
Total thermal input to pulse jet is 40 to 45 kW
Combustion-chamber outer wall temperature can be
maintained at 370°C (700°F), using forced-air-flow
cool ing
4. Tail-pipe inner surface temperature can be maintained
below 150°C (300°F) by using forced-flow water cooling
on its exterior surface
5. Cooling devices for the pulse jet remove ^12 kW of the
total thermal input of 40 to 45 kW.
-------
90
All of the above results were obtained in testing at ambient pressure.
Based on the unit which yielded the above results, a pulse-jet has been
fabricated for subsequent testing at elevated pressures.
Resonant Manifold. The resonant manifold system (RMS), shown
schematically in relationship to the pulse jet in Fig. 23, is also under
development. As with the pulse jet, the manifold is being tested at ambient
pressure prior to high-pressure testing. The RMS, in conjunction with the
pulse jet, has been tuned at ambient pressure with a sliding piston, whose
location can be adjusted through the downstream opening of the resonant
manifold.
/-
t
RESONANT MANIFOLD
/-
COOLING COILS
WATER IN
*— L
RESONANT M
EXHAUSTINC
£— SLIDING PISTON ' 15cm-- -p-SOUND DUCTH r^=^ ,. — ,
-pJ ] 1 1 > M
u u o u O O O
\NIFOLD . X/2
' DUCT 51-56 cm
FREQ: 550 Hz
dB: 146
V2
102-114 cm
FREQ: 350 Hz
dB: 144
3
\
I
I
-MICR
PHON
(POS
ARB
U "" U ~~i \ LI
AIR IN \
WATER \- CC
OUT
L- PULSE- JET TAIL PIPE
0-
E
TION
TRARY)
AIR COOLING JACKET
FUEL
AIR
COMBUSTION
CHAMBER
Fig. 23. Schematic of the Pulse Jet/Resonant
Manifold System
Two locations of the resonant manifold and the corresponding half-
wavelengths (as well as their respective frequencies) are also indicated in
Fig. 23. They were determined by calculation and were confirmed by experimental
measurement.
The entire length of the RMS is cooled by water flowing through a
copper coil wound on the outside wall of the manifold. The outer wall temper-
ature was maintained between 210 and 270°C. Approximately 16 kW of energy is
removed by cooling of the manifold.
The variation in sound intensity between a tuned and untuned RM
length can be as much as 10 dB. The RMS exhausting duct is not tuned with
respect to the sliding piston to muffle the sound from the duct; the sound for
the acoustic dust-conditioning system will be extracted from the opposing
(upstream) sound duct(s) (see Fig. 23).
Fabrication of components for the high-pressure manifold has been
completed. The design of the unit is basically that of the ambient-pressure
unit illustrated in Fig. 23.
-------
91
Pressure and Flow Controls. Installation of the pulse-jet/RMS as an
integral part of the ANL FBC unit for testing requires that pressure and
flow control between the two systems be controllable so that the two systems
can be isolated from each other in case of emergency. The control system has
been designed and is currently being set up at the University of Toronto for
testing.
As soon as all three units (the pulse-jet, the resonant manifold,
and the pressure and flow control system) have been operated satisfactorily as
a unit at elevated pressure, the system will be transported to ANL for testing
of its effectiveness in reducing particulate emissions from the ANL combustor.
c. High-Efficiency Cyclone
A high-efficiency, controlled-vortex cyclone (TAN-JET) has been
obtained from the Donaldson Co. and has been installed in the flue gas system
of the ANL combustor as a secondary cyclone. Experiments to evaluate the
performance of the unit as a particulate-removal device are being initiated.
In addition to testing of the cyclone's performance, the unit is intended to
permit testing of the granular-bed filter under conditions of reduced loading
from that previously used in granular-bed experiments. In addition, the
cyclone will be extremely helpful in evaluating the capability of the pulse-jet
acoustic agglomeration system for increasing the particle-removal efficiency
of mechanical collectors.
MISCELLANEOUS STUDIES
1. Pulsed L-Valve Tests
(W. M. Swift)
a. Introduction
The L-valve is a type of a nonmechanical valve which can be used to
control the flow of particulate solids into either dense-phase or lean-phase
media, such as fluidized beds or pneumatic conveying lines. This valve, which
uses a relatively small amount of aeration gas to control the flow of solids
through the valve, has several advantages over mechanical valves: It is very
inexpensive, has no moving parts subject to wear or seizure, and can be used
in high-temperature and/or high-pressure environments.
In an L-valve, solids flow by gravity from the hopper into the top
of the L-valve. By use of a timer and a solenoid valve, aeration gas is
pulsed or supplied continuously at right angles to the solids downward flow.
The gas transports the solids horizontally in dense-phase flow to the point
where the solids are discharged into a pneumatic transport line or a fluidized-
bed.
The L-valve has been used in various applications for a number of
years. Only recently, however, were the effects of geometrical and particle
parameters on the performance of an L-valve reported in the literature.™
Specifically, the effects of pipe length and diameter, particle size and
density, and the location of the aeration gas inlet on L-valve performance
were reported.
-------
92
ANL is interested in applying the L-valve to feed systems for
fluidized-bed combustors. In particular, there is interest in adapting the
L-valve to controlling solids feed rates for ANL's fluidized-bed combustion and
sorbent-regeneration process development units. The process development
units, however, require relatively low solids feed rates as compared with the
high solids throughputs usually achieved with an L-valve. A 15.4-cm-dia
atmospheric combustor, for example, requires feed rates of approximately 3.5
kg/h of coal and 1.1 kg/h of dolomite; in comparison, a feed rate as high as
9000 kg/h is reported in the literature-^ for a 7.6-cm (^3-in.) L-valve with
an aeration-gas flow rate of only 0.9 L/s (^1.8 ft-Vmin) .
In an existing application of an L-valve at ANL for feeding sulfated
sorbent to a regeneration reactor, the required feed rates are obtained by
pulsing the L-valve aeration gas. By use of a 12.5-mm (3/8-in. pipe) L-valve,
feed rates of 27 kg/h and less are obtained, using cycle times of >2 s and an
aeration period of ^0.5 s/cycle. It is calculated that a feed rate of ^3.5
kg/h would require a cycle time of ^16 s. Thus, at low feed rates, feeding
would be highly discontinuous.
A test program was performed to develop an L-valve feeder capable of
operating at low solids feed rates and employing very short L-valve cycle
times. The design performance requirements were that the valve be capable of
feeding pulverized coal at the rate of 3.4 kg/h and of feeding dolomite at a
rate of 1.1 kg/h. A minimum cycle rate of 1 cycle/s for coal feeding was
specified to limit the coal feed increments to no greater than 5% of the
burnout rate of 20 s assumed for the combustor. Cycle rates as low as 0.2
cycle/s were considered acceptable for dolomite feeding. L-valve aeration
duty was to be held between 10 and 40% of the total cycle time. For coal
feeding, a feed-rate variability of less than 2% between any two consecutive
100-s feed periods and less than 10% between any two 100-s periods greater
than one hour apart were requirements for acceptable L-valve operation. The
results of this study have been reported in detail in a recent quarterly
report (ANL/CEN/FE-78-3). The following represents a summary of the significant
findings.
b. Equipment and Procedure
The L-valve was a standard 9.52-mm (3/8 in.) stainless steel tubing
cross fitting. The solids flowed from the bottom of the solids hopper in
gravity flow down through a 25.4-mm (1-in.) downcomer line to the L-valve.
The aeration gas entered the L-valve at right angles to the solids downward
flow and transported the solids in dense-phase flow to a point where they
again could fall under gravity flow into a pneumatic transport line. A
rotameter in the aeration-gas line qualitatively indicated gas flow during the
timed aeration pulses. The solenoid valve controlling the flow of L-valve
aeration gas was operated by a timer which had a cycle-time capability of 0.1
to 9.9 s in 0.1-s increments; the solenoid-open capability ranged from 0.1 to
0.9 s in 0.1-s intervals.
A solid 9.52~mm rod was inserted into the L-valve through the
downward-oriented leg of the cross fitting. The depth of penetration of the
rod into the L-valve was adjustable, providing additional control of the
solids flow rate.
-------
93
Triangle coal (%85 wt % between 90 and 710 ym) and virgin Greer
limestone (^75 wt % between 600 and 1400 ym) were used in the L-valve feeder
tests. Tests to measure feeding rates were made using either the coal or a
mixture of three parts by weight coal and one part by weight limestone.
c. Test Results
Tests Al to A3 represent three consecutive 100-s coal feeding tests
using identical conditions—an aeration cycle time of 1.0 s and an aeration
pulse of 0.2 s. The variance in feed rate between tests A2 and Al was 2.0%;
between tests A3 and A2, the variance was 3.3%.
In coal feeding tests Bl to B4, the conditions were identical to
those for Al to A3 except that the aeration-pulse duration was 0.1 s. The
maximum variation in feed rate between any two consecutive feed intervals was
1.2% The maximum deviation in feed rate of any feed period from the average
feed rate was also 1.2%.
An examination of test results for series A and B illustrates the
sensitivity of the feed rate to changes in aeration pulse time. Decreasing
the aeration pulse from 0.2 s to 0.1 s at a total cycle time of 1.0 s decreased
the feed rate by 25%—i.e., from 5.6 kg/h to 4.2 kg/h. To provide a finer
control of feed rate, a timer capable of 0.01-s increments in both cycle and
pulse duration time would be desirable.
Test series C and D (three experiments in each series) were done
under conditions identical to those in series B except that the cycle time was
1.1 s in series C and 1.2 s in series D. Cycle times above 1.0 s exceed the
specification originally set for each coal feed increment. At 1.2-s cycle
time, however, the coal feed rate of 3.5 kg/h was very close to the design
feed rate of 3.4 kg/h. Possibly, the same result could have been achieved at
a 1.0-s cycle time if the timer had been capable of aeration pulses shorter
than 0.1 s. It was noted that the feed rate (in terms of g/aeration pulse)
decreased slightly with small increases in cycle time.
For test series E (2 tests), the rod was inserted into the bottom of
the cross, extending to the centerline of the horiontal run of the L-valve.
At the same test conditions as were used in the D-series runs (1.2-s cycle
time; 0.1-s aeration pulse), the feed rate in El was approximately 20% lower
with the rod inserted. In run E2, with a cycle time of 1.0 s and an aeration
pulse of 0.1 s, the design feed rate of 3.4 kg/h was obtained.
The remaining experiments tested the premixing of coal and sorbent
(mixtures contained 75 wt % coal) to achieve the design combined feed rate of
4.5 kg/h. No feed tests were made with dolomite alone, but feed rates near
the 1.1 kg/h specified could very likely have been achieved by using the
specified maximum cycle time of 5 s.
Tests Fl through F4 (with cycle times of 1.0 or 0.9 s and aeration
pulse durations of 0.1 s) were the initial feed tests with the coal-limestone
mixture. After test Fl with a cycle time of 1.0 s and a feed rate of 4.1
kg/h, the cycle time was reduced to 0.9 s and a feed rate of 4.4 kg/h was
obtained. This feed rate is very close to the design feed conditions of
-------
94
4.5 kg/h. The F-series test results also illustrate the sensitivity of the
feeder to changes in cycle time. Decreasing the cycle time by 10% (from 1.0
to 0.9 s) increased the feed rate by
Three H-series experiments, also performed with the coal-limestone
mixture, a 0.9-s cycle time, and a 0.1-s aeration pulse, consisted of consecutive
tests of 1200-s duration. The maximum variation in feed rate between any two
tests was less than 1%. The results were also in good agreement with the
results of tests performed at test durations of 100 s.
In the final set of tests, series J with 100-s duration, 1.0-s
cycle time, and 0.1 or 0.2-s aeration pulse, the effect of decreasing the
pressure drop across the solenoid valve was investigated. In the previous
tests, the pressure in a surge tank in the aeration gas supply line was 0,190
kPa (28 psia) with the L-valve feeding to ambient pressure. For the J-series
experiments, the surge tank pressure was lowered to ^14 kPa . A comparison of
experimental results shows that decreasing the pressure in the surge tank in
one test decreased the feed rate o,10% — from 4.1 kg/h to 3.7 kg/h. By then
increasing the aeration pulse from 0.1 to 0.2 s in three other tests, the feed
rate was increased by o,25% to-\,4.6 kg/h.
In conclusion, the test results demonstrate the potential of a
pulsed L-valve feeder for providing low flow rates of solid materials.
Problems of plugging in the L-valve were not encountered, even though the
largest limestone particle fed, 0,1400 ytn, was less than a factor of 6 smaller
than the ID of the L-valve. Variations in solids feed rates at fixed test
conditions were generally within the rigid specifications set forth in the
test plan. With additional refinement of the L-valve timer to allow it to
control the cycle time and pulse duration to 0.01-s intervals, the L-valve
could be easily adapted to on-line control.
-------
95
APPENDIX A
Compositions (wt %) of Limestones and Dolomites
Limestone
ANL-4801
ANL-4901
ANL-4902
ANL-4903
ANL-5001
ANL-5101
ANL-5102
ANL-5201
ANL-5202
ANL-5203
ANL-5204
ANL-5205
ANL-5206
ANL-5207
ANL-5301
ANL-5302
ANL-5303
ANL-5304
ANL-5401
ANL-5402
ANL-5403
ANL-5501
ANL-5601
ANL-5602
ANL-5603
ANL-5801
ANL-6101
ANL-6301
ANL-6401
ANL-6501
ANL-6701
ANL-6702
ANL-7401
ANL-8001
ANL-8101
ANL-8301
ANL-8701
ANL-8901
ANL-8902
ANL-8903
ANL-9201
ANL-9401
ANL-9402
ANL-9501
(Tymochtee)
(1337)
(Dolowhite)
(1351)
(Greer)
(1360)
(Chaney)
(1343)
(1336)
(Grove
or 1359)
CaCC>3
48.7
49.9
49.7
49.2
50.4
51.8
51.2
52.2
52.7
52.6
52.9
52.1
52.7
52.3
53.4
53.9
53.7
53.5
54.5
54.9
54.0
55.6
56.8
56.6
56.3
48.7
61.2
63.2
64.2
65.9
67.3
67.5
74.7
80.4
81.6
83.9
87.0
89.8
89.6
89.3
92.6
94.1
94.7
95.3
M8CC3 Fe203 U20, Si02
40
43
43
44
43
43
43
43
42
36
45
41
42
37
45
44
42
3
42
27
44
43
45
43
41
40
28
32
29
31
31
0
10
3
11
13
1
2
3
1
5
1
0
1
.2
.0
.8
.6
.0
.3
.4
.0
.0
.4
.8
.0
.2
.5
.4
.9
.8
.7
.9
.3
.1
.3
.6
.8
.7
.2
.7
.6
.5
.5
.5
.91
.2
.5
.6
.4
.2
.2
.02
.21
.3
.0
.87
.3
(Cont
0.66
1.69
0.43
0.52
0.37
0.41
0.62
0.26
0.34
0.58
0.25
0.07
0.13
2.35
0.07
0.16
0.30
2.33
0.07
2.73
0.08
0.23
0.09
0.33
0.46
0.66
5.56
0.39
0.33
0.38
0.37
0.08
1.07
1.24
0.86
0.14
3.4
0.66
0.37
0.98
0.2
0.35
0.14
0.09
'd.)
1
0
0
0
0
1
0
0
0
1
0
0
0
1
0
0
0
0
0
1
0
0
0
0
0
1
0
0
0
0
0
0
1
3
0
0
2
1
0
0
0
0
0
0
.81
.93
.87
.90
.98
.46
.43
.68
.87
.13
.04
.16
.25
.0
.08
.43
.31
.73
.34
.10
.17
.18
.01
.18
.60
.81
.51
.28
.69
.28
.22
.18
.82
.18
.19
.42
.0
.04
.79
.83
.42
.51
.23
.25
7
1
5
8
3
3
1
1
I
8
0
3
1
3
0
1
2
36
1
10
0
2
0
1
1
7
3
2
5
2
1
30
11
10
1
1
7
4
5
5
1
3
0
0
.0
.69
.71
.77
.52
.61
.2
.22
.60
.77
.54
.79
.14
.64
.69
.35
.94
.12
.04
.14
.62
.97
.18
.05
.08
.0
.15
.54
.10
.08
.25
.22
.2
.34
.86
.24
.1
.0
.06
.5
.26
.18
.63
.77
Na20
0.09
0.05
0.03
0.05
0.34
0.07
0.05
0.07
0.11
0.04
0.04
0.04
0.05
0.09
0.05
0.03
0.04
0.16
0.04
0.08
0.03
0.03
0.02
0.03
0.2
0.09
0.13
0.04
0.15
0.04
0.04
0.08
0.19
0.23
0.10
0.02
<0. 1
—
0.03
0.12
0.10
0.05
0.02
0.03
K20
0.38
0.12
0.30
0.50
0.10
—
0.19
0.13
0.15
0.18
0.01
0.01
0.02
0.16
—
0.07
0.10
0.41
0.1
0.43
0.07
0.03
—
0.09
0.22
0.38
—
0.05
0.31
0.05
0.15
0.13
0.36
—
—
0.02
«._
0.29
0.22
0.10
0.02
—
-------
96
Appendix A (Cont'd.)
Limestone
ANL-9502
ANL-9503
ANL-9504
ANL-9505
ANL-9601 (2203)
ANL-9602
ANL-9603
ANL-9701 (Germany
Valley)
ANL-9702
ANL-9703
ANL-9704
ANL-9705
ANL-9706
ANL-9801
ANL-9802
ANL-9803
ANL-9901
ANL-9902
ANL-9903
0,003
95.6
95.5
95.8
95.1
96.5
96.2
96.4
97.8
97.5
97.6
97.8
97.3
97.3
98.3
98.2
98.0
99.1
99.1
99.8
MgC03
3.36
0.76
0.58
3.47
3.3
0.43
1.56
0.6
0.68
0.58
0.34
0.53
0.98
0.6
0.47
1.26
0.60
0.38
0.53
Fe203
0.05
0.05
0.30
0.90
0.17
0.12
0.10
0.10
0.05
0.19
0.30
0.17
0.08
0.15
0.18
0.07
0.05
0.05
0.02
A1203
0.11
0.13
0.36
0.29
<0.01
0.21
0.30
1.8
0.05
0.50
0.05
0.35
0.06
0.16
0.10
0.14
0.06
0.04
0.16
Si02
0.42
0.15
2.71
0.70
0.20
1.19
0.70
0.2
0.21
1.08
0.15
0.20
0.23
0.20
0.29
1.47
0.17
0.13
0.24
Na20
0.02
0.02
0.05
0.04
0.05
0.03
0.05
0.25
0.01
0.03
0.04
0.05
0.03
0.04
0.04
0.02
<0.013
0.04
0.20
K20
0.04
<0.01
0.13
0.03
0.01
0.01
0.11
—
0.01
0.17
0.02
0.01
0.01
0.20
0.01
0.02
<0.01
0.01
0.01
-------
97
APPENDIX B
Limestone Designations and Suppliers
Limestone
Limestone
Supplier
Location
ANL-4801
ANL-4901
ANL-4902
ANL-4903
ANL-5001
ANL-5101 (Tymochtee)
ANL-5102
ANL-5201
ANL-5202
ANL-5203
ANL-5204
ANL-5205
ANL-5206
ANL-5207
ANL-5303 (1337)
ANL-5302
ANL-5303
ANL-5304
ANL-5401
ANL-5402
ANL-5403
ANL-5501
ANL-5601 (Dolowhite)
ANL-5602
ANL-5603
ANL-6101 (1351)
ANL-6301
ANL-6401
ANL-6501
ANL-6701
ANL-6702
Midwest Aggregates Corp.
Vulcan Materials
Black Creek Limestone Co.
Harris Limestone Co.
E. F. Duff and Sons
Porter Limestone Co.
Midwest Aggregates Corp.
Midwest Aggregates Corp.
Bullitt County Stone Co.
James River Limestone Co.
Mayville White Lime Works
Mayville White Lime Works
Quapaw Co.
Chas Pfizer Co.
Vulcan Materials Co.
May Stone and Sand
Vulcan Materials Co.
Delta Mining Co.
Raid Quarries (Medusa Aggregates)
U.S. Steel Corp.
Lime Products Corp.
Kaiser Refractories
G. & W. H. Corson, Inc.
York Stone Co.
Jeffery Limestone Co.
Meshberger Stone, Inc.
G. & W. H. Corson, Inc.
Meshberger Stone, Inc.
Osmundson Bros.
Vulcan Materials Co.
Fort Wayne, IN
Monroe, WI
McCook Complex, IL
Piedmont, MO
Huntsville, OH
Rockton, IL
Fort Wayne, IN
Fort Wayne, IN
Shepardville, KY
Buchanan, VA
Mayville, WI
Mayville, WI
Drumwright, OK
Gibonsville, OH
Birmingham, AL
Fort Wayne, IN
Parsons, TN
Mill Creek, OK
Burlington, IA
Chicago, IL
Union, ME
Salinas, CA
Plymouth Meeting, PA
York, PA
Parma, MI
Columbus, IN
Plymouth Meeting, PA
Columbus, IN
Adams, MN
Parsons, TN
(Cont'd.)
-------
98
APPENDIX B (Cont'd.)
Limestone
Limestone
Supplier
Location
ANL-7401
ANL-8001 (Greer)
ANL-8101 (1360)
ANL-8301
ANL-8701 (Chaney)
ANL-8901 (1343)
ANL-8902
ANL-8903
ANL-9201 (1336)
ANL-9401
ANL-9402
ANL-9501 (Grove or 1359)
ANL-9502
ANL-9503
ANL-9504
ANL-9505
ANL-9601 (2203)
ANL-9602
ANL-9603
ANL-9701
(Germany Valley)
ANL-9702
ANL-9703
ANL-9704
ANL-9705
ANL-9706
ANL-9801
ANL-9802
ANL-9803
ANL-9901
ANL-9902
ANL-9903
Fort Calhoun Limestone Co.
Greer Limestone Co.
Monmouth Stone Co.
Vulcan Materials Co.
Hooper Brothers Quarry
Vulcan Materials Co.
Civil Bend Bethany Falls
Georgia Marble Co.
Pete Lien and Sons
Lime Products Corp.
Grove Lime Co.
American Aggregates
Southern Materials
Rose Equipment, Inc.
Midwest Minerals
Columbia Quarry Co.
Delta Mining Corp.
G. & W. H. Corson, Inc.
Greer Limestone Co.
Hemphill Brothers
Austin White Lime Co.
Midwest Limestone Co.
Western Materials Co.
U.S. Steel Corp.
Iowa Limestone Co.
Chem. Lime Inc.
Rigsby and Barnard Quarry
Western Materials Co.
Southern Materials
Calcium Carbonate Co.
Fort Calhoun, NE
Morgantown, WV
Monmouth, IL
Birmingham, AL
Weeping Water, NE
Parsons, TN
Bethany, MO
Tate, GA
Rapid City, SD
Union, ME
Stephens City, VA
Indianapolis, IN
Ocala, FL
Weeping Water, NE
Pittsburgh, KS
Valmeyer, IL
Mill Creek, OK
Plymouth Meeting, PA
Morgantown, WV
Seattle, WA
McNeil, TX
Gilmore, City, IA
Orleans, IL
Chicago, IL
Alden, IA
Clifton, TX
Cave In Rock, IL
Orleans, IN
Ocala, FL
Fort Dodge, IA
-------
99
REFERENCES
1. J. A. Shearer, I. Johnson, and C. B. Turner, The Effect of Sodium Chloride
on the Reaction of S02/C-2 Mixtures with Limestones and Dolomites, Argonne
National Laboratory Report ANL/CEN/FE-78-8 (July 1978).
2. D. L. Keairns et^ _al_. , Fluidized-Bed Combustion Process Evaluation, Phase II,
Pressurized Fluidized-Bed Coal Combustion Development, Report EPA-650/2-75-
027-C, Westinghouse Research Laboratory, Pittsburgh, Pennsylvania
(Sept. 1975)
2a. D. A. Wenz, I. Johnson, and R. D. Wolson, J. Chem. Eng. Data 14(2), 250-252
(1969).
3. G. J. Vogel e^ a\_. , Regeneration of Sulfated Limestone from FBCs and
Corrosive Effects of Sulfation Accelerators in FBCs, Annual Report, July
1977-September 1978, Argonne National Laboratory Report ANL/CEN/FE-78-13.
4. G. J. Vogel et_ a}^. , A Development Program on Pressurized Fluidized-Bed
Combustion. Annual Report July 1, 1974-June 30, 1975, Argonne National
Laboratory Report ANL/ES-CEN-1011.
5. T. A. Hendrickson, Synthetic Fuels Data Handbook, Cameron Engineers, Inc.,
Denver, p. 7 (1975).
6. R. B. Snyder, W. I. Wilson, Irving Johnson, and A. A. Jonke, Synthetic
Sorbents for Removal of Sulfur Dioxide in Fluidized-Bed Combustion,
Argonne National Laboratory Report ANL/CEN/FE-77-1.
7. R. B. Snyder, W. Ira Wilson, and Irving Johnson, The Prediction of
Limestone Requirements for SC>2 Emission Control in Atmospheric Fluidized
Bed Combustion, presented at the 5th International Conference on Fluidized-
Bed Combustion, Washington, DC, December 1977.
8. John E. Mesko, Multicell Fluidized-Bed Boiler Design, Construction, and
Test Program, ERDA Report FE-1237-Q76-3 (Mar. 1976).
9. S. S. Strom e^ al. , Preliminary Evaluation of Atmospheric Pressure
Fluidized-Bed Combustion Applied to Electric Utility Large Steam Generators,
EPRI Report RP 412-1, interim report, Alliance, Ohio (Feb. 1976).
10. Robert L. Gall, Joseph S. Mei, and John S. Wilson, Fluidized-Bed Combustion
Test of Low-Quality Fuels—1. Anthracite Refuse, Morgantown Energy
Research Center, Morgantown, West Virginia (May 1977).
11. R. C. Hoke £t_ al., Studies of the Pressurized Fluidized-Bed Coal Combustion
Process, Report EPA-600/7-76-011, Exxon Research and Engineering Company,
Linden, New Jersey (Sept. 1976).
12. Neil H. Coates (Ed.), Fluidized-Bed Combustion, Mitre Corporation,
McLean, Virginia (May 1976).
13. E. P- O'Neil, D. L. Keairns, and W. F. Kittle, Thermochim. Acta 14, 209
(1976). —
-------
100
14. M. Hartman and R. Coughlin, A.I.C.h.E. J. 22_, 490 (1976).
15. E. B. Robinson, R. D. Glenn, S. Ehrlich, J. W. Bishop, and J. S. Gordin,
Study of Characterization and Control of Air Pollutants from a Fluidized-
Bed Combustion Unit - The Carbon Burnup Cell, EPA Contract CPA 70-10,
Pope, Evans, and Robbins (Feb. 1972).
16. H. B. Lange and C. L. Chen, S02 Absorption in Fluidized Bed Combustion of
Coal—Effect of Limestone Particle Size, EPRI Project No. RP 719-1, The
Babcock and Wilcox Co., Alliance, Ohio (Mar. 1977).
17. Reduction of Atmospheric Pollution, National Coal Board, London England,
Reference No. DHB 060971 (Sept. 1971).
18. A. A. Jonke et_ al. , Reduction of Atmospheric Pollution by the Application
of Fluidized-Bed Combustion, Argonne National Laboratory Report ANL/ES-
CEN-1004 (June 1971).
19. W. T. Reid, External Corrosion and Deposits-Boilers and Gas Turbines,
American Elsevier Publishing Company, New York, p. 140 (1971).
20. P. H. Tufle and W. Beckering, A Proposed Mechanism for Ash Fouling
Burning Northern Great Plains Lignite, J. Eng. Power, Trans. ASME, 407
(1975).
21. G. H. Gronhovd, W. Beckering, and P. H. Tufle, Study of Factors Affecting
Ash Deposition from Lignite and Other Coals, ASME Winter Annual Meeting,
Los Angeles, Calif. Nov. 16-20, 1969. p. 9.
22. L. E. Paulson and W. W. Fowkes, Changes in Ash Composition of North
Dakota Lignite Treated by Ion Exchange, U.S. Bureau of Mines RI 7176,
p. 18 (1968).
23. K. E. Phillips, Energy Conversion from Coal Utilization CPU-400 Technology,
Annual Report for July 1974 - June 1975, Combustion Power Co., Inc., Report
FE-1536-A1.
24. R. R. Ruch, H. J. Gluskoter, and N. F. Shimp, Occurrence and Distribution
of Potentially Volatile Trace Elements in Coal, Final Report, Illinois
State Geological Survey, No. 72 (Aug. 1974).
25. An Investigation of Hot Corrosion and Erosion Occurring in a Fluid Bed
Combustion-Gas Turbine Cycle Using Coal as Fuel, FE/1536-2, Westinghouse
Research Laboratory (May 1977).
26. H. E. Lunt, Hot Corrosion in Gas Turbine, preprint, presented at the
Winter Annual Meeting of ASME, Atlanta, Georgia, Nov. 27-Dec. 2, 1977.
27. J. Szekely, J. W. Evans, and H. T. Sohn, Gas-Solid Reactions, Academic
Press, Inc., New York, Chap. 2 (1976).
-------
101
28. W. M. Swift, G. J. Vogel, A. F. Panek, Potential of Fluidized-Bed
Combustion for Reducing Trace-Element Emissions, Paper No. 75-46.3,
presented at the 68th Annual Meeting of the Air Pollution Control
Association, Boston, Mass., June 15-20, 1975.
29. W. M. Swift, G. J. Vogel, A. F. Panek, and A. A. Jonke, Trace-Element
Mass Balances around a Bench-Scale Combustor, Proceedings of the Fourth
International Conference on Fluidized-Bed Combustion, the MITRE Corpor-
ation, McLean, Virginia (1976).
30. J. W. Kaakinen, R. M. Jordan, M. H. Lawasani, and R. E. West, Trace
Element Behavior in Coal-Fired Power Plant, Environ. Sci. Technol. jK9),
862 (1975).
31. D. H. Klein, A. W. Andren, J. A. Carter, J. F. Emery, C. Feldman,
W. Fulkerson, W. S. Lyon, J. C. Ogle, Y. Talmi, R. I. Van Hook, and
N. Bolton, Pathways of Thirty-Seven Trace Elements through Coal-Fired
Power Plants, Environ. Sci. Technol. 9(10), 973 (1975).
32. C. E. Billings, A. M. Sacco, W. R. Matson, R. M. Griffin, W. R. Coniglio,
and R. A. Hand ley; Mercury Balance on a Large Pulverized Coal-Fired
Furnace, J. Air Pollut. Control Assoc. 2J3(9), 773 (1973).
33. Effects of Trace Contaminants from Combustion, Proceedings of a Workshop,
Aug. 2-6, 1976 (Report ERDA 77-64, available from NTIS, Springfield,
Virginia).
34. D. F. S. Natusch, J. R. Wallace, and C. A. Evans, Jr., Toxic Trace
Elements: Preferential Concentration in Respirable Particles, Science
183 (4121), 292 (1974).
35. W. M. Farmer, Measurements of Particle Size, Number Density, and Velocity
Using a Laser Interferometer, Appl. Optics _10, 2603 (1972).
36. J. C. Montagna, G. W. Smith, F. G. Teats, G. J. Vogel, and A. A. Jonke,
Evaluation of On-Line Light-Scattering Particle Size Analyzers for
Measurements at High Temperature and Pressure, Argonne National Laboratory
Report ANL/CEN/FE-77-7 (Feb. 1978).
37. E. L. Weiss and H. N. Frock, Rapid Analysis of Particle Size Distribution
by Laser Light Scattering, Powder Technol. 14, 287 (1976).
38. D. L. Keairns et^ aU , Fluidized-Bed Combustion Process Evaluation,
EPA-650/2-75-027-C, Environmental Protection Agency (1975).
39. T. M. Knowlton and I. Hirsan, L-Valves Characterized for Solids Flow,
Hydrocarbon Process. 57, 149 (Mar. 1978). ~ —'
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-600/7-79-203
2.
3. RECIPIENT'S ACCESSION1 NO.
4. TITLE AND SUBTITLE
Support Studies in Fluidized-bed Combustion-
1978 Annual Report
5. REPORT DATE
August 1979
6. PERFORMING ORGANIZATION CODE
7. AUTHORIS)i j0hnson, G. J. Vogel, S. H. D. Lee, D. S. Moul-
ton, F.F.Nunes, J.A.Shearer. G.W.Smith, E.B.Smyk,
R. B.Snyder, W. M.Swift. F. G. Teats (cont block 15)
MING <
8. PERFORMING ORGANIZATION REPORT NO
ANL/CEN/FE-78-10
9. PERFORMING ORGANIZATION NAJvfE~AND ADDRESS
Argonne National Laboratory
9700 South Cass Avenue
Argonne, Illinois 60439
10. PROGRAM ELEMENT NO.
INE825
11. CONTRACT/GRANT NO.
IAG-D5-E681 (EPA) and
W31-109-Eng-38 (DoE)
12. SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development*
Industrial Environmental Research Laboratory
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND PERIOD COVERED
Annual; 7/77 - 9/78
14. SPONSORING AGENCY CODE
EPA/600/13
f NOTES (*)DoE is cosponsor. Project officers: D.A.Kirchgessner (EPA);
John Geffken (DoE). EPA-600/7-76-019, -77-138, and -79-157 are related reports.
(Cont. fm block 7: C.B.Turner, W.I.Wilson, and A.A. Jonke.)
15. SUPPLEMENTARY NOTES
16. ABSTRACT
repOr^ gjves results of laboratory- and process-scale EPA studies
supporting the national development of atmospheric and pressurized fluidized-bed
combustion (PFBC) of coal. Program objectives are: (1) to develop basic information
needed to optimize the use of limestone for SO2 emission control from FBC; and (2)
to develop information on processes to treat high-temperature high-pressure gases
from PFBC, enabling them to be used to operate gas turbines. Limestone sulfation
enhancement experiments showed CaC12 and MgC12 to be more effective than NaCl
in increasing the pore size and hence the sulfation capability of calcined limestones.
Experiments to evaluate coal-pyrolysis char as a feedstock for FBC were conducted
on Wyoming subbituminous coal pyrolyzed at about 650 C. Char combustion efficien-
cies were 94-99%. The use of oil shale for SO2 emission control in FBC was eval-
uated. Considerably larger quantities of shale (than limetone or dolomite) would be
required because of shale's lower CaO content. Studies are being conducted to
assist in preventing hot-corrosion of turbine blades by alkali-metal compounds from
PFBC. Results of one set of experiments show that various mineral additives are
highly effective in retaining Na and K in the bed: 90-95% of the elements can be re-
tained.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
c. COS AT I Field/Group
Pollution Oil Shale
Fluidized Bed Processing
Combustion Charcoal
Coal Gas Turbines
Calcium Carbonates Corrosion
Sulfur Oxides Chemical Analysis
Pollution Control
Stationary Sources
Char
Trace Elements
13 B
13H,07A
21B
2 ID
07B
08G
13G
11M
07D
18. DISTRIBUTION STATEMENT
Release to Public
19. SECURITY CLASS (ThisReport)
Unclassified
21. NO. OF PAGES
110
20. SECURITY CLASS (Thispage)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
------- |