JUiL
mm
SELECTED STUDIES
ON
ALKALINE ADDITIVES
SULFUR DIOXIDE CONTROL
U. S. ENVIRONMENTAL PROTECTION AGENCY
-------
SELECTED STUDIES ON ALKALINE ADDITIVES
FOR SULFUR DIOXIDE CONTROL
by
R. H. Borgwardt, D. C. Drehmel,
T. A. Kittleman, D. R. Mayfield,
and J. S. Bowen
ENVIRONMENTAL PROTECTION AGENCY
National Environmental Research Center
Research Triangle Park, North Carolina
December 1971
-------
The APTD (Air Pollution Technical Data) series of reports is
issued by the Office of Air Programs, Environmental Protection
Agency, to report technical data of interest to a limited num-
ber of readers. Copies of APTD reports are available free of
charge to Federal employees, current contractors and grantees,
and nonprofit organizations - as supplies permit - from the
Office of Technical Information and Publications, Environmental
Protection Agency, Research Triangle Park, North Carolina 27711
or from the National Technical Information Services, 5285 Port
Royal Road, Springfield, Virginia 22151.
Office of Air Programs Publication No. APTD-0737
ii
-------
TABLE OF CONTENTS
A. Summary 1
B. General Introduction 5
C. Discussion of Projects 9
1. Kinetics of the Reaction of SO-
with Calcined Limestone
a. Introduction 10
b. Experimental 11
c. Discussion of Results 15
d. Conclusions 28
e. Nomenclature 29
2. Properties of Carbonate Rocks
Related to S02 Reactivity
a. Introduction 31
b. Experimental 33
c. Results 37
d. Discussion of Results 45
e. Conclusions 64
f. Nomenclature 67
3. A field Study of the Role of
Overburning
a. Introduction 69
b. Experimental 71
c. Discussion of Results 76
d t Conclusions 87
4. Methods for Testing the Degree
of Overturning
a. Introduction 89
b. Experimental 92
c. Discussion of Results 95
d. Conclusions 112
iii
-------
5. Capacity of Limestones for
Sorption of SO-
a. Introduction
b. Experimental 115
c. Discussion of Results 119
d. Conclusions 130
D. General Conclusions 132
E. Application of Results to Limestone
Processes 135
F. Recommendations 138
G. References 140
Appendix A Kinetic Data of Carbonate
Rock Types 143
Appendix Bl Procedure for Dead-Burning
Tests 158
Appendix B2 Statistical Analysis of
Dead-Burning Data 162
Appendix B3 Crossplots of Data from
the Dead-Burning Study 184
Appendix Cl Calcined and Uncalcined
Limestones Ranked According
to Bed Weight Gain 197
Appendix C2 EPA-APCO Limestone Inventory 205
Appendix D Sorption of S02 by Waste Kiln
Dust from Portland Cement
Manufacturing Operations 218
Appendix E Copper Oxide Sorption of SO- 224
Appendix F Char Sorption of S0» and
Regeneration 228
iv
-------
TABLE OF FIGURES
1-1 Differential Reactor 13
2 Sorption of S02 by 1351 at Various
Reactor Temperatures 17
3 Sorption of S02 by Different
Particle Sizes of 1351 17
4 Log r Versus Log C to Estimate "m" 19
5 Arrhenius Plots for the Reaction of
S02 with Four Limestones 19
6 Reaction Rate Versus Particle
Diameter 24
7 Sorption of SO, by Calcined
Limestones 24
8 Relation of Reaction Rate and
Sulfate Loading 26
2-1 Comparison of Reactivity of
Calcines with S02 at 980°C 38
2 Sorption of S02 by Different
Particle Sizes 40
3 Estimation of Initial Rate and
Effectiveness Factor 41
4 Comparison of Reactivity with
B.E.T. Surface Area 49
5 Reaction Rate at Constant Sulfation
as a Function of Particle Diameter 52
6 Reaction Rate Versus Surface Area 55
7 Effect of Mode of Calcination
on Reactivity 60
8 Relationship between Effectiveness
Factor and Pore Size 65
-------
3-1 Boiler Plan and Sample Locations 75
2 Reaction of SO, with Calcined
Limestone 78
3 Rate of Hydration of Calcined
Dolomite 80
4-1 Typical Firing Cycle 94
2 Acid Titration of #2061 97
3 Surface Area Versus Flue Gas
Sorption 99
4 Pore Volume Versus Flue Gas
Sorption 100
5 Small Pore Volume Versus Flue
Gas Sorption 101
6 Small Pore Volume Versus S02
Capacity 102
7 Flue Gas Sorption Versus S02
Capacity 103
8 C02 Test: Effect of Reaction
Temperature 105
9 S02 Test: Effect of Reaction
Temperature 106
10 C02 Test: Effect of Reaction
Time 107
11 S02 Test: Effect of Reaction
Time 108
12 Slaking Test Main Effects 109
13 COn Test: Comparison of Measured
Reactivities for Sulfated and
Unsulfated Limes 111
14 Reactivities of Sulfated and
Unsulfated Limes Measured by
Hydration Weight Gain 113
Vi
-------
5-1 Flow Diagram for Fixed Bed
Reactor System 117
2 Reaction Temperature and Form
of the Reactant 120
3 Frequency Distribution of
Capacities of 86 Stones Tested 120
vii
-------
LIST OF TABLES
1-1 Properties of Calcines 12
2 Kinetic Parameters for Sorption
by Calcined Limestones 22
2-1 Stone Identification and
Description 32
2 Chemical Analyses 34
3 Kinetic Parameters for S0£
Sorption by Calcines 39
4 Parameters for SC^ Sorption by
Calcines as a Function of Particle
Size 43
5 Summary of Physical Properties
of Calcines 46
6 Physical Properties of Raw
Limestones 50
7 Effect of Calcination Temperature
on Physical Properties of Calcite
Spar 57
8 Comparison of Reactivities of
150/170 Mesh Stones at 980°C 59
3-1 Composition of Additives 73
2 Sulfate and Carbonate Content of
Boiler Samples 81
3 Specific Volume of Pores and
Degree of Sulfation 84
4 Physical Properties of Boiler
Samples as Determined by Oil
Absorption 86
4-1 Composition of Limestones 93
2 Correlation Coefficients 98
viil
-------
5-1 Effect of Calcination 122
2 Correlations between Physical
Properties and 980°C Capacities 125
3 Capacity of Various Classes of
Limestone 128
IX
-------
A. SUMMARY
Concentrated on limestone-sulfur dioxide (S07) reactions,
alkaline additives research has defined the kinetics and capacity
as well as the effects of everburning. The kinetics of S09
sorption by limestone calcines has been determined with a high
temperature differential bed gas solids contactor. It has been
shown that for small particles the reaction is first order and
chemical rate controlled with activation energies in the range
of 8.1 to 18.1 k cal/g. mole. In general, reaction occurs
initially throughout the particle volume and the internal dif-
fusion resistances become limiting only after conversion of at
least 20% calcium oxide. The reaction rate was weakly dependent
on particle size and strongly depressed by sulfate loading.
For average particle diameters less than 500 microns, the rate
was essentially independent of particle size. The effect of
increasing conversion on rate is explained as an exponential
relationship between the frequency factor and sulfate loading.
Experiments on a wide range of carbonate rocks using the dif-
ferential reactor confirm £he importance of chemical reaction
rate controlling for S02 sorption. Furthermore, the strong
influence of physical properties on limestone reactivity has
been delineated. Small pores and large surface areas lead to
high reaction rates while large pores lead to high capacity.
-------
Pores smaller than 0.1 micron plug rapidly with, reaction products
and are apparently responsible for shell formation on certain
stones. The results- confirm the exponential decay of reaction
rate during the chemical reaction-controlled phase, which
generally extends from 50 to 90 percent conversion of CaO in
95 micron particles at 980°C. Stones of varying geological
type yield calcines of distinctly different physical structures
which show correspondingly large differences in both rate of
reaction and total capacity for S0_ sorption.
The loss of reactivity due to everburning has been studied
using an oil fired boiler belonging to the Florida Power Cor-
poration (in St. Petersburg), During these field tests, samples
of lime were collected from the flue gas at various points in
the boiler and analyzed to determine the degree of calcination,
extent of sulfation, and changes in physical properties related
to chemical reactivity. The boiler samples were found to be
considerably less reactive with S0_ than stone calcined in the
laboratory. The low porosity and high density of the injected
lime indicates that everburning is at least partly responsible
for the low boiler desulfurization achieved when additives are
injected with fuel.
The effects of everburning for calcination temperatures ranging
from 1700 to 3200°F were determined by measuring more than a
dozen chemical and physical properties of these calcines. While
-------
density increased, other reactivities (e.g., SO- sorption, CO-
sorption, extent of slaking), pore volume, and surface area
decreased markedly with calcination temperature. Loss of
reactivity is as high as 80% for sorption of SO- from flue gas
in 120 seconds when the calcination temperature is raised from
1700°F to 2600°F. Moreover, a strong intercorrelation was
noted between S0? reactivities, surface area and pore volume.
This supports the conclusions that (1) reactivity is a function
of surface area and pore volume and (2) loss of reactivity (or
dead-burning) is attributable to growth of CaO crystals which
decreases the surface area and pore volume. The best test for
determining the loss of reactivity of partially sulfated lime
involves measuring the weight gain of hydrated samples.
Eighty-six carbonate rock samples were tested in a fixed-bed
reactor to determine their capacity to react with flue gas
containing sulfur dioxide. Although most of the work was
performed with the carbonate and the oxide at standard test
conditions, supplementary tests were made on hydrates, oxides
and carbonates over a wide range of reaction temperatures and
calcination conditions. At 980°C the average sulfur loading
of the carbonate feed was -practically equivalent to the loading
for the precalcined samples. Differences among the samples
were only slightly related to chemical composition; porosity.
as measured by mercury pore volume, best explained variations
in capacities of the samples. Chalks and oolitic samples were
-------
the most efficient absorbents, and magnesite and Iceland spar,
the least efficient of the stones tested.
In addition to limestone as a material for sorbing SO., other
materials have also been considered. Interest had been shown
in kiln dust from Portland cement manufacturing as a SO. sorbent.
Tests conducted with kiln dust samples exposed to flue gas in a
differential reactor and an aqueous batch scrubber indicated
that kiln dust has absorptivity and reactivity comparable to
limestone, but no better. Metal oxides such as cupric oxide
have also been suggested as potential SO. sorbents. Thermo-
gravimetric analysis (T6A) of a sample of cupric oxide exposed
to pure SO. had a maximum weight gain of 48 weight percent at
738°C. However, at the temperature of interest (315°C) the
cupric oxide had only a 2-3 percent weight gain. Finally,
the reactivity of char with sulfur oxides within the Westvaco
sorption-regeneration scheme has been studied. At the time
ten percent of the feed SO. concentration broke through the
bed of char, loading on the char was 15.8 wt. percent SO..
Loaded S0« in the form of H.SO, was regenerable with H_S to
form sulfur.
-------
B. GENERAL INTRODUCTION
Limestone processes for controlling sulfur oxides, especially
the dry limestone injection process, have long been of great
interest because of their simplicity and low cost. The Division
of Process Control Engineering at an early date selected the
dry limestone injection process for development and large scale
demonstration of optimum performance on a power boiler unit of
the 150 MW class. This is in accord with a report of the
(26)*
National Academy of Engineering , which recommended the
development of throw-away processes for S0~ control on the
highest priority basis. In support of this goal, the activities
of the Process Research Section have been directed toward specific
problem areas relating to the limestone processes, with major
emphasis over the past two and one-half years on acquiring
information on the dry injection process which will be required
to apply the process on a wide scale. Little was known about
the mechanism and rate of reaction between limestone and sulfur
oxides, nor had the differences between limestones been related
to their potential reactivity. That this information was vital
for optimization of the process can easily be seen. The lime-
stone injected into a power boiler must calcine Cevolve CO_ to
become lime) and react with most of the sulfur oxides present -
all within two seconds or less. If the limestone is injected
* Refers to bibliography listed under G._ REFERENCES.
-------
too close to the fireball, the lime produced will be dead-burned
and become unreactive. If the limestone is injected higher or
farther from the fireball, the residence time for the lime
particles in the zone of the boiler favoring reaction with
sulfur oxides will be seriously reduced. Hence, an understanding
of the kinetics of lime-SO_ reactions and the characteristics
of dead-burning were necessary. In addition, data on the
capacity of limestones was of interest to evaluate the maximum
potential of limestones to absorb sulfur oxides, and to identify
those unsuitable for injection due to markedly low capacity.
From the general objectives of defining the capacity and kinetics
of limestone-SO_ reactions as well as determining the nature of
everburning, there evolved five discrete investigations:
1. Kinetics of the Reactions of SO- with Calcined Limestone
2. Physical Properties of Carbonate Rocks Related to SO-
Reactivity
3. A Field Study of the Role of Overburning
4. Methods for Testing the Degree of Overburning of
Calcined Limestones
5. Capacity of Limestones for Sorption of S0~
The first of these was undertaken to determine the mechanism
and rate of reaction of limestones after calcination under
standardized conditions. Effects of reaction temperature,
-------
particle size, and extent of reaction on the reaction rate were
investigated. The second study characterized ten diverse types
of naturally occurring carbonate rocks for their S09 reactivity.
It was the goal of this study to identify those physical or
petrographic properties (such as pore size, surface area and
crystal structure) which contribute to greatest S0« reactivity,
and to identify those types of limestones which perform most
effectively as sorbents for sulfur dioxide.
For the field study of everburning, limestone was injected with
the fuel in an oil-fired furnace. In so doing, the limestone
was given the maximum solids residence time and high tempera-
tures for fast calcination and fast reaction. If SO- removal
efficiency were to be poor, it would be the result of calcination
at excessively high temperatures leading to dead-burning. Other
laboratory tests were made to confirm everburning of the lime-
stones. To develop a test for dead-burning in the presence of
fly ash as well as partial sulfation, which could be used to
quantify the influence of the effect of dead-burning in full
scale evaluation of the process, a fourth study was initiated.
Not only was a dead-burning test desired, but the nature of
dead-burning itself in terms of physical and chemical property
changes needed to be defined.
-------
The last Of the above listed investigations was responsible for
determining the ability of a large number of limestones to
achieve a high absorption efficiency. Thus, a comprehensive
limestone inventory was established and a ranking within this
inventory as to SO. reaction capacity was identified so that
selective recommendation of limestones would be possible.
Other materials besides limestone have been studied as sorbents
for controlling sulfur oxides. Among them are kiln dust, copper
oxides, and char. In general, these investigations were designed
to test the feasibility of proposed processes and were not intended
to be in depth experimental programs. Brief discussions of the
results may be found in the appendices. The kiln dust tests
are discussed in Appendix D; the copper oxide study, in Appendix E,
and the char, in Appendix F.
-------
C. DISCUSSION OF PROJECTS
In this section are given individual detailed discussions of
each of five project areas which have been investigated during
the past two and one-half years (July 1968 to December 1970)
by the Process Research Section. The intent of each of these
projects as discussed fully in the general introduction was
to delineate some aspect of limestone reactivity with respect
to SO- sorption. Each of the five parts of this section
identifies its own' objectives and conclusions as well as the
details of the experimental procedure and results. General
conclusions drawn from the whole of the experimental effort
may be found in the section immediately following.
-------
10
1. Kinetics of the Reaction of SO, with Calcined Limestone
a. Introduc tion
Processes in which limestone and dolomite are used to desulfurize
flue gas are being intensively investigated under the sponsor-
ship of the Air Pollution Control Office. Such processes include
the dry injection of pulverized stone into boiler furnaces, the
wet scrubbing of flue gases and combinations of dry injection
and wet scrubbing, and the use of fluidized bed contactors,
fluid bed combustion, and thin fixed beds. An understanding of
the chemical reactions associated with these processes is desirable
to obtain most efficient performance when such process is applied.
It is generally assumed that limestone absorbs SO~ by a mechanism
involving two consecutive steps - dissociation of the calcium
carbonate, followed by reaction of CaO with sulfur dioxide. It
is expected that the rate of the second step will be important
in any of the proposed pollution control processes, and especially
in the dry injection process . Several investigators ' '
have determined the saturation capacities of a large number of
naturally occurring limestones and dolomites under various
conditions of reaction with SO^. Other studies under way will
define the rate of reaction of uncalcined limestones in the
(8)
disperse phase . The purpose of the investigation reported
here was to determine the rate of reaction of limestones after
calcination under standardized conditions.
-------
11
b. Experimental
Bench scale experiments were carried out to measure the rates
of reaction for several selected limestone calcines when ex-
posed to a fly ash-free flue gas of controlled composition.
Table 1-1 shows the chemical composition and primary physical
characteristics of the calcined stones used in this work.
Complete geological descriptions of the stones are given in a
separate report . Calcination was carried out in 180 g.
batches in an Inconel kiln 12.5 cm. long and 8 cm. in diameter,
rotated at 1 r.p.m. It was heated to 980°C. in an electric
furnace and then charged with 10/28 mesh stone. The kiln was
maintained at 980°C. and purged with air for 2 hours to remove
C02 during calcination. Conversion to the oxide was complete
under these conditions (C0_ < 0.5%). The calcined stone was
cooled, crushed, and screened into size ranges of 12/16, 42/65,
and 150/170 mesh (Tyler). The calcined samples were stored
in airtight containers until used.
The rate of reaction with SO- was determined in a differential
I
reactor (Figure 1-1) constructed of Inconel alloy. In this
type of reactor the thin layer of solid and high gas flow
prevent gas-phase concentration gradients around the reacting
solid. The gases, preheated in an outer annular section, pass
downward through the inner reactor tube about 73 cm. long and
3.42 cm. in inside diameter. The Sample is supported on a
30-mesh Inconel screen in a removable carrier. The carrier
-------
12
Table 1-1. Properties of Calcines
Bulk . b
density, Porosity,
Stone LOI %a %CaO %MgO %Fe2°3 %Si02 g./cc. cc./cc.
1337 47.4 55 43 0.33 0.92 1.41 0.58
1351 42.4 54 28.5 7.0 8.2 1.59 0.56
1343 42.8 94 0.8 0.66 2.98 1.88 0.45
1360 43.8 81 13.0 1.25 3.65 1.51 0.56
a Weight loss on calcination at 980°C.
150/170 mesh particle size.
-------
13
FLUE GAS
:-—N2 GAS
TO INFRARED
TO ORSAT
Figure 1-1. Differential Reactor .
1. Teflon solenoid valve 2. Heating tapes 3. Thermometer 4. Preheat
lection 5. Thermocouple 6. Sample 7. Reactor tube 8. Sample carrier 9.
Heating furnace 10. N» Purge exhaust 11. Alundum filter 12. Constant-
temperature oven 13. Orifice 14. Manometer 15. Flow control valve
-------
14
is sealed against a flange in the center of the reactor tube
so that the entire gas flow passes through the solid during
exposure.
In this investigation a sample consisted of 30 mg. of calcined
stone, distributed uniformly over the 2.65 cm. diameter screen.
For small particle sizes a disk of woven refractory fabric was
placed on the screen and a 1 cm. thickness of refractory
(fused quartz) gauze was placed on the fabric. The lime
particles were dispersed into the gauze.
The mass flow rate of gas through the screen was maintained
constant at 0.075 g./(sq. cm.)(sec.), which at 870°C. corre-
sponds to a superficial velocity of 240 cm. per sec. A high
gas velocity reduced gas film resistance to a negligible
value, so that mass transfer to the particle surface did not
affect rate measurements. The gas fed to the reactor was a
flue gas generated by combustion of fuel oil containing carbon
disulfide. The composition of the flue gas was 10.5% CO.,
3.4% 02, 9.9% H20, 0.27% S02> 0.003% S03, and 75.9% N£ by
volume. The sulfur dioxide concentration was monitored con-
tinuously with an infrared analyzer after removal of the
water vapor from the sample stream.
The reactor was mounted in an electric furnace containing
three heating sections. The center section was energized
-------
15
by a proportional controller acting on a thermocouple located
3.4 cm. above the screen supporting the lime sample. The other
two sections were equipped with variable transformers set by
thermocouples in the top and bottom of the reactor tube to
maintain a uniform temperature over the full length of the
reactor and preheater assembly. The thermocouples were cali-
brated in situ against a multiple shielded, high-velocity
thermocouple. A multipoint recorder continuously monitored
reactor temperatures.
Before a run was started, the carrier and sample were allowed
to heat up for 5 minutes to the reactor temperature. The
time of exposure of the solid to the gas stream was controlled
by solenoid valves that started the gas flow at the beginning
of the run and purged the reactor with nitrogen at the end of
the run. The sample was removed from the carrier after expo-
sure (along with the refractory gauze, if used) and analyzed
for sulfate. The exposed sample was dissolved in water by
soaking with ion exchange resin, filtered, and the filtrate
titrated in 80% isopropyl alcohol with barium perchlorate
using thorin indicator.
c. Discussion of Results
The chemical reaction between limestone and sulfur dioxide at
high temperature in the presence of excess oxygen is:
CaO 4- S02 + 1/2 02 ^ CaS(>4 (1)
-------
16
Equilibrium data for the reaction have been summarized by
(34)
researchers at Battelle Memorial Institutev . The reaction
proceeds to the right at temperatures up to 1230°C. at partial
pressures of SO- corresponding to flue gas concentrations of
about 3000 p.p.m. Equilibrium would also permit the MgO com-
ponent of dolomite to react at temperatures below 840°C., and
it has been reported to participate in the reaction with SO^
(35)
in the presence of iron oxide impurities . More recent
(4)
investigations have shown that SO- reacts only slightly
with MgO at 430° to 700°C. in a fluidized bed, reaction
occurring preferentially with calcium even when Fe2°3 content
is as high as 7%. Calcined magnesite (MgCCL) and calcined bru-
cite [Mg(OH)_] were shown to have low capacities for sorption
(2)
of sulfur dioxide. Dolomites injected into a pilot furnace
have been examined by X-ray diffraction analyses, but showed
no MgSO, as a reaction product, although small amounts were
detectable by DTA methods. No distinction is made in this
study between CaO and MgO, although the data support the con-
clusion that only CaO has a significant reaction rate under the
conditions investigated.
When reaction 1 takes place in flue gas containing high concen-
trations of carbon dioxide, equilibrium also favors a competing
reaction(34) below 770°C.:
CaO + C02 ,=&: CaC03 (2)
Typical experimental results for the sorption of S02 are shown
in Figure 1-2; the mg. of S03 found in 150/170-mesh particles
-------
17
Figure 1-2
60 80
TIME, seconds
100 120
Sorption of sulfur dioxide by dolomite 1351 at various
reactor temperatures
SO» concentration - 3000 p.p.m. (dry basis), particle size = 150/170
Figure 1-3
10
60 80
TIME, seconds
Sorption of sulfur dioxide by different particle sizes of
dolomite 1351
SO> concentration - 3000 p.p.m. (dry basis), temperature = 870° C.
-------
18
(D = 0.0096 cm.) after reaction is plotted against exposure
time at various reaction temperatures. Figure 1-3 shows a simi-
lar plot for different particle sizes at a reaction temperature
of 870°C. Total conversion of the CaO in this dolomite would
correspond to an ordinate value of 23.2 mg. These figures
illustrate the strong sensitivity of the reaction to tempera-
ture and the surprisingly low sensitivity to particle size,
which were characteristic of all the stones examined.
The rate of sorption was measured as the tangent to the smooth
curve drawn through the data and is defined as:
r -
r ~ w dt
where W is the grams of calcined stone exposed in the reactor,
and n' is the gram moles of SO- in the stone at time t. (Note:
Nomenclature is summarized under Subsection e. at the end of
this section) .
The data were correlated according to the rate expression for
(31)
chemical reaction in a porous solid :
since dn'/dt = -dn/dt
W dt = p~ kvC
The effect of SO- concentration, C, on the rate of sorption is
shown in Figure 1-4. The SO- concentration was varied between
-------
0.03
Figure 1-4
SLOPE, m 1.088
0.02 0.1 1
S02 CONCENTRATION
-------
20
58 and 6000 p.p.m. by changing the carbon disulfide content
of the fuel oil burned in the furnace. The rate was measured
at a conversion of 10.5% of the CaO in 12/16-mesh stone reacted
at 870°C. The line fitted to these data by the method of least
mean squares has a slope of 1.008 or m = 1, indicating that
the reaction is first order with respect to the concentration
of SO- in the gas phase. For the remainder of the experimental
work, which is reported below, the SO^ concentration was fixed
at 3000 p.p.m. (dry basis) or 2.9 X 10 g. moles/cc. (wet
basis, 870°C., 1 atm.).
The rate constant k is a function of temperature and also some
function of nf/W, the sulfate loading; it decreases as the reaction
progresses and the solid reactant is consumed. The temperature
dependency was correlated by the Arrhenius equation:
, . -E/RT ,,,
kv = Ae (6)
An Arrhenius plot for each of the four calcined stones is shown
in Figure 1-5 for reactions with S02 at temperatures between
650° and 980°C. The rates were measured at a sulfate loading
_q
of 1 X 10 g. mole/g. of 150/170-mesh particle size sample.
This loading corresponds to approximately 10% conversion of the
CaO. The data show a linear correlation between log r and 1/T,
as specified by Equations 5 and 6. The apparent activation
energy determined from the slope of these plots was distinctly
different for each stone, ranging from 8.1 to 18.1 k cal/g. mole.
-------
21
When rates were measured at higher CaO conversions - up to 20% -
the plots shifted toward the abscissa, but remained parallel
to the lines shown in Figure 1-5, thus indicating no significant
change in the activation energy.
The high sensitivity of the rate to temperature suggests chemical
reaction to be the predominant rate-controlling resistance during
the initial period of S0_ sorption by small particles. The
apparent activation energy for sorption, controlled solely by
bulk diffusion, would be only 3.4 k cal./g. mole. A summary
of the empirical kinetic parameters estimated for the four stones
is given in Table 1-2.
The data in Figure 1-5 further indicate that reduction in rate
of reaction with S02 as a result of competition with C02 was
not important at 650°C. At 540°C., however, there was evidence
that reaction 2 was significant and the Arrhenius plots could
not be extrapolated to that temperature. The plots also failed
when the reaction temperature was raised from 980° to 1100°C.,
sorption rates decreasing at the higher temperature. Subsequent
experiments in which the calcine was heated for 10 minutes at
1100°C. and then reacted at 870°C. showed the same difference
in rate when compared to a sample which was not exposed to the
high temperature. It was concluded that the loss of reactivity
was due to the changes in porosity and bulk density which occur
when lime is "hard-burned"^ .
-------
22
Table 1-2. Kinetic Parameters for Sorption
by Calcined Limestones
*
Activation Reaction rate Frequency
energy, E, constant, k , factor. A,
Stone cal./g. mole sec.~^ V sec.~l
1337 10,000 4.8 X 103 2.4 X 105
1351 18,100 7.2 X 103 9.0 X 106
1343 14,200 4.0 X 103 1.1 X 106
1360 8,100 2.3 X 1Q3 5.5 X 10^
a —3
Evaluated at sulfate loading of 1 X 10 g. moles/g.,
150/170 mesh particle size; 980°C.
-------
23
To ensure that gas film diffusion was not influencing the rate
measurements, a supplementary experiment was made with a 0.95-cm.
diameter sample carrier which permitted exposure of the solid
at high gas velocity. Samples of dolomite 1337 (D = 0.13 cm.)
were exposed for 2 minutes at 980°C. The amount of sulfate
found in the stone was only 5.3% greater at 2400 cm./sec. gas
velocity than found at 225 cm./sec. The negligible effect of
gas velocity verified that the rate of sorption was not limited
by mass transfer to the solid surface.
The effect of particle size on reaction rate is shown in Figure 1-6,
in which the value of r at a sulfate loading of 2 X 10 g. mole/g.
is plotted against the inverse of particle diameter. If the
particles are assumed to be spherical, the total exterior surface
of a given mass of stone (specific surface) would increase with
1/D when the particle size is reduced. If the reaction were
occurring only at the outer surface, the plot shown in Figure 1-6
would be expected to be a straight line through the origin. It
is clear that the rate was not proportional to specific surface
and for fine particles was essentially independent of it. The
results suggest that some reaction takes place within the interior
structure of the solid and that the relative importance of the
internal reaction becomes greater as the particle size decreases.
These observations are similar to the effects associated with
highly porous catalysts and are consistent with the fact that
the pore space in calcined limestones usually accounts for 50%
-------
s
5: 3
o
<
ut
24
Figure 1-6
MESH SIZE
12/16 42/65
1337
20 40 60 80 100 120
Reaction rate vs. l/BP
Sulfate loading = 2 X 10-' g. moles/g.
SO
Ul
3
8
&
0.5
Figure 1-7
1351
Dp * 0.0096 cm
S 10
100 1.000
TIME, seconds
10,000
Sorption of sulfur dioxide by calcined limestones
-------
25
or more of the total volume of the particles. The porosities
(or fraction of particle volume that is pore space) for the
four calcined stones are given in Table 1-1.
Figure 1-7 shows the conversion versus time response for dif-
ferent particle sizes of stones 1351 and 1343 over long periods
of exposure at 870°C. An analysis of the response according to
(32)
the method of Shen and Smith was made to test for intra-
particle diffusion. This model is based on diffusion through
the product crust as the rate limiting mechanism and a nonporous
solid reactant in which reaction occurs only at the interface
of the unreacted core. The data from this investigation could
not be correlated by the shell diffusion model given by Shen
and Smith's Equation 31. Figure 1-6 shows that the model for
shell diffusion, which predicts that the rate of sorption at
_ £
a given sulfate loading will increase with 1/D , is clearly
inconsistent with data on particle size versus reaction rate.
The data could be correlated empirically by a plot of log r
against sulfate loading, n'/W, as shown in Figure 1-8. The
data were linear for all particle sizes of each of the four
stones examined at the reaction temperature of 870°C. The
observed response can be interpreted in terms of a change in
the frequency factor, A, of Equation 5. The frequency factor,
which relates the reaction rate to the number of molecular
-------
26
Figure 1-8
1C
a
I
Dp =0.13
Op = 0.0096
Op = 0.051
0 12 34 5
7 8 9 10
1337
Relation of reaction rate and sulfate loading for dolomite
-------
27
collisions occurring per unit volume per unit time, is
dependent upon the amount of SO- and the amount of unreacted
CaO present (as well as other factors such as surface effects
and probability of reaction of the molecules) . As the reaction
progresses and CaO is consumed, the frequency factor will be
expected to decrease in some manner related to the amount of
sulfate formed. If r and A are the rate and frequency factor
o o -i ./
at zero sulfate loading, then, from Equations 3, 5, and 6,
nAoC -E/RT
~-
designating the slope of the straight line fitted to the data
in Figure 1-8 as 3, the equation for the rate of reaction can
be written as a function of sulfate loading:
, n AC E „ n1 , r /QN
lo* -^ --- RT = ^ W + 10g ro (8)
Equations 7 and 8 reduce to
A = A e-en'/W (9)
o
The observed relationship between rate and sulfate loading thus
indicates an exponential decrease in the frequency factor. The
data of Figure 1-8 show that coefficient 3 is a function of
particle size, the rate of reaction of large particles being
more sensitive to sulfate loading than small particles. Esti-
mates of the effectiveness factor, obtained by extrapolation
of the straight lines of Figure 1-8 to zero sulfation, indicate
-------
28
n - 1 for particles smaller than 0.05 cm. as shown by the
approximately equal intercept of the two lines on the ordinate.
The lower intercept for 5 =0.13 cm. indicates n < 1> implying
P
the presence of pore diffusion resistance of significant magnitude
for the larger size particles.
The results of this study show that the rate of sorption of S02
by calcined limestones is dependent to a very large extent upon
the kinetics of the chemical reaction, particularly at small
particle sizes, and that the rate of reaction predominates as
the overall rate-controlling resistance for conversion of at
least the first 20% of the CaO.
d. Conclusions
Differential reactor techniques were used to measure the rate
of reaction of sulfur dioxide with four natural specimens of
limestone, after calcination at standardized conditions. The
rates were measured as a function of S0« concentration, particle
size, and CaO conversion between 540° and 1100°C. A first-
order chemical reaction was the predominant resistance limiting
the rate of sorption of SO^ by small particles. The activation
energy was dependent upon the type of stone, and ranged between
8.1 and 18.1 k cal./g. mole. The rate was essentially independent
of particle size for D < 0.05 cm. Reaction occurs initially
throughout the particle volume under the isothermal reaction
-------
29
conditions studied and the internal diffusion resistances become
limiting only after conversion of at least 20% CaO. The reaction
rate decreased rapidly with increasing conversion, explained by
an exponential relationship between the frequency factor and the
sulfate loading.
e. Nomenclature
A = frequency factor, sec. (g. moles/cc.)
A = frequency factor at zero solid conversion,
sec.~-'-(g. moles/cc.)m~
C = gas phase concentration of sulfur dioxide, g. moles/cc.
D = mean particle diameter, cm.
E = activation energy, cal./g. mole
k = reaction rate constant per unit volume of solid,
V sec. ~1(g. moles/cc.)m~1
m
= order of reaction with respect to sulfur dioxide
dn/dt = rate of change of S0_ in the gas phase, g. moles/sec.
nf = sulfate in solid as S03, g. moles
R = gas constant, 1.987 (cal./g. mole°K.)
r = rate of formation of SO, in solid, g. moles/(g. sec.)
r = rate at zero solid conversion, g. moles/(g. sec.)
o
V = volume of solid, including intraparticle pores, cc.
W = weight of solid sample, g.
T = temperature, °K.
t = time, sec.
g = empirical correlation coefficient defined by Equation 9
-------
30
= effectiveness factor, ratio of reaction rate to
the rate that would be obtained if entire volume
of particle participated equally in reaction
= bulk (particle) density of solid, g./cc.
-------
31
2. Properties of Carbonate Rocks Related to SO,, Reactivity
__— Z
a. Introduction
This report summarizes an experimental study of the S0_ sorption
characteristics of ten diverse types of carbonate rocks (lime-
stones, dolomites, magnesites, etc.). Specimens representative
of a broad spectrum of naturally occurring stones were selected
for detailed petrographic, mineralogical and chemical examin-
ations under Contract CPA 22-69-65 with the Illinois State Geo-
logical Survey (ISGS). It is the goal of that contract to
establish criteria, based on petrographic and mineralogical
properties, for selecting carbonate rocks of greatest potential
for the desulfurization of stack gases. Tests conducted by
APCO, to evaluate the isothermal rates of reaction of calcines
prepared from these stones are reported here. The rate of
reaction of the raw stone under non-isothermal conditions will
be determined by Battelle Memorial Institute under Contract
PH 86-67-115 using a dispersed phase reactor.
A brief description of the rock types selected for this study
is shown in Table 2-1. This was taken from the ISGS Quarterly
Report, dated December 4, 1969. Two additional stones were
also tested which are not included in that report: Stone No.
2129 (ISGS Type 11), a Michigan marl, which is an unconsolidated
form of CaCO,, and Stone No. 1336 (ISGS Type 10), which is a
-------
32
Table 2-1
Stone Identification and Description
Type 1. Calcite, variety Iceland spar, transparent, colorless
and optically near perfect; cleavage rhombs up to 1 inch.
Type 2. Calcite, very coarse calcite spar, translucent milky in
color due to abundance of crystalline imperfections;
cleavage rhombs up to 1 inch.
Type 3. Calcite limestone, with few scattered fine-grained
dolomite rhombs; coarse unequant granular and porous;
consists of recrystallized crinoid fragments and a few
finely recrystallized bryozoans. Gray.
Type 4. Calcitic limestone, very fine, equant granular, and
dense; light brownish gray.
Type:5. High purity dolomite, medium-grained, granular and
porous; gray reef type of dolomite.
Type 6. 81 percent dolomite, contains clay and fine quartz silt
impurities; medium-grained, granular and microporous;
buff color, non reef type of dolomite.
Type 7. Magnesite, high purity, very fine, equant granular and
microporous; near white.
Type 8. Aragonite sand, contains traces of magnesium calcite;
consists of oolitic and concentrically banded fossiliferous
sand-sized particles. The particles have a banded
structure and consist mostly of fibers of aragonite.
Type 9. Calcitic dolomite (70% dolomite, 18% calcite) with
limonite along grain boundaries and a few scattered
particles of chert; fine-grained, moderately equigranular,
and microporous; mixture of brown and gray colored stone.
-------
33
marble. These two stones were tested because of the unusually
high reactivity found for the marl in pilot scale coal-fired
furnace tests conducted by Babcock & Wilcox (Contract PH 86-67-127)
and the contrastingly low reactivity found for the marble in
similar tests. Chemical compositions of all stones are reproduced
in Table 2-2.
b. Experimental
The reactivity of the eleven types of limestone was measured
and correlated by the methods outlined in the previous section
(C 1 b.) and in the paper, "Kinetics of the Reaction of SO-
with Calcined Limestone" . Calcines prepared in a small
rotary kiln (2 hours at 980°C.) were crushed and screened
into 3 particle size ranges: 12/16 mesh (5 = 0.13 cm.),
42/65 mesh (D = 0.025 cm.) and 150/170 mesh (D = 0.0096 cm.).
P P
Thirty-milligram samples of the calcines were exposed for varying
lengths of time to flue gas containing 3000 p.p.m. SO-, dry
8 3
basis (2.63 X 10 g. mole/cm. , wet basis, 980°C.). Using
the smallest particle size, reaction rates were determined at
reactor temperatures of 650°C. (1200°F), 760°C. (1400°F), and
870°C. (1800°F). The rates were evaluated and compared at a
_o
sulfate loading of 2.5 mg. SO-/30 mg. calcine (10 g. mole/g.).
The sorption curves and Arrhenius plots for these data are given
in Appendix A.
-------
Table 2-2
Chemical Analyses in Weight Percent
ISGS Stone Type
APCO Stone No.
Mineralogical
Type
SiO,
Ti02
A1203
Fe203
FeO
MnO
MgO
CaO
Na20
K20
P 0
C02
S03
SrO
Cl
C (organic)
H20 (100°C)
Ign. Loss
1 2
2201 2202
Iceland calcite
spar spar
ND
ND
ND
ND
.13
ND
ND
55.3
0.03
0.02
b
trace
43.95
0.01
0.14
ND
_
-
43.49
ND
ND
0.01
0,19°
—
0.06
ND
55.5
0.07
0.02
^ ,
43.35
0.17
0.002
0.04
_
-
43.15
3
2203
calcite
ND
ND
ND
0,20C
—
0.10
1.86
53.4
0.08
0.02
—
43.75
0.20
0.009
0.03
—
-
43.67
4
2204
calcite
1.53
ND
0.01
0.31C
—
0.09
0.00
54.8
0.06
0.04
w
43.35
0.15
0.019
ND
-
-
43.15
5
2205
dolomite
0.03
ND
0.02
0.34C
—
0.02
21.40
30.30
0.10
0.03
fm
47.30
0.13
0.019
0.09
—
-
47.24
6
2206
dolomite
11.8
0.02
1.77
0.13
0.41
0.02
17.4
26.5
0.06
0.90
0.02
40.27
0.03
0.04
trace
—
j
40.46d
7
2207
magnesite
0.47
ND
0.08
ND
0.07
ND
44.2
2.93
0.03
0.03
trace
50.96
0.01
0.01
ND
—
-
51.56s
8 9 10 11
2208 1701f 1336 2129
calcitic
aragonite dolomite marble marl
0.19
ND
0.27
ND
0.01
ND
ND
55.2
0.29
0.03
0.01
42.10
0.37
0.10
0.24
-
-
43.33
5.88
0.15
0.69
2.82
1.75
0.21
15.33
30.82
0.28
0.22
0.10
40.68
0.42
0.04
ND
-
-
41.85
0.85 3.63
0.05
0.20 0.95
0.15 ND
-
0.01
1.4 ND
53.7 46.6
0.14
0.18
0.67
43.4 37!l8 £
0.50
-
- -
5.60
4.04
43.4 47.06
ND - not detected: Limits of detection for Si02, 0.03; Ti02 and MnO, 0.01; A1203, 0.05; Fe203, 0.01;
MgO, 0.10; Cl, 0.02.
b Trace of P20tj is approximately 0.005%.
c Percentage of total iron expressed as Fe203.
d Includes 0.30% H20+.
e Includes 0.61% H20+.
^ Stone No. 1701 came from the same source as Stone No. 1351, studied by various investigators.
-------
35
Previous data were correlated by the expression
1 dn' kv „
Because of inaccuracies involved in measuring particle densities,
P , of finely divided calcines, the rate data are correlated in
this report on the basis of total surface area, S , which is
O
more easily and accurately determined. Since k = k p S ,
v s Hp g'
where k is the reaction rate constant per unit surface with
S
units of cm. /sec. Values of S were obtained from B.E.T.
g
2
measurements and are expressed in cm. /gm. Equation (2) is
the simplified model used for data correlation in this report.
The effectiveness factor is used here to represent the degree
to which reaction occurs within the internal structure of the
solid. The maximum value of n = 1 indicates that reaction occurs
equally throughout the internal pore structure. Low effectiveness
factors, n « 1 are associated with strong pore diffusion resis-
tances and indicate that internal structure does not participate
in the reaction, i.e., that reaction takes place primarily on
the outside surface of individual particles. By analogy with
catalytic reactions, n would be expected to be a function of
particle size, and the ratio of reaction rate/pore diffusion
-------
36
rate (and hence, pore size). All other factors being equal,
n decreases with (1) increasing particle size, (2) increasing
temperature or reaction rate and (3) small pore size.
Three different particle sizes of each calcine were exposed at
980°C. for periods of 5 seconds to 2 hours to obtain additional
parameters which have not been measured previously. These include
effectiveness factors, n; reaction rates at zero sulfation, r ;
the $ coefficient (sensitivity of S0~ sorption rate to sulfate
loading); and total CaO utilization. The estimation of these
parameters is based on plots of log reaction rate versus sulfate
loading, which have been shown to give a linear correlation over
a wide range of solid conversions .
Total CaO utilization was determined for each particle size
by exposure for 2 hours at 980°C.
Samples of calcines were submitted to International Minerals
and Chemical Company for complete physical characterization
(pore size distribution, pore volume and B.E.T. surface) and
these data are summarized in the discussion of results. Calcines
were also submitted to Illinois State Geological Survey for
study by scanning electorn microscope.
-------
37
c. Results
Large differences were found in SO- sorption rates for the calcines
prepared from the various types of limestone examined in this study.
Figure 2-1 compares the reactivities of 150/170 mesh particles
at 980°C. In general, the marble and spars showed the lowest
rates of sorption, whereas the dolomites and marl showed the
highest.
The kinetic parameters for the reaction are summarized in Tahle 2-3.
The activation energies, determined from the Arrhenius plots
given in Appendix A, fall within the range previously reported
for the reaction of calcined stone with S0~, and again showed
distinct differences between stones. The lowest activation
energies were found for Iceland spar (Type 1), and an impure
dolomite (Type 6). The Iceland spar showed other clear effects
of strong diffusional resistance. The high impurity content
of the dolomite may account for its kinetic behavior.
Particle size had a marked effect on sorption rate for nearly
all stones at 980°C. Figure 2-2 illustrates typical sorption
curves for three different particle sizes of one calcine.
Rate measurements obtained from such plots were correlated in
the manner shown in Figure 2-3. Extrapolation of the rate data
to zero sulfation permitted the initial rate, r , to be evaluated
independent of the effects of sulfation. The reaction rate
-------
38
2O
Figure 2-1.
60 80
TIME (SEC)
100
12O
I4O
Comparison of Reactivity of Calcines with SCK at 980°C.
Particle size = 150/170 mesh.
-------
39
Table 2-3. Kinetic Parameters for S09 Sorption by Calcines
(150/170 mesh particles, 980°C.)
ISGS
Stone
Type
1
2
3
4
5
6
8
9
11
APCO
Stone
No.
2201
2202
2203
2204
2205
2206
2208
1701
2129
Activation
Energy, E,
cal./g. mole
9,500
26,500
15,500
12,500
19,500
9,200
14,400
18,100
_
Reaction rate
constant, k ,
cm. /sec.
0.186
0.737
0.228
0.219
0.186
0.068
0.152
0.147
0.16
Frequency
factor, A
, 0
cm. /sec.
8.50
3.18 X
1.19 X
31.2
4.84 X
2.78
47.3
2.20 X
_
>
io4
io2
io2
io2
-------
40
30
27
150/170 MESH
MESH
100
200
500
600
700
300 400
TIME, SECONDS
Figure 2-2.
Sorption of SO- by Different Particle Sizes of Type 4 Calcine
at 980°C.
-------
41
40
30
20
T 1 1 I
10
JB
o>
UJ
*
tr
UJ
oc
0.2
150/170
mesh
_l I
24 6 8 10
SULFATE LOADING, (gm moles/flm ) * I03
12
Figure 2-3.
Estimation of Initial Rate, r , and Effectiveness Factor, n, for the
o
Sorption of S02 by Type 4 Calcine,
-------
42
constants given In Table 2-3 were computed from r for the
150/170 mesh particles, based on the surface area of the unreacted
calcine. The effectiveness factors given in Table 2-4 were
estimated as the ratio of r for the particle size in question
to the value of r for 150/170 mesh particles. This assumes
n - 1 for the smallest particle size. This assumption appears
to be well justified by the experimental evidence for all stones
except the Iceland spar (Type 1) and the marble (Type 10), as
will be discussed later.
The empirical relationship shown in Figure 2-3, which was
previously reported to correlate reactivity at 870°C., was
found in this study to be also valid at 980*C. Effectiveness
factors, however, which had been evaluated at n - 1 for
particles of 42/65 mesh at 870°C., were lower at the higher
temperature used in this evaluation. At 980°C. the value of
n was generally in the range of 0.7 for 42/65 mesh and about
0.3 for 12/16 mesh particles. Log r versus sulfate loading
was linear for small particles to at least 50% CaO conversion
at 980°C.; for two stones, marl (Type 11) and aragonite (Type 8)
results were linear over the full range of conversion.
The capacities of Iceland spar (Type 1) and marble (Type 10)
were especially sensitive to particle size, utilization increasing
by a factor of 2 when particle size was reduced from 12/16 mesh
to 42/65 mesh, and increasing again by a factor of 3 when particle
size was reduced further from 42/65 to 150/170 mesh.
-------
Table 2-4.Parameters for SO Sorption by Calcines as a Function of Particle Size (980°C.)
Stone Particle
Initial rate, r ,
Effectiveness Coefficient 6, Total
Total CaO
Type size, mesh gm.moles/gm.sec.( x 10 ) factor, n
( x 10~ ) Capacity^3' Utilization^
1
2
3
4
12/16
42/65
150/170
12/16
42/65
150/170
12/16
42/65
150/170
12/16
42/65
150/170
0.0090
1.04
5.00
0.274
1.41
1.61
0.673
0.849
1.08
0.649
1.02
1.39
< 0.0018
< 0.21
0.17
0.85
0.62
0.79
0.47
0.74
25.4
30.4
13.4
12.3
10.5
7.75
6.62
3.83
2.77
12,5
4.80
2.89
2.6
6.0
18
17
18
21
20
25
33
6.8
18
33
6.2
13.6
43
40
43
50
49
61
80
16
43
80
U)
-------
Stone Particle
Initial rate, r ,
Effectiveness Coefficient g, Total
Total CaO
type
5
6
7
8
Marl
1336
size, mesh
12/16
42/65
150/170
12/16
42/65
150/170
150/170
42/65
150/170
12/16
42/65
150/170
12/16
42/65
150/170
4
g.moles/g.sec. ( x 10 )
0.610
0.935
1.91
0.195
0.457
0.728
—
0.316
0.316
0.894
0.894
0.894
0.50
_2
factor, n ( x. 10 )
0.32 17.9
0.49 3.93
3.03
0.27 4.58
0.63 3.00
1.92
—
1 2.38
2.38
1 2.14
1 2.14
2.14
, (a)
capacity
9.0
22
22
19
19
19
1.3
36
39
35
35
35
2.0
4.8
14.3
utilization
37
90
90
100
100
100
50
86
93
100
100
100
4.9
11.7
35
(a)
(b)
Milligrams S0_ absorbed by 30 mg. calcine, 2 hr. exposure.
Percent conversion of CaO to sulfate, 2 hr. .exposure.
-------
45
Type 7, a magnesite, showed no SO^ sorption beyond that attribu-
table to CaO impurities for any reaction temperature from 540°C.
to 980°C. To ensure that poor performance was not a result of
dead-burning at the 980°C. calcination temperature, additional
tests were made in which the raw stone was calcined in-situ at
reaction temperature. The maximum reactivity was found at
650°C., at which temperature 6.5 mg. of S0_ was absorbed by
30 mg. calcine in 100 seconds. At 760°C. sorption dropped to
2.6 mg. per 100 seconds exposure. At 540°C., as at 980°C.,
pickup was only about 1 mg. of SO..
•J
d. Discussion of Results
(1) Pore Structure
Since 50 percent or more of the total particle volume of calcined
limestone consists of pore space, it would be expected that the
characteristics of the internal particle structure, such as pore
volume, pore diameter and surface area, would significantly
influence reactivity with S0«. This would be especially anti-
cipated where large particles are concerned. The primary physical
properties of the calcines, determined by American Instrument
Company and International Minerals & Chemicals Company, are
summarized in Table 2-5. These measurements show marked differ-
ences between the properties of the calcines prepared from the
different types of stone. Mean pore size, for example, varied
from 0.07 micron (Type 1) to 4 microns (Type 8): a factor of 60.
-------
46
Table 2-5
Summary of Physical Properties of Calcines
Stone Particle Mean, pere diam., ' Pore volume, B.E.TA Surface,
Type
1
2
3
4
5
6
7
8
9
10
(marble)
11
(marl)
size, mesh microns
12/16 0.075
42/65
150/170
12/16 2.1
42/65
150/170
12/16 0.60
42/65
150/170
12/16 0.44
42/65
150/170
12/16 0.27
42/65
150/170
12/16 0.30
42/65
150/170
12/16 0.01
42/65
150/170
12/16
42/65 4.0
150/170
12/16 0.42
42/65
150/170
12/16 0.065
12/16 1.6
42/65
150/170
cc./g.
0.26
0.23
0.26
0.34(C)
0.27
0.28
0.31
0.29
0.26
0.32
0.31
0.32
0.39
0.37
0.50
0.40
0.40
0.40
0.34
0.39(C)
0.34
0.37
0.39
0.31
0.032
1.19
1.28
m /g.
9.8
9.9
10.2
0.83
0.7
0.7
1.5
1.6
1.8
1.9
2.1
2.4
3.6
3.4
3.9
3.6
3.9
4.1
22.7
29.0
37.8
0.65
0.79
3.4
3.7
3.1
0.63
3.5
2.2
3.5
by mercury intrusion •'excluding voids > ly excluding voids > lOy
-------
47
Surface areas were also quite different, ranging from 0.6 to
2
10 m. /g. for different limestones. Complete pore spectra were
obtained for each calcine; however, since the distribution of
pore sizes in a given calcine covered a narrow range (indicating
uniform-sized pores), only the mean pore diameters listed in
Table 2-5 are used in making comparisons.
For a given calcines pore structures showed no apparent dependence
on particle size. The data of Table 2-5, for example, show that
the total pore volume of smaller particles was not consistently
or appreciably different from that of the larger particles.
Mean pore size also showed no trend with particle size. A small
2
increase in surface area, about 0.4 m. /gm., is evident when
particles are reduced from 12/16 to 150/170 mesh. This is
attributed to the new surface exposed at the points of fracture
when the solid is crushed to make smaller particles.
Previous work has indicated that SO- sorption by small particles
is controlled by chemical reaction rate. If diffusion into the
internal pore structure is fast relative to the reaction rate,
the entire B.E.T. surface area will participate in the reaction.
Under such conditions n = 1 and, from equation (2) the rate of
reaction of different calcines (compared at a given sulfate
loading, temperature, and SO- concentration) would be expected
to be directly proportional to B.E.T. surface area, S . The
O
-------
48
rate curves of Figure 2-1 are reproduced in Figure 2-4 with
the B.E.T. surface of each calcine (determined prior to
reaction) shown in parenthesis. With three exceptions the
reactivity does increase with surface as expected.
Before pursuing this analysis further, some discussion of these
three exceptions, Type 6, Type 1 and Type 10, is in order. In
Type 6, considerable amounts of impurities are present, the
calcine containing about 23 percent inerts. These inerts, in
addition to reducing the amount of CaO per unit weight of calcine,
can presumably block S02 from access to the reactive CaO in a
manner similar to CaSO,, thus affecting the initial reaction
rate as though the stone were partially sulfated. Furthermore,
physical measurements of the raw stones (Table 2-6) show an
unusually high surface area for Type 6, indicating that much of
the pore structure is not attributable to CaO, but rather to
the impurities. As a result of these considerations Type 6
is not further considered in this analysis.
The two other stones, Type 1 and Type 10, have no chemical
impurities which could account for their low sorption rates.
Examination of the data in Table 2-5 reveals that these two
calcines are physically similar with regard to their small
pore size. Both have mean pore diameters less than 0.1 micron.
These small pores, which account for nearly all of the surface
area, are apparently the cause of the slow SO. sorption rates.
-------
49
20
15k
ui
z
o
i
10k
MARL(in-situ)
(3.5?)
T5
(3.9)
1336
(0.63)J
3 20 40 60 80 100 120 140
TIME (SEC)
Figure 2-4.
Comparison of Reactivity with B.E.T. Surface Area (in parentheses,m /gm.)
-------
50
Table 2-6
Physical Properties of Raw Limestones, 10/28 Mesh
Stone
Type
1
2
3
4
5
6
7
8
10
11
Mean Pore Diam.
Microns
40
35
5
2
20
0.3
0.15
0.2
20
1
Pore Volume,
cc^/gm.
0.040
0.019
0.041
0.041
0.016
0.051
0.054
0.069Ca)
0.016
0.89
B.E.T.2Surface
m /gm.
0.19
0.17
0.32
0.46
0.18
2.87
1.54
3.26
0.09
4.48
-42 mesh particle size, voids larger than 20 microns
omitted.
-------
51
Several different sources of evidence show that these small
pores lead to high diffusion resistance and plug easily with
reaction products. For example, the data of Table 2-5 show
an unusually strong sensitivity of capacity to particle size
for these two stones. In no other case were such large increases
in utilization found when particle size was reduced. A test
of sorption capacity with -325 mesh Stone Type 10 (D = < 44y)
showed that 83 percent CaO utilization can be achieved with
very fine particles. This compares to 4.8 percent utilization
with 12/16 mesh stone, which indicates clearly that reaction
occurs with this stone only at the outside surface of the particles.
Comparison of reaction rates (at a given sulfate loading) is made
for several stones in Figure 2-5. Most stones had responses
characteristic of pore diffusion resistance for particles down
to 150/170 mesh (pore diffusion corresponds to a straight line
through the origin). Marl (Type 11) and aragonite (Type 8)
showed complete absence of diffusional effects (horizontal
line). Iceland spar (Type 1} and marble (Type 10), however,
showed the concave-upward response which is characteristic of
shell diffusion on a non-porous solid. These two stones are
the only specimens of approximately 15 examined at APCO in which
shell formation could be confirmed by S0_ sorption characteristics.
This result is logically attributable to quick pluggage of the
small pores with reaction products. The conclusion that SO^
sorption by Type 1 involves shell formation is in accord with
-------
52
MESH SIZE
12/16
42/65
150/170
STONE SULRftTE LOADING
TYPE (gm.motes/ gm.)x^O3
- 4
X
u
o>
«n
*>
O Tt
E 3
E
o«
UJ
tr
z 9
O 2
111
-------
53
the results reported by TVA for Iceland spar calcite as
observed by direct microscopic examination of reacted particles
(March 1968 progress report to APCO).
From the experimental evidence discussed above, it is concluded
that the Iceland spar (Type 1) and marble (Type 10) are, in
effect, non-porous solids. This results in the case of the
marble from a combination of low porosity and small pore size.
Although the Iceland spar has sufficient porosity, the mouths
of the pores are of uniform small size which plug quickly with
reaction product. In both cases the reaction with SCL is
confined to the outside particle surface. It should be noted
that these two stones are not typical. Mercury intrusion tests
on calcines prepared from 85 different limestones generally
available in the United States showed average mean pore sizes
of about 0.3 micron. Calcines with mean pore sizes of 0.1
micron or less are rare, when prepared under these conditions.
Referring again to Figure 2-2, the correlation of reaction rate
data with B.E.T. surface was made by comparison of rates at a
sulfate loading of 7 mg. S0» per 30 mg. calcine. Comparison
at a constant degree of sulfation is necessary to ensure that
the effect of sulfation does not mask out the effect of surface
area, since it has been established that rate decays exponen-
tially with sulfation. The rates evaluated at constant sulfation
are plotted against B.E.T. surface (of 150/170 mesh particles
-------
54
prior to reaction) in Figure 2-6, for reaction at 980°C.
Additional data which is available on other stones are also
included. It is apparent from Figure 2-6 that the reaction
rate does increase with B.E.T. surface area in a manner which
is approximately linear. This result confirms again that
chemical reaction controls the rate of sorption of S0» by
small particles and indicates that n - 1 for D < 0.01 cm.
P
at 980°C.
The presumption that small pores plug more rapidly than large
pores (made with regard to the discussion of Type 1 and Type 10)
was confirmed by an experiment in which samples of a calcine
were reacted with SO- for varying lengths of time, and the pore
spectra determined by mercury intrusion measurements on each
reacted sample. Comparison of the changes in spectra showed
considerable difference in the rate of decrease in pore volume
for pores of different size. The volume of small pores
(between 1 and O.ly) decreased with sulfation at a rate about
2.3 times greater than the rate of decrease in volume of pores
larger than 1 micron.
Surface Biffusion. Ionic Diffusion
A short experiment was made to test for the presence of surface
diffusion resistances and also to test for effects of S0~
diffusion through the CaO lattice as possible rate-limiting steps.
For this purpose, Type 2 calcine (which has large pores and
therefore would be expected to show such effects most strongly)
-------
55
O
UJ
UJ
c»
E
O
10
ow
(O
9
E
A
UJ
.20
.18
.16
.14
.12
.10
.08
2 .06
h-
o
UJ
(T
.04
.02 -
T2bB
T3
1343
T5
T9
1234
B.E.T. SURFACE AREA OF CALCINE, m^/gm.
Figure 2-6.
Reaction Rate vs. Surface Area of 150/170 mesh Particles. Rates
Evaluated at 980°C. and Sulfate Loading of 29.1 x 10~ gm.moles/gm
-------
56
was exposed in the reactor to SO. for short periods followed
by long periods of soaking at reaction temperature but without
SO- present. For example, a sample was exposed for 35 seconds
to flue gas, purged and held at 980°C. for 30 minutes and then
exposed again to SO- for 35 seconds. The amount of sulfate in
the sample was then compared to that absorbed by another sample
exposed continuously for 70 seconds. The results showed no
difference between the two modes of exposure, indicating that
diffusion of sulfate into the solid during long periods of
soaking does not increase the reactivity and therefore solid
diffusion would not appear to be a significant factor. This
experiment was repeated several times with different exposure
times and multiples of exposures with the same result.
Calcination Temperature
Four samples of Type 2 calcite spar CIO/28 mesh) were calcined
in the rotary kiln for 2 hours at temperatures of 930°C. (1700°F),
980°C. (1800°F), 1040°C. (1900°F) and 1115°C. (2040°F). The
physical properties of the calcines thus prepared are shown in
Table 2-7. The maximum surface area was obtained at 980°C. and
the surface area decreased sharply at temperatures above 980°C.
The SO- reactivity shown in the table was determined at 980°C.
with 150/170 mesh particles, by exposing five samples for 300
seconds. The average amount of S02 absorbed again correlated well
with the B.E.T. surface area. The observed reduction in SO,,
-------
57
reactivity is clearly a case of loss of surface area (due to
everburning). Since this stone contains no silica impurities,
chemical side reactions are obviously not responsible for loss
of reactivity.
Table 2-7
Effect of Calcination Temperature on
Physical Properties of Type 2 Calcite
Spar, 12/16 Mesh
Gale.
Temp.
1700°F
1800°F
1900°F
2040°F
Mean Pore
Diara. Microns
1.5
2.0
2.2
3.0
True Density
gm./cm.
3.46
3.43
3.67
3.46
Pore Vol.
cc/gm.
0.35
0.40
0.38
0.26
BET Surface
m2/gm.
0.65
0.83
0.59
0.32
S02 , s
Reactivity
10.7
12.5
10.1
6.2
(a)
mg. 863 absorbed by 30 mg. 150/170 mesh particles in
300 seconds at 980°C.
In-Situ Calcination
The kinetic parameters presented in Tables 2-3 and 2-4 were
determined with stone calcined in a rotary kiln at 980°C. for two
hours. In order to ensure that the (reactivity of such samples
was not altered as a result of their being coaled and stored
for some time prior to reaction, a series of tests was made
in which the stone was calcined in the reactor immediately
preceding exposure to SO-. This was done by placing 150/170
mesh particles of raw stone in the reactor carrier and inserting
the carrier into the heated reactor at 980°C. The amount of
raw stone placed in the carrier was equivalent to 30 mg. calcine
-------
58
(about 52 mg. uncalcined). The sample was thus subjected to shock
calcination when the carrier was inserted into the reactor. After
5 minutes flue gas was admitted to the reactor and the sample
exposed for 100 seconds.
Ten such runs were made with each stone; the average S02 sorption
is compared in Table 2-8 with the sorption found for kiln cal-
cined stone. Although several stones showed significant changes
in reactivity, depending on the mode of calcination, the relative
ranking of the reactivity was generally the same as that obtained
for kiln-calcined material. Several samples, Type 8, Type 10
and Type 11, were more reactive when tested in this manner, the
marl (Type 11) being the most striking example. Several tests
were made with marl in which the sample was left in the reactor
for 2 hours after shock calcination prior to S0? exposure.
These samples showed no loss of reactivity compared to samples
heated for only 5 minutes, thus indicating that the 2 hour cal-
cination used in preparing kiln calcines was not responsible
for the loss of reactivity. Figure 2-7 compares the reactivity
of marl for the two modes of calcination. No other stone
showed such an extreme enhancement of reactivity when calcined
in-site.
-------
59
Table 2-8
(a.)
Comparison of Reactivitiesv ' of 150/170 mesh Stones at 980°C.
Stone
Type
1
2
3
4
5
6
7
8
10
11
2061
Raw Stone
Calcined in- situ
8.8
10.3
11.3
14.4
10.0
10.9
1.1
8.5
5.5
20.4
11.8
Standard Deviation
of 10 in-situ tests
± 0.8
± 0.9
± 0.9
± 0.7
± 0.7
± 0.8
± 0.4
± 0.7
± 0.4
± 2.3
± 0.8
Kiln Calcined
for 2 hours at 980°C
7.7
9.0
13.0
13.7
15.6
10.9
1.3
6.5
2.3
12.5
-
!
Milligrams SO- absorbed by 30 mg. calcine in 100 seconds.
-------
60
20
15
L
o
O»
E
I0
IT
o
CALCINED IN-SITU
CALCINED IN ROTARY
KILN
IN-SITU 650° C. REACTION
26% CALCINED
20 4O 60 80 100 120 140
TIME (SEC)
Figure 2-7.
Effect of Mode of Calcination on Reactivity of Michigan Marl.
-------
61
Reaction Model
It is concluded from the above discussion that a simple model
which assumes chemical reaction to be the rate-limiting resistance
will adequately correlate the data for pure calcines which have
pore diameters larger than 0.2 microns. The expression given
by equation (2) defines the rate in terms of the initial
(unsulfated) characteristics of the calcine:
1 dn' _ „
r=— -rr=kSC n
W dt s g so2
Previous data have indicated, and the present work has confirmed,
that reaction rate decays exponentially as sulfation progresses.
The relationship between rate and sulfate loading, nT/W, can
be described empirically by:
r = e-Bn'/W (3)
This change has been interpreted in terms of a decrease in the
frequency factor A , since
. . -E/RT
k = Ae
s
the experimental results can be expressed by:
A=A e-*n'/W
o
An alternative interpretation of the observed relationship
between rate and sulfation could be made on the basis of changes
in k S n as reaction progresses. It is known that the surface
s g
area and pore volume of the calcine/reaction-product mixture
decreases as sulfate loading increases. This loss of surface,
-------
62
however, is not necessarily due to changes in CaO surface, but
more probably results from the reduced pore size brought about
by the accumulation of reaction products on the pore walls.
It may be argued that the intrinsic surface of CaO is not in
fact altered at all by the presence of CaSO, product. Note
that the linear relationship between rate and surface area at
constant sulfate loading (shown in Figure 2-6) will not
account for the exponential decay of reaction rate.
A comparison of the experimental 3 values also suggests that
changes in the effectiveness factor, due to pore plugging,
likewise will not account for the observed rate decay. The
data for marl and aragonite, both of which have effectiveness
factors of unity, show equivalent slopes of $ ~ 2.2. These
two stones have pores considerably larger than the other stones,
but 3 values of similar magnitude. The fact that the sensitivity
of rate to sulfation is similar despite the difference in pore
sizes, indicates that the accumulation of products in the pores
is not limiting the rate of S02 diffusion into the solid as
long as the pores do not plug. The fact that aragonite and
marl are free of pore diffusion is confirmed by the fact that
particle size does not affect the rate at any degree of sul-
fation up to about 80 percent. It is concluded from this
reasoning that the observed decay in rate of reaction does
not result from pore plugging. A more likely mechanism which
would more closely agree with the observed responses would be
-------
63
a blockage of individual CaO sites by larger CaSO, molecules.
It is concluded from these considerations that the change of
rate with sulfation is best described in terms of the reaction
rate constant and the preferred expression of the overall model
would be in the form:
1 dnf . -en'/W -E/RT _ ,..
W dF = Ao e e CS02 n (4)
Values of A , 3 and n determined experimentally for the stones
tested in this study are given in Table 2-4. This model
adequately describes the sorption of SO- by calcines under
isothermal conditions up to about 50 percent conversion of the
CaO. The limits of applicability of the model, presumed to be
the point at which pore plugging begins (and diffusional effects
become rate limiting rather than chemical reaction), are:
Maximum Percent CaO
Stone Conversion to which
Type Equation (4) applies
1 29 (n < 1)
2 50
3 48
4 50
5 62
6 63
8 77
10 19 (n < 1)
11 86
Note that the 5 purest stones (Tl, T3, T4, T5 and T8) show
fairly close agreement on the value of k . The average "intrinsic
-------
6'4
rate constant" of these five stones has a value:
k = 0.194 ± 0.029 cm./sec.
s
The effectiveness factor, n> is strongly dependent upon pore
size, as would be expected. Figure 2-8 shows the relationship
between the effectiveness factor (evaluated at zero sulfation)
and pore size for 42/65 mesh calcines at 980°C. Pores larger
than about ly correspond to n - 1 for this particle size; when
particle size was increased to 12/16 mesh, TI dropped to about
0.3. Estimates based on these data using the Thiele diffusion
modulus indicate that n values in the region of 1.0 should be
expected for 150/170 mesh particles.
e. Conclusions
The isothermal rates of reaction of S0? differ widely among
calcines prepared from carbonate rocks of varying geological
type. Physical properties of calcines prepared from different
geological types of stones also differ widely.
Many of the observed differences in S0_ reaction rate can be
logically interpreted in terms of the physical characteristics
of the calcine, particularly the size of the pores.
The total S09-sorption capacity of the limestone calcines
studied increased with pore size.
-------
MARL T8
(4,1)
T3
Ui
O.I 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 I.I 1.2 1.3
MEAN PORE DIAMETER,MICRONS
.4 1.5 1.6 1.7
-------
66
The rate of reaction increases with decreasing pore size until
a critical pore diameter of about 0.1 micron is reached.
Presumably pores smaller than 0.1 micron are rapidly blocked
by reaction products. These blocked pores inhibit diffusion
of SO- into the solid interior, resulting in a lower rate.
Maximum rate results when B.E.T. surface area is in the region
2
of 3.5 m. Ig. (corresponding to pore diameters of 0.2 - 0.3y)
under isothermal reaction conditions.
Chemical reaction is the sole limiting resistance for particles
smaller than 0.01 cm. at temperatures up to 980*0. when pores
are larger than 0.2jj. The isothermal reaction of pure limestones
with SO- show an intrinsic rate constant (per unit of surface)
on the order of 0.2 cm./sec.
At conversions of about 50 percent the rate of sorption of S0~
by 90p particles of normal pore size changes from chemical
reaction to a combination of chemical reaction and diffusion.
Both rate and capacity of S0_ sorption are highly dependent upon
particle size. The effect of particle size is not the same for
all stones but is determined primarily by the size of the pores.
Small pores lead to the highest sensitivity between the reac-
tivity of calcines and particle size. Calcines with very large
pores may show no dependence of reactivity upon particle size.
-------
67
No evidence of surface diffusion or ionic diffusion could be
established as a major limitation on SO- sorption.
The relative ranking of calcines prepared from the eleven types
of stones listed in Table 2-2 was established in the following
order with respect to isothermal rate of reaction with SO- at
980°C.: Til (marl) > T5 > T9 > T4 > T3 > T6 > T2 > Tl > T8 >
T10 (marble). In this rating marl was calcined in the reactor.
T7 (magnesite) reacts only slowly with S0_ at any temperature
between 540° and 980°C., under isothermal conditions, and
consequently, T7 was omitted from this rating.
f. Nomenclature
A frequency factor, cm./sec.
A frequency factor at zero solid conversion, cm./sec.
C gas phase concentration of sulfur dioxide, g. moles/cc.
DU —
D mean particle diameter, cm.
P
E activation energy, cal./g. mole
k reaction rate cpnstant per unit surface, cm./sec.
k reaction rate constant per unit volume of solid,
sec. (g. moles/cc.)
dn/dt rate of change of SO- in the gas phase, g. moles/sec.
n' sulfate in solid as SO-, g. moles
R gas constant, 1.987 (cal./g. mole °K.)
r rate of formation of SO- in solid, g. mole/g.(sec.)
-------
68
r rate at zero solid conversion, g. moles/g. (sec.)
2
S B.E.T. surface area of calcine, cm /g.
O
W weight-of calcine sample, g.
T temperature, °K.
t time, sec.
3 empirical correlation coefficient defined by Equation 3
n effectiveness factor, ratio of reaction rate to the rate
that would be obtained if the entire volume of the particle
participated equally in reaction
p bulk density of calcine (particle density), g./cc.
-------
69
3. AField Study of the Role of Overturning of Limestone
a. Introduction
Full-scale boiler-injection tests to evaluate the dry injection
process for removing sulfur dioxide (SCL) from the flue gases
of power plants were begun in late 1969. These tests called
for the injection of additives into the boiler at several points
downstream from the combustion zone. However, a. comprehensive
(33)
TVA report pointed out the desirability of injection directly
into the burners by mixing the additives with the fuel. This
approach was reported to have several advantages over separate
injection. (1) It would allow maximum residence time for the
solid to mix and react with the gases. (2) It would provide
optimum distribution of additive in the high-temperature zone.
(3) It would reduce maintenance problems by eliminating the
need for much of the injection equipment. (4) It would allow
easy control of the additive-to-fuel ratio, and would auto-
matically adjust the ratio to changes in boiler load.
The decision to inject separately during the full-scale dry
injection process evaluation tests, instead of mixing with the
fuel, was based on the results of several experiments in which
additives fed to coal-fired furnaces with the pulverized coal
(13)
produced little desulfurization. For example, Goldschmidt
found that dolomite hydrate removed only 8.9 percent of the S02
when added in stoichiometric proportion to the sulfur. Finely
-------
70
pulverized dolomite injected with coal removed 12 percent of
the SO- when fed to a pilot furnace compared to 25 percent
(2)
removal when the dolomite was fed separately . Limestone
injected through the upper row of coal burners (2700°F), but
separately from the fuel, removed only 20 percent of the S0_
(19)
at twice stoichiometric feed rate . However, the same author
reported desulfurization efficiencies of 50 to 60 percent when
the stone was injected at 2100°F.
It is generally believed that the low efficiency of desulfuri-
zation achieved in the high temperature injection tests was a
result of the phenomenon of everburning. Overturning is the
process whereby a relatively unreactive lime is produced when
calcination occurs at high temperature and is known to involve
both physical and chemical changes in the stone. Overburned
lime can range from the qualitative extremes of hard-burned
to dead-burned depending upon the severity of calcination
conditions. Among the physical changes associated with over-
burned lime, as compared to lime calcined under moderate (or
soft) conditions, are loss of porosity, increased density ,
loss.of surface area, growth of crystallites, and larger mean
(23)
pore diameter . The extreme condition of everburning results
! :
in a lime that has lost all porosity and the bulk density of
which approaches the absolute density of calcium oxi4e, 3.40
grams per cubic centimeter. This dead-burned lime, which is
chemically inert, is produced at temperatures of 3QOO°F or
more' '
-------
71
In addition to physical changes, chemical changes can also occur
in limestone when it is overburned. These changes result from
side reactions in which CaO unites with impurities in the stone
to form silicates, aluminates, and ferrite, which make calcium
unavailable for reaction with S0_. These chemical changes are
of particular concern where limestone injection is contemplated
(34)
for coal-fired boilers because of the possibility of reaction
of the additives with coal ash, which is composed primarily of
the oxides of silicon, iron, and aluminum.
The purpose of this field study was to determine whether additives
are overburned when injected with fuel. This information is
needed because no examination of the stone was made in previous
tests that definitely established that the materials were hard-
burned or dead-burned. Our object was to take the simplest case,
in which the chemical effects of everburning would not be a factor
(by using pure additive and an oil-fired furnace), and determine
whether the physical properties and chemical reactivity of the
lime produced during boiler calcination is significantly different
from that of stone calcined under controlled conditions in the
laboratory.
b. Experimental
Two separate series of injection tests were made using four
different additives (two limestones and two dolomites). The
test boiler was a 300,000 lb/hr., 900 psig, 900°F., Babcock &
-------
72
Wilcox* type FH integral boiler, located at the Bayboro Station
of the Florida Power Corporation in St. Petersburg, Florida.
The boiler fired-No. 6 fuel oil containing 2.3 percent sulfur
at a rate of 10,000 pounds of oil per hour at an operating
load of 150,000 pounds of steam per hour.
The compositions of the stone, determined spectrographically
by Bituminous Coal Research, Inc., are given in Table 3-1.
The possibility of tying up calcium in side reactions was mini-
mized by using pure forms of limestone and dolomite. Because
the ash content of the fuel oil was very low (0.08 percent),
the reaction of the additives with ash was considered an insignif-
icant factor. By using an oil-fired boiler, it was possible to
approach the problem specifically in terms of the physical changes
in the lime that might be related to its reactivity with S0».
During the first series of tests, the effect of boiler load on
the degree of burning of additives injected with the fuel was
investigated. The dry, pulverized additives were mixed with
fuel oil and fed by a constant-displacement pump into the fuel
line that leads to the burners. The feed rate was stoichio-
metrically equal to the sulfur content of the oil on the basis
of the CaO and MgO content of the lime. Each additive was fed
under conditions of both high boiler load (300,00 Ib. steam
per hour) and low boiler load (150,000 Ib. steam per hour).
* Mention of company or product name does not constitute
endorsement by the Air Pollution Control Office.
-------
73
Table 3-1. COMPOSITION OF ADDITIVES, IGNITED BASIS
Component
CaO
MgO
Si02
A100_
2 3
Fe-0,
2 3
Ti°2
Na20
K20
Test series 1
Dolomite
56.0
37.0
4.55
0.60
0.86
0.03
0.25
0.1
Limestone
96.0
1.0
1.13
0.41
0.20
0.03
0.02
0.1
Test series 2
Dolomite j Limestone
57.0
39.0
1.0
0.2
0.02
0.03
0.02
0.1
95.0
1.9
2.15
0.2
0.02
0.03
0.02
0.1
-------
74
At the boiler outlet ahead of the air preheater particulate
samples were drawn by means of stainless steel probes that
fed into miniature cyclones. The temperature of the cyclone
was 450°F. The sample collector was purged with dry nitrogen
to prevent contact of solids with the flue gas during sampling.
Each run consisted of a sample period of about 4 hours.
During the second series of tests, the effects of particle size,
iron content, residence time, and injection temperature on the
amount of S02 absorbed by the additives were investigated.
Stabilized dispersions of limestone and dolomite in light fuel
oil were prepared by Basic Chemicals in two particles size dis-
tributions. Most of the work was done with the finer dispersion,
in which 50 percent (by weight) of the particles were smaller than
2.4 microns and 10 percent larger than 5 microns. The other dis-
persion, which will be referred to in the following discussion as
"coarse", contained particles 50 percent of which were smaller
than 8 microns and 20 percent of which were larger that 30 microns.
A special preparation of fine dolomite was also made containing
4 percent mill scale as an articicial source of Fe»0-. The
additive dispersions were injected at two points: at the burners
with the fuel, and downstream from the combustion zone at a
temperature of 2430°F. The sampling points were located around
the boiler as shown in Figure 3-1 (during separate injection
tests the first sample point was used as the injection site
and samples were taken at the fourth sample position only).
-------
75
2ND SAMPLE
SCREEN TUBES
1ST SAMPLE
3RD SAMPLE
SUPERHEATER TUBES
4TH SAMPLE
BOILER EXIT
(4 PROBES)
rMBUSTION
ZONE
BURNERS
BOILER PLAN AND SAMPLE LOCATIONS, TOP VIEW
Figure 3-1.
-------
76
Water-cooled sampling probes were used at the first and second
positions, but otherwise the sampling equipment was the same
as discussed earlier. The residence time of the particulate
in the probe during sampling was less than 0.1 second.
The temperatures shown in Figure 3-1 were measured at high boiler
load with a water-cooled, high-velocity thermocouple. At low
load the temperatures at sample positions 1 and 2 were 2270°
and 1970°F., respectively.
c. Discussion of Results
Preliminary observations based on the first series of tests
showed that only a small amount of the lime had reacted with
SO- in the boiler. Samples collected at the boiler outlet
during high-load conditions contained 7 percent SO-. The amount
of sulfate found in the stone was consistently greater for
samples collected at low boiler load than for samples collected
at high load, but it never exceeded 14 percent utilization of
the CaO and MgO. Carbonate analyses (evolution method) showed
that calcination of the stone was complete when fed to the
burners during either mode of boiler operation.
The chemical reactivity of the boiler samples was determined
in the laboratory in two ways: as the rate of reaction with
SO- and as the rate of hydration. The S02 reactivity was
measured in a differential reactor by exposing the samples
-------
77
to flue gas containing 0.27 percent S0?. Figure 3-2 shows the
amount of sulfate formed in 30-milligram samples as a function
of exposure time when reacted at 1600°F. These data are for the
limestone fed during the first test series. Data for the same
stone calcined under soft burn conditions (1700°F., 30 minutes)
are shown for comparison. The rate of reaction, measured as
the slope of the curve at a given value of sulfation, was in
all cases greater for the laboratory calcine than for the limes
calcined in the boiler. For the examples shown, the rate of
reaction of SO- with the boiler-calcined limes were one-fifth
the rate of reaction with the soft calcined lime when compared
at a sulfation of 4 mg SO-/30 mg calcine. On the basis of
this comparison, it may be concluded that calcination in the
furnace flame resulted in a lime that was considerably less reac-
tive than stone calcined under less severe conditions.
The total S0_-sorption capacity of stones calcined in the boiler
was measured by exposing samples in the laboratory reactor for
\
2 hours at a temperature of 1800°F. Under these conditions 75
percent of the theoretical conversion was obtained for the
dolomite samples (CaO component), and 65 percent for the lime-
stone. These utilization values were not significantly different
from those obtained with stones calcined under soft conditions.
Because the primary difference was in the rate of SO- sorption,
rather than capacity, it was concluded that the limes were
hard-burned but not dead-burned.
-------
78
14
12
10
'8
d°
CO
6
I I I I I I
BOILER SAMPLE
TEST SERIES
I I I I I T
I I I I I I I t I I I I I
20 4O 6O SO 100
EXPOSURE TIME, SECONDS
120
14
12
10
o
m
£
1
*
I I \ II\ \
LABORATORY CALCINE
_ (30 MIN., 1700°F)
I I I I I I I I I I I
140
0 2O 4O 60 8O IOO 120 140
EXPOSURE TIME, SECONDS
Figure 3-2. Reaction of S02 with Calcined Limestone at
1600°F in Flue Gas, 0.27 percent S02.
-------
79
The rate of hydration test, a standard technique used in the
lime industry for evaluating the degree of burning, was performed
for four boiler samples by G. & W. H. Corson, Inc. A series of
standard laboratory calcines was prepared from the raw stone at
temperatures of 1800°, 2100°, 2400°, and 2700°F. The percent
of hydration versus time was measured for each calcine and compared
to that of the boiler samples. Figure 3-3 summarizes the results
of tests with the dolomite run in the first test series. The
laboratory calcines show decreasing reactivity as calcination
temperature was raised, in accordance with the effects of hard-
burning. The rate of hydration of the boiler limes was lower
than the hardest-burned laboratory calcine prepared at 2700°F.
Limestone samples showed similar results (e.g., 50 percent
hydration of the boiler calcined limes required 202 minutes,
compared to 21 minutes for the 2700°F. laboratory calcine and
1.2 minutes for the 1800°F. laboratory calcine). Although the
accuracy of this test is confounded by the presence of sulfate,
which may also influence the hydration rate, the results
qualitatively confirmed that the boiler-calcined limes were
overburned.
Table 3-2 is a compilation of the chemical analysis data for
samples obtained from the second series of tests. Comparison of
the carbonate and sulfate analyses of samples collected at
sample position 1 with those of samples collected at position 4
(runs 3, 4 and 5) shows that both calcination and sulfation of
the lime were completed in the furnace section before it reached
-------
80
100
80
H
yj
o
60
40
o
o
0 20
I8OO°F
A BOILER SAMPLES
0 LABORATORY CALCINATION
I
10
100
1000
TIME, MINUTES
Figure 3-3. Rate of Hydration of Calcined Dolomite: Test Series 1
-------
81
Table 3-2. TEST SERIES 2: SULFATE AND CARBONATE CONTENT OF BOILER SAMPLES
Run
1
2
3
4
5
6
7
8
9
10
Boiler load, Sample
Additive 000 Ibs. stearo/hr position
i
Coarse dolomite
Fine dolomite
Fine dolomite
Fine dolomite
Fine limestone
Coarse limestone
Fine limestone
Separate injection-
Fine dolomite
Separate injection-
Fine dolomite + Fe_0
Separate injection-
Calcined limestone
305
275
250
150
250
250
150
250
250
250
4
4
4
2
1
4
3
1
4
3
2
1
4
4
4
4
4
Weight percent
Sulf ate ' Carbonate
As SO.^Y As C02
6.5
3.6
3.7
3.2
4.7
13.2
13.6
12.5
4.3
4.2
3.8
5.1
7.7
13.7
13.3
17.0
6.4
1.3
2.0
1.8
0.4
0.4
0.7
1.7
0.3
1.3
0.7
0.04
0.9
2.0
2.1
10.6
30.3
2.4
(a) Calcined basis
-------
82
the first sample position, at which the temperature was 2430°F.
No significant amount of S02 reacted with the lime while it
passed through the boiler section, even though the lime was
completely calcined. These facts show that the rate of calci-
nation was not limiting S0_ sorption, (i.e., the calcination
step was completed prior to entry into the boiler, but there
was no further sulfation even though the S0» capacity of CaO
is greatest in that temperature zone). It was concluded that
the deactivation occurred in the highest temperature region of
the furnace. The additional residence time was not the cause
of higher sulfation at low boiler load.
The coarse additives fed during runs 1 and 6 (Table 3-2) absorbed
more sulfate than did the finer additives. This effect was also
evident in comparing the results of series 1 and 2, greater S02
sorption being attained in the former case where coarser feed
was used. This might be a result of less rapid calcination of
the larger particles, a factor known to affect the density of
lime: higher densities being associated with high rates of
(12)
calcinationv .
The physical properties of boiler samples obtained from both
the first and second series of tests were examined and compared
to those of soft-calcined stones. Mercury porosimetry measure-
ments on 80 limestones and dolomites calcined in the APCO
laboratory show that most of the pore volume is in a range of
-------
83
pore size between 1.7 and 0.1 micron. Pores in this size range
typically comprise about 85 percent of the total pore volume in
soft-calcined limes. Furthermore, for a given stone, the greatest
mercury intrusion occurs over a narrow range of pressure, indi-
cating that the pores are of uniform size. The mean diameter
of these uniform pores can vary from stone to stone but is
usually about 0.3 micron. Porosimetry measurements of the boiler
samples ±n no case showed the large volume of pores of uniform
size found in the laboratory calcines. The specific volume of
pores in the range of 1 to 0.1 micron is compared in Table 3-3
for limes calcined in the boiler during test series 1 and labor-
atory calcines prepared from the same stones. Pores larger
than 1 micron were necessarily excluded from this comparison
because of the fact that interparticle voids cannot be differen-
tiated from intraparticle pores by mercury porosimetry measure-
ments, and the former represent most of the total void volume
in the finely powdered boiler samples. The data of Table 3-3
show that the boiler samples had markedly fewer pores in the
particular size range characteristic of soft calcined lime.
The bulk density and porosity of limes obtained from the second
series of tests were estimated by oil sorption tests. This
test gives a good indication of the total volume of intraparticle
pores when the amount of oil sorbed by the interior of the
particles is large compared to the amount retained on the
-------
84
Table 3-3. SPECIFIC VOLUME OF PORES I TO 0.1 MICRON
DIAMETER (BY MERCURY INTRUSION) AND DEGREE
OF SULFATION OF ADDITIVES: TEST SERIES 1
Dolomite
Laboratory calcine
Boiler low load
Boiler high load
Specific pore
Volume, cc/g
0.37
0.078
0.063
Sulfation.
% SO
none
16.6
7.3
•3—
Limestone
Laboratory calcine
Boiler low load
Boiler high load
0.31
0.16
0.15
none
11.5
7.1
-------
85
particle surface. The powder is immersed in No. 2 fuel oil,
filtered and dried on absorbent paper. The volume of oil
contained in the pores is determined by multiplying the weight
gain by the specific volume of the oil. Laboratory calcines
prepared from the same stone were also carried through the
procedure as a basis for comparison. As shown in Table 3-4,
the boiler samples had high densities and low porosities.
These effects are both associated with hard-burning during
calcination . Comparison of the high- and low-load samples
showed that the greater sulfation achieved at low load could
not be explained on the basis of differences in porosity.
Tliis is confirmed by the data in Figure 3-2, which shows that
samples collected at different boiler loads did not have sign-
ificantly different rates of reaction with S0_. It was concluded
that the degree of everburning was not related to boiler load
and that improved sulfation at low load was the result of some
other variable - possibly differences in oxygen content of the
flue gases - associated with the higher excess air fed at low
boiler load to improve heat transfer. The adverse effect of low
oxygen on the desulfurization reaction has been demonstrated in
pilot-scale experiments on coal-fired furnaces for both separate
injection ' and addition with the fuel^ .
Several supplementary tests were made injecting additives
separately from the fuel by spraying the suspension into the
-------
86
Table 3-4. PHYSICAL PROPERTIES OF BOILER SAMPLES AS DETERMINED
BY OIL SORPTION: TEST SERIES 2, SAMPLE POSITION 4
Dolomite, laboratory
Fine dolomite, run 2,
Fine dolomite, run 4,
Coarse dolomite, run
Limestone, laboratory
Fine limestone, run 5
Fine limestone, run 7
Coarse limestone, run
calcination^3'
high load
low load
1, high load
calcination
, high load
, low load
6, high load
Bulk density,
gm/cc
1.62
2.18
2.24
2.05
1.52
2.07
2.00
2.19
True density,
gm/cc
3.36
3.32
3.37
3.30
3.35
3.27
3.07
3.17
Porosity,
cc/cc
0.52
0.35
0.34
0.39
0.55
0.37
0.34
0.33
(a) Calcined 2 hours at 1700°F
-------
87
furnace at position 1 at high boiler load. A comparison of runs 3
and 8 in Table 3-2 shows that fine dolomite absorbed 3 times as
much sulfate under these conditions as when fed to the burners.
This was true even though the residence time was shorter and
calcination was not complete. The addition of 4 percent iron
oxide to the dolomite that was injected separately improved the
S02 sorption slightly, as shown by run 9. A test in which
precalcined lime was injected separately gave poor results;
only half as much SO. was absorbed by this material as was absorbed
by uncalcined dolomite. This observation verifies the findings
(2 8)
of others ' that no advantage is gained by precalcination
before injection.
The predominant form of sulfur in the lime was sulfate whether
injection was made with the fuel or separately. Measureable
amounts of sulfide (averaging 0.27 percent by weight as S) were
found in the 13 samples fed with the fuel. Limes from the
separate injection tests contained about 0.7 percent sulfite
I
(as S02) but no sulfide.
d. Conclusions
The results of these tests have shown that the dry-limestone
process should not be applied by injection of the sorbent with
the fuel. Instead, additives must be injected separately to
-------
88
achieve most efficient utilization of the limestone. Overturning
is at least partly responsible for the low efficiencies found
when additives are fed to the burners. The lime produced by
injection with the fuel is much less reactive with S0_ than
lime that is not calcined in the combustion zone. This loss of
reactivity is associated with changes in the physical structure
of the stone during calcination in the combustion zone. These
physical changes are sufficient to deactivate the lime, regardless
of further side reactions with impurities or ash. Additives
injected into coal-fired furnaces with the fuel would therefore
be expected to yield only less-satisfactory results than those
reported here.
Boiler load was found to be an important variable affecting
desulfurization when additives were fed with the fuel. This
is apparently due to the higher excess air used during low load
and is not a result of differences in flame temperature or
residence time.
These field tests indicate that there is an optimum particle
size as well as an optimum injection temperature. The results
under the conditions of these tests show that injection
temperatures somewhat higher than 2400°F. would be best for
2-micron particles.
-------
89
4. Methods for Testing the Degree of Overburning
of Calcined Limestones
a. Introduction
Currently, the dry injection process is being evaluated in a
full scale demonstration of performance at the TVA Shawnee Steam
Power Plant, Paducah, Kentucky. Optimum conditions for sulfur
dioxide abatement will be determined as well as the effect on
operation of the power plant and the economics of the process.
(22 32)
Under the process evaluation planned by TVA ' , carbonate
rock (limestone or dolomite) is ground to 70% less than 200 mesh
and is injected into the furnace above the fireball at tempera-
tures ranging from 2400°F to 2900°F. The carbonate rock calcines
(evolves CCO and the resultant lime reacts with the sulfur
oxides in the presence of oxygen to form calcium sulfate. The
calcium sulfate is removed with the fly ash in the standard dust
collection systems. It has been supposed that the petrographic
nature of the lime accounts for its level of chemical reactivity,
and moreover, that the reactivity of the lime is dependent in
part on its calcination conditions. An otherwise reactive
calcitic or dolomitic lime may become unreactive due to dead-
burning or everburning of the limestone. Causes of this loss
of reactivity may be physical changes or side reactions with
impurities in the carbonate rock or both.
-------
90
As a limestone is calcined at higher temperatures, it loses
reactivity as well as surface area and porosity . Volumetric
t-j 23)
shrinkage as high as 52.6% has been noted ' for an oolitic
stone calcined at 2450°F . Loss of surface area with increased
calcination temperature has been reported * ' * such
2 2
as a decline from 62.9 m /g. to 0.20 m /g. in specific surface
area for a corresponding increase from 750°C to 1300°C in ealc-
(24)
ination temperature for Iceland spar calcite . Corresponding
to the volumetric shrinkage there is an increase in particle or
apparent density ' ' from a low of 1»5 gm/cc for lime pre-
pared at low temperatures up to 3.0 gm/cc for limes prepared at
very high temperatures.
Changes in porosity (i.e., fraction of lime particle which is
pore space) with increased calcination temperature include loss
of porosity(3' 6> 7> 21> 24' 29J and increase in the mean pore
(24 29)
diameter ' . Thus, the large pores grow at the expense
of the small pores which contribute greater porosity. In con-
junction with the growth of pores there is a loss of porosity
attributable to the growth of large crystals by assimilating
small crystals ' . As a result of increased calcination
temperature, the reaction rate with sulfur dioxide or hydro^-
chloric acid is retarded ' , the C02 absorption is lessened^ ,
the slaking rate is slower , and the extent of sulfation
. , . (20, 25)
declines '
-------
91
Several tests for the reactivity of calcium oxide have been
widely used already. Two of these tests, the ASTM slaking
rate test and the G. and W. H. Corson X-ray hydration test rely
on decreased reactivity of overburned lime with water. The
former test follows the reaction rate from the temperature rise
during slaking in an insulating bottle. Another rate test,
the coarse grain titration test, depends on titration at one-
minute intervals with hydrochloric acid to construct characteristic
titration curves. The titration curve for a lime calcined at
high temperature will rise far more slowly than that for a lime
(27)
calcined at low temperature. A fourth test proposed by Y. Ohno
would expose the test sample to CO- and judge the degree of
everburning from the extent of recarbonization.
In order to determine if the limestone injection at the Shawnee
Plant led to optimum calcination of the limestone, it was necessary
to identify a test for lime reactivity using small samples of
limes diluted with fly ash. In addition to considering the tests
listed above, many other possible tests were evaluated. The
procedure was to determine various chemical and physical pro-
perties of different limestones calcined at a range of tempera-
tures. The more promising tests were then applied to samples
from tests at Babcock & Wilcox and from preliminary injection
runs at the Shawnee Plant.
-------
92
b. Experimental
The carbonate rocks tested were two calcites and a dolomite which
have been identified as Stone Nos. 2061, 2062, and 2069, respec-
tively. An analysis of these is listed in Table 4-1. Two sets
of these stones were sent to G.. and W. H. Corson, Inc., Plymouth
Meeting, Pa., where they were batch calcined at 1700, 2000, 2300,
2600, and 3200°F in a rotary kiln. The stones were calcined
for two hours at temperature after the initial heat-up period
(see Figure 4-1). The first set (Set I) was returned as cal-
cined; the second set (Set II) was ground to minus 170 mesh at
Corson1s. All grinding operations were done under nitrogen and
samples were shipped in cans sealed under nitrogen.
Some of the tests were performed by contractors. General
Technologies Corporation was sent samples of the first set for
infrared (IR) analysis. To American Instrument Corporation
(AMINCO) went minus 70, plus 140 mesh (70/140) samples of the
first set and 170/270 samples of the second set. Test results
reported by AMINCO were surface area by the B.E.T. (Brunauer,
Emmett, Teller) method and mercury penetration porosimetry
data. The latter yielded information such as density, porosity,
distribution of pore size, and median pore size.
The tests conducted in the APCO laboratory may be divided into
two groups - chemical reactivity tests and physical property
tests. The chemical reactivity experiments were as follows:
-------
93
Table 4-1. Composition of Limestones
Analysis, weight per cent
Stone Number
Component
Si°2
Al_0_
2 3
Fe00,
2 3
MgO
CaO
Ti°2
Na2°
K20
MnO
2061
1.40
<0.2
0.27
1.77
95
0.03
<0.02
<0.1
<0.03
2062
2.25
<0.2
0.31
2.42
94
0.03
<0.02
<0.1
<0.03
2069
2.63
0.72
0.70
36.0
58.0
0.05
0.06
0.47
<0.03
-------
4.0
3.5
3.0
2.5
Figure 4-1
TYPICAL FIRING CYCLE
tn
2.0
ui
1.0
0.5
i . i
1000 1500 2000
TEMPERATURE °T.
2500 3000 3500
-------
95
1. Absorption of sulfur dioxide from flue gas
2. Absorption of sulfur dioxide from pure SO
3. Absorption of carbon dioxide
4. Absorption of steam
5. Modified coarse-grain acid titration
6. Hydration-weight gain.
The physical property tests were density determinations made
in one case by using an air pycnometer and in the second, by
immersion in oil. A description of the procedure for these
experiments is given in Appendix B.
Preliminary to the final results all proposed overburning or
dead-burning tests were checked for their feasibility in the
present TVA Shawnee injection test application. The most
severe limitation was. that the sample size could not be larger
than several grams. Both the slaking rate test ' and the
coarse grain titration called for sample sizes an order of
magnitude larger. In the latter case, it was found that the
test could be modified and still give reproducible results.
Unfortunately, this was not the case for the slaking rate test
which had to be eliminated for consideration in this application.
c. Discussion of Results
(1) Effect of Calcination Temperature
The results of the various physical and chemical tests show
that apparent densities, median pore diameter, and average IR
-------
96
band shift increased with increasing calcination temperature
while all other properties decreased.
Results of the acid titration test for Stone No. 2061 are shown
in Figure 4-2. All of the chemical and physical properties were
significantly affected by calcination temperature; however, the
most dramatic changes were in SO- absorption from flue gas, pure
S02 absorption, pure CO- absorption, hydration-weight gain,
B.E.T. surface area, and pore volume. All of these properties
decrease in about the same way except for CO- absorption which
drops abruptly at 2000°F. This behavior has been noted before
(27)
by Y. Ohno and has been attributed to the "shrinking,
melting, and fusing of the crystals' , thus, obstructing
carbon dioxide penetration into the lime.
Statistical analysis of the data shows that such properties as
S0_ absorption from flue gas, pure SO- absorption, B.E.T. surface
area, and pore volume change significantly with calcination
temperature but not with stone sample. Hence, the data of all
three stones of a set of calcines were taken together to calc-
ulate correlation coefficients between different properties.
The results shown in Table 4-2 demonstrate the high degree to
which sulfur dioxide reactivities, surface area, and pore volume
are intercorrelated (also refer to Figures 4-3 to 4-7 which are
the crossplots for the data from Set I). This evidence supports
the conclusion that loss in surface area and pore volume at
higher calcination temperatures account for loss in reactivity.
-------
o:
.3
".{
i
Z
9
o
o
o
Figure 4-2
ACID TITRATION OF ^2061
CALCINE TEMPERATURE
+ 1700° F
D 2000° F
o 2300° F
A 2600° F.
X 3200° F
1 t 1 1 1 1 1 1 1 1 1 1
1 1 1
68 10 12 14 16 18 20 22 24 26 28 3O
TIME IN MINUTES
-------
98
Table 4-2: Correlation Coefficients
Flue Gas
Absorption
mg./30 mg.
Pure SO,, % gain
Absorption
B.E.T. Surface
Area, M /g.
Pore Volume
cc/g.
0
0
-.
3 «0 •
t-l rf» 00
P»« < B
0.807
0.863
0.858
Set I
Set II
-------
CT1
p^
en
*
M K
Q£flD BURN STUDY/CROSS PLOT NO 31
B.E.T. SURFACE AREA VS. FLUE GAS CAPACITY
*
(\J
K
-, *
LJ
(X
L_
cn"""".
n
UJ
••o
QQO
| ^ | | f
1.00 .3.00 5.00 7.00 9.00 11.00 13.00 15.00 37.00
FLUE GRS CflPRClTY CMG/30MG3--X
-------
u?_
ir\ —
UJ
cc
o
o
[OD BURN STUDY/CROSS PLOT NO 35
HG PORE VOLUME VS. FLUE GAS CAPACITY
*
m
§
~ • —)• T—" • i 1 1 • • i - • r—• ]
1.00 3.00 5.00 7,00 9.00 11.00 33-00 J5.00 17.00
FLUE GflS CflPflClTY CMG/30MG]=X
-------
r-
•T -,
r~\
O
\
u
C_J
21 '•
ID
_J
O
i
Q_
_J
_J
a:
So
4-5= DERD BURN STUDY/CROSS PLOT NO 36
SMALL PORE* VOLUME VS. FLUE GAS CAPACITY
)K
* *SMALL PORE - PORE HAVING
A DIAMETER LESS
* THAN TWO MICRONS
K
j'.OO 3.00 S.OO '/.DO 9.00 11.00 13.00 15,00 J7.00
FLUE GflS CflPflCITY CMG/30MG)=X
-------
LD
\
LJ
C_J
Z)
_1
O
"°-
_ I
COO
4-6: DEflD BURN STUDY/CROSS PLOT NO 47
SMALL PORE* VOLUME VS. S02 CAPACITY
JK
K TSMALL PORE- PORE HAVING
A DIAMETER LESS
THAN TWO MICRONS
• I I I I I I I I
7.00 15.00 21.00 31.00 39.00 47.00 55.00 63.00 71.00
502 CflPflCITY CPERCENTD-X
-------
O
S-i
c— 1
o
o
>-
I!
<—>
l~
•ZL
LJ
LJ
OC
.
to
Hwr.4-7, CERD BURN STUDY/CROSS PLOT NO 27
PURE S0t CAPACITY VS. FLUE GAS CAPACITY
o
u>
X.
r r~ ~r~ " i ""~ r \~ i
.00 3.00 c>.00 7.0(1 9.00 11.00 13.00 IV 00 .17.00
FLUE GflS CnPnCITY IMG/30MGD-X
-------
104
(2) Effect of Test Conditions
Upon review of the data on the effect of calcination temperature,
three reactivity tests were selected for further study; they
were pure C0_ absorption and pure S0» absorption and hydration-
weight gain. Possible parameters for the first two included
reaction time, reaction temperature, gas flow rate, and particle
size. The effect of particle size was studied first by finding
the absorption for samples of 70/140, 140/200, 200/270, and
-270 mesh ranges.
Carbon dioxide absorption was almost independent of particle
size but sulfur dioxide absorption was slightly lower (less than
10 per cent) for the largest size ranges. On the other hand,
carbon dioxide absorption was affected drastically by the temp-
erature of the absorption test whereas sulfur dioxide absorption
was not (see Figures 4-8 and 4-9). Both reactivity tests were
independent of gas flow rate and weakly dependent on reaction
time after ten minutes exposure (refer to Figures 4-10 and 4-11).
Possible parameters for the hydration-weight gain test were slaking
time, slaking volume, drying time, and drying temperature. Effects
of the last two are not significant while the other effects are
i '
illustrated in Figure 4-12. Increasing either slaking time or
slaking volume of distilled water increases the hydration-weight
gain even for extremely hard-burned stones.
-------
4O
105
Figure 4-8
C02 TEST
CALCINED AT 1700° F.
CO
30
h-
Z
UJ
u
cc
UJ
(L
10
x CALCINED AT 2000° R
I
500
600 TOO
REACTION TEMPERATURE °C.
800
-------
106
Figure 4-9
S02 COMBUSTION BOAT
TEST
80
70
< 60
CD
h-
CD
CALCINED AT 1700 °R
o
QC
UJ
Q.
40
30
CALCINED AT 2000° F.
20
10
1400 1500 1600 1700 1800 I90O 2000
REACTION TEMPERATURE °F.
-------
107
Figure 4-10
C02 TEST
CALCINED AT 1700° F.
o
CALCINED AT 20OO° F.
X
120
REACTION TIME-MINUTES
-------
80
70
§ 60
50
8 40
tr
u
CL
30
20
10
108
Figure 4-11
S02 COMBUSTION BOAT
TEST
CALCINED AT 1700° F.
CALCINED AT 2000° F
X.
J_
_L
J.
10 20 3O 40 50
REACTION TIME- MINUTES
60
70
-------
109
Figure 4-12
SLAKING TEST MAIN EFFECTS
30
28
26
24
22
20
18
1 l6
I l4
i 12
g 10
o
er 8
6 -
4-
2 -
I
2000° F
CALCINE
3200° F
/ CALCINE
X - I ML SOAK
o - 5 ML SOAK
0.5 1.0 1.5 2.0
SOAKING TIME - HOURS
2.5
-------
110
(3) Effect of Partial Sulfation and Fly Ash
Besides the samples already listed above, certain samples exposed
to SO -containing flue gases were received from Babcock and Wilcox
X
(B&W). Under a contract to APCO, B&W injected the Stones No. 2061,
No. 2062, and No. 2069 into their pilot scale coal-fired furnace.
Samples of lime-fly ash mixtures from the cyclone downstream of
the furnace were tested by all of the methods mentioned above.
For the tests based on steam absorption and acid titration, the
measured reactivity was very low or there was none at all.
Results for physical property tests of these samples were dif-
ficult to interpret in comparison to results for pure limes.
To verify the exact relationship between the extent of sulfation
and the degree of dead-burning measurements, the following experi-
ment was conducted. Samples of Stone No. 2061 were exposed to
either air or flue gas at 1800°F for varying lengths of time.
The reactivity of the samples was determined using the C0_
absorption test and the hydration weight gain test. The measured
reactivity for samples exposed to flue gas was corrected for
the calcium sulfate content, i.e., the calcium oxide content
which has already reacted and is unavailable for the dead-burning
test reactivity measurement. After this correction, the degree
of dead-burning as measured for the samples exposed to air and
for those exposed to flue gas should be the same. This is the
case for the hydration weight gain test but not for the C09
absorption test. As shown in Figure 4-13 there is a large
-------
lii
iu 80
70
60
!«
40
LU
O
X 30
O
5
1 20
cc
10
Figure 4-13: COMPARISON OF MEASURED
REACTIVITIES FOR SULFATED AND
UNSULFATED LIMES
AIR EXPOSURE-
UNSULFATED SAMPLES
F.LUE GAS EXPOSURE-
SULFATED SAMPLES
\5 30 45 60 75 90
TIME OF EXPOSURE, MINUTES
IO5 I2O
-------
112
deviation between the measured reactivity for the air exposed
samples and that for the flue gas exposed samples when reactivity
is determined by- CCL absorption. As shown in Figure 4-14 there
is no such deviation for the hydration weight gain test; thus,
this test is independent of the extent of sulfation.
d. Conclusions
Four reactivity tests - namely, flue gas absorption, SO,, absorption,
C02 absorption, and hydration weight gain are suitable as dead-
burning tests for samples of limited size diluted with fly ash.
The later three are sufficiently uncomplicated to be used as
field tests. For sulfated samples, the hydration weight gain
test is recommended since it is independent of the degree of
sulfation. Excellent correlation between SO- reactivities,
B.E.T. surface area, and pore volume support the conclusion
that dead-burning can be accounted for by loss of surface area
and pore volume.
-------
113
Figure 4-14
40
36
32
. 24
X 20
o
16
Si
g 12
X
8
REACTIVITIES OF SULFATED AND
UNSULFATED LIMES MEASURED BY
HYDRATION WEIGHT GAIN
REACTIVITY BY HYDRATION
EXTENT OF
SULFATION
X-AIR
0-FLUE GAS
EXPOSURE
A-PER CENT S02
20
18
16
14
12
10
8
15 30 45 60 75
TIME OF EXPOSURE, MINUTES
90
-------
114
5. Capacity of Limestone for Sorption of SO,.
a. Introduction
The low cost and widespread availability of naturally occurring
carbonate rocks, particularly limestones*, has led to interest
in their use as reactants for desulfurization of combustion gases.
(2B)
Potter, Harrington, and Spaite have presented some of the
background and engineering considerations for development of a
limestone desulfurization process. In the development of the
process, it became apparent that some limestones are better
absorbents than others and that selective recommendation of
limestones would be necessary to optimize full scale field trials
(29)
in power boilers. Potter has presented results of the
present study which was designed to determine (1) the differences
in the sulfur dioxide reaction characteristics of a large number
of limestones, and (2) the physical and chemical properties
responsible for these differences.
A fixed-bed reaction test was selected because of its relative
simplicity. However, the fixed-bed approach is not easily
suited to reaction rate studies because of the changes in
conversion and S0_ partial pressure throughout the bed: There-
fore, the work has been limited to determining the ability of
a sample to achieve a high absorption efficiency. In this
* In this report "limestone" includes all carbonate rocks
containing magnesium and calcium. High purity CaCO is
termed "calcite". 3
-------
115
report, absorption is measured as bed weight gain and is
generally expressed as capacity.
Using the results of preliminary tests of ten samples ,
standard calcination and reaction conditions for comparing
capacities were selected. Since the early work showed little
correlation between sulfur oxide capacity and chemical com-
position of the limestone, in the work reported here characteri-
zation of physical properties and classification of samples
was emphasized.
b. Exp er imental
(1) Sample Preparation and Analyses
Crushed samples of limestone were screened to obtain a -18 +20
U. S. mesh cut (mean opening size 0.92 mm). After a crushed
stone was thoroughly blended, a representative sample was
analyzed spectrochemically. The carbonate content of the rock
was determined by dissolving a portion of the raw stone in
IN HC1.
The calcined material for the standard screening studies was
prepared in a bench scale calciner consisting of an Inconel
pipe rotating in a muffle furnace. A 180 g. charge was fed to
the preheated kiln and held at 980°C for 4 hours. To form
hydrates, hot water was slowly added to the calcined stone.
After the cake was dried to constant weight at 110°C, it was
-------
116
crushed and scteened to obtain -d8 +20 U.S. mesh hydrate pellets.
Thirty-nine of the calcined stones were tested for B.E.T. surface
area and mercury pore distribution. The bulk specific gravity
was determined by an oil sorption test (ASTM test C-127-54)
modified for use with No. 2 fuel oil.
(2) Capacity Determination
The fixed bed reactor system used to determine sulfur oxide
capacity is shown in Figure 5-1. No. 2 fuel oil (C:H = 6.8 by
weight) blended with controlled amounts of carbon disulfide (CS»)
was burned at 17% excess air to give a flue gas with the following
average composition by volume: 10.5% CO-, 3.4% 0-, 9.9% H20,
0.27% S02, 0.003% S03, and 75.9% N2>
The 35-mm diameter Inconel reactor was preheated to the desired
temperature, 980°C for the standard test condition, before the
sample was dumped from the top of the reactor onto a 30 mesh
screen in the center of the reactor. A 20-gram charge of cal-
cined stone is stable at 980°C. For uncalcined and hydrated
samples a weight equivalent to 20 grams of calcined stone was
used. The flue gas was passed through the bed at a rate of 425
i :
standard liters per hour for 3-1/2 hours. In an oxidizing I
atmosphere at 980°C the net reaction is
CaO, N + 1/2 0. + SO- + CaSO.
2(g) 2(g) 4Cs)
-------
117
Figure 5-1
TBIP ERA TUBE CONTROL
CONDENSATE
DIFFERENTIAL
PRESSURE VALVE
Flow diagram for fixed-bed reactor system
-------
118
The S0_ content of the effluent was monitored continuously
with an infrared analyzer. At the end of the exposure period,
air was admitted to the reactor to inhibit side reactions.
The final weight of the bed was then determined. The sample
size, flow rate, and contact time were selected so that even the
most reactive stones had nearly ceased reaction. Therefore,
the data give a comparison of the saturation conversions of
the samples.
The bed weight gain of a 20-gram calcined sample (expressed as
g. SO-/100 g. of calcined stone) was selected as the index of
sulfur oxide capacity. The bed weight gain, also referred to
as "loading" or "capacity", is the weight of SO- removed from
the flue gas. No side reactions with carbon dioxide or water
were observed during the cooling and handling of the reacted
sample. In order to determine the reproducibility of the test,
single replicate tests were made on 22 different samples. The
standard deviation was found to be 2.98 g./lOO g. Since loadings
varied between zero and 85 g./lOO g., the test is considered to
be sufficiently sensitive and precise.
-------
119
c. Discussion of Results
(1) Form of Sorbent and Reaction Temperature
The carbonate, oxide, and hydroxide forms of two samples were
tested at 430, 705, 980, and 1095°C. The results are given in
Figure 5-2 for a high-calcium stone and a dolomite, designated
A and B, respectively. This work shows the relative performance
and temperature sensitivity of the different forms of the stone
and also the relation between optimum capacity and capacity at
980°C, the temperature at which the bulk of the experimental
work was done.
The hydrates have greater capacity than the oxides and carbonates
throughout the temperature region examined. In an air pollution
control process, the additional capacity of hydrates has the
greatest relative advantage at low temperatures where reduced
gas volumes, lower viscosity, and cheaper materials of construction
might offset the cost of sorbent preparation. Below 900°C the
capacity of the calcined stone is greater than that of the carbo-
nate. Since the low temperature (400 to 800°C) loadings are
substantially different from the high temperature loadings,
conclusions presented here are most valid for temperatures in
the region of 1000°C.
Below 1000°C the capacity of the carbonates falls off rapidly
with decreasing temperature. Above 1000°C the loading generally
decreased with increasing temperature. The reduced capacity at
-------
120
Figure 5-2
REACTION I(«»l**IUH[, '
Reaction temperature and form of reactant
Figure 5-3
11
•« 1
5
I
*" *
I
8 ,
* '
__
-
_
-
1
0 1 10 IS J
REACTION TEMPERATURE. MO'C
—
. js
AVERAGE CAPACITY OF U SAilPL ES.
\\
T|1-T_,
5
4
"V •
2.25 9 SPs p* 100 g
psa PTisa
^
;•
i
•:;:
—
— ]
•- t
0 H » 35 01 *S SO SS W 41 70 75 ID 1
AVERAGE CAPACITY, ff 10, ,., IOC ,
Frequency distribution of capacities
of 86 stones tested
-------
121
higher temperatures may result from changes in physical properties.
For example, significant crystallite size growth has been reported
above 1000°C . Reduced capacity may also be due to thermo-
dynamic limitations: little CaSO, formation is expected above
f O / \
1230°C in combustion flue gases . Disregarding these reactant
considerations, there is little incentive for higher temperature
desulfurization because increased gas volumes and viscosity
coupled with more costly materials of construction would result
in higher operating costs. The 980°C standard test temperature
seems to be near the most practical temperature for a high-
temperature, SCL control process based on the use of uncalcined
limestone.
(2) Calcination
A series of fixed-bed calcinations were made at 1316°C for 16 hours
to evaluate the effect of high-intensity calcination on the SO
X.
capacity and physical properties of calcined stones. The results
are given in Table 5-1. There was in all cases a loss in surface
area and an increase in the mean mercury intrusion pore size
when the calcination temperature was increased from 980 to 1316°C.
Loss in pore volume and surface area probably resulted from shrin-
(23)
kage and crystallite growth . The variations in SO, loading
that resulted from changes in calcination procedure are well
explained by the volume of pores larger than 0.3y. The minimum
pore diameter measured by the mercury porosimeter is 0.017u.
-------
Table 5-1. EFFECT OF CALCINATION
Volume of pores
mple
C
D
E
F
G
Hg pore volume
(cm3/g)
985°C
0.31
0.24
0,34
0.38
0.43
1316°C
0.21
0.25
0.07
0.32
0.14
B.E.T. surface area
(m2/g)
985°C
9.2
8.4
4.1
1.7
2.0
1316°C
1.1
1.1
1.0
0.1
0.6
Mean Hg pore size
(microns)
985°C
0.06
0,10
0.78
0.92
0.62
1316°C
0.35
1.3
6.0
18.0
3.5
larger than 0.3y
(cm3/g)
985°C
0.03
0.10
0.33
0.35
0.43
1316°C
0.18
0.18
0.07
0.32
0.14
SOX capacity
(g S03/100 g
985'C
13.6
26.8
29.0
47.0
63.0
1316°'
20.1
31.5
9.15
18.1
24.6
to
fo
-------
123
Those samples whose volume of large pores was reduced had a
corresponding loss of capacity, whereas the two samples whose
S0~ loading was increased as a result of high-temperature calc-
ination had an increased volume of pores larger than 0.3p.
Each of the 81 samples was tested in the fixed-bed reactor in
both the uncalcined and calcined conditions. The correlation
coefficients* between capacities of uncalcined samples and
capacities of calcined samples was +0.734. The mean loadings
of uncalcined and calcined samples were 41.7 and 43.1 g. SCL/
100 g. of sample, respectively. These values suggest that at
980°C the relative capacity of a sample is, on the average,
the same in either the uncalcined or calcined state. Analysis
of variance tests on 33 combined samples and on 8 individual
classes did not show statistically significant differences
between uncalcined and calcined samples. Because of the equi-
valence of the capacities of uncalcined and calcined stones,
averages of tests on uncalcined and calcined samples are used
for each stone in the balance of this discussion.
* The correlation coefficient, r, is an index of the strength
of the linear relationship between the changes of two variables.
For example, r = 0.65 means that 65% of the time a positive
change in y is accompanied by a positive change in x.
-------
124
(3) Capacity Correlations
The average SO, loadings of 86 samples reacted at 980°C are
summarized in a frequency distribution chart (Figure 5-3).
It is apparent that the capacities of various limestones for
sulfur oxides differ widely. The average loading of 42.3 g./
100 g. for the -18 +20 mesh particles corresponds to 30%
utilization for pure CaO. However, since the average CaO
content of the* samples was only 65%, the actual average utili-
zation of the CaO was 45%. Incomplete conversion of the
reactant suggests diffusion limitation by the surface layer
of reaction products. Electron microprobe analysis has verified
/Q\
the existence of such a layer . However, our studies with
different particle sizes suggested that the depth of penetration
of sulfur is not constant. Further work is necessary to define
the limiting mechanism.
The factors that best explain the wide variations in SO- loading
at the standard test condition are summarized in Table 5-2.
With the exception of ignition loss and amount soluble in HC1,
the physical and chemical analyses were performed on the calcined
material. Since a large number of samples was tested and
analyzed, it was found that the values of the correlation coef-
ficient did not change greatly with further increases in the
number of samples considered. The absolute values of r are not
as important as their relative magnitudes.
-------
125
Table 5-2. CORRELATIONS BETWEEN PHYSICAL
PROPERTIES AND 980°C CAPACITIES
Independent
variable
% CaO
% MgO
% Loss on
Ignition
% of Raw stone
soluble in HC1
B.E.T. surface
area (m /g)
Mean Hg pore
size (micron)
Area of pores
>0.3y (m2/g)
% Fe203
Bulk specific
gravity
Volume of pores
>0.3n (cm3/g)
Total Hg pore
volume (cm^/g)
% Oil
absorption
Number of
samples analyzed
59
59
39
39
39
39
39
59
39
39
39
39
Correlation
coefficient
-0.018
-0.043
-0.035
-0.202
-0.246
+0.398a
+0.400a
+0.429b
-0.507b
+0.620b
+0.635b
+0.643b
Significant at 95% confidence level
Significant at 99% confidence level
-------
126
Surprisingly, the chemical composition and carbonate content
(i.e., loss on ignition and amount soluble in HC1) explain
little of the variation in performance of the samples. Of
the chemical components only the iron is significantly related
to capacity at or above the 90% confidence level. Whether the
contribution of the iron is chemical or physical is not clear
at the present time.
The fact that the nitrogen adsorption surface area does not
correlate with capacity indicates that pores in the angstrom
size range do not significantly contribute to the fixed-bed
reaction. On the other hand, a number of physical tests
show that pores in the micron range account for most of the
reactivity. The correlation of the mean mercury pore size
indicates the relative importance of larger pores. The area
of pores larger than 0.3p was calculated by integration of
mercury intrusion data . A trial-and-error calculation with
12 of the samples showed that the arbitrary 0.3u integration
limit gave the best correlation with loading. Clearly, this
partial surface area is more directly related to capacity than
is the B.E.T. area.
Oil immersion led to significant correlations with capacity
when expressed either as bulk specific gravity or as percent
absorption. The total mercury pore volume was the most useful
-------
127
parameter for explaining differences among samples because
(1) unlike oil absorption, mercury intrusion is not limited
to use on large particles and (2) correlation calculations have
shown that the bulk specific gravity, oil absorption, and volume
and area of larger pores are significantly related to the total
mercury pore volume. Therefore, the iron content, mean mercury
intrusion pore size, and the total mercury pore volume are
considered to be the most fundamental variables for explaining
differences in the capacities of samples. When these three
variables were used in a stepwise regression analysis, the
combined correlation coefficient was + 0.775.
(4) Classification of Samples
The carbonate form of limestone is the most likely feed
material in an SO,, removal process using limestone at temper-
atures higher than the calcination temperature. Since most
of the physical properties in the preceding section were
determined on precalcined material, data on eight classes of
t
limestone are given in Table 5-3 to show the capacities of
various types of carbonate rock at the standard test condition.
The classes were selected to give the greatest possible
variety of crystal structures, chemical compositions, and
geological origins.
Iceland spar was included because this high-purity calcite
is composed of very large crystallites. A taagnesite was
-------
128
Table 5-3. CAPACITY OF VARIOUS CLASSES OF LIMESTONE
Classification
Iceland spar
Magnesite
Marble
Calcite
Dolomite
Aragonite
Oolite
Chalk
Number of
samples in class
Variance
Mean capacity1
(g S03/100 g)
1
1
5
7
7
3
4
5
2.90
0.0545
6.55
2.14
10.32
6.45
3.97
6.54
17.3
19.7
32.0
32.7
43.8
52.1
57.1
66.4
At 980°C
-------
129
tested to determine the reactivity of MgCCL. Replicate runs
were made on single samples of Iceland spar and magnesite to
establish their capacities under the same conditions. Hard,
highly crystalline marble samples provide interesting compari-
sons with the soft, fine-grained chalks. The round-grained,
precipitated oolitic samples represent another extreme in
crystal structure. The orthorhombic aragonites were included
because most limestones are rhombohedral. Galcites were
defined as those having at least 95% CaO, and the dolomites
were defined as samples with 40% to 44% MgCCL and 54% to
CaCCL . The results of these classifications are prest
in Table 5-3 in order of increasing mean capacity.
Iceland spar ±s apparently an inferior form of carbonate rock
for absorption of sulfur oxides. A t-test on the various
classes showed that we can be 95% confident that Iceland spar
has lower capacity than the average calcite. The low capacity
of magnesites was also shown by the t-test. Thermodynamically
(34)
MgO can react with SO,, , however, the rate of reaction is
apparently very slow.
It was determined, with 90% confidence, that for the test
conditions used the mean capacities of calcite and dolomite
differ. The CaO utilization of the dolomites is about twice
that of the calcites. Oolites and chalks appear to have
desirable properties for reaction with sulfur oxides.
-------
130
Although the five chalks had a higher mean capacity than the
oolites, a t-test showed that differences between these two
groups were not statistically significant.
Further refinement in the classification of samples could be
achieved if mineralogical and petrographic characterizations
were made. For example, crystallite sizes may be important
in predicting the capacity of a stone. This idea is supported
by the data in Table 5-3; the highly reactive chalks are
typically fine materials, whereas Iceland spar and marble are
relatively coarse materials.
d. Conclusions
1. Between 900 and 1000°C raw and precalcined samples
have the same sulfur oxide absorption efficiency.
At these temperatures calcined stone does not lose
capacity through the changes in physical properties
that accompany overcalcination at higher temperatures.
2. The wide differences in capacity are best explained
by reaction in pores of the size that are measured
by mercury intrusion. Chemical composition is of
only secondary importance.
3. Under the chosen test conditions chalk and oolitic
samples have the greatest sulfur oxide saturation
-------
131
capacity. Iceland spar, magnesite, and marble are
less reactive forms of limestone.
In addition to the data reported above a large amount of data
has been accumulated for other limestones and dolomites.
Approximately 150 limestones and dolomites have been inventoried
by the Process Research Section. Data are available for most
of these sorbents which include chemical compositions, physical
properties (B.E.T. surface area, pore volume and distribution,
etc.), and S09 sorption characterization (bed weight gain, CaO
utilization, etc.). All such data is being compiled by the
Process Research Section and will be available to interested
groups.
-------
132
D. GENERAL CONCLUSIONS
1. Chemical reaction is the sole limiting resistance for S0_
sorption when particles are smaller than 100 microns at tem-
peratures up to 980°C if pores are larger than 0.2 microns.
The isothermal reaction of pure limestones with S0~ show an
intrinsic rate constant (per unit of surface) on the order of
0.2 cm/sec.
2. The rate of reaction increases with decreasing pore size
until a critical pore diameter of about 0.1 micron is reached.
Presumably pores smaller than 0.1 micron are rapidly blocked
by reaction products. Maximum rate results when B.E.T. surface
2
area is in the region of 3.5 m /g (corresponding to pore
diameters of 0.2 - 0.3y) under isothermal reaction conditions.
3. The total S02-sorption capacity increases with increasing
pore size, Furthermore, the capacity of limestones is correlated
to the area of pores greater than 0.3 microns but not to B.E.T.
surface area.
4. Both rate and capacity of SO- sorption are highly dependent
upon particle size. The effect of particle size is not the same
for all stones, but is determined by the size of the pores.
-------
133
Small pores lead to the highest sensitivity between the react-
ivity of calcines and particle size. Calcines with very large
pores may show no dependence of reactivity upon particle size.
5. The relative ranking of calcines with respect to isothermal
rate of reaction with S02 at 980°C is as follows: marl > high
purity limestone > Iceland spar > aragonite > marble. Magne-
site reacts only slowly with S02 at any temperature between
540°C and 980°C.
6. The relative ranking of limestones with respect to fixed
bed capacity at 980°C with -18 +20 mesh particles is as follows:
chalk > oolite > dolomite > calcite > marble > magnesite > Iceland spar.
7. Loss of reactivity because of excessively high calcination
temperatures (known as dead-burning) may be attributed to the
growth of CaO crystals and the subsequent loss of surface area
pore volume. Since loss of surface area and loss of pore volume
are similar functions of increasing calcination temperature,
both reaction rate and capacity of overburned limes are lost
proportionally.
8. Injection of limestone with the fuel into a boiler leads
to dead-burning and poor S0_ removal efficiency. Samples from
field tests where limestone was injected with the fuel in an
-------
134
oil-fired boiler demonstrated the high densities and low
porosities characteristic of overburned limes.
9. Among many physical and chemical properties which are good
indications of the degree of dead-burning, CO, weight gain, SO
fL
weight gain and hydration weight gain are suitable for samples
of limited size diluted with fly ash. For partially sulfated
samples, the hydration weight gain test is recommended since
it is independent of sulfate loading up to 20% S0« in the test
sample.
-------
135
E. APPLICATION OF RESULTS TO LIMESTONE PROCESSES
In the dry limestone injection process, sorption of S0« will
be required of initially uncalcined limestone undergoing cal-
cination and reaction in a nonisothermal environment. In this
regard, the conditions under which the process is to be run
and conditions under which experimental kinetic data of this
project were collected are quite different. For instance, the
i
kinetic data were obtained for precalcined limestone reacted
under isothermal conditions. This does not mean that the data
are not applicable. Polythermal reactions may be regarded as
a series of isothermal reactions over short time increments
provided that the rate of temperature change with time may be
predicted and that there is no change in mechanism of the
reaction. Rate of temperature change may be predicted with
heat transfer correlations and a heat balance. While a multi-
tude of reaction mechanisms are possible, the high injection
temperature of limestone in the dry injection process would
preclude the sulfite reactions and suggest the dominant reaction
is direct conversion to calcium sulfate. Thus, the dominant
reaction mechanism would not change.
The second difference between experimental conditions and boiler
injection conditions was precalcined versus uncalcined limestone.
Once again, the reaction mechanism could be significantly different
-------
136
except if most of the limestone first calcines and then reacts
to form the sulfate. In this case, the difference is one of
calcination conditions, i.e., calcination at a low temperature
for two hours versus calcination at a high temperature for a
few tenths of a second. Effects of varying calcination condi-
tions are reflected in the surface area of the calcine and,
subsequently, in its reactivity.
The importance and functional relationship of the surface area
to the sulfur oxide reactivity of limes has been shown in the
investigations discussed above (for example, refer to conclusions
Subsection e, of Section C2). It remains to describe the effect
of limestone type and injection conditions on the development
of surface area during calcination. To this end work at Battelle
Memorial Institute has been committed using a dispersed phase
reactor.
There is little doubt that dead-burning is an important phenomenon
in limestone processes and in particular the dry injection process.
The above studies have shown that calcination at high tempera-
tures yields a dense lime with little surface area and little
sulfur dioxide reactivity. Not only is injection of lime'stone
with the fuel into a boiler to be avoided, b"t also injection
should be sufficiently far from the fireball that particle
temperatures never exceed 2000°F. Most tests for dead-burning
-------
137
are confounded by the presence of fly ash or of partial sul-
fation. Although modification of procedures may be required,
a simple determination of the weight gain from hydration shows
the most promise as a dead-burning test for the dry injection
process.
Selection of carbonate rock type for limestone processes depends
on the nature of the process. For fluid bed combustion where
limestone remains as a chemically active fluid bed to react
completely, the proper measure of its reactivity is its capacity.
Rankings of limestone and limestone types for this application
are given in Table 5-3 of Section C5 and in Appendix Cl. For
the dry injection process, limestone may react only during very
short residence times, and the proper ranking is derived from
reaction rate measurements as given in Figure 2-4^ Section C2.
-------
138
F. RECOMMENDATIONS
1. Sulfur oxide abatement processes involving injection of
limestone into a power boiler should be significantly enhanced
when highly reactive types of limestones are used; especially,
aragonite and marl. Testing of these materials in demonstration
scale applications is recommended.
2. It has been shown that dead-burning is an important factor
in the dry limestone injection process. Not only must injection
of limestone with the fuel be avoided but also injection should
be into a temperature zone allowing greatest residence time
without everburning.
3. Two fine-grained limestone materials have shown outstanding
S0» reactivity; specifically of the stones tested, marl with the
highest rate of reaction and chalk with the largest capacity for
absorption of S0_. These materials should be studied in depth
to delineate those properties contributing to high reactivity.
4. Much of the background information on limestone-SO_ reactivity
may be applicable to limestone dissolution rates in water and
subsequently to limestone slurry-S02 reactivity. Those prop-
erties responsible for high dissolution rates of limestone in
water should be identified as well as the relative dissolution
rates of characteristic limestone types.
-------
139
5. Limestone may well be an efficient sorbent for sulfur
emissions found in reducing conditions. The kinetics of
limestone - hydrogen sulfide reactions should be determined
besides screening limestone types for reactivity.
-------
140
G. REFERENCES
1. ASTM Standard C110-67
2. Attig, R. C., "Dispersed-Phase Additive Tests for SO- Control",
Interim Report LR:68:4078-01:9 by Babcock and Wilcox Co. for
contract PH 86-67-127 (Dec. 10, 1968)
3. Azbe, V. J.t "Fundamental Mechanics of Calcination, Hydration",
Lime Manufacture May 1939
4. Bertrand, R. 'R. et al, "Fluid Bed Studies of the Limestone
Based Flue Gas Desulfurization Process", Progress Reports 9
and 10 by Esso Research and Engineering Co., for contract
PH 86-67-130 (1968)
5. Borgwardt, R. H., Environmental Science and Technology, 4_ (1)
59 (1970)
6. Borgwardt, R. H. et al, "The Dry Limestone Process for SO-
Control: A Field Study of the Role of Overburning", APCA,
New York, N. Y. (June 22-26, 1969)
7. Boynton, R. S., "Chemistry and Technology of Lime and Limestones",
Wilev, New York 1966
8. Coutant, R. W. et al, "Investigation of the Reactivity of
Limestone and Dolomite for Capturing S0? from Flue Gas", Reports
by Battelle Memorial Institute for contract PH 86-67-115
(Aug. 30, 1968 and Feb. 12, 1969)
9. Cunningham, W. A., I and EC, 43 (3) 635-8
10. Eades, J. L. et al, "Scanning Electron Microscope Study of
Development and Distribution of Pore Spaces in Calcium Oxide",
2nd Annual SEM Symp., IITRI (April 1969)
11. Eckhard, S., Z. Analyt Chem. 209:156 (1965)
12. Fischer, H. C., J. Amer. Ceramic Soc., 38:7, 245-51 (1955)
13. Goldschmidt, K., VDI Berichte. 6:21, 84 (1968)
14. Harrington, R. E. et al, Am. Ind. Hyg. Assoc. J., 29, 52-8 (1968)
15. Harvey, R. D., Environmental Geology Notes, 21, Illinois State
Geological Survey (1968)
-------
141
16. Hatfield, J. D. et al, Monthly Report for contract TVA-29232A
(October 1968)
17. Haynes, W., "Pilot Injection Studies at the Bureau of Mines",
report presented at PHS Limestone Conference (Dec. 4-8, 1967)
18. Hedin, R., "Structural Processes in the Dissociation of Calcium
Carbonate", National Lime Assoc., Azbe Award #2 (1961)
19. Ishihara, Y., "Removal of S02 from Flue Gases by Lime Injection
Method"., report of Central Research Institute of Electric Power
Industry, Komae, Kaitatama, Tokyo, Japan presented at PHS
Limestone Conference (Dec. 4-8, 1967)
20. Kim, Y. K., "Sulfation of Limestone Calcines", PHS Symp.,
Cincinnati, Ohio (Feb. 5-6, 1969)
21. Lougher, E. H., "Identification of Test Methods for Determining
the Degree of Burning of Limestones", Final Report (Nov. 13, 1968)
22. Ludwig, J. H. et al, Chem. Engr. Progress, 63_ (6), 82-4 (1967)
23. Mayer, R. P. et al, "Physical Characterization of Limestone
and Lime", National Lime Association, Washington, D. C. (1964)
24. McClellan, G. H. et al, "Scanning Electron Microscope Study
of the Textural Evolution of Limestone Calcines", ASTM
25. Mullins, R. C. et al, "Effects of Calcination Conditions on
the Properties of Lime", ASTM
26. National Research Council (National Academy of Engineering),
"Abatement of Sulfur Oxide Emissions from Stationary Combustion
Sources", 1970
27. Ohno, Y., Gypsum and Lime, 28_, 22-28 (1957)
28. Potter, A. E. et al, Air Engineering, 22-6, (April 1968)
29. Potter, A. E., Ceramic Bulletin, Am. Ceramic Soc., 48 (9),
855-8 (1969)
30. Rootare, H. M. et al, J. Phys. Chem., H., 2733 (1967)
31. Satterfield, C. N. and Sherwood, T. K., "The Role of Diffusion
in Catalysis", Addison-Wesley, Reading, Mass. (1963)
32. Shen, J. and Smith, J. M., Ind. Eng. Chem. Fundam., 4_, 293-301
(1965)
-------
142
33. Tennessee Valley Authority, "Sulfur Oxide Removal from Power
Plant Stack Gas Conceptual Design and Cost Study" (1968)
34. Ward, J. 0. et al, "Fundamental Study of the Fixation of Lime
and Magnesia", Report by Battelle Memorial Institute for
Contract PH 86-66-108 (June 30, 1966)
35. Wicker, K., Mitt. Ver. Grosskesselbes. 83, 74-82 (1963)
36. Wuhner, J., "On the Reactivity of Lime from Different Kiln
Systems", National Lime Association, Azbe Award #5 (1965)
-------
143
APPENDIX A
Kinetic Data of Carbonate Rock Types
-------
144
KINETIC DATA OP CARBONATE ROCK TYPES
The graphs appearing in this appendix present the kinetic data
obtained in differential reactor test runs in accord with the
discussions of Section C2 of the report. The carbonate rock
types indicated on the graphs are identified in Table 2-1 of
Section C2.
-------
145
20
18
16
1
3
S? 12
O
fO
JT
•K 10
CO
8 6
I I
I I I I
Type 1. 150/170 mesh calcined 925"C _
1 I I I I I 1 I I I I I 1
20 40 60 80 100 120 140
1 l T T T I I I I II I
Type 1. 150/170 mesh calcined 2040"F _
20
18
16
14
12
10
8
6
4
I I I I I I I I I I I I I
Typel. 150/170 mesh calcined 325'C
870 8C_
1 l I I I I I I l I I l
100 120 140 ' 0
60 80 100 120 140
I l l I l I I
Type 2. 150/170 mesh _
-------
146
g
3
3
a
S
1
S
-s
a
S
20
U
16
14
12
10
S
6
4
2
I 1 I 1 1 1 1 1 1 1 I 1 1
Type 2. 150/170 mesh calcined 2040 "F _
-
-
— —
_ _
_ _
- —
— ~
-
_ —
-
1800 °F""
n n ir~
-.V^***"^ 1200 'Fj
i i i I I I I i r
Type 2. 650 °C 150/170 mesh
I I l l l I I 1
100 200 300 400 500 600 700 800
20
II
t»
3
12
10
8
6
4
I I I l I
Type 3. 150/170 mesh _
I I I I I I
Type 4. 150/170 mesh _
S80°C
I 1 1 I I I I I I 1 I
Tint, IK
20 40 60 80 100 120 140
Tim, sec
-------
147
20
18
16
| 14
JG
3
r 12
S
J-
1
1 8
10
18
16
| "
3
£ 12
O
n
I "
3 8
&e i
^5 o
4
2
1 I i I I I
TypeS. ISO/170 mesh
nc'
I I I I I I I I I I I
GO 80 100 120 140
I I I I I I T I T I I I I
Type6. 150/170 mesh
-i
lit!'
20
40 60 80 100 120 140
i r T
i riii r
TypeS. 150/170 mesh
870 °C
650 °C
i I I I I II JII I ' I '
20
40 GO
TiM,SK
100 120 140
120 14C
-------
148
20
18
IS
I I I I T
s
10
I .
I I I I I T
Types. 150/170 mesh -
20
60 80
TiM, MC
100 120 140
20
II
16
14
12
10
I I
Calcined in-situ
"" to S02 exposure
980 °C
I I 1 T
•
MICHIGAN MARL-
150 170 mesh
1
Calcined in rotary kiln
at 980'C
I 1 I I I 1 I i i I I i
20 40
60 80
HM.MC
100 120 HO
-------
100 c=r
10 =
0.1
^ i ii Him 111 HUH ! r
150 170 mesh
11
mi inn
Type 1. 980 *C
T r nmirTTTmnr
42 65 mesh
rn=
i i m.iii i 11 Him i 11 nun i n
i 11 nun i 11 imii i i nun i 11
i 10 IN 1000 10.000 1
10 100
TIME,sec
1000 10.000 1 10 100 1000 10.000
•P-
Vf)
100
10
N'
0.1
I mill! I I I Illlll I 11 mill I 11113
150 170 mesh
Type 2. 980 °C
"mwnrm
42 65 mesh
1 10 100
L_LLIHJJL_JL
1000 10,000 1 10
mini ! n nun i ii
I 111(1111 I 1 Illlltl
II16 mesh
II HUH I I I
100 1000
TIME, we
10,000 1 10 100 1000 10.000
N' = mg. SO absorbed by 30 rag. calcine
-------
100
io
N'
0.1
ISO 170 mesh
Type 3. 980 "C
nriTiniir 7T
42 65 mesh
immn i
12 IS mtsh
1 10 100 1000 10.000 1
10 100
TIME, set
1000 10,000 ]
1000 10000
Type 4. 980 *C
10
0.1
E I I ITTI1IT I I I mill I I Illllll I
150 170 mesh
~i iiiiiiii TTimiir TTTwiiii 11113
42 65 mesh
I II Ililll II Illllll I I Illllll 1 I Illllll L_l_illili I I I HUH I ill.11 HI I II
12 IGmesfi
11 nun i 11 nun i ii
1 10 100 1000 10.000 1
100 1000 10,000 1
TIME. sec.
10 100 1000 10,000
N =• mg. SO absorbed b
y 30 mg'. calcine
-------
100 =r
10 =
N'
0.1
150 170 mesh
Type 5. 980 °C
7 j inrT'r
n~Trn=3
F
1 10 tOO 1000 10.000 1
42 65 i>esh
TmTn r1^^ <
U 16 "esh
100
TIME.sec
1000 10,000 1 10 100 1000 10.000
100
N1
0.)
111 nun i i iimn nr
150/170 mesl)
Type 6. 980 °C
42 65 mesh
ii inn i 11 nun ..MI nun i i
! I IIIUI TTTTl
12 16 roesti
I I 1112
1 10 100 1000 10,000 1
10 100
TIKE, sec
1000 10,000 1
10
I Illlllll.J.IlJJilil
too loon 10.000
N1 = nij.',. SO absorbed by 30 nip,, calcine
-------
100
10
N'
0.1
TypeS. 980 °C
ETT
150 170 mesh
TTTfflU I ! I Illlll T
42 65 mesh
1 10 100 1000 10,000 1 10
100
TIME, set
1000 10,000
MICHIGAN MARL 980 °C
100
It
N'
0.1
TT
•65 mesh
Mllll! I I III!!'
1Z 16 mtsh
1 10 100 1000 10.000 1
10 100
TIME, sec
Ul
i MiiiiH i muni i i i i
1000 10,000 1 10 100 1000 10.000
N' = rag. S0~ absorbed by 30 mg. calcine
-------
153
30
27
24
.1 21
O
T5
<»0 ID
E 18
o
CO
£ 15
.g
1 T
T 1 1
Typel. 980 °C
12
E
CO
55' 9
150/170 mesh
42/65 mesh -
100 200 300 400 500 600 700
30
27
24
21
18
15
I I
Type 2. 980 C
150 170 mesh
100 200 300 400 50u 600 700
TIME, sec
100 200 300 400
TIME, sec
500 600 700
-------
I I 1 I t
Type 6. 980 C
i l I I
MICHIGAN MARL 980'C
100 200 300 400
TIME, sec
500 600 700
100 200 300 400 500 600 700
TIME, sec
-------
155
I I I I 1 I I I I I I -
Type 2. 980 C 42 65 mesh -
I
.. 0.01
I I
I I I
Type 2. 980 C 12 16 mesh
0.0001
— I I I I I I I I I I I
0.01
0.001
15 20
S03l rag
30 35
0.001
150.170 mesh
12/16 mesh
42/65 mesh
5 10 S03,mj 25 30
-------
156
III\IIIIIIIF;
Typel. 980 C 150/17a mesh Z
I I I I 1
tilt
f I 1 I :
Typel. 980 C 42/65 mesh:
0.001 =
10 15
20
25 30
35
0.0001
I
4
5
7
I I _
Typel. 980°C 12/lfrmeshI
o
5
fa
0.001
OS
O
O
S5
OS
0.0001
r \ T^ i i i j
Type 2. 980 C 150 170 mesh -
2
$03, mg
0.001
-------
157
I I I I I I I
Types. 980"C
12.16 mesh
I I } I I I I I I I
1.0 cr
i I I I I I I
I I I TJ
Types. 980'C Z!
150 170 mesn
I t I I I 1 1 1
0.001
0
5
10
15
20
I I \ I —
TypeS. 980 °CZ
FlIIIIiI 1I T T T T
MICHIGAN MARL 980 C
0.1
t I Ill
0.01
10
15 20
SOs, mg
25 30
35
0.001
1 I t ..j_..J_
10
15 20
SOa, mg
25 30
35
-------
158
APPENDIX B
Bl. Procedure for Dead-Burning
Tests
B2. Statistical Analysis of
Dead-Burning Data
B3. Crossplots of Data from the
Dead-Burning Study
-------
159
Bl. PROCEDURE FOR EXPERIMENTAL DETERMINATION OR COMPUTATION
OF REPORTED DATA
1. Flue Gas Absorption - Method of R. H. Borgwardt
a. Thirty milligrams of screened material are placed
on a rock wool mat and into a differential reactor.
b. Exposure is for 120 seconds at 1800°F.
c. Recovered sample is analyzed for sulfate and results
are reported as MgS03/30 mg.
2. Pure S02 Absorption
a. One gram of screened material is put in a combustion
boat and into a horizontal tube furnace with quartz
combustion tube. (See Figure Bl-1).
b. Exposure is for 30 minutes at 1800°F with a gas
flow of 2 SCFH. The sample is in a complete
atmosphere of S0?.
c. Results are reported as per cent weight gain.
3. Pure CO Absorption
a. One gram of screened material is put in a combustion
boat and into a horizontal tube furnace with alumina
combustion tube.
b. Exposure is for 60 minutes at 1400°F with a gas
flow of 2 SCFH. The sample is in a complete
atmosphere of C02.
c. Results are reported as per cent weight gain.
4. Steam Absorption
a. One gram of screened material is put in a combustion
boat and into a horizontal tube furnace.
b. Steam from a steam generator is fed to a preheat
coil and into the stainless steel combustion tube.
c. Exposure is for 5 minutes at 500°C with results
reported as per cent weight gain.
-------
160
5. Acid Titration
a. One gram of screened material is put in a 400 ml
beaker and pH probes positioned in place.
b. At time zero, 200 ml of distilled water is added
with vigorous agitation.
c. Titrations to pH 8 at minute intervals with IN HC1
give results of cumulative acid versus time.
6. B.E.T. Surface Area - as reported by AMINCO.
7. Total Pore Volume - as reported by AMINCO.
8. Small Pore Volume - using the mercury penetration
porosimetry curve, volume of pores with diameter less
than 2 microns is calculated.
9. Mercury Intrusion Density - as reported by AMINCO.
10. Average IR Band Shift - as reported by GTC ; using
infrared analysis of samples prepared by the KBr pellet
technique, the average shift of three bands is calculated.
11. Median Pore Diameter - using the mercury penetration
porosimetry curve, the diameter corresponding to intrusion
of 50% of the total volume is reported.
12. Air PycnometerDensity - standard technique.
13. Oil Absorption Density - No. 2 fuel oil is displaced
in a graduated cylinder.
14. Hydration-Weight Gain
a. One gram of material is slaked wxth one milliliter
of distilled water for one hour.
b. The sample is dried for one hour at 260°C.
c. The percent weight gain is reported.
-------
161
["I .X-ROTAMETER'
EXPOSURE GAS
SOURCE
y /Tr0.0!*805.!1-0/*. -B0.AT- jn-_ .V 1
HORIZONTAL TUBE FURNACE
TO VENT
SCRUBBING !
SYSTEM I
Figure Bl-1
COMBUSTION BOAT
TEST APPARATUS
-------
162
B2. SOME STATISTICAL ANALYSES OF "LIMESTONE DEAD-BURNING" DATA
I. SUMMARY
1. Data
As given in Tables B2-1 and B2-2 the data are in two batches
on limestones from three sources. A measurement was taken for
each source at five preset calcination temperatures^ The
measurements for Batch II (B2) have eight components besides
temperature, and these eight will be termed the "basic set"
of variables. The measurements for Batch I (Bl) have five
additional components. This set of thirteen variables will
be called the "extended set" of variables.
2. Purpose
The amount of SO? absorbed from the flue gas is a decreasing
function of temperature in the range of temperatures for this
study. Obviously, a number of other components are also highly
correlated to the flue gas measurement and hence they are all
controlled by temperature. The purpose of this investigation
generally was to search for other, not so obvious, relationships
among the variables which may shed light on the dead-burning
effect.
-------
163
TABLE B2-la
Properties of Fredonia White
Set I, 70/140 mesh
Calcine Temperature
1700 2000 2300
2600
3200
Flue Gas Absorption
1800°F, 2 min, mg/30 mg
Pure SO- Absorption*
1800°F. 30 min, % gain
Pure CO- Absorption*
1400 F, 60 min, % gain
Steam Absorption
500 C, 5 min, % gain
B.E.T. Surface Area
m2/g
Total Pore Volume
cc/g
Small Pore Volume
cc/g
Density by Mercury
Intrusion g/cc
Average IR Band
Shift, %
8.6
66.0
37.4
29.7
3.1
5.9
40.7
8.7
22.5
1.6
2,5
21.9
5.8
12.5
1.0
1.6
10.4
2.6
10.8
0.8
1.0
8.0
1.7
7.2
0.3
.61 .51
.28 .14
.33
.09
2.82 2.44 3.23
62
80
.25
.04
3.34
95
.33
.04
3.30
100
* Note: Listed values for 140/200 mesh
-------
164
TABLE B2-lb: Properties of Cedar Bluff
Set I, 70/140 mesh
Calcine Temperature
Flue Gas Absorption
1800°F, 2 min, mg/30 mg
Pure S0_ Absorption*
1800°F, 30 min, % gain
Pure C0_ Absorption*
1400°F, 60 min, % gain
Steam Absorption
500°C, 5 min, % gain
B.E.T. Surface Area
m2/g
Total Pore Volume
cc/g
Small Pore Volume
cc/g
Density by Mercury
Intrusion, g/cc
Average IR Band
Shift, %
1700
6.70
62.2
25.8
26.6
1.9
.70
.28
2.81
2000
3.76
27.3
8.7
17.0
1.2
.37
2300
3.81
26.7
6.6
15.5
0.9
.49
2600
2.60
18.8
5.6
12.3
0.8
.49
3200
2.51
13.2
3.6
10.2
0.5-
.37
.10
.08
2.95 3.32
25
44
.05
3.30
67
.02
3.19
100
*Note: Listed values for 140/200 mesh
-------
165
TABLE B2-lc: Properties of James River
Set I, 70/140 mesh
Calcine Temperature
1700
2000
2300
2600
3200
Flue Gas Absorption 9.56 9.04 6.60 2.28 1.80
1800 F, 2 min, mg/30 mg
Pure SO, Absorption* 67.4 62.1 46.2 13.2 7.6
1800 F,"30 min, % gain
Pure CO,, Absorption* 22.6 12.6 8.3 2.0 1.5
1400 F,"60 min, % gain
Steam Absorption 16.8 13.7 10.1 4.6 2.8
500 C, 5 min, % gain
B2E.T. Surface Area 4.1 2.2 1.5 — —
m /g
Total Pore Volume .69 .66 .55 .46 .37
cc/g
Small Pore Volume .35 .31 .21 .01 .01
cc/g
Density by Mercury 2.84 2.74 3.39 3.44 3.31
Intrusion, g/cc
Average IR Band 0 26 57 83 100
Shift, %
*Note: Listed values for 140/200 mesh
-------
166
TABLE B2-2a: Properties of Fredonia White
Set II, 170/270 mesh
Calcine Temperature
Flue Gas Absorption
1800°F, 2 min, mg SO.,/30 mg
«3
Flue Gas Absorption
1800°F, 2 hrs, mg SO.,/30 mg
Pure S07 Absorption
1800°F, 30 min, % gain
Pure C02 Absorption
1400*F, 60 min, % gain
B.E.T. Surface Area
m^/g
Total Pore Volume
cc/g
Small Pore Volume
cc/g
Density by Mercury
Intrusion g/cc
Extent of Slaking
i _ t n •»_. at * __
1700
7.04
34.3
90.0
54.0
2.2
.77
.31
3.47
33.6
2000
3.52
22.6
16.8
9.8
1.1
.37
.07
2.91
28.1
2300
2.48
13.7
10.1
2.5
0.9
.30
.05
3.11
23.1
2600
2.08
11.2
6.0
3.9
1.3
.31
.08
2.92
20.2
3200
1.92
5.6
3.2
0.7
0.7
..22
.02
3.09
11.7
-------
167
TABLE B2-2b: Properties of Cedar Bluff
Set II, 170/270 mesh
Calcine Temperature
1700 2000 2300 2600 3200
Flue Gas Absorption 4.72 6.04 3.88 3.36 2.64
1800 F, 2 min, mg/30 mg
Pure SO- Abosrption 31.0 39.4 25.5 21.5 9.8
1800 F, 30 min, % gain
Pure CO- Absorption 11.0 6.7 5.4 6.0 2.9
1400°F, 30 min, % gain
B2E.T. Surface Area 1.4 1.2 0.9 1.1 0.5
m /g
Total Pore Volume .59 .48 .36 .34 .33
cc/g
Small Pore Volume .09 .05 .05 .06 .02
cc/g
Density by Mercury 3.27 3.15 3.17 3.16 3.21
Intrusion, g/cc
Extent of Slaking 32.1 30.4 26.9 27.0 10.4
1 ml, 1 hour, % gain
-------
168
TABLE B2-2c: Properties of James River
Set II, 120/270 mesh
Calcine Temperature
Flue Gas Absorption
1800°F, 2 min, mg/30 mg
Pure S0~ Absorption
1800°F, 30 min, % gain
Pure C0? Absorption
1400°F, 30 min, % gain
B.E.T. Surface Area
m2/g
Total Pore Volume
cc/g
Small Pore Volume
cc/g
Density by Mercury
Intrusion, g/cc
Extent of Slaking
1 ml, 1 hour, % gain
1700
8.92
69.9
12.5
2.4
.62
.30
3.09
17.3
2000
4.92
49.4
6.7
1.3
.37
.15
3.16
10.4
2300
4.72
34.4
4.3
1.0
.37
.11
3.16
13.2
2600
2.92
10.3
2.8
1.0
.35
.07
3.25
2.6
3200
1.80
8.0
1.7
0.7
.26
.05
3.24
5.2
-------
169
3. Analyses
Two computational methods of inspecting the data were tried:
1) principal components, and 2) regression with the flue gas
measurement as the dependent variable.
II. GRAPHS AND CORRELATIONS
Correlation coefficient matrices are given for the measurements
in Bl and B2 on the computer printout as marked by the tabs.
The most notable difference between Bl and B2 is that mercury
intrusion density (MERDEN) in Bl is negatively correlated with
the temperature related variables whereas in B2 it is positively
correlated. An inspection of the graphs of MERDEN vs. temperature
indicated this pictorially. The B2 measurement from source 1,
Fredonia, at 2000°F is a major contributor, but even without
it the slope would be negative.
Some anomalies are evident in the ACID measurement, Bl, where
the three high temperature measurements for source 3, James River,
are much lower than the others.
III. PRINCIPAL COMPONENTS
Temperature is a controlled variable, and variables in both the
basic set and the additional set are dependent upon temperature,
-------
170
hence correlational techniques such as principal components
analyses must be used and interpreted with appropriate caution.
The magnitude of any "correlation" cited in this report is not
meaningful itself, since its size is controlled by the selection
of temperatures, but it may be meaningful relative to another
such correlation from other variables, batches, etc.
For this situation we expect one large temperature component
with heavy loadings on those variables most related to temperature.
The size of this component is not of itself meaningful for reasons
mentioned above. The second and higher components, since they
are orthogonal to this first major temperature component, hopefully
will exhibit factors perhaps not controlled by temperature and
yield useful information on the variable relationships.
1. Batch I, Basic Set of Variables
In the computer printouts as marked by tabs are the results of
the principal components computation on the Bl data for the
basic set of variables. The.first component accounts for 86%
of the total variation and all variables are weighted more or
less equally with MERDEN, here coded as (MIDEN), and median pore
diameter (MDPDIA). MDPDIA have negative weights as expected
because of their negative correlation with the other variables.
An inspection of case ranking by that component in Table B2-3
shows this to be a good, but not perfect temperature component.
-------
WT2- 3
flat oh No. *
Variables: 8 BASIC SET
Component No. 1
Rank
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Rank
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Source
123
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Source
123
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Temperature
17 20 23 26 32
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Temperature
17 20 23 26 32
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Component No. 2
Rank
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Rank
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Source
123
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Source
123
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Temperature
17 20 23 26 32
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Temperature
17 20 23 26 32
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Component No. 3
Component No. 4
-------
172
The second component shows heavy weight, (-), for MERDEN with some
weight, (-), for mercury pore volume (MERPVO). There is no clear
linear relationship to temperature or source.
Component three is largely a source component with CO- and BET
carrying positive weight and MERPVO a negative weight. It is
doubtful whether the fourth component is meaningful.
2. Batch II, Basic Set of Variables
The first component for the B2 data is similar to that for the
Bl data except that MERDEN enters with the same sign as the
temperature related variables instead of the reversed sign as
for Bl. This is expected because of the change in sign of the
correlation for MERDEN from Bl to B2. This component is a strong
temperature component (Table B2-4). Note that the direction
of the eigenvector is reversed from that of Bl, hence temperatures
are ordered from high to low.
The second component again has MERDEN as the major contributor
but MERPVO is no longer important. Instead the MDPDIA seems to
be the contributor in its place in both the 2nd and 3rd component.
The third component is clearly the source component.
3. Batch I, Extended Set of Variables
The same computation was done on Bl with the extended set of
variables less flue gas. The first component is a pretty good
-------
Batch No. II
Tables J32-
Variables: 8 BASIC SET
Component No. 1
Rank
1
2
3
4
5
6
7
8
9
10
13
12
13
14
15
Rank
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Source
123
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Source
123
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Temperature
17 20 23 26 32
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Temperature
17 20 23 26 32
X
X
x
X
X
X
X
X
X
X
X
X
X
X
X
Component No. 2
Rank
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Rank
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Source
123
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Source
123
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Temperature
17 20 23 26 32
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Temperature
17 20 23 26 32
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Component No. 3
Component No. 4
u>
-------
174
temperature component as it was for the basic set, but the addi-
tional variables change the later components. In Table B2-5
the strong source difference for ACID shows itself in component
two where ACID has the heaviest weight and easily separates
James River from Fredonia and Cedar Bluff. Notice the large
gap in component values between case no. 1 and case no. 11. The
generally low values for AIKPY for James River influence this
also. OILAB in component three aoes for Cedar Bluff what ACID
did for James River in component 2. MERDEN (MINT) here tries
to influence the components as it did for the basic set, but
the source effect on the additional variables dilutes its efforts.
4. Summary of Principal Components Analysis
The temperature component appeared as expected in the first
component. Source, another controlled variable just as tem-
perature, appeared strongly in the second or third component,
and may have obscured some of the variable relationships. The
one variable that does seem to operate somewhat independently
of the temperature component is MERDEN. MERPVO does this also
for Bl, but not for B2 where MDPDIA appears in a similar role.
-------
Batch No.
Variables: 12 EXTENDED SET
Component No. 1
Rank
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Rank
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Source
123
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Source
123
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Temperature
17 20 23 26 32
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Temperature
17 20 23 26 32
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Component No.
Rank
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Rank
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Source
123
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Source
123
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Temperature
17 20 23 26 32
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Temperature
17 20 23 26 32
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
-vl
Ui
Component No. 3
Component No. 4
-------
176
IV. REGRESSION
Linear least square regressions were run on each batch with
flue gas as the dependent variable. Only constant and first
degree terms were used in the model. Batch I was analyzed first
using the basic set of variables (seven independent), and then
using the extended set (eleven independent). The stepwise re-
gression program in the UCLA Biomedical Series, BMD02R, and
the Linear Least Squares program of Wood and Toman were used.
Printouts are marked by tabs with comments below.
1. Stepwise Progress Tables
Tables B2-6, B2-7, B2-8 trace the progress of the stepwise re-
gressions for Bl basic set, for B2 basic set, and for Bl extended
set respectively. In these tables the independent variables
are listed in the leftmost column. Across the top are the step
numbers and underneath them the estimated root mean square for
error at that step. In the table for each step the independent
variables are ranked by their importance as predictor variables
as indicated by their F values.(F to enter for variables not in
the equation and F to remove for variables already in. Variables
already in the equation are ranked ahead of those not yet'in,
and they have their ranks underlined.)
Generally the relative predictive value of an independent variable
as given by its rank is lowered at a particular step if the
-------
177
TABLE B2-6
Trace of Stepwise Regression
Batch I, Basic Set
Step
so2
co2
BET
MINT
MPVOL
SFVOL
MDPV
a
0
a.
2
3
7
5
2
3
1
i
' 2
3
7
6
5
*
.61
2
1,
2
3.
-
-
-
4
.56
3
I
2
3
-
6
*
5
•53
4
1
2_
1
7
6
4
5
•52
5
1
2
1
6
5
i
-
.51
6
l
i
1
6
7
i
2
.47
7
JL
2
3
6
7
i
2
.50
TABLE B2-7
Trace of Stepwise Regression
Batch II, Basic Set
so2
co2
BET
MINT
MPVOL
SPVOL
MDPD
A
a
0
1
6
2
7
3
4
5
1
1
2
^
3
6
-
5
.82
2
1
i
4
5
3
-
-
.63
3
1
2
5
*
3.
-
-
A3
4
i
2
-
4
3
-
5
.38
5
1
i
7
i
i
6
2
.38
6
i
i
7
i
3
6
2
•38
-------
178
TABLE B2-8
Trace of Stepwise Regression
Batch T, Extended Set
Step
0
so2 i
COO 7
STEAM 9
ACID 12
BET k
A T"D"CV T ^
*i_L£\jri XU
OILAB n
MINT 8
MPVOL 5
SMPV 2
MDPV 3
I.R. 6
/\
cr
1
1
3
2
6
7
11
*
12
10
9
8
5
.61
2
1
9
£
12
5
6
7
3
11
*
8
10
.53
3
l
-
2
-
5
7
6
1
8
4
-
-
.50
if
1
-
2
7
5
-
9
1
6
i
8
-
.W
5
1
8
2
7
i
9
-
2
6
1
-
10
•35
6
1
11
2.
7
i
9
8
5
6
3
10
-
.34
7
1
-
£
7
i
8
-
5
6
1
9
-
•33
8
1
10
2
6
i
8
-
5
1
3
-
9
.29
9
l
10
2
6
i
8
-
5
7
1
-
9
•30
10
i
10
3
8
if
7
11
2
6
2
12
2
.32
11
1
10
3
1
i
6
11
5
8
2.
12
2
.35
12
3.
5
-
9
2
i
l
10
-
1
8
6
.11
-------
179
reduction in the error sum of squares due to that variable
(or potential reduction for a variable not yet entered) is
made much smaller or "stolen" by the variable just entered,
i.e., the two variables yield redundant information. The
value will be raised if this (potential) reduction is not lowered
very much while the error sum of square is considerably reduced
by the entering variable, i.e., the information for the two
variables is not very redundant. If two independent variables
are completely uncorrelated the reduction due to one variable
will not be lowered at all by the other one's entry, hence its
rank will probably rise.
2. Residual Error
In Bl the additional variables in the extended step contribute
significantly to the regression equation as indicated by the
lowering of 6*, the root mean square for error, from about .50
to .30. The drastic reduction of & to .11 in the 12th step
of the Bl, extended set, regression is probably specious due
to the very low degree of freedom for error.
3. Action of the Independent Variables
Several things should be noticed from this regression.
a. SO, enters first and is consistently the most
important predictor even though its importance
is reduced little by little as correlated variables
-------
180
enter the equation. It is not as important in
B2 as in Bl; CO- and MPVOL seem to contribute
more in B2 to compensate.
b. CO- is worth more (relatively, compared to the
other variables) in the presence of SO- than it
is by itself. In all of the regressions it comes
from a low ranking to second or third with the
entry af SO-. Its information is apparently less
redundant and more complementary to S0_, than
others such as SMPV, MINT, MPVOL, MDDIAM.
c. SMPV is largely redundant with SO since it loses
heavily with the introduction of SO,,. For Bl,
extended set, it finally works up to have first
ranking at the last step. The drastic shuffling
of rankings at this last step, just as the lowering
of o, are suspect because of the very low degrees
of freedom for error.
d. BET's influence is reasonably stable in Bl. It is
probably a good predictor variable, but of secondary
importance to SO-. In B2 it yields to the mercury
penetration variables.
e. MINT and to a lesser extent SMPVOL have an interesting
action in Bl. With the basic set of variables they
have little importance after S02 and (X>2 are intro-
duced. However, with the extended set, wherein
-------
181
STEAM enters at the second step in place of CO ,
we see that MINT assumes a high rank. This suggests
that the MINT, STEAM variables must be used together
as a somewhat equivalent to, but better predictor
than CO..
f. MPVOL apparently works with CO- in B2, but no such
action is seen in Bl where it is constantly low
in rank.
4. Other Notes
a. In Bl the Cedar Bluff 2300°F measurement shows a
rather large residual, but when it was eliminated
there was no significant change in the regression
results.
b. In B2 the Fredonia 1700° measurement has a large
WSSD (Wood-Toman) indicating that the location of
the independent variable is far removed from the
other 14 measurements. Elimination of this measure-
ment has little effect on the regression equation
indicating that the high value of FLUE GAS at that
point is well predicated by the high values of the
other variable.
5. Conclusions from the Regression Analysis
It appears obvious that the capacity measurements are the strongest
predictors. There are at least two useful degrees of freedom
-------
182
among them which here seem to be given by either the S0_ and
STEAM combination or the SO- and C0_ combination.
After that the mercury penetration measurements add some
additional information with probably two or more degrees of
freedom. MEKDEN AND SMPVOL (together with STEAM) or MPVOL seem
most promising.
BET stands as a useful variable in Bl somewhat independent of
the rest, but in B2 it is redundant to the mercury penetration
measurements.
V. COMPARISON OF REGRESSION AND PRINCIPAL COMPONENTS ANALYSES
Generally, those variables which appeared with some force in the
non-temperature components proved to be rather poor predictors
in the regression equations. For example, in Bl, basic set,
MERDEN and MERPVO were strong in the second component but law
in predictor value. Similarly in B2, basic set, MERDEN and
MDPDIAM act similarly whereas MERPVO does not appear in the
second component, but it is instead a good predictor variable
for flue gas. This suggests that such variables have a variance
component in this process which is unrelated to the basic
capacity measurements.
In Bl extended set, the additional variables AIRPY, OILAB, and
ACID displayed the same characteristic by carrying heavy weight
-------
183
in later components but having little predictor value. These
variance components seem to be caused by source differences.
-------
184
B3. CROSSPLOTS OF DATA FROM THE DEAD-BURNING STUDY
As outlined in Section C4, three limestones numbered 2061,
2062, and 2069 were calcined at a range of temperatures and
tested for chemical and physical properties. The response
for each of these properties to calcination temperature is
given in the following crossplots.
-------
o
o
<*"»-,
o
o
'"•o
CDO
CO
_
-------
a
a
r-
CD-
CD
o
03
>-
II,
O
DC
a:
ex.
CEo
O
en
UJ
cc
Q_0
r-*"
DEflD BURN STUDY/CROSS PLOT NO 02
co
* *
*
*
37.00 39.00 21.00 23.00 25,00 27.00 23.00 31.00 31.00
CflLCINE TEMPERRTURE CHUNDREDS OF DEGREES
-------
o
O
O
O
>-
1!
o
o
in.
CJ>
cc
UJ
o_
o
CE
0_
(X
o
CJ
cc
ID
Q_0
DEflD BURN STUDY/CROSS PLOT NO 03
oo
X
17.00 19.00 21.00 23 00 25,00 27.00 .29.00 31.00 33.00
CflLCINE TEMPERRTURE CHUNDREDS OF DEGREES
-------
o
o
o
o
o
o
*
O
O
LJ
CJ
t—
h—I
u
DERD BURN STUDY/CROSS PLOT NO 04
*
x
oo
K
K
5^'
21 w
cc
u
COO
17.00 19.00 21.00 23 00 25 00 27.00 29-00 31.00 33.00
CflLCINE TEMPERATURE (HUNDREDS OF DEGREES F5=X
-------
en
en .,,
W I
en
a
cr
u.
QC,
LJ
•Q
OERD BURN STUDY/CROSS PLOT NO 06
31*
X
K
)K
vo
Mf-
^i I n i i r ^^
17.00 19,00 21,00 23.00 25,00 27.00 29.00 31.00 33-00
CRLCINE TEMPERflTURE CHUNDREDS OF DEGREES F)=X
-------
en'
en
03
cn'
to
110-1
CJ
CJ
in
en
~ZL
UJ
cn
•—foi
DEflD BURN STUDY/CROSS PLOT NO 07
*
*
VO
17.00 19.00 21.00 23 00 25.00 27.00. 29.00 31.00 33.00
CflLCINE TEMPERflTURE (HUNDREDS OF DEGREES F:3=X
-------
O1
en
en
C71
in
II
r->
O
CO
2
LJ
to
CD
cr
DERD*BURN STINGY/CROSS PLOT NO*U8
X
17,00 J9.00 21.00 23.00 25,00 27.00 29,00 31.00 33.00
tfiLCINE TEMPERflTURE CHUNDREDS OF DEGREES H=X
-------
en
en
in
in'
DEflD BURN STUDY/CROSS PLOT NO 09
X
X
X
vo
JK
17.00 39.00 21.00 23.00 25.00 27.00 29.00 31.00 33.00
CflLCINE TEMPERflTURE CHUNDREDS OF DEGREES
-------
in_
II
f—>
CD
LJ
21
ID
_J
O
UJ
oc
DEflD BURN STUDY/CROSS PLOT NO 10
* 5
*
en
I I 1 1 \ \ 1 1
17.00 19.00 21.00 23.00 25.00 27.00 29.00 31.00 33.00
CflLCINE TEMPERRTURE CHUNDREDS OF DEGREES H=X
-------
r-
3- .
K
H
LJ
CJ
O
§°
Q_
DERD BURN STUDY/CROSS PLOT NO 11
X
X
K
__ *
cos
17.00 19.00 21.00 23.00 25,00 27.00 29,00 31.00 31 00
"" TEMPERflTURE [HUNDREDS OF DEGREES " "
-------
D
O
o
o
D
ft
A
CD
0
I—/
cc
CE
i — i
Q
LdO
CE
»— «
CD
LJ
OEflD BURN STUDY/CROSS PLOT NO 12
K
vo
Ui
*
i i i i r i i i
17.00 19.00 21.00 23.00 25.00 27.00 29.00 31.00 33 00
CflLCINE TEMPERflTURE CHUNDREDS OF DEGREES H=X
-------
o
o
8.
o
o
I
o
o.
o
03
CJ
DC
LJ
CO
Q-8
UJ
CD
CES
DERD BURN STUDY/CROSS PLOT NO 13
*
fsi
37,00 19.00 21.00 23.PO 25,00 27.00 29.00 31.00 33 00
CRLCINE: TEMPERRTURE (HUNDREDS or DEGREES FD=X
-------
197
APPENDIX C
Cl. Calcined and Uncalcined Limestones
Ranked According to Bed Weight Gain
C2. EPA-APCO Limestone Inventory
-------
198
Cl. Calcined and Uncalcined Limestones Ranked Accordingto
Bed Weight Gain
The following limestones were tested in a fixed-bed reactor
to determine their capacity to react with flue gas containing
sulfur dioxide. The equipment and experimental procedure has
(29)
been documented by Potter and is also described in Section C5
earlier in this report. Essentially, a fixed bed of 20 grams
of calcine is exposed to flue gas at 980°C for 3 1/2 hours.
The flue gas, having an average composition of 10.5% C0», 3.4%
02, 9.9% H20, 0.27% S02, 0.003% S03, and 75i9% N2, flows through
the bed at a rate of 425 standard liters per hour.
The bed weight gain of a 20 gram, -18 +20 U.S. mesh, calcined
sample (expressed as g S0~/100 g of calcined stone) was selected
as the index of sulfur oxide capacity. For uncalcined samples
a weight equivalent to 20 g of calcined stone was used. The bed
weight gain, also referred to as "loading" or "capacity", is the
weight of SO- removed from the flue gas. Data for samples pre-
calcined at 980°C for 4 hours are presented in Table Cl-1. Data
for uncalcined stones are shown in Table Cl-2.
The time to 20% breakthrough is the time required for the S0'2
concentration of the gas effluent from the fixed bed of limestone
to reach 20% of the inlet SO™ concentration.
-------
199
CaO utilization is the percent CaO which reacted based upon
the bed weight gain and percent CaO in the limestone sample.
Spectrochemical analyses were provided by Bituminous Coal Research,
Inc. (BCR). The limestone numbers in this table were originally
assigned by BCR and limestones were referenced by a "BCR Number".
However, all limestones in the EPA-APCO limestone inventory are
now referenced by a "Lime&tone Number".
-------
TABLE Cl-1. CALCINED LIMESTONES RANKED ACCORDING TO BED WEIGHT GAIN
Composition, weight %
Limes tone
Number
1371
1376
1336
1682
1341
1381
1379
1375
1382
1340
1677
1353
1354
1699
134
1342
1684
1696
1695
1343
1384
1694
1369
1363
1697
CaO
55.00
13.90
95.00
96.00
56.00
88.00
83.00
29.50
95.00
57.00
42.50
55.00
57.00
96.00
57.00
63.00
57.00
57.00
58.00
94.00
97.00
56.00
96.00
66.00
92.00
MgO
11.80
80.00
2.55
0.50
42.00
7.80
3.80
60.00
1.61
37.00
1.50
42.00
37.00
1.10
37.00
26.00
39.00
35.50
31.50
0.85
1.10
25.00
1.40
1.42
3.70
Fe 0
/ J
1.65
1.63
0.26
0.13
0.46
0.58
0.66
1.65
0.23
2.63
0.95
1.00
0.71
0.20
2.63
0.23
0.84
1.22
0.93
0.66
0.20
1.10
0.20
2.50
0.45
^3
4.65
0.86
0.36
0.20
0.12
0.46
1.27
1.12
0.20
0.27
1.35
0.19
0.43
0.30
0.27
0.30
0.30
0.60
1.32
0.73
0.20
1.40
0.30
6.60
0.66
SiO
*
21.20
3.20
1.49
1.40
0.91
2.35
6.60
4.55
1.12
1.95
53.00
0.96
5.30
1.26
9.5
9.50
2.53
2.00
4.50
2.98
1.00
13.30
1.16
21.00
2.40
CaO+MgO
66.80
93.90
97.55
96.50
98.00
95.80
86.80
89,50
96.61
94.00
44.00
97.00
94.00
97.10
89.00
89.00
96.00
92.50
89.50
94.85
98.10
81.00
97.40
67.42
95.70
CaO Utili-
zation %
00.00
4.92
6.60
9.40
17.00
13.60
17.10
48.65
15.20
27.20
36.60
30.70
31.70
18.60
30.00
29.80
32.80
33.30
33.50
21.60
21.10
37.50
22.70
32.20
24.90
Time to 20%
B . T. , min.
00.00
15.00
5.75
15.00
7.50
36.00
12.00
27.00
21.00
30.00
32.00
26.00
25.00
41.00
17.00
33.00
30.00
40.00
28.00
54.00
28.00
30.00
49.00
60.00
47.00
Bed Wt. Gain,'
grams
00.00
1.37
1.82
2.52
2.71
3.41
4.08
4:11
4.17
4.41
4.44
4.85
5.12
5.14
5.34
5,
5,
5.
35
39
47
5.55
5.79
5.84
6.01
6.30
6.38
6.55
S3
o
o
. Breakthrough time is time in minutes required for S02 concentration of effluent gas to reach 20%
of inlet S02 concentration.
Gain in weight of bed having initial weight equivalent to 20 -grams of calcined stone.
-------
TABLE Cl-1. CALCINED LIMESTONES RANKED ACCORDING TO BED WEIGHT GAIN (continued)
Composition, weight %
1 Bed Wt. Gain,2
grams
Limestone
Number
1702
1362
1358
1352
1368
1383
1364
1678
1356
1380
1382
1691
1690
1355
1686
1361
1357
1346
1367
1373
1680
1693
1343
1345
1688
CaO
«""•«••"'«•
60.00
56.00
59,00
56.00
82.00
76.00
57.00
28.50
53.00
57.00
95.00
77.00
34.00
.62.00
56.50
75.00
77.00
46.00
58.00
85,00
49.00
93.00
56.00
63.00
55.00
MgO
32.50
3.50
10.00
31.00
2.17
1.29
3.50
21.00
36.00
41.00
1.61
2.77
23.50
1.82
41.50
7.25
3.35
28.00
28.20
5.30
33.50
0.96
42.00
17.00
42.00
Fe^O
2 3
1.35
18.90
4.00
1.75
1.54
3,65
2.90
1.90
7.50
0.56
0.23
0.86
1.59
2.36
0.41
5.10
1.85
2.55
9.30
1.06
1.48
0.40
0.38
1.54
0.86
Al«0,
STT3
0.79
3.65
4.20
2.48
2.20
3.10
8.50
3.15
0.42
0.30
0.20
2.03
2.30
5.60
0.30
2.37
2.80
3.40
0.35
0,95
0.94
0.82
0.22
2.48
0.30
SiO
/
4.10
15.50
20.40
7.40
9.70
13.10
25.60
41.00
1.50
1.53
1.12
15.30
34.60
26.70
1.32
8.20
14.00
17.60
3.85
4.20
13.20
4.00
0.89
13.60
1.00
CaO+MeO
92.50
59.50
69.00
87.00
84.17
77.29
60.50
49.50
89.00
98.00
96.61
79.77
57.50
63.82
98.00
82.15
80.35
74.00
86.20
90.30
82.50
93.96
98.00
80.00
97.00
CaO Utili-
zation %
45.80
44.50
43.70
46.70
32.80
30.40
49.80
107.00
56.80
53.50
32.00
39.70
90.00
50.80
57.50
43.80
42.80
75.50
61.00
40.50
76.90
40.60
69.50
62.90
71.50
Time to 20%
B. T. ? min.
35.00
68.00
58.00
60.00
60.00
68.00
57.00
67.50
62.00
60.00
80.00
66.00
67.00
68.00
52.00
82.00
72.00
96.00
82.50
107.00
95.00
75.00
104.00
79.00
87.00
7.08
7.12
7.42
7.49
7.76
8.02
8.08
8.66
8.66
8.70
8.70
8.74
8.75
8.97
9.21
9.39
9.41
9.99
10.14
10.16
10.76
10.82
11.17
11.29
11.35
-------
1701
TABLE Cl-1. CALCINED LIMESTONES RANKED ACCORDING TO BED WEIGHT GAIN (continued)
Composition, weight %
Limes tone
Nuirih er
H \.iui^ c *•
1374
1378
1930
1698
1365
1370
1366
1351
±-J <~t _L.
1685
1679
CaO
65.00
56.00
55.00
72.00
52.50
53.00
66.00
54.00
86.00
91.00
MgO
**a~
18.50
41.00
43.00
19.50
22.00
36.30
16.50
28.50
8.20
2.40
Fe-0-
— 2—3
1.70
0.37
0.36
2.35
15.50
0.79
7.10
7.00
0.69
0.59
^•£0
— 5P3
1.28
0.23
0.14
0.39
1.81
1.21
1.27
1.55
0.25
1.00
SiO
12.20
2.42
0.78
5.85
7.25
7.50
7.50
8.20
2.72
4.10
CaO+MgO
89.50
97.00
98.00
91.50
74.50
89.30
82.50
82.50
94.50
93.40
CaO Utili-
zation %
60.10
74.30
75.90
58.40
84.20
84.80
68.60
86.00
55.00
53.20
Time
B. T.
107
87
132
110
114
100
115
117
138
118
to 20%"
, min.
.00
.00
.00
.00
.00
.00
.00
.00
.50
.00
Bed Wt. Gain,
grams
11
11
12
12
12
12
12
13
13
13
.47
.90
.14
.59
.63
.72
.92
.07
.51
.84
60.00 23.50
7.40
0.73
6.80
83,50
81.00
112.00
13.93
NJ
O
-------
TABLE Cl-2. UNCALCINED LIMESTONES RANKED ACCORDING TO BED WEIGHT GAIN
Composition, weight %
Limes tone
Number
1371
1376
1341
1368
1355
1375
1379
1336
1369
1699
1695
1694
1697
1363
1702
1354
1686
1343
1362
1696
1358
1684
1381
1688
1352
CaO Utili- Time to 20%
CaO
55.00
13.90
56.00
82.00
62.00
29.50
83.00
95.00
96.00
96.00
58.00
56.00
92.00
66.00
60.00
57.00
56.50
94.00
56.00
57.00
59.00
57.00
88.00
55.00
56.00
MgO
11.80
80.00
42.00
2.17
1.82
60.00
3.80
2.55
1.40
1.10
31.50
25.00
3.70
1.42
32.50
37.00
41.50
0.85
3.50
35.50
10.00
39.00
7.80
42.00
31.00
Fe 0
• — «£— J
1.65
1.63
0.46
1.54
2.36
1.65
0.66
0.26
0.20
0.20
0.93
1.10
0.45
2.50
1.35
0.71
0.41
0.66
18.90
1.22
4.00
0.84
0.58
0.86
1.75
A1~P_
2 o
4.65
0.86
0.12
2.20
5.60
1.12
1.27
0.36
0.30
0.30
1.32
1.40
0.66
6.60
0.79
0.43
0.30
0.73
3.65
0.60
4.20
0.30
0.46
0.30
2.48
SiO.
L
21.20
3.20
0.91
9.70
26.70
4.55
6.60
1.49
1.16
1.26
4.50
13.30
2.40
21.00
4.10
5.30
1.32
2.98
15.50
2.00
20.40
2.53
2.35
1.00
7.40
CaO+MgO
66.80
93.90
98.00
84.17
63.82
89.50
86.80
97.55
97.40
97.10
89.50
81.00
95.70
67.42
92.50
94.00
98.00
94.85
59.50
92.50
69.00
96.00
95.80
97.00
87.00
zation %
0.67
67.20
18.00
14.30
19.00
41.00
14.30
15.33
16.20
17.20
32.20
33.70
20.80-
29.20
42.90
37.20
40.60
22.70
39.80
39.20
39.20
41.20
27.40
44.00
43.96
B. T. , min.
00.00
21.00
7.50
19.00
52.00
16.00
19.00
23.30
35.00
31.00
21.00
34.00
32.00
31.00
27.00
36.00
67.00
41.00
45.00
12.00
45.00
35.00
45.00
40.00
26.00
Bed Wt. Gain,2
grams
00.00
2.67
2.89
3.36
3.41
3.47
3.54
4.18
4.49
5.00
Ni
5.35 8
5.40
5.47
5.49
6.03
6.03
6.05
6.09
6.37
6.39
6.68
6.73
6.92
6.96
7.03
-------
TABLE Cl-2. UNCALCINED LIMESTONES RANKED ACCORDING TO BED WEIGHT GAIN, (continued)
Composition, weight %
•Limestone
Number
1342
1373
1353
1677
1693
1680
1357
1678
1681
1380
1356
1361
1682
1384
1691
1690
1930
1345
1346
1366
1701
1367
1698
1364
1365
1351
1679
1374
CaO
63.00
85.00
55.00
42.50
93.00
49.00
77.00
28.50
72.00
57.00
53.00
75.00
96.00
97.00
77.00
34.00
55 00
63.00
46.00
66.00
60.00
58.00
72.00
57.00
52.50
54.00
91.00
65.00
MgO
26.00
5.30
42.00
1.50
0.96
33.50
3.35
21.00
3.45
41.00
36.00
7.25
0.50
1.10
2.77
23.50
43.00
17.00
28.00
16.50
23.50
28.20
19.50
3.50
22.00
28.50
2,40
18.50
FeJD
0.23
1.06
1.00
0.95
0.40
1.48
1.85
1.90
0.96
0.56
7.50
5.10
0.13
0.20
0.86
1.59
0.36
1.54
2.55
7.10
7.40
9.30
2.35
2.90
15.50
7.00
0.59
1.70
—2^3
0.30
0.95
0.19
1.35
0.82
0.94
2.80
3.15
1.20
0.30
0.42
2.37
0.20
0.20
2.03
2.30
0.14
2.48
3.40
1.27
0.73
0.35
0.39
8.50
1.81
1.55
1.00
1.28
SiO
9.50
4.20
0.96
53.00
4.00
13.20
14.00
41.00
16.90
1.53
1.50
8.20
1.40
1.00
15.30
34.60
0.78
13.60
17.60
7.50
6.80
3.85
5'. 8 5
25.60
7.25
8.20
4.10
12.20
CaO+MgO
89.00
90.30
97.00
44.00
93.96
82.50
80.35
49.50
75.45
98.00
89.00
82.15
96.50
98.10
79.77
57.50
98.00
80.00
74.00
82.50
83.50
86.20
91.50
60.50
74.50
82.50
93.40
89.50
CaO Utili-
zation %
39.80
29.20
47.20
62.00
28.40
55.50
35.80
100.00
39.60
51.20
54.90
40.00
32.40
32.70
43.80
101.80
66.30
58.30
81.50
58.20
60.80
70.00
55.20
85.30
90.20
93.60
55.00
.80.20
Time
B. T.
70
64
31
45
64
48
42
45
62
49
137
75
62
43
87
52
99
79
112
113
105
82
102
62
165
150
167
115
to 20%
, min.
.00
.00
.00
.00
.00
.00
.00
.00
.00
.00
.00
.00
.00
.00
.00
.00
.50
.00
.00
.00
.00
.00
.00
.00
.00
.00
.00
.00
Bed Wt. Gain,
grams
7
7
7
7
7
7
7
8
8
8
8
8
8
9
9
10
10
10
10
11
11
11
11
13
13
14
14
15
.17
.32
.45
.55
.58
.80
.91
.12
.14
.27
.32
.59
.96
.08
.63
.04
.46
.49
.78
.00
.34
.60
.85
.90
.93
.22
.42
.04
-------
205
C2. EPA-APCO Limestone Inventory
The following is a listing of the current limestone inventory
maintained by EPA-APCO. These samples are located in Cincinnati,
Ohio, at two main locations: (1) Research Laboratory Branch,
Fairfax Facility, 3914 Virginia Avenue, and (2) U.S. Post Office,
tyain Building, Room 37C.
The limestone samples are listed according to limestone number,
source, identification of material, and the amount of material
on hand as of August 1970.
-------
206
EPA-APCO LIMESTONE INVENTORY - August, 1970
Limestone
Number
1335
1336
1337
(reordered
as 1930)
1338
1339
1340
1341
1342
1343
(reordered
as 1700)
1344
1345
1346
Source
Hills Material Co.
Rapid City, South Dakota
Supt. of Quarry -
John Holmes
Georgia Marble Co.
Tate, Georgia
Mr. Hall
Min., Pigm., & Metals,
Chas. Pfizer Co.
Gibsonburg, Ohio
Mr. Plantz
Number Not Used
(changed to 1693)
Number Not Used
Nantahala Talc & Limestone
Lexington, Kentucky
Mr. Ferebee
New England Lime Co.
Div. of Chas. Pfizer
& Co., Inc.
Canaan, Connecticut
Mr. Ingram
Conklin Limestone Co.
Lincoln, Rhode Island
C. E. Conklin
Hooper Bros. Quarries
Weeping Water, Nebraska
Mr. Hooper
Number Not Used
New Point Stone Company
Batesville, Indiana
Norman A. Wanstrath
Mankato Aglime & Rock Co.
Mankota, Minnesota
Amount of
Identification Material on
of Material Hand. (Ibs.)
Limestone Dust 3
//2 White crushed
marble
#10 glass house
stone
Aglime
Limestone
Lime
132
38
83
59
67
1/8" down aglime 101
258
137
-------
207
EPA-APCO LIMESTONE INVENTORY - August, 1970
Limestone
Number
1347
1348
1349
1350
1351
(reordered
as 1701)
1352
(reordered
as 1702)
1353
1354
1355
1356
1357
1358
Source
Number Not Used
J. E, Baker Co.
Millersville, Ohio
TVA Paradise Steam Plant
Drakesboro, Kentucky
Owen C. Janow
Batesville White Lime
Batesville, Arkansas
General Manager - Mr. Cobb
Jeffrey Limestone Co.
Parma, Michigan
John C. Jeffrey
Millard Limestone Co.
Annville, Pennsylvania
J. E. Baker Co.
York, Pennsylvania
Mr. Paul
Montevallo Limestone Co.
Montevallo, Alabama
Elkins Limestone Co.
P. 0. Box 1228
Elkins, West Virginia
Darrel Hankey
Valley Dolomite
Bonne Terre, Missouri
Louis Huber
TVA Paradise Steam- Plant
Drakesboro, Kentucky
Owen C. Janow
Hanna Coal Company
Adena, Ohio
Identification
of Material
Amount of
Material on
Hand, (Ibs.)
Ground Raw Stone
Greenville Quarry
I '
205
244
147
130
210
Ground Stone
Dolomite stone
Ag Lime
Granular limestone 67
8 mesh sand
232
Ag Lime
83
Greenville Quarry 176
II, Limestone gravel
Ag Lime
133
-------
208
EPA-APCO LIMESTONE INVENTORY - August. 1970
Limestone
Number
1359
(reordered
as 1699)
1360
(reordered
as 1698)
1372
1373
1374
1375
1376
1377
1378
1379
1380
Source
Grove Lime Company
Stephens City, Virginia
Monmouth Stone Co.
Monmouth, Illinois
Dan Kistler
L. F. Rooney
Dept. of Geology
Indiana University
Bloomington, Indiana
L. F. Rooney
Dept. of Geology
Indiana University
Bloomington, Indiana
L. F. Rooney
Dept. of Geology
Indiana University
Bloomington, Indiana
Basic Chemicals
Cleveland, Ohio
Basic Chemicals
Cleveland, Ohio
TVA Russfield
Nashville, Tennessee
Verplanks Coal & Dock
Fettysburg, Michigan
Limestone Products of Amer.
P. 0. Box 490
Newtown, New Jersey 07860
Mr. Thompson
Rockwell Lime Co.
Manitowac, Wisconsin
Pres. - Michael Brisch
Amount of
Identification Material on
of Material Hand, (Ibs.)
Ag Lime
Sample B
Rockford, Indiana
Sample C
North Vernon,
Indiana
Sample D
Position 8
140
29
72
84
180
Fluxing Fines
Crestite #3
119
108
78
136
190
Dolomitic Lime- 148
stone 1/4" to 1/8"
-------
209
EPA-APCO LIMESTONE INVENTORY - August, 1970
Limestone
Number
1381
1382
1383
1384
1677
1678
(reordered
as 1931)
1679
1680
1681
1682
1683
1684
Source
Teeter Stone, Inc.
Get tysburg, Pennsylvania
Hemphill Bsos., Inc.
Seattle, Washington
Colarusso & Sons
Hudson, New York
Southern Materials
P. 0. Box 218
Ocala, Florida
V. Pres. - H. B. Roberts, Jr,
Lone Star Materials
P. 0. Box 918
Austin, Texas
Noble W. Prentice
Dolomite Products
Buffalo fie Howard Road
Rochester, New York
Indus Limestone
Immokalee, Florida
John S. Lane & Son, Inc.
Meriden, Connecticut
Countyline Stone Co.
County Line Road
Akron, New York 14001
Pres. - John W. Buyers
California Rock & Gravel
1800 Hobart Building
582 Market Street
San Francisco, California
Pres. - F. N. Woods III
Union Carbide Corp.
New York, New York
G. & W. H. Corson, Inc.
Plymouth Meeting, Pa. 19462
Wm. Webster
Amount of
Identification Material on
of Material Hand, (Ibs.)
158
116
120
Coarse Granular 200
CaCO,
Limestone
Screenings
201
Crushed Limestone
5/32" + #35
3/4" limestone
114
88
172
249
127
12
170
-------
210
EPA-APCO LIMESTONE INVENTORY - August, 1970
Limestone
Number
1685
1686
1687
1688
1689
(reorder
of 1361)
1690
1691
1692
1693
1694
1695
Source
Raid Quarries, Inc.
F. & M. Bank Building
Budington, Iowa
Paul R. Orr
Mayville White Limeworks
Mayville, Wisconsin 53050
James 0. Smith
Identification
of Material
Aglime
Amount of
Material on
Hand. (Ibs.)
155
Fertilizer Grit
Company
Frount & 8th Sts.
Quincy, Illinois 62301
Dir. of R&D - Vernon R. Heaton
Quincy Granular
Limestone
234
217
Marblehead Lime Co.
300 W. Washington St.
Chicago, Illinois 60606
John F. Romanyak
Tennessee Valley Authority
Paradise Steam Plant
Drakesboro, Kentucky
Gene Farmer
McConville, Inc.
Ogdensburg, New York
Warren Brothers Roads Co.
Syracuse, New York
Partin Limestone Products
San Bernadino, California
Pete Lien & Sons
Rapid City, South Dakota
Willingham - Little
Ga. Marble Co., Stone Div.
Lohitestone, Georgia
Willingham - Little
Ga. Marble Co. Stone Div.
Jasper, Georgia
Dolomitic lime-
185
158
Very Fine Ground
Pool Covering
Gravel
Georgia White
Grandlux
Willingham White
Size #10
210
260
234
76
214
166
-------
211
EPA-APCO LIMESTONE INVENTORY - August. 1970
Limestone
Number
1696
1697
1698
(reorder
of 1360)
1699
(reorder
of 1359)
1700
(reorder
of 1343)
1701
(reorder
of 1351)
1702
(reorder
of 1352)
1921
1930
(reorder
of 1337)
1931
(reorder
of 1678)
2057
2058
Source
James River Hydrate
Supply Co.
Buchanah, Virginia
Marble Cliff Quarries Co.
Carntown, Kentucky
Monmouth Stone Co.
Monmouth, Illinois
Grove Lime Co.
Stephens City, Virginia
Hooper Bros. Quarries
Weeping Water, Nebraska
Jeffrey Limestone
Parma, Michigan
John C. Jeffrey
A. E. Millard Limestone Co.
Annville, Pa. 17003
Mr. Schredder
Chas. Pfizer & Co., Inc.
260 Columbia Street
Adams, Massachusetts
Chas. Pfizer & Co.,
Gibsonburg, Ohio
Inc.
Dolomite Products
Buffalo & Howard Road
Rochester, New York
TVA Prototype Sample
Kentucky Stone
Irvington, Kentucky
TVA Prototype Sample
Kentucky Stone
Russellville, Kentucky
Identification
of Material
#3
Aglime
Aglime
#2 Dolomite
Limestone
Amount of
Material on
Hand, (Ibs.)
196
249
40
158
205
138
139
Calcite crystals
95 brand
Limestone
Limestone
98
53
138
16
16
-------
212
EPA-APCO LIMESTONE INVENTORY - August. 1970
Limestone
Blue Limestone
White Limestone
Limestone
Number Source
2059 TVA Prototype Sample
Road Material, Inc.
2060 TVA Prototype Sample
Fredonia Valley
2061 TVA Prototype Sample
Fredonia Valley
2062 TVA Prototype Sample
Cedar Bluff Limestone
2063 TVA Prototype Sample Limestone
National Gypsum
2064 TVA Prototype Sample
Alabaster Limestone
2065 TVA Prototype Sample
Longview Limestone
2066 TVA Prototype Sample
Hoover Limestone
2067 TVA Prototype Sample
Rigsby and Bernard Limestone
2068 TVA Prototype Sample
Williams Limestone
2069 TVA Prototype Sample
James River Limestone
2070 TVA Prototype Sample
Greenville Limestone
2071 TVA Prototype Sample
Marble Cliff Limestone
Amount of
Identification Material on
of Material Hand, (Ibs.)
Sample 1697 from
Marble Cliff,
Caiutown, Kentucky
16
16
50
16
16
16
20
15
Sample 1696 from 16
James River Hydrate
2072 TVA Prototype Sample
Lambert & Lambert Limestone
16
21
-------
213
EPA-APCO LIMESTONE INVENTORY - August. 1970
Limestone
Number
2073
2074
2075
2076
2077
2078
2079
2080
2081
2082
2083
2084
2085
Source
TVA
TVA
TVA
TVA
Longhorn Portland Cement Div.
Kaiser Cement and Gypsum
San Antonio, Texas 78218
Mr. Art Knutson
Indiana Limestone Co.
P. 0. Box 72
Bedford, Indiana
Mr. Gary (G. W.) Gaiser
Indiana Limestone Co.
P. 0. Box 72
Bedford, Indiana
Mr. Gaiser
Ohio Geological Survey
Columbus, Ohio
Mr. Horace Collins
State Geological of Kansas
Lamerence, Kansas 66045
Mr. R. G. Hardy
Ward's Natural Science
Establishment
Rochester, New York
Ward's
Rochester, New York
Ward's
Identification
of Material
Ward's
Austin
Chalk
Standard Bluff
Rustic Gram
Amount of
Material on
Hand. (Ibs.)
Marl (Tufa)
Northeastern Ohio
Kansas Chalk
Iceland spar
Second grade
Iceland spar
First grade
Dolomite
Selasvann, Norway
White Chalk
Dover Cliffs
England
83
58
70
98
55
-------
214
EPA-APCO LIMESTONE INVENTORY - August. 1970
Limestone
Number
2086
2087
2105
2106
2107
2108
2109
2110
2111
2112
2113
Source
TVA
Spring Valley Limestone
Lone Star Cement
P. 0. Box 839
Demopolis, Alabama
Mr. Mike Ried
Ward's
Ward *s
Filer's
P. 0. Box 995
Loma Linda, Calif. 92354
Identification
of Material
Amount of
Material on
Hand, (Ibs.)
Colbert County
Limestone
Selma Chalk
100
94
Sharp pseudo- 20 pieces
hexagonal aragonite
crystals from Spain
New York marl 1
California, tan
banded aragonite
Eckert Educational Mineral
Research Co.
1244 East Colfax
Denver, Colorado 80218
Lois Hurianek (303) 272-8943
Massive Colorado
Aragonite
12
Soil Conservation Service
U.S. Dept. of Agriculture
Midtown Plaza
Syracuse, N. Y.
Mr. Bernard Ellis
Soil Conservation Service
U.S.D.A.
Soil Conservation Service
U.S.D.A.
Soil Conservation Service
U.S.D.A.
Marl, Canastota,
Madison Cok, N.Y.
10
Marl, Canastota
Madison Co., N.Y.
Marl, Prattsburg,
Stuben Co., N.Y.
Marl, Gorham,
Ontario Co., N.Y.
35
36
27
Missouri Lead Operating Co. Lead doped Dolomite, 40
Boss, Missouri 65440
Mr. J. H. Davis
5/8"
-------
215
EPA-AKO LIMESTONE INVENTOR* - August. 1970
Limestone
Number
2114
2115
2116
2117
2118
2119
2120
2121
2122
2123
2124
2125
2126
2127
2128
Source
Ward's
Ward's
Ward's
Ward's
Ward's
Ward's
Ward's
Ward's
Ward's
Ward's
Ward's
Ward's
Ward's
Ward's
Ward's
Amount of
Identification Material on
of Material Hand, (Ibs.)
English Oolitic
Limestone
Indiana, Bull,
Oolitic limestone
Pennsylvania,dark
gray, oolitic
limestone
Chocolate, 1
Tennessee, Marble
Black, Vermont, 2
Marble
Gray-White
Mississippi Chalk
Fine crystalline 3
White Vermont Marble
Coarse crystalline 5
White, Georgia Marble
Coarse crystalline 2
Pink, Georgia
Coarse, White 2
dolomitic, New York
Marble
Fine white dolomitic 1
Mass. Marble
Travertine,
Tivoli, Italy
Calcareous, Tufa 2
Mumford, N. Y.
Cherty Limestone 2
Leroy, New York
Oolitic Limestone
Bedford, Indiana
-------
216
EPA-APCO LIMESTONE INVENTORY - August. 1970
Limestone
Number
2129
2201
2202
2203
2204
2205
2206
2207
2208
(1351)
Source
Babcock & Wilcox
(from S. K. Vorres
R. R. #2
Fremont, Michigan 47912)
Ecjunga, Durango, Mexico
Supplier: Filer's
Redlands, California (1)
Hillside Mine Dump
Rosiclare, Illinois (1)
Columbia Quarry
Company Mine
Valmeyer, Illinois (1)
Allied Stone Co. Quarry
Milan, Illinois (1)
Midway Stone Quarry
Osborne, Illinois (1)
Abandoned Quarry
Bourbonnais, Illinois (1)
Red Mountain District
Santa Clara County,
California (i)
Identification
of Material
Newaygo County
Michigan marl
Amount of
Material on
Hand. (Ibs.)
20
Calcite, Iceland 1
spar, IGS Type 1
(2)
Calcite, 20
IGS Type 2
Limestone, coarse- 30
grained, high
purity, ISG Type 3
Limestone, fine- 30
grained, high
purity, IGS Type 4
Dolomite, Reef type, 30
high purity, IGS
Type 5
Dolomite,
IGS Type 6
Magnesite, fine-
grained, high
purity, IGS Type 7
North Cat Cay, Bahama Islands Aragonite,
Supplier: Ocean Industries, IGS Type 8
Ft. Lauderdale, Florida
30
10
18
Jeffery Limestone Company Dolomite,
Quarry IGS Type 9
Parma, Michigan (1)
Same, as Lime-
stone 1351
(1) Supplier: R. D. Harvey, Illinois State Geological Survey
(2) ISG - Illinois State Geological Survey
-------
217
EPA-APCO LIMESTONE INVENTORY - August. 1970
Limestone
Number
2210
2211
Source
Chrzanow, Poland
Supplier: Power Metrology
Research Organization
Energopomiar,
R. T. Chrusciel, Glivice,
Poland
Nova Huta Plant
Krakow, Poland
Supplier: Power Metrology
Research Organization
Energopomiar,
R. T. Chrusciel, Glivice,
Poland
Identification
of Material
Dolomite for
Nova Huta Iron
& Steel Plant,
uncalcined
Limestone No. 2210
Calcined in Nova
Huta Plant, iron
and steel process
Amount of
Material on
Hand. (Ibs.)
-------
218
APPENDIX D
Sorption of S0_ by Waste Kiln Dust from
Portland Cement Manufacturing Operations
-------
219
Sorption of SO^ by Waste Kiln Dust from
Portland Cement Manufacturing Operations
In addition to limestone as a material for sorbing S0_ from
power plant flue gases by dry-injection processes, other materials
have also been considered. Some interest has been shown in kiln
dust as a S0? sorbent. Kiln dust consists of the fines and con-
densed volatiles blown from the raw mix as it passes through a
rotary kiln in portland cement manufacturing processes. This
dust is collected from kiln exhaust systems in cyclones and
baghouses. In order to maintain an acceptable low alkali content
in the cement product, the kiln dust cannot be recombined with
the product.
Table D-l lists chemical analyses for three kiln dust samples.
Table D-2 lists the reactivity of these kiln dust samples to flue
gas in a differential reactor. Details of the design and operation
of this reactor -are given in the paper, "Kinetics of the Reaction
of SO with Limestone"1. Results for a 100 second exposure to
flue gas are reported in milligrams S03 per 30 milligrams calcined.
When as-received or uncalcined material is tested, the charge
weight used is equivalent to 30 milligrams calcined. Absorptivity
of the kiln dust is comparable to but no better than Stone No. 2061,
an oolitic calcite. Moreover, the Na20 content of kiln dust may
be as high as 2.5%,2 whereas, it is only a trace quantity (<0.02%)
1 Environmental Science & Technology, Jan. 1970, p. 59-63.
2 Nebgen, J. W., et al., Midwest Research Institute Prospectus,
No. RC-199, May 5, 1970, p. 2
-------
220
Table D-l
Analysis o£ Kiln Dust^ ^
% Loss on Ignition @>1800°F
% Silica (Si02)
% Iron Oxide (Fe,03)
% Aluminum Oxide (Al^O.,)
% Calcium Oxide (CaO)
% Magnesium Oxide (MgO)
North Western
States* '
23.6
15.5
2.6
6.6
49,5
0.9
Medusa
17.3
15.1
3.1
7.3
52.9
3.1
(4)
American
27.2
15.9
3.0
4.0
47.5
2.2
^ ' Analysis by Gilbert Associates, Inc., 2249 Fairview St.,
Reading, Pennsylvania 19606. Analyses for alkali content
not requested.
(2)
Sample provided by North Western States Portland Cement Co.,
Mason City, Iowa.
C3)
Sample provided by Medusa Portland Cement Co., Box 5668,
Cleveland, Ohio 44101.
(4)
Sample provided by American Cement Corp.
-------
221
Table D-2
Reactivity of Kiln Dust in a Differential Reactor1
Temperature
°F
soa
1000
1200
1400
1600
1800
2
Stone No.
2061
(1.5)*
(3.1)*
4.6
6.4
8.0
9.4
mg 50^/30 mg calcine
Samples from Cement Companies
North Western
States Medusa
2.3
2.1
2.9
6.2
6.5
7.2
3.7
3.1
4.4
6.2
6.4
5.9
(uncalcined)
American
—
0.5
1.8
3.0
2.7
4.8
* extrapolated
Note: 1. 100 second exposure to flue gas of following compositions
Component Volume %
C02 10.5
°2 3'4
H20 9.9
S02 O-27
S03 0.003
N2 75.9
2, Sample supplied by Fredonia Valley Quarries, Kentucky
(White limestone).
-------
222
in most limes. Unfortunately, the presence of sodium in fly
ash decreases fusion temperature of fly ash which leads to
increased quantity and strength of deposits encountered around
the superheater and reheater tubes in a power boiler. Thus,
injection of kiln dust to remove S0_ before these tubes would
likely lead to unacceptable periods of "downtime" for tube
cleaning and loss of heat transfer efficiency.
Table D-3 indicates the reactivity of several kiln dust samples
in an aqueous batch scrubber. A description of this apparatus
is given by Potter . The procedure is to bubble flue gas
through a stirred slurry of the tested material and to monitor
the pH of the solution and the S0? content of the scrubbed gas.
An important datum is the breakthrough time, i.e., the length
of time during which the slurry successfully removes all of
the sulfur oxides. The pH of the solution at breakthrough as
well as the breakthrough time in minutes is given. Reactivity
of the kiln dust is comparable to finely ground uncalcined
limestone.
International Symposium for Lime and Limestone Scrubbing,
Pensacola, Florida, March 16-20, 1970.
-------
223
Table D-3
Reactivity of Kiln Dust in anAqueous Scrubber
North Western
States Kiln Dust
Uncalcined
Calcined
Medusa Kiln Dust
Uncalcined
Calcined
Stone No. 1700^
Dnealcined
Calcined
Breakthrough
Time, min.
44
41
41
39
45
61
pH at
Breakthrough
4.3
4.5
4.4
4.3
5.3
5.5
A calcitic aglime supplied by Hooper Brothers Quarries,
Weeping Water, Nebraska.
-------
224
APPENDIX E
Copper Oxide Sorption of SO-
-------
225
Sorption of SO^ by Copper Oxides
The Air Preheater Company and the Owens-Corning Fiberglas
Corporation have established the fabric filterhouse to be an
effective chemical contactor for removal of sulfur dioxide from
flue gas streams. It has been suggested that one potential
application for the filterhouse would be to remove S09 from
the flue gases from the reverbatory furnace in the copper
smelting industry. The idea would be to inject either cupric
oxide or the solid effluent from the roaster or reverberatory
furnace into a filterhouse along with the effluent gases from
the reverberatory furnace. After sorbing S0« the solid reaction
products would then be fed back into the roaster feed stream
where the sorbed S0? would be driven off in the roaster to
enrich the roaster effluent gases which contain five to ten
percent S0~ by volume. The roaster effluent gases containing
SO are normally sent to a sulfuric acid manufacturing process.
An exploratory investigation was made to show the reactivity
of cupric oxide to sulfur dioxide. Chemical grade cupric oxide
(Baker A.C.S. grade) was exposed to pure S02 in a thermogravi-
metric analyzer (an Aminco Modular Thermo-Grav). In the
thermogravimetric analysis (TGA) mode it automatically records
sample weight change as a function of sample temperature at
-------
226
pre-selected heating rates. In this case the TGA was performed
with 100 milligrams of cupric oxide exposed to a 100 cc/min SO-
purge and heated at a rate of 6°C/min. Figure E-l shows the
reactivity of cupric oxide with SO-.
As shown in Figure E-l the cupric oxide reached a maximum weight
gain of 50% at a temperature of about 720°C, after which the
reacted sample underwent decomposition. The significance of
the break in the graph between 560 - 695°C is not known.
The significant information to be obtained from Figure E-l is
that at 3l5°C (600°F) there is only a 2-3 percent weight gain
by the cupric oxide. Since commercially available fabric
filters have a maximum operating temperature of 600°F, cupric
oxide would not be art effective sorbent for a fabric filterhouse
application unless fabric filters with higher operating
temperatures can be developed.
-------
50
10
N>
50 150 240 340 445 500 620 695 750 870 950
TEMPERATURE, °C.
Figure E-l. Thermogravimetrlc Analysis of Reactivity of Cupric Oxide to SO-
-------
228
APPENDIX F
Char Sorption of SO and Regeneration
-------
229
CHAR SORPTION OF S02 AND REGENERATION
A number of processes under development for the control of
sulfur oxides emoloy the catalytic properties of activated
carbon to oxidize sulfur dioxide to sulfur trioxide on the
carbon surface and to retain the sulfuric acid formed from the
sulfur trioxide in the pores of the activated carbon. Thermal
regeneration of the carbon results in significant consumption
of the rather expensive activated carbon. In order to avoid
this problem associated with the regeneration of the spent
carbon, the West Virginia Pulp and. Paper Company (Westvaco)
has suggested a reductive type regeneration in which hydrogen
sulfide is used* The hydrogen sulfide is reacted first with
a portion of the sulfuric acid to yield elemental sulfur. A
part of the sulfur then reacts with the remainder of the
sulfurtc acid, presumably preventing reaction of the sulfuric
acid with the carbon and thereby averting consumption of the
carbon-
Several experiments to check the char sorption-regeneration
scheme proposed by Westvaco were run. The following aspects
of this scheme were tested: (1> S02 sorption capacity of the
char at 25Q°F, (2) the reduction of H2S04-loaded char by ^S
at 300°F, 03) the reduction of H2S04 with sulfur at 600°F.
-------
230
The attached data summarize the results obtained in these
experiments. The following additional comments are offered:
SO Sorption
£.
3
A fixed bed containing 9.38 grams (15 cm ) of char (supplied by
Westvaco) was exposed to flue gas containing 2800 ppm SO- (dry
basis) at 250°F. The gas flow was 505 ml. per minute (21°C)
through the bed, for a space velocity of 2000 hr. . Figure F-l
gives the'breakthrough curve for the run, as measured with an
infrared analyzer continuously recording the SO- concentration
in the reactor effluent gas stream. S07 concentration in the
effluent gas stream reached 10 per cent of the inlet concen-
tration (10% breakthrough) in 6.65 hours, at which time the
char loading was 15.8 wt. percent SO-. Exposure to flue gas
was ended at 8.75 hours. The weight gained by the char during
the total exposure period was 2.80 grams. Integration of the
IR curve indicated an SO- sorption equivalent to 2.83 grams ELSO,.
Reduction
Three samples of char from the experiment described above were
exposed to pure H_S at 300 °F using the apparatus shown in Figure F-2.
The gas volumes were measured in 100 ml burettes using saturated
salt solution (acidified with HC1) as the displacement fluid.
Different volumes of H-S were fed to each char sample, equivalent
to 33%, 88% and 100% stoichiometric to H-SO, based on the assumed
reaction
-------
3000 -
Space Velocity » 2000 hr
-1
2500
S02 Feed Level, 2800 ppm
-------
232
DRYING
TUBE
I
CHAR
SAMPLE
300° F.
CHL BATH
HOT PLATE
FEED BURETTE
GAS COLLECTION
BURETTE
Figure F-2. Apparatus used for H SO,/H2S Reduction Experiments
-------
233
The data obtained are summarized in Table F-l, along with three
additional runs made with as-received char which had not been
exposed to flue gas. Figures F-3, 4 and 5 show the relationship
between the amount of H S fed to the char (at constant rate)
and the amount of gas which passed through the char bed. The
gas collected includes the air displaced from the reactor tube
which had a volume of 28 ml.
H0SO,/S Reaction
""""".£ •~~Xjr~"i—•• ••• • ' •
The char samples obtained from the above H.S treatment were
each heated to 600°F under flowing nitrogen using the apparatus
shown in Figure F-6. The presumed reaction for this stage is
H2S04 + 1/2 S •»• 3/2 S02 + H20
The sulfur dioxide collected in the peroxide absorbers was
analyzed to determine the total S0_ evolved. When S0_ evolution
was completed, the char was removed and extracted in a Soxlet
apparatus with carbon disulfide to recover elemental sulfur.
Results of these tests are listed in Tables F-2 and 3. Finally,
the three samples of as-received char which were exposed to H-S
(runs A, B, C), were composited and extracted with carbon disul-
fide. Sulfur recovery was 0.288 grams for the composite of nine
grams of unloaded char for a baseline loading of 0.032 grams of
sulfur per gram of unloaded char. This may be compared with
the results for runs 1, 2, and 3 in Table F-3 where the maximum
sulfur recovery was 1.52 grams of sulfur on 3.56 grams of loaded
char or 0.745 grains of sulfur per gram of unloaded char (i.e.,
-------
234
Table F-l: H2S04/H2S Reaction Stage at 300°F
Weight H0SO,-loaded char put in reactor tube » 3.82 grams each run
~~VL *f
Feed rate of H S to char during reaction =1.6 ml/min (20°C, 1 atm)
Total gas
Total H2& passed thru
added, ml. char, ml.
Run (Stoich.) (20°C, 1 atm.) (20°C, 1 atm.)
1 (33%)
2 (88%)
3 (100%)
Weight as-received
Feed rate of H-S to
A
B
C
219
575
655
char
char
219
575
655
28
34
76
put in reactor tube =
during reaction » 1.
80
262
381
Wt. gain Wt. gain
by char, by drying
gm. tube, gnu
0.187
0.576
0.713
2.92 grams
6 ml/min.
0.052
0.069
0.046
0.107
0.042
0.035
0.035
0.058
0.052
-------
Total H2S Fed « 33% Stoichiometric
H-SO, Loaded Char
30
60 90 120 150
H2S Fed, ml. (20°C, 1 atm.)
Figure F-3. 300°F H2S Reduction
180
210
240
-------
360 -
Total H2S Fed - 88% Stolchiometric
Char as received
HnSO, Loaded Char
10
60
120
180 240 300 360 420
H2S Fed, ml. (20°C, 1 ata.)
Figure F-4. 300°F ttJS Reduction
460 540
600
-------
360 -
Total H2S Fed « 100% Stoichlattetric
H-SO, Loaded Char
10
H2S Fed, ml. '(20eC, 1 atm.)
Figure F-5. 300°F H.S Reduction
660
-------
238
CHAR
SAMPLE
DRYING
TUBE
ELECTRIC FURNACE
600°F
H202 ABSORBERS
Figure F-6. Apparatus used for H-SO, Reaction Experiments
-------
239
Table F-2: H2SC>4/S Reaction Stage at 600°F
Char from
Run No.
1 (33%)
2 (88%)
3 (100%)
Wt. char to
reactor, gm.
3.85
4.21
4.30
SO- recovered,
gm.
0.568
0.227
0.169
Wt. loss of
char during
reaction, gm
0.900
0.789
0.769
Table F-3: Sulfur Extraction
Char from
Run No. (Table 2)
Wt. char to
Extractor, gm.
Sulfur
recovered, gm.
1 (33%)
2 (88%)
3 (100%)
2.96
3.42
3.56
0.183
0.449
1.52
-------
240
23 times the baseline loading). The work showed conclusively
that ELSO, can be reduced to elemental sulfur in yields exceeding
80 percent at a constant temperature of 300°F, and that the pro-
duct of that reduction can be recovered by solvent extraction
methods.
------- |