r
PB 202 407
INVESTIGATION OF THE REACTIVITIES OF LIMESTONE TO REMOVE SUL
FUR DIOXIDE FROM FLUE GAS
J. D. Hatfield, et al
Tennessee Valley Authority
Muscle Shoals, Alabama
30 June 1970
J
DISTRIBUTED BY:
P\r	^
LKJULfe}
National Technical Information Service
U. S. DEPARTMENT OF COMMERCE
5285 Port Royal Road, Springfield Va. 22151
This documsat has been approved for public release and sale.

-------
fA-Jzaa. +Q7
HAP MIC DATA
SHEET
L Report No.
APTD-0699
4. Title aad Subtitle
Investigation of the Reactivities of Limestone to
Remove Sulfur Dioxide From Flue Gas
3- Recipient's Accessio« No.
5- Rep or" *'t:c
June 50, 1970
7. A-horfs) T~. IT Hatfield, T. r Ki¦,
G. H. McCle1lan
R. C~! Mu 1 lins , and
8- Performing Organization Rept.
No.
9. Performing Organization Name and Address
Division of Chemical Development
Tennessee Valley Authority
Muscle Shoals, Alabama 35660
10. Project/Task/Worlc Unit No.
11. Coatract /Gram No.
TV-29232A and
TV-30530A
12. Sponsoring (
Air Pollution Control Office
Environmental Protection Agency
Post Office Box 120S5
Research Triangle Park, North Carolina 27709
1J. Type of Report a Period
Covered
l Supplementary Nme, DISCLAIMER - Th i s report was furoishid to the Office of
Air Programs by Tenneasee_Valley Authority, Division of Chemical
IS.
«r rrograms Dy lennessee valley Authority, Division of Chemi
iDrn!t?!«;.;us?'g;5tt;ieg'sgre»v?yws^ fcric
•¦a 35660,
14- Abstracts
An Investigation haa bean conducted, concerning the propartlea of a line
•tone that ara Important In detaralnlng lta effectiveness when Injected
Into a conbuatlon chamber to decrease the ealaelon of SO2 In the stack
gaa fros burning sulfur-containing fuel In a povar plant. To obtain
thla Information, reactivities of thirty-five atonea were seasured and
correlated with their alneraloglcal and cryatallographlc propertlea.
¦atea and capacltlea of sulfation were taken aa aeaaures of reactivity.
The report la divided Into three parta: reaction aechanlaae and
klnetlca, properties of calclnea and aulfated llaeatonea, and
•valuation of the reactivities of selected llaeatonea toward SO,.
17. Key Ttxds and Document Analysis. 17a. Descriptors
Desulfurizatlon
Flue gases
Limestone
Chemical reactivity
Reaction kinetics
Sulfation
Dlsproportionation
Calcines
17b. Identifiers 'Open-Ended Ten
Calcination
17c. COSATI Field/Group 13B, 7D
18. Availabiliry Statement
Un licni ted
19. Security Class (This
Report)
Page
UNCLASSIFIED
21. No. of Pages
136
22. Ptii
FORM NTIS-S5 (IO-70)
USCOMM-DC 40320-P7 1

-------
Tennessee Valley Authority
Division of Chemical Development
Fundamental Research Branch
INVESTIGATION OP THE REACTIVITIES OF LIMESTONE TO
REMDVE SULFUR DIOXIDE FROM FLUE GAS
By
J. D. Hatfield, Y. K. Kim, R. C. Mullins, and 0. H. McClellan
Muscle Shoals, Alabama 33660
INVESTIGATION OF THE REACTIVITIES OF LIMESTONE TO
REMOVE SULFUR DIOXIDE FROM FLUE GAS
SUMMARY
To provide a guide for the selection of limestones for dry
injection to absorb sulfur oxides from power plant combustion systems,
the reactivities of thirty-five stones were measured and correlated
with their mineralogical and cystfil 1 ographic properties.
Rates and capacities of sulfation were taken as measures of
reactivity. Both Isothermal and polythennal measurements were made In
a standard laboratory TGA apparatus that had an upper temperatur limit
of 940* C. The mlnerological and crystallographic properties were
determined by petrographlc and electron microscopy and by x-ray methods.
Calcination of the limestone proceeds from the outside surface
toward the center of each particle. The rate of calcination is affected
by the reaction temperature, the particle size of the stone, the
crystallite size of the carbonate mineral in the stone, the relative
amounts of magnesium and calcium In the stone, and the partial pressures
of CO^ and S02 In the furnace atmosphere. The reaction of CaO with S02
and 02 at 915° C. is first order with respect to the partial pressure
of SOp in the range 1 to 8.U$ by volume, zero order with respect to the
partial pressure of 02 in the range 1 to 10$ by volume, and zero order
with respect to the loading of product until the layer of CaS04 coopletely
1

-------
shields the available reaction surface. Below &»0° C., the formation
of sulfite predominates over that of sulfate, and the oxidation of
sulfite to sulfate begins at about 720° C.; the sulfite also begins to
decompose and to disproportionate in this temperature range. Disproportion-
ation of sulfite at 880° C. Is prevented by the presence of oxygen or C02,
either of which promotes the formation of sulfate. The first-order
specific rate constants for the decomposition of CaSCXj and CaS04 and the
loss of reactivity (dead-burning) of a high calcitic (Colbert) limestone
were expressed in Arrhenius form.
Hater vapor affects the rate of reaction of S02 with MgO but
not with CaO; the reaction with MgO is too slow to be important wder
injection conditions. The direct substitution of S02 for C02 in the
limestone also is too slow to be significant under injection conditions.
The sulfation reaction occurs simultaneously with the calcination reaction
when the rate of the calcination reaction is low-, sulfation is most
rapid, however, when there is a relatively large surface of unreacted
CaO. The products of the sulfation reaction retard the calcination
reaction when the two reactions occur simultaneously.
The physical properties of a calcine and its reactivity with
S02 are influenced markedly by the texture of the parent limestone from
which the calcine is formed. An Iceland spar, of mean crystallite size
Q
7125 A., calcined slowly below 950 C. because of its high thermal
stability, and yielded a calcine with small pores and small crystallites
of CaO. A very pure calcitic limestone, of mean crystallite size 3875 A.,
ii
yielded larger crystallites of CaO than did the Iceland spar under the
conditions, but the pore volume was larger and more favorably distributed among
different sized pores for reaction with S02. Between 1000° and 1100°C. the
calcines of both materials shoved a minimum in pore volume, which was
Interpreted as the end of sintering and the beginning of recrystalli-
tation of the CaO. During sintering, small pores coalesce into larger
pores with small changes in the crystallite size of the oxide; during
recry stall! zation the crystallites grow into massive single crystals
of CaO at rates proportional to the crystallite size of the original
carbonate. The Iceland spar was less reactive with S0g than was the
calcitic limestone under all the experimental conditions tested
because of its slower calcination, its unfavorable pore-size distribu-
tion and pore volume, and its more rapid recrystallization. Electron
microscope photographs of cleaved sulfated particles of thcae stones showed
clearly the reaction zone between oxide and sulfate and the growth of the
crystallites of sulfate at the outer surface.
A dolomite of mean crystallite size 2075 A* was much superior
to a calcite of mean crystallite size 3875 A. in the absorption of SO^
when calcined for 1 or k minutes at 1100* or 1225® C. The properties of
the calcines that influence the absorption of SO3 were affected by the
tine and temperature of calcination, by the particle size, and by the
type of stone. Several interactions of factors affecting the various
properties of the calcines were identified, including some three-factor
interactions. It is shown that the temperature of calcination of the
dolomite and the time of calcination of the calcite were the more critical
factors that affect the SO3 absorption capacity of the two stones.
Hi

-------
Studies were made of the calcination, isothermal sulfation, and
polythermal calcination-sulfation of 55 limestones, and the distinguishing
characteristics of each reaction were correlated with the chemical aqd
mlneralogical properties of the stones. Six measures of the capacity
and rate of reaction with SO2 vere correlated with the chemical and
mlneralogical properties of the stones, as well as vlth the reaction
parameters. Multiple correlations of the capacity and rate parameters
vlth the properties of the stones yielded equations from which the
reactivity of a stone could be predicted fran its chemical and mlneralogical
characteristics. The equations vere tested vlth five stones considered
for use In the Shavnee tests, using the reaction rate data at, five
temperatures obtained by Battelle in the dlspersod-phase reactor. The
prediction of the polythermal capacity for absorption of SOg from the
chemical and mlneralogical properties of the stones correlated veil vith
the Battelle reaction rates at temperatures below 9^0° C. (172U* F.),
but not at higher teaperatures. No other predicted reactivity measure
correlated significantly vlth the Battelle rate at any temperature.
The equations that express the polythermal capacity for absorp-
tion of sulfur oxides, v3, ng. 8O3/5O mg. limestone, are
v3 = -I.57IZ4 + 5-685*5 - 32-l^z© + 5*60327 - O.528OZ9 + 2.8l6zi0
log v3 = 0.05llUzs - O.O668824 4 0.29U2zs + O.I85UZ7 - O.OlUO^zo + 0.12002!o
where z2 = (< rhanbic carbonates)/(#CaO)
z4 = mlcrcscopic grain size (coded)
£5 s unit-cell length, a, of the carbonate mineral, A.
o x-ray crystallite size of the carbonate mineral, microns
iv
Zt a log (NasO content of limestone, «t»)
zq s sum of contents of lllite, montmorillonlte, tremolite,
llmonite, feldspar, and muscovlte,
Z10 = llJoonite)/('i total FegOs)
The best estimate of the value of v3 is the arithmetic average of the
values calculated from these two equations.
Although the data obtained at the temperatures accessible
in this experimental study do not extrapolate veil to the higher
temperatures that are encountered in actual power-plant combustion
chambers, the properties of stones that vere shovn to be desirable at
low temperatures probably vlll be advantageous at the furnace temperatures
also. Perhaps the major difference to be expected betveen the results
obtained at low temperatures and those anticipated at high temperatures
le the marked increase In the rate and extent of dead-burning at the
high temperatures, as Indicated by the results of the x-ray examinations.
No vay of avoiding the detrimental effect of rapid dead-burning at
temperatures above about 1050° C. is Indicated by our data. Our data
Indicate that stones that contain the smallest amounts of Impurity
minerals and vhose rhombic carbonates are primarily calcitic and are
present as the smallest crystallites will be the most effective for
injection into the combustion chamber. Increasing the sodium content of
the stone Increases its reactivity, whereas iron, if present, should be
in the form of limonite or goethite rather than in the lattice of the stone.

-------
INVESTIGATION OF THE REACTIVITIES OF LIMESTONE TO
REMOVE SULFUR DIOXIDE FROM FLUB GAS
COHTHSTS
Page
Introduction		1
Reaction Mechanism and Kinetics 		2
Calcination			2
Sulfation 		6
Effects of and Composition of Gas		6
Effect of Temperature 		8
Oxidation		10
Dlsproportionation 		10
Decomposition of Sulfation Products 		11
Effect of Water Vapor on the Sulfation of CaO
and MgO		11
Dead-Burning		1}
Substitution 		1U
Properties of Calcines and Their Sulfated Products 		19
Effect of Texture of Calcites 		19
Electron Microscope Studies of Reaction Surfaces ....	Uj
Properties of Calcite "and Dolomite Calcines		!»8
Absorption of SO^		51
Other Properties of the Calcines		60
Page
Comparison of Llnestones		6b
Introduction 		&i
Selection of Limestones		66
Calcination		71
Sulfation		73
Isothermal		T)
Polythermal Calcinatlon-Sulfation 		75
Correlations 		j€
Multiple Correlations 		95
Application to Stones to be Used in Shawnee Tests ....	99
References	105
APPENDIX A: Mechanism of Calcination Reaction 		107
APP1HDIX B: Calculation of Kinetic Parameters for
Calcination	109
APPENDIX C: Kinetic Interpretation of Differences Among
Stones	Ill
APPENDIX D: Determination by Regression Analysis of the
Significant Factors That Affect the S02
Absorption Properties of Limestone 		115
APPENDIX E: Multivariate Analysis--A Look at Principal-
Components Prediction	118
APPENDIX F: Relative Importance of Properties of Limestones
For Reaction With Sulfur Oxides 		122
APPENDIX C: The Textural Evolution of Limestone Calcines . .

-------
TABLES
Page
I. Sorption of SOa by Lime	 11
II. Statistical Summary of Pore-Size Distributions of
Calcines	 22
III. Bulk Densities of Calcines	 28
IV. Properties of Sulfation Products	 33
V. Properties of Calcines Under Different Calcination
Conditions	 50
VI. Analysis of Effects of Fbctors and Their Interactions
on the Properties of Calcines	 52
vn. Analysis of Variance for Partial Duplicates of Eight
Absorption Samples Using the Average of Three
Methods of Analyses 	 55
VIII. Chemical and Mineralogieal Properties of Limestones . . 67
IX. Accessory Mineral Content of Limestones	 69
X. Calcination and Isothermal Sulfation of Limestones ... 71*
XI. Polytherm&l Calcination and Sulfation of Limestones . . 77
XII. Capacity of Limestones for SOs and Rate of Sulfation . . 60
XIII. Intercorrelation of Chemical end Mineralogieal
Properties	
XIV. Intercorrelation of Calcination and Sulfation Parameters 8U
XV. Correlation of Calcination and Sulfation Parameters With
Chemical and Mineralogieal Properties of Limestones . . 87
Page
XVI. Correlation of Capacities and Rates of Reaction With
Chemical and Mineralogieal Properties of Limestones . . 88
XVII. Correlation of Capacities and Rates of Reaction Vlth
Calcination and Sulfation Parameters of Limestones ... 88
XVIII. Correlation of Initial Reaction Rate With Chemical
Properties and Texture of Limestones	 91
XIX. Correlation of Combination Reaction Rate With Chemical
Properties and Textures of Limestones 	 9U
XX. Coefficients of Significant Factors in Multiple
Correlation Equations 	 98
XXI. Predicted Reactivity Parameters of Limestones to he
Tested at Shawnee	 100
XXII. Correlation With TVA Results of Reaction Parameters
Calculated From Battelle Data 	 102
XXIII. Changes in Limestone Properties Required to Produce
Measured Changes In Reactivity 	 123
FIGURES
1.	Kinetic parameters for Calcination, k = k Concentration on Reaction Velocity .... k
k. Effect of Temperature on the Absorption of S02 by Calcines
of Minus Zh- Plus 28-Mesh Colbert Limestone From Gas
Containing ^.5$ SOa	 9
5- Thermal Decomposition of Calcium Sulfite and Calcium
Sulfate 	 9

-------
Page
6. Effect of Hater Vapor on the Sulfation of MgO and
Limestone Calcine	 12
7- Effect of Calcination Temperature on the Specific Rate
of Loss of Reactivity With S02	 12
8.	Calcination and Sulfation at 090° C. Without C02 .... 17
9.	Calcination and Sulfation at 9?0° C. Kith 9.54 C02 ... 17
10.	Calcination and Sulfation at 8550 C. With 16$ C0a ... 17
11.	Mean Pore Size of 16-Hour Calcines	 23
12.	Frequency Distribution of Pores in 16-Hour Limestone
Calcines	 25
15- Frequency Distribution of Pores in 16-Hour Iceland
Spar Calcines 	 25
lU. Pore Volumes of 16-Hour Limestone Calcines	 26
15.	Pore Volumes of 16-Hour Iceland Spar Calcines	 26
16.	Mean Crystallite Sizes of 16-Hour Calcines 	 29
17- Crystallite Sizes of Calcines Heated Continuously at
6° C. per Minute	 29
18.	Surface Fractures on the Cleaved Surface of a Calcined
Single Crystal of Calcito. 12.000X	
19.	Reaction Zone of Calcined Single Crystal of Calcite
After Exposure to Sulfur Dioxide. 12,OOOX	• . U6
90. Isopleths of SO3 Absorption as Functions of the Time
and Temperature of Calcination 	 58
21. Calcination and Sulfation of a Calcite, BCR 1679 .... 72
Page
22. Calcination and Sulfation of a Dolcalte, BCR I69O ...	72
2}. Comparison of Calculated Reaction Parameters With
Battelle'8 Apparent Reaction Rates 		103
2t». Relation Between Reduced Time and Degree of Calcination
of Minus 2h- Plus 28-Mesh Colbert Limestone		103
25.	Effects of Amount of Injected Limestone and Kinetic
Parameters on the Desulfurization of Stack Gas		11U
26.	Number of Roots Required to Explain Variations Among
13 Measured Properties of Limestones		120

-------
INVESTIGATION OF THE REACTIVITIES OF LIMESTONE TO
REMOVE SJLFUR DIOXIDE FROM FLUE GAS
by
J. D. Hatfield, Y. K. Kim, R. C. Mullins, and G. H. McClellan
Division of Chemical Development
Tennessee Valley Authority
Muscle Shoals, Alabama 35660
INTRODUCTION
This investigation, conducted under AFCO (formerly HAPCA)
Contracts TV-29232A and TV-}05J0f, is concerned with the properties of
a limestone that are Important In determining its effectiveness when
injected Into a combustion chamber to decrease the emission of S0s in the
stack gas from burning sulfur-containing fuel In a power plant. The
report is divided Into three parts: reaction mechanisms and kinetics,
properties of calcines and sulfated limestones, and evaluation of the
reactivities of selected limestones toward SO£. This report covers vork
done during the period July 1, 1968, through June 30, 1970.
2
REACTION MECHANICS AND KINETICS
In a study of the properties of limestones that determine
their reactivities with sulfur oxides in flue gas, it was necessary to
study and define the reactions that occur. The following is a summary of
the reactions that occur when limestone is injected into a power plant
combustion chamber to decrease the emission of sulfur dioxide in the
stack gases.
Calcination
The reactions
CaCC>3 -• CaO + CO;.	(1)
MgCOj -• MgO 4 C02	(2)
occur when limestone is heated to a sufficiently high temperature.
FOr most limestones, decomposition begins in the temperature range
600° to 700° C. (1100® to 1J00° F.) and accelerates as the temperature
is raised. The rate of calcination is affected by the temperature, by
the particle size of the limestone, by the microscopic or x-ray crystallite
size of the carbonate fraction, by the proportion of MgCOa and CaCOs
In the limestone, and by the partial pressure of C02 in the calcining
atmosphere. Sulfur dioxide, if present in the gas, also affects the
rate of calcination.
Calcination is a phase-boundary reaction and proceeds from the
outside surface into the center. The order of the decomposition reaction
of a quite pure calclte is 2/3 vhen the particles are relatively large
(minus 24- plus 28-mesh), as shown in Appendix A. The order of the

-------
3
decomposition reaction of oplcl"™ carbonate has been reported to be between
zero and one ihr different preparations asd under different experimental
conditions; for this study ve have used the value of 2/3 because of the
demonstrated dependence of the rate of reaction upon the surface of the
limestone in the particle size range used.
The details of calculating the kinetic parameters of calcination fran
polythertnal the mo gravimetric analysis (TGA) runs for the different
limestones are given In Appendix B. The activation energy for calcination,
E, of the stones ranged from JO to 75 kcal./mole when their minus
2k- plus 28-mesh fractions were calcined in a nitrogen atmosphere. The
frequency factor, ko, (sometimes called the pre-exponentlal term In the
Arrhenlus expression) was related to the activation energy for all
calcites and dolomites by the expression
log ko = -1.233 + 0.2HAE	(3)
The consistency of the data with this expression is shown in Figure 1.
The use of equation 3 to obtain a one-parameter characteristic
of a given limestone may be sanevhat misleading In the ordinary use of
activation energy for reaction. Normally, a high cultivation energy
Indicates that a higher temperature is required to produce a given
reaction rate than is needed for a lover activation energy. The change
in the frequency factor, ko, with the activation energy, however, over-
comes this effect, and the stones with the higher activation energies
decompose most readily. Van Heek and Jifntgen (lA) shoved the inter-
relationship between ko, E, the order of the reaction, and the heating
rate in affecting the decomposition of carbonates. Equation 3» therefore,
o-oujcrrt
&-OOLOMTTC
A-Cju.cnr-00L£MTc
m-MMBmsnz-oolcmic
8 io
Figure 1. Kinetic Parameters for Calcination, k = ko exp(-E/RT).
PlmiTf P. Effort nf	O-. on
Reaction Velocity at 915° C. with
Minus 2U- Plus 28-Mesh Colbert
Llmestvne.
*	^ -* * AJAAVWW WJ. WVg
Concentration on Reaction
Velocity.

-------
5
quantifies this relationship for limestones when the order is 2/3 end
the heating rate is 5°C. per minute, the conditions used in this
study.
Dolomite .ideally contains CaC03 and MgCO^ equimolar amounts,
with the two compounds in alternate layers in the structure. Magnesium
carbonate decomposes at a lover temperature than does calcium carbonate,
and it is possible to calcine the MgCOs portion of dolomite and leave the
CaCOs unchanged. In most dolomites, however, calcium and other cations,
such as Fe and Kn, ore substituted for some of the magnesium in the
magneslte layers, and there may also be substitution of magnesium for
calcium in the calclte layers. These isomorphous substitutions affect the
dimensions of the unit cells of both calclte and dolomite and also affect
the grain size of the stones. The substitution of Mg for Ca in either dolomite
or calclte Is limited to about 5$, whereas that of Ca fbr Mg in dolomite
can be as high at 20* (20). It is to be expected that a higher temperature
will be required to complete the calcination of these substituted stones,
and that both the maximum rate of decomposition and the temperature at
which the maximum rate is obtained will be related to the relative amounts
of Mg and Ca. The grain texture and the crystallite size within the grains
of the stone also affect the rate of calcination and the temperature of its
itiatImum rate, however, and the calcination parameters are complicated functions
of the mincralogical and chemical properties of the stone.
6
Sulfation
The sulfation studies were mAde by TGA with minus 2h- plus
28-mesh (Tyler) Colbert (high-calclunO limestone that had been calcined at 900°C.
for 30 minutes. The calcined material was treated with gases containing
different amounts of S02 and 02 to determine the order of the reaction.
The reactions are
CaO + S02 -~ CaSQs	(If)
CaS0(3 + 0.502 -~ CaS04	(5)
CaO + S02 + 0«502 -» CaS04	(6)
in which equation 6 is the sum of equations U and 5.
Effects of Loading and Composition of Gas: The kinetic
equation fron\ which is determined the order of reaction with respect
to the partial-pressure of S02, the partial pressure of 03, and the
loading or fraction sulfated is
if = KPsiPom*-n	(7)
where	wc = initial weight of calcine
w - weight of calcine -~ absorbed sulfur compounds at time t
K «s specific rate constant
Pg = partial pressure of S02
P0 - partial pressure of 02
i = order of reaction with respect to S02
m = order of reaction with respect to 02
n = order of reaction with respect to loading

-------
7
In each run Pg and P0 were kept constant so that the Integrated form vas
log v = -i- log + —i— log t	(8)
D+i	n+i
where Vj Is the weight gain at t = 1 minute.
Results for several crmhlnations of S02 and 0a contents In
reactions with the calcine at 915° C. are shown In Figure 2; the results
are plotted according to equation 8. In all the tests the slope
d log w/dt was unity until a sufficiently thick layer was built up on the
available reaction surface. This demonstrates that n = 0, or l/(n+i) = 1,
for the first reaction stage, which is of primary (and perhaps only)
concern when limestone is injected into a combustion chamber.
In .the absence of oxygen, test 8.UA of Figure 2, the initial
rate, as shown by the slope, was the same as that with oxygen (test 8.!*),
but the ratio of weight gains at t = 1 of runs 8.U and 8.UA was the same
as that of the molecular weights of SO3 and SCfe. Hence, test' 8. It A was
proceeding according to reaction h, but test 8.U was proceeding according
to reaction 6. It Is postulated that reaction ') occurs first and is the
rate-limiting step; reaction 5 occurs at 915° C. at a relatively high
rate when oxygen is present.
The order of the reaction with respect to the S02 content of the
gas was determined by plotting the values of Wj against the concentration
of S02 in the gas, as shown in Figure 5 - The linear relation Indicates
that I = 1 in the range of S02 concentration from 1 to 8.U$; Borgwardt (U)
shoved a similar relation in the range O.OO58 to 0.6$ S02.
8
The value of m In equation 7 was determined by varying the 02
content of the gas from 1 to 1Of> with the S02 content constant at 2$.
The results of each test fell on the curve for 2$ SOa in Figure 2,
which indicates that the concentration of oxygen in this range has no
effect, so that m = 0 in the range 1 to 10$ 02.
Effect of Temperature: A series of tests was made with gas
containing U. 5$ 02, ^-5$ S02, and 91$ to determine the effect of
temperature on the sulfation reactions; the results are shown in Figure
U. The data show a large increase in the extent of reaction as the
temperature was raised from 392° to 655°C. and as the time was increased
from about 2 to 10 minutes, a relatively small increase in the extent of
the reaction between 635° a™1 717°C°., and an increase in weight gain
between 717° and 800°C. These phenomena are consistent with the assump-
tion that the sulfite reaction predominates at low temperature and that
there is a competition between the decomposition of sulfite and its oxida-
tion as the temperature is raised. Ho detectable change in rate or
extent of reaction was noted when the temperature was raised above 800°Cv,
but this was due to equipment limitations.
The data in Figure U were evaluated kinetically by determining
the activation energy for the overall reaction for the first 25$ of
reaction, and the instantaneous rate of reaction at completion. This
degree of reaction was selected because initial slopes could not be
measured precisely, whereas the times and slopes at 25 ~ completion could
be determined with considerable precision. The activation energies
were I5.6 kcal/mole fbr the overall first quarter of reaction and If.8
Xcal./mole for the Instantaneous reaction at 25^ completion. It is

-------
9
Figure fr. Effect of Temperature on the Absorption
of SOa by Calcines of Minus 2U- Plus 28-Meah Colbert
Limestone From Oas Containing U.5# 80s.
Figure S- Thermal Decomposition of
Calcium Sulfite and Calcium Sulfate.
10
probable that the activation energy for the initial reaction between
S02 and CaO for this particular limestone will be lover than those
determined at 25$ completion.
Oxidation: An effort was made to study the oxidation of
calcium sulfite, reaction 5, by thermogravimetric analysis, but the
weight gain at the low temperatures used was not large enough to
indicate any significant reaction; at high temperature the calcium
sulfite decomposed and complicated evaluation of the kinetics.
Dlsporportlonatlon
The disproportlonation of calcium sulfite
UCaS03 -» 3CaS"V + CaS	(9)
was studied at 880° C. (1615° F.) and is svmmarized in Table I. Lime
was treated with S02 alone and in mixtures with 02, COa, or both, and the
distribution of the sorbed sulfur was determined by chemical analysis.
In the presence of oxygen, only calcium sulfate was formed; when no
oxygen was present, considerable amounts of sulfite and sulfide appeared
In the product. In the presence of carbon dioxide but no oxygen, less
sulfide was formed than when both carbon dioxide and oxygen were absent,
Indicating that carbon dioxide oxidizes both sulfite and sulfide to sulfate-
The last two tests of Table I show that sulfide, If formed by disproportlona-
tion, does not persist in the presence of oxygen even at low temperatures. The
Increase in sulfite content as the temperature was raised from ;80° to
635° C. indicates that the reaction of lime with S02 increases faster
in this temperature range than does the oxidation reaction.

-------
11
TABLE I
Sorption of SOa fry T.tmo
Ten®.,
°C.
Cumuli.
of
gas. f,
Distribution
0
CO
U
SOp
Oa
CO?
so«--
SO,"
S"
880
0.5
3
15
>99
<0.05
<0.05

0.5
0
0
71
2
27

0.5
0
15
93
1
6

0.5
3
0
>99
<0.05
<0.05
635
0.5
3
0
58
US
<0.01
579
0.5
3
0
75
25
<0.01
Decomposition of Sulfation Products: The decomposition of
CaSQs and CaS04
CsSQa -• CaO + S02	(10)
CaS04 -» CaO + S02 + 0.502	(11)
vas studied by analysis of samples that had been heated at constant
temperature for different times. The results are shown In Figure 5,
and the rate constants for the first-order decomposition reactions are
expressed In Arrhenius form as
CaSQj: k = 1-3 x 10s exp (-6T,500/RT) sec"1	(12)
CaS04: k = 6.1 x 10s exp (-69,500/RT) sec"1	(13)
Effect of Water Vapor on the Sulfation of CaO and MgO: A
sample of Colbert limestone that had been calcined at 900° C. vas compared
with reagent powdered Mg(0H)2 that had been calcined at 610° C. to form
MgO. Each oxide was treated at 900° C. with a gas containing S02,
02, and either zero or 10$ water vapor. As shown in Figure 6, water
vapor did not affect the first stage of the reaction of lime with S02,
but the second stage was promoted slightly. The MgO did not react with
S02 in the absence of water, and its rate of reaction in the presence of
12
Mfiure_6. Effect of Water Vapor on the
Sulfation of MgO and Limestone Calcine.
Ttwourusf. -c
iioo	itoo	1100	1000
wis	gao
1000/T
riflure J. Effect of Calcination Temperature on
the Specific Rate of Loss of Reactivity with S02.

-------
13
water was so low that It would be expected to be of no value under Injection
conditions; the product of the reaction was MgSOt.
Dead-Burning
The loss of reactivity with S02 of lime at temperatures above the
optimum has been ascribed to growth of the CaO crystallites and to the
sintering that occurs. Calcination at these high temperatures results in
a decrease in the porosity and an increase In the compactness of the calcine,
in combination of smnll crystals into larger ones, and In formation of a film
of certain impurities on the surface of the calcine. The reaction may be
indicated as
CaO -» (CaO)	(1U)
where the parentheses Indicate a less reactive species from whatever cause.
Calcines were prepared at different temperatures in the range
1117° to 1317° C. for different times in the range 1 to 30 minutes, anil
measurements were made of the capacity of the calcines to absorb 80a
at 920° C. frctn an atmosphere containing 1$ 80s, hf 02, and 924 H2. The
limestone was a very pure coarse-grained calclte, 99$ CaCOa, designated
samplo 89C, that was selected from a group of agricultural stones obtained
ftrcm the Department of Agriculture; the clear grains were handplcked to
obtain a calclte as free as possible from impurities already present In
mufti! amounts. The results of the sulfations were treated as those of a
first-order reaction—the rate of loss of reactivity with SOs vas assumed
to be proportional to the remaining reactivity. The reactivity and
capacity of a soft calcine at 900° C. in an atmosphere containing
15^ C02 and 8536 Hj, were arbitrarily taken as the marimim reactivity
1U
and total capacity of the stone. The degree of dead-burning
or loss of reactivity, «, was defined as
, . OS - GT,t
Gg
where Gg = S02 capacity of soft calcined (900° C.) stone
» S02 capacity of stone calcined at T° C. for t minutes
At each constant temperature, the plot of log (1 - c) against time was
straight line, as the Integrated first-order model predicts, as in
the expression
In (1 - c) » K^t	(16)
where kj[j is the specific rate confltant for dead-burning of this particular
stone.
An Arrhenius plot of the logarithm of the dead-burning constant
against the reciprocal of the absolute temperature Is shown In Figure
7, from Which Is obtained, for this particular limestone,
kjD = 3-3 x 108 exp (-71,000/RT) sec*1	(17)
Substitution
The reaction
CaCOs + S02 -• CaSOg + C02	(18)
In which the C02 of calclte Is replaced by S02 is theroodynamlcally
favorable at temperatures below those of the decomposition of calclte,
but TGA Is not sufficiently sensitive to detect any weight change until
decomposition is pronounced. Tan Heek (15) found a significant substitution
reaction before calcination began and found that the extent of the substitution

-------
15
reaction was influenced greatly by the supply of SO^ and varied ancmg
1 limestones and povdery precipitated carbonates. Coutant (5) also
demonstrated an Interaction of sulfation with calcination and a correla-
tion between the degree of calcination of the atones and the fraction of
sulfite sulfur in the product after short times of exposure. When the
temperature is high enough, however, calcination of the limestone proceeds
rapidly in the presence of S02. A given loss in weight is obtained at
a higher temperature in the presence of S02 than in an inert atmosphere.
There is considerable reaction with S02 during calcination when the
heating rate is 5° C. per minute, but the proportions formed by
substitution, equation 18, and by reaction with the calcine, equation h,
are unknown. It is probable that the bulk of the product, CaS03 and
eventually CaS04, originates from reaction with the calcine, both
simultaneously with and subsequent to the calcination reaction.
Ib test the relative degrees of sulfation and loss of C02
from the limestones, several series of runs were made with minus 2k-
plus 28-mesh Colbert limestone at two temperatures, 855° and 090° C.
The CO 2 content of the input gas ranged from 0 to l6jt, but the Kfe
and 02 contents were kept constant at 5$ find J4, respectively. The
limestone sample, 166 mg., was distributed evenly in a 80-mg. plug
of quartz wool that was placed in a Gooch crucible that had been cut off
to form a container 20 mm. in diameter and 8 mm. high. A perforated
cover was placed on the container to permit ingress and egress of gases.
The assemblv was placed in the TGA apparatus in a C02 atmosphere, and the
16
furnace, preheated to the desired temperature, was elevated into position.
When thermal equilibrium was established (there was no loss or gain of
weight in the COg atmosphere), the reaction tube was evacuated abruptly
and the flow of simulated stack gas was started. At the >nrt of a pre-
determined time of exposure, the sample container was removed from the
heating zone and swept with an atmosphere of C02 to quench the calcination
and sulfation reactions; this treatment, however, led to seme unavoidable
recarbonatlon of the unsulfated lime. The sample was analysed for
sulfur and C02, and the degrees of sulfation and calcination were
calculated from the TGA weight change.
A similar procedure was used to determine the change in weight
on calcination in the absence of S02; the S02 in the simulated stack gas
vas replaced by H2. When calcination was complete the composition of the
stack gas was changed to the proper proportions of C02, S02, 02, and N2,
and measurements were made of the sulfation reaction.
In the absence of C02 at 890° C., Figure 8, the rate of
calcination in the presence of S02 vas almost the same as that in its
absence, but the product of the sulfation reaction prevented complete
calcination. Some sulfation occurred during calcination and continued
after calcination ceased. Precalcination had only a slight effect on
the rate of sulfation because calcination was much more rapid than
sulfation.
The addition of 9-5$ C02 to the simulated stack gas retarded
both calcination and sulfation reactions, Figure 9- Calcination was
only 85^ complete in the presence of S02, and the sulfation reaction
was affected also. At 855° C. In an atmosphere of 16$ C02, Figure 10,

-------
17
Figure 8* Calcination and Sulfation
at 09O°C. Without C02.
Figure 9- Calcination and Sulfation
at890°C. With 9.5 C0a.
• » 10 12
TIME, WtNUTCS
Figure 10. Calcination and Sulfation at
055°C. With 16$ C02.
18
the rate of calcination was considerably lover in the presence of SO^
than in its absence, and the calcination leveled off at about 50$
completion in the presence of S02. The sulfation reaction vas also
much slower and only 50$ of the calcine vas sulfated. The net weight
change In this test vas very small.
While these results are not conclusive as to the relative
importance of the substitution reaction, it Is obvious that both the
rate and the amount of sulfation depend heavily upon the availability
of calcine to react vith the SO^• It is possible that in polythexmal
heating SO2 replaces a small amount of C02 before calcination begins,
but it is evident that when conditions are such that calcination Is slow
in the absence of SO2, sulfation also is slow in the presence of SOg.
Under otherwise identical conditions, Increasing the amount of calcine
increases the sulfation reaction. The retardation of calcination by
80s may reflect the formation of a shell of product on the surface that
hinderB the escape of COg in the surface-controlled phase-boundary
reaction.

-------
19
PROPEHnES 07 CALdHES AHD THEIR SULFATH) PRODUCTS
In this section the properties of calcines and their sulfated
products are compared on the basis of texture and type of stone.
Results of x-ray, electron microscope, and porosity studies of two
caldtes of different texture are described that show the differences
in the physical properties of their calcines and their reactivities vlth
SOj,. Results of electron microscope studies of reaction surfaces of
single crystals are presented to demonstrate the unreacted core mechanism
and to describe the reaction front. The effects of the time and tempera-
ture of calcination and the particle size of the initial stone on the
properties of the calcines of a pure calclte and a rather pure dolomite
are described also.
Effect of Texture of Calcltes
In the study of the effects of the crystallite size of the raw
material on its calcination and sulfation, the calcltes studied were
a high-purity calcltic limestone (USDA Reference Sample 89) and clear crystals
of Iceland spar (Hard's natural Science Establishment, Inc.). Both materials
were crushed and screened to produce a minus 20- plus 2k-meah fraction
from which samples were handpicked to minimize the effect of minor
accessory minerals. The cell constants of both materials showed that
there had been no significant isomorphism, and that both materials were
pure calcltes as defined by Goldsmith and Graf (2) ¦
20
A J-gnus fpnuflr of each calclte was calcined in a platinum
boat for 16 hours at a constant temperature in the range 750° to lJOO" C.
The samples were inserted Into the preheated furnace to effect initial
shock calcination; the long retention time was sufficient far complete
caldnwt.lnn at each temperature.
Dte cwTrlneil samples were placed Immediately in a drying oven
at 110* C. to prevent hydration and recarbonation. Samples for study by
x-ray diffraction at roan temperature were ground to minus 200-mesh under
nitrogen In a dry beat. The samples for electron microscopy were
replicated Immediately upon removal from the drying oven. All the
electron micrographs were prepared by tvo-step replication that yielded
platinum-shadowed carbon replicas. Each sample was replicated in
triplicate to minimi re sampling and Observational errors.
A series of calcination studies was made an minus 200-mesh
samplea of each, material with a high-temperature x-ray diffractometer
furnace which permits the examination of hot calcines. The results
were compared *ith those obtained with the shock-calcined samples.
The dqgree of calcination of the samples was determined by
x-ray examination which can detect residual or uncalcined calcium
carbonate. The Dean crystallite size was determined by the x-ray
line-broadening technique. Only the relative sizes of crystallites
are important in these studies, and the values reported may not have
real physical significance.

-------
TABLE II
Statistical Summary of Fore»81te flletrlbutiona of Calcines

(Prepared by shock calcination and heating for 16 hours;




pore sizes determined by electron microscopy)


Calcination
Mean
Standard
Median
Mode of
Number


Temperature,
pore size,
deviation,
pore size
, pore size,
of pores


°C.
A.
A.
A.
A.
measured
Skevness
i-Valuea



Calcltlc limestone



750
635
269
600
1»00
2U6
2.19
_
850
1188
627
1100
1300
252
0.97
12.78
900
119U
588
1100
1100
253
1.05
0.11
950
1U0U
658
1}00
1300
210
1.17
3.6 3
1000
2183
U.83
2000
2100
250
1.55
8.52
1050
1998
920
1800
1500
250
0.90
1.96
1100
1958
963
1800
1500
201
0.87
o.w
1200
1655
ioai
11*00
1300
25l»
1.19
3.23
1300
IU89
837
1300
1300
25U
1.13
2.oe



Iceland
spar



750
720
3
«	*
5	i
V
6	>1
o	,0
o
CO	Tj
8	2
o	a
*	1
¦d
8
?
$
n
o
3
C
O
U
fc
*4
11
4) h
<1 S
-	3
•s	fl
5	g
B	"3
£ 3
S £
S
9
2
:i9)
the
+>
e
0


0
tH
C
CI
43
0
wi
%
a
£
£
CO
ft)
M
U
O
•H
O
3
O
i
w
•d
<3
0
0
a>
&
ft.
O
3
i
0
.4
2
a
a
* S
1 ',M
¦d 1
V	H
. US
S .
4»| 4»
*1
O
!
s
«rt
+>
8
•i
3
4
•0
£
I Ml
1
«
&
i
3 '
5
I
+>
I
*
o
s
£
S
0
N
(0
01
fc
o
p.
s
g
5
ft>
o
8
•H
CO
£
&
s
S
4»
e
s
«a
5
&
s
Q
a)
O
4>
>>
&
%
£
*1 v
i 1
¦3 I
0\
® "S

«> ja
§ a>
s I
3
•p
s
3
fc
s
I

-------
Figure 11. Mean Pore Size of 16-Hour Calcines.
sb
Inflection in the curve between 950° and 1100° C. indicates a change in
the rate of grovth of the mean pore size of the Iceland spar. The nean
pore size of the calcines of the Iceland spar Increases between 1100°
and L2O0"C., after which a decrease begins such as that observed with
the limestone.
Selected cumulative frequency curves for the pore sizes of
calcines of the limestone and Iceland spar are shown In Figures 12 and
13. The data in Figure 12 show changes in pore size of the limestone
calcines as Indicated in the statistical summary of these data. Between
750" and 1000° C. there is a general increase of pore sizes with rising
temperature, but at temperatures above 1000° C. this trend Is reversed
and the fraction of small pores increases with rise in calcination
temperature. This trend in the formation of pores was unexpected and
in marked contrast with the behavior of Iceland spar.
The development of pores in the Iceland spar calcines, Figure
13, shows a continuous increase in pore size with rising temperature of
calcination. The data are grouped around three temperature ranges and
Indicate a soft burn below 900° C., a medium burn below 1100° C., and
a hard burn above 1100° C. The observed decrease in porosity is consistent
with the observations of Hedin (15) who calcined Iceland spar at different
temperatures.
The distribution nf rmri*s in	rtf +>*£ calcines determined
by mercury porosimetry. Raising the temperature of calcination of the
limestone decreased the total pore volume and shifted the distribution
toward larger pores in all the limestone calcines except that made at
1100° C. (Figure 11+). This sample showed a higher pore volume and a

-------
25
Figure 12. Frequency Distribution of Pores in 16-Hour
Limestone Calcines.
40© 6O0
1000	tOOO	4000 60OO 11X000
PORE SIZE, A.
Figure 13. Frequency Distribution of Pores in 16-Hour
Iceland Spar Calcines.
26
PORE DIAMETER, MICRONS
Figure 1U. Pore Volumes of 16-Hour Limestone Calcines.
PORE OIAMETER. MICRONS
Figure 15. Pore Volumes of 16-Hour Iceland Spar Calcines.

-------
27
more even distribution of pores in the 17-5- to 0.175-micron region
than the 1000* C. calcine. The Iceland spar data (Figure 15) show less
than *i»i^ the total pore volume of limestone calcines prepared at the
sane temperatures, and almost all the pores in the 750* and 850° C.
calcines are smaller than 0.2 micron. The total pore volume of the
spar calcines decreases and the pore distribution shifts toward larger
values with rising calcination temperature, except for the 1200" C.
calcine. The total pore volume of this calcine was larger than that
of the 1100° C. spar calcine; this may have resulted frcm the mechanism
that caused the larger pore volume in the 1100° C. limestone calcine.
The hulk densities determined by mercury displacement are presented
ill Table III.
The crystallite sizes determined by x-ray of the shock
calcines that vere quenched in air are shown in Figure 16. The lime-
stone calcines had larger CaO crystallites than the spar calcines at
temperatures below 950° 0. Between 950° and 1050° C. both materials
show a decrease in crystallite size with a minimum at 1000° C. Further
rise in calcination temperature causes the expected Increase In crystallite
size; the crystallite size of the spar calcines reaches a maximum at 1100° C.
and then decreases, whereas that of the limestone calcine3 Increases up to
1200° C. before decreasing. Mayer and Stove (lj) studied a lime pre-
pared In a rotary kiln at 1350° C. and found that the crystallite size of
the calcium oxide ranged frcm 0.1 to 100 microns (103 to 10s A.).
TABLE III
Balk Densities of Calcines
(Prepared by shock calcination and heating for
16 hours. Bulk density determined by mercury
displacement with all pores larger than
17-5 microns filled with mercury)
Density, g./cc., of calcine
Calcination
frcm indicated
calcite
temperature,
Calcltic
Iceland
°C.
limestone
soar
750
1.50
1.66
850
1.51
1.69
950
1.96
-
1000
2.02
2.68
1100
2-32
3.06
1200
2.56
2.6U
1300

3-Q3

-------
29
Figure 16. Mean Crystallite Sizes of 16-Hour Calcines.
Figure IT. Crystallite Sizes of Calcines
Heated Continuously at 6° C. per Minute.
50
The decrease in crystallite size of the shock-calcined naterials
at temperatures above 1200° C. was considered a result of the experimental
procedure of quenching the heated samples. Samples of both materials
vere then heated.in a high-temperature diffractometer furnace over the
same-temperature range so that the phase changes and crystallite growth
could be studied without quenching the samples. Several samples of each
material were ground to minus 200-meeh and a portion of each sample was
placed on the stage, heated rapidly to 600° C.J and held there for 1
hour. The temperature programmer then was started, and the sample was
heated at 6° C. per minute to 11*00° c. with continuous recording of the
diffraction maxim from which the crystallite sizes were determined.
The results in Figure 17 show that the continuously rising temperature
caused a continuous increase in crystallite size when there vas no
quenching, from which it was concluded that either the shock calcination
or the rapid air quenching after heating for 16 hours shattered or
at least highly strained the crystals. The effects of shock calcination
are not well known, and they are difficult to reproduce for direct x-ray
study. The long period of calcination after rapid dissociation should
effectively anneal the sample. Quenching, however, probably had a major
effect on the crystallite size, and attempts were made to evaluate its
importance. A sample of each material was heated to lUOO° C. with the
temperature programmer and held there for h hours, by which time
crystallite growth had ceased. The power was turned off and each sample
was cooled to roam temperature in less than ^ minutes. The x-ray maxima
of the undisturbed sample after cooling showed the same relative decrease
in crystallite size as in Figure 16. Quenching of the sample thus is

-------
51
one cause of the inflection in the crystallite-size corves observed
at high temperature.
The rapid crystallite growth, particularly at temperatures
above 1100* C., seems to follow the exponential growth carve proposed
by Fischer (8). Because of experimental limitations, however, Fischer's
data were collected at temperatures below 875° C. and on quenched samples.
His maximum rate of crystallite growth Is about 100 A./hr., whereas
values calculated from Figure IT, with a heating rate of 6° C./mln.,
can be more than 1000 A./hr.
The reaction of the calcines with sulfur dioxide was Investi-
gated at 790°, 1050°, and 1200° C. under two different experimental
conditions: calcination for 16 hours and subsequent sulfation, and
simultaneous calcination and sulfation. The calcination-sulfation
tests were a logical extension of the calcination study and provided
data for comparison with other fixed-bed tests. The simultaneous
calcination-sulfation reactions have been studied in fixed-bed tests
and the combination of shock calcination with sulfation was thought
to be the most realistic laboratory simulation of dry limestone injection
that could be obtained without specialized ecpjlpment. The temperatures
of the tests were selected to produce soft-, medium-, and hard-burned
limes that would have different reactivities vlth sulfur dioxide.
The calcines for the consecutive tests were prepared by shock
calcination and further heating for 16 hours. The hot calcines then
were subjected for 15 minutes to a mixture of air and I# S02. The
samples calcined at IO5O0 C. were exposed to the sulfur dioxide for
32
5 and }0 minutes also to measure the rate of change of the crystallite
site of the reaction product and to prepare seaples at different stages
of sulfation for study in the electron microscope. The crystallites of
the calcium sulfate grew larger vlth rising temperature, and those on
the limestone calcines were consistently larger than those on Iceland
spar, as shown in Table IV.
The a amples simultaneously calcined and sulfated at 750° c.
for l^ minutes were Incompletely calcined; neither lime nor calcium
sulfate was detectable by x-ray In the treated Iceland spar, but both
were present in the treated limestone.
The sulfated samples were examined In the electron microscope
to gather fundamental data on the reaction and to obtain a qualitative
comparison of the results of consecutive and simultaneous treatments.
The sulfur dioxide reacts Initially with lime to form individual
crystals of CaS04 that form Interlocking aggregates. On further
sulfation these crystallites cover the entire surface of the particles,
and the growth of adjacent crystals results In a general smoothing of
the surface as the crystals unite into larger aggregates. At higher
temperatures, these larger aggregates continue to grow and form a
tight, smooth surface. Even with similar thicknesses of sulfated
layers, the permeability of the surface changes considerably as the
surface texture develops. Once the surface coating f*ormn; *n_ff"sion
processes are Influenced by the texture of the sulfate layer.

-------
TABLE IV
Properties of Sulfation Products
Sulfation conditions Crystallite
Temp.,
Time,
size of
Ratio
of areas
of
•c.
mln.
CaSO*. A.
CaCOa
CaO
CaBOj

Sulfation of limestone calcines

750
13
1725
0
1.00
2.20
1050
5
2950
0
1.00
0.35
IO50
15
2325
0
1.00
0.U2
1050
>0
2U50
0
1.00
0.68
1200
15
3075
0
1.00
0.13

Sulfation of Iceland
spar calcines

750
15
1350
0
1.00
0.t)6
1050
3
1600
0
1.00
0.05
1050
15
20*0
0
1.00
0.03
1050
30
1950
0
1.00
0.35
1200
13
2275
0
1.00
0.07

Simultaneous
calcination and
sulfation of limestone
750
15
2400
1.00
o.oe
0.07
1090
15
2550
1.00
0.(9
O.65
1200
15
3150
0
1.00
1.73

Simultaneous calcination and sulfation of
Iceland s;
par
750
13

1.00
0
0
1050
15
2825
1.00
Q.2U
0.19
1200
15
1950
0
1.00
0.33
5^
A second general consideration was the nechanism of the
sulfation reaction. Two principal models have been proposed: continuous-
reaction and unreacted-core. In the continuous-reaction model, the
sulfation reaction is thought to occur throughout the internal structure
of the lime particles, and not Just on the external surface. In the
unrcacted-core model, the sulfur dioxide reacts initially with the
external or geometric surface of the lime particles and foms a coating.
This initial phase is followed by a reaction between sulfur dioxide that
has diffused through the outer surface layer of calcium sulfate to the
lime in the unrcacted core. Examination shoved that the sulfation takes
place by advance of the sulfate front, and no diffusion reaction is
apparent along the shrinkage crackt. The unreacted-core or shell
model thus appears to be correct, at least as it Applies to these stones.
Itoe surface of a particle of calcitic li-T.estone simultaneously
calcined and sulfated at 750* C. for 15 minutes is co-upletely covered
with the calcium sulfate, and the individual crystals are discrete.
Crystals of calcium sulfate formed by consecutive calcination-sulfation
reactions arc smaller than those formed by simultaneous reaction, an
observation in agreement with the x-ray crystallite-size data on the
samples.
In contrast to the veil-sulfated surface of the limestone
calcine, the surface of Iceland spar simultaneously calcined and sulfated
at 750" C. showed only slight sulfation. X-ray examination did not
detect any lime or sulfate in the sample, although evidence of slight
reaction was seen in the electron microscope. The surface of a particle

-------
55
of calcined and subsequently sulfated Iceland spar shoved a good
sulfation reaction, although It was less than that of the limestone.
The comparatively smal1 size of the crystals of the observed sulfate
is in agreement with the x-ray crystallite-size data.
At 750° the limestone reacts better with sulfur dioxide
than does Iceland spar under either consecutive or simultaneous conditions.
The limestone behaved about the same under the two conditions, but the
Iceland spar sulfated better when it had been calcined because the
dissociation was caaplete after 16 hours at 75^° whereas there was
no detectable calcination in the 15-rainute exposure under the simultaneous
reaction conditions.
The limpstone that had been calcined and sulfated simultaneously
for 15 minutes at IO5O0 C. had surfaces on which the reaction had proceeded
to the extent that the surface had become quite smooth and was composed
of large aggregates of crystallites. The calcined material also sulfated
quite well in 15 minutes at IO5O0 C., but its textural development indicated
less sulfation of the oxide than in the simultaneous reactions. Thus,
30 minutes was required for the limestone- calcine to sulfate to the same
degree that was obtained in 15 minutes in the simultaneous treatment.
The surface of Iceland spar simultaneously calcined and sulfated
for 15 minutes at IO5O0 C. indicated a greater degree of reaction than
that at 750° C. The surface was completely covered by the sulfate, and
intergrowth of crystals had smoothed the surface. An Iceland spar
sample calcined at IO5O0 C. and then sulfated for 5 minutes had under-
gone a good surface reaction, and shoved the unusual development of a
few large single crystals of CaS04.
36
The data indicate that at 1050e C. the Iceland spar is
beginning to sulfate quite veil but still is less reactive than the
limestone. These results are in agreement with the x-ray crystallite-
size data which show that the crystallites of the calcium sulfate on the
Iceland spar calcines are smaller than those on limestone calcines.
Samples prepared at 1200® C. were generally too featureless to warrant
presentation, although the general conclusions are applicable in that
temperature region with one important exception. Thermal etching of
the sulfation product at 1200° C. indicated simultaneous formation and
decomposition of the calcium sulfate.
A comparison of the reactivities of the two calcites with
sulfur dioxide can be made from the areas of the x-ray maxima. Tte
ratios of areas for sulfated samples presented in Table IV show that
limestone calcines react with sulfur dioxide better than the Iceland
spar calcines under the conditions tested. Calcined materials shoved
a general decrease in reactivity with sulfur dioxide as the temperature
was raised. IJhen simultaneously calcined and sulfated, both materials
reacted better at the higher temperatures because of the increased
degree of calcination during the short 15-minute exposure. The Iceland
spar sample calcined at 10^0° C. shoved a relatively large amount of
unreactcd calcium oxide compared to the limestone sample prepared under
the same conditions. The Iceland spar oxide must have been inactive
because of extensive recrystallization of the lime, or because its poor
pore-size distribution caused it to be unavailable for reaction, or
because of both of these interrelated factors. The 1200r C. samples
were completely calcined in the 15-minute calcination-sulfation experiments.

-------
57
All the isothermal shod;- calcination data show that the
properties of the calcined products of the limestone-are distinctly
different frcm those of the Iceland spar, even when prepared under the
same conditions. The only significant difference in the starting
materials was the crystallite size of the uncaldned stones. The mean
crystallite size of the Iceland spar was 7125 A. and that of the
limestone 5875 A., and the higher thermal stability of the Iceland spar
undoubtedly reflects its larger crystallite size. The low-temperature
(below 950° C.) calcinations completely dissociated the spar, but only
very slowly, and produced a line with frrrml 1 pores and small crystallites.
The unfavorable pore-size distribution would require that reactions of
lime with sulfur dioxide take place as the gas diffused into the mnnl 1
pores or through the sulfate layer formed during the early stages of
sulfation.
The limestone calcines prepared at temperatures below 95^° 0.
had larger crystallites than the spar calcines, but tended to have
larger pore volumes (generally about twice as large) and a more favorable
distribution of pore sizes. Potter (21) and others have shown that
pores larger than 0.3 micron correlate well with the capacity to absorb
sulfur dioxide in fixed beds. It appears that smaller pores require
diffusion times too long to be of practical significance In actual injection tests.
During calcination at temperatures below 1000° C., the pores
grow by coalescing of small pores into larger pores, which causes the
large pores to grow at the expense of small pores as can be seen by the
porosimetry data (Figures lk and 15) and reach a maximum at about 1000° C.
38
Between 1000" and 3100° C., both materials show a decrease in crystallite
size and a minimum in total pore volume. The coincidence of these changes
in physical properties is interpreted as the end of sintering of the
oxide as described by Hcdin (1^) and the beginning of recrystalllzation
reactionsr The'small crystallites formed initially during calcination
have grown by encompassing oxide forming at sites within a limited
domain. The size of the region increases with rising temperature and
gives rise to larger crystallites. This type of growth reaches a
maximum at about 950° C. and a second type of growth begins that is
characterized initially by the development of smooth areas containing
large oxide crystals that show evidence of growth spirals and oxide
crystal morphology. As the temperature continues to rise, geometrically
outlined pores and massive oxide single crystals develop, as evidenced
by the exponential growth of crystallite size at temperatures above
1100° C. noted In continuously heated samples.
Fischer (j) observed that his limestone calcines showed a
minimum porosity at 1200° C., an Increased porosity at 1J20® C., and
then a decrease In porosity. In this study, the limestone showed a
minimum porosity at 1000° C. and the spar a minimum porosity at 1100° C.
The pores in these samples are distributed so that big pores have
become quantitatively the most significant, and the	pores are
those Interstitial or geometric pores that form as the calcium oxide
crystals grow. In the electron microscope study, the emphasis is upon
these small pores, so that the decrease in mean pore size observed in
the limestone calcined at temperatures above 1000° C. is consistent with

-------
39
the interpretation of the calcination process presented previously. A
final decrease in porosity occurs &6 Fischer (J) reported, re-emphasizing
the insignificant contributions of the wrml 1 pores to the total pore
volume in calcines that are largely recrystallized.
The Iceland spar calcines show trends different from those of
the limestone calcines, and their interpretation is less apparent. The
principal difference between the high-temperature recrystallizatlons of
Iceland spar and of limestone appears between IO5O0 and 1100" C. where
the relative crystallite size of the oxide produced from the limestone
Increases !Of>, which Is only slightly greater them the mensuration
error, whereas the crystallite size of the Iceland spar oxide Increases
more than 60$. Over the same temperature range the mean pore size of
the Iceland spar calcine increases to 1882 A. but that of the limestone
calcine decreases from 2102 A. at 1000° C. to 1958 A. at 1100° C. TheSe
data Indicate that Iceland spar calcined for 16 hours at 1100° C. is
almost completely recrystallized but that the limestone is not. This
Interpretation is supported by the physical appearance of the surfaces
of the samples as well as by a minimum in the total pore volume. The
limestone showed some straight-edged geometric pores characteristic of
samples undergoing recrystallization, but only minor growth spirals and
little development of oxide crystal morphology. In contrast to the
limestone, the Iceland spar surface showed good growth spirals and marked
development of crystal morphology, in addition to numerous geometric pores
to
The limestone crystallite size is	in the 1200° C.
calcine The limestone recrystallizes over a wider range of temperature
(100C° to 1200° C ) . i Torms smaller crystallites than does the Iceland
spf" No physica) • ence has been observed that can explain this
difference, but it must be related to the nudeatlon of the oxide during
recrystAl 1.ization and to the texture of the calcine.
The final dc -vase in crystallite size of the 16-hour calcines
can result from too rapid cooling of the samples. In an attempt to
confirm this observation, samples were cooled very slowly in the tube
fUrnace and x-ray furnace from iUOO® C. to near room temperature. All
these samples showed a very reproducible decrease in crystallite size
as indicated by Figures 16 and 17. No satisfactory explanation for this
observation has tccn proposed, but the effect is real.
The sulfation reaction of lime with sulfur dioxide has been
described in	of continuous-reaction and of unreacted-core models.
Borgvardt (J) has presented data for the continuous-reaction model that
indicates that the sulfation reaction takes place throughout the Interior
of the particle and not just on the outer surface. The data of Hatfield
and Kim (12) support the unreacted-core or shell-formation model as the
jnechanism of the sulfation reaction. The electron microscope studies
show that in a fixed-bed test the reaction proceeds by the unreacted-
core mechanism. Examination of a number of particles shows that the
shell is seldom thicker than TO microns, even after exposure for several
hours, and the crystallites of the sulfate at the outer surface are
generally larger than those at the reaction front. Calcines so treated with

-------
41
sulfur dioxide have such an impervious surface coating that they can tie
s' ored for months at room temperature with no detectable hydration or
r.-carbonatlon of 1' lime in the unreacted core.
Hie 11m." one calcines were more reactive with sulfur dioxide
than the Iceland spar calcines under all the experimental conditions.
The difference In reactivity may reflect differences in calcination
behavior, in the K»£«.*»i.-lumes and size distributions of the calcines,
end in the temperatures at which recrystallizatlon is complete. The
data in Table IV show that at 750° C. the calcined limestone was more
than k times as reactive as the'Iceland spar calcine. At 1050° C.,
the reactivity of the Iceland spar calcine vas one-tenth that at 750° C.,
Whereas that of the limestone calcine vas one-fifth that at 750° C.
This larger decrees: in reactivity of the Iceland spar calcine probably
Indicates a higher degree of recrystallizatlon. At 1200° C., both
calcines lost rfeoi-.tivity.
The difference between the two materials Is especially marked
in the simultaneously calcined and sulfated samples. After 15 minutes
at 1050° C. in an atmosphere containing S0s> the ratio of areas
CaS04:CaC03:CaO is 0.19:1.00:0.2U for the Iceland spar and 0.65:1.00:0.0)
for the limestone. The lime produced frcn the Iceland spar does not
react with sulfur dioxide as veil as the lime produced frcm the limestone
as can be seen by comparing the relative areas of CaS0« and CaO in the
three-component samples. This may reflect diffusion problems resulting
from the pore-size distribution In the samples, or frcm the recrystalliza
tion of the lime which decreases its reactivity.
h2
Finally, it should be noted that, although both stones were
screened to minus 20- plus 2lt-mesh at the beginning of the experiment,
no attempt vas made to control the particle size of the limes. The
Iceland spar shoved exceptional physical pliysical stability on heating,
and there was practically no decrepitation of the material on shock
calcining at any temperature. The limestone dusted badly during calcina-
tion at all temperatures. The surface area contributed by the decrepi-
tation of the limestone probably was a factor in its higher reactivity v:ith
sulfur dioxide and may indeed be a desirable property ofstones selected for in-
jection, although this effect may disappear whai the stone is very finely Ground.
The physical and chemical properties of a calcine are influenced
markedly by the crystallite size of the parent linestone. The high thermal
stability of Iceland spar reflects its relatively large crystallite size,
and its calcines react slowly with sulfur dioxide.
In addition to the decomposition of the calcium carbonate,
calcination effects a recrystallizatlon of the resultant calcium oxide
which is accelerated by rising temperature and which decreases the
reactivity of the calcine with sulfur dioxide.
Sulfation of lime takes place in two stages: reaction of
sulfur dioxide and oxygen with the exposed surface of lime, followed by
diffusion of the gaseous reactants through the resultant shell of calcium
sulfate to the core of unreacted lime. Pores In the calcine smaller than
about 0.1 micron do not admit sulfur dioxide rapidly enough to have
significant effect on the sulfation reaction at short retention times.

-------
hj>
Calcines react best with sulfur dioxide at temperatures below
1030* C., but simultaneous calcination and sulfation, which is required
when powdered limestone is injected into power plant fXiraaces, is
best carried out at temperatures of IO5O0 C. and above.
Electron Microscope Studies of Reaction Surfaces
Although many studies of the reaction of SOg with calcined
calcium carbonates are made with natural limestones, these materials
are not suitable for precise laboratory experiments because vide
variations in their chemical and physical properties influence their
behavior. To minimize these variations in properties and their influence
on experimental conditions, a number of tests were made on the calcination
and reaction of single crystals of Iceland spar.
Several, crystals of the spar, about 12 mesh (1600 microns)
in size, were calcined at 900° c. in the TGA apparatus until they
reached a constant weight. The calcined crystals were then exposed to
S02 for about kO minutes and examined with the electron microscope.
The crystals were cleaved to expose the unreacted oxide cores
and the reaction zone between the calcium sulfate and calcium oxide and
the cleaved surfaces were replicated by the two-step polystyrene or
faxfilm tape technique. In spite of its coarse crystallinity, the oxide
rapidly hydrated on exposure to the atmosphere, so that replicas had to
be made immediately after cleaving. The plastic replicas were platinum
shadowed and carbon coated in the evaporator. The plastic portions of
the replicas were removed by washing with a suitable solvent, and the
shadowed carbon replicas were then examined with the electron microscope.
kk
An area of calcium oxide in the unreacted core of the specimen
is shown in Figure 18. These large calcite crystals did not react
extensively with the SO^, and the unreacted oxide comprised about 8o£ of
the specimen. The surface features shown in Figure 18 are typical of
the unreacted areas. The texture is generally massive, as might be
expected along the cleavage plane of such a single crystal pscudomorph.
Die fine dark lines on the surface represent shrinkage cracks, fractures,
and coalescing pores that were caused by the calcination. These features
follow regular geometric patterns and probably are controlled in part by
the structure of the calcium oxide pseudomorph after calcite. The coarse
textural features on the surface indicate local irregularities or grain
boundaries of crystals that are not in structural continuity with the
crystals that comprise most of the surface.
The reaction zone between the CaS04 and CaO in the crystals is
shown in Figure 19- The particle labeled A near the top of Figure 19
is the same particle labeled A near the bottom of Figure 18. These tvo
figures show that in a distance of about 25 microns there is change
from unreacted calcium oxide to a well-defined calcium sulfate zone.
The line B-C in Figure 19 shows the approximate position of the calcium
sulfate-calciuQ oxide reaction front.
These electron microscope studies thus confirm mercury poro-
simeter measurements which show that 90^ of the surface area of the
particles is accounted for by pores smaller than O.O35 micron. Measure-
ments on Figure 18 show that the widths of the shrinkage cracks and

-------
1*5
. r;
¦M .'V,'VV'J .1. ¦! I..M ty	'i ¦-
\y	iV •-¦ '	\
•'• • •¦ • ; "r — * ' vT •':
' . Vo * *	\ • % / ¦<¦*¦' y
U-^T	-v., V, \> Vn.;/^
!'V-' a.. :-\.v-r\..
c* '.
>r
»•<
&*>\ ,V-
w&g- a :;
\ *
* y	v.~^\ . \f'\ . ^.-•v .'•••.
• • ¦ v^- »	. • ¦ • V. V.
fer- y • \ •	¦, .1. ;•:; *
•»* ^ - *	' • ¦•• ¦• ! • I
%L.	+. -i* ¦ •• ' .7	-	.7 V J • V>Hir f /> "M
1	^ •'N/V'	v»v ,'V* /¦" vi,k-.v>'
i#;v	'• <*0&y \V^4\ • .'¦
¦ •. •; VV**/- ^ xv:- .:.3l
^	1	t	( ¦ . •/. W»'i	t l . »	*1 , V '	* * . A1
f	% ***'' » •
rm*./7-t^ •	%..• .>k~—4-""
iJ>K .' ••• x#£.. Jtel ..--A- ' .,/•¦ '>
M ii.r	/	• ' \	\* ^,r\/vf » *. .lA."'.	. •!

0
L
X
1

jL

±
Figure 10
Surface Fractures on the Cleaved Surface of a Calcined
Single Crystal of Calcite. 12,000X

-------

ft * Ml »v
%''
V *0/),
v.Jy. %

r
' w\ n «w . T• )	V ,¦<** • w/) ^ ;
. *^S "•  , L.i	\ vi> », t/rt
•« ••••- \.- .• -ffrh-\	t!-'t"'1 V!
1 —¦ • »...- '«		 "., ¦> 4^/, .	.) . .
' leSV' ''••¦''•¦¦	•••c-i" VVH
o
L
5
I
Figure 19
Reaction Zone of Calcined. Single Crystal of Calcite
after Exposure to Sulfur Dioxide. 12,000X
[The line BC divides the region of unreacted CaO (upper)
from that of product CaS04]
jj> H

-------
I»-T
coalescing pores are no wider than 0.030 micron, and many are much smaller.
The agreement of the measurements for the surface-area and size of con-
tributing pores made by the two methods thus is not likely to be
coincidental.
The electron micrographs clearly show that the reaction of
the SO2 with the line does not follow these i fractures. Some of
the fractures shown in Figure 19 can be followed well into the sulfate
reaction zone. The geometric surface area of the particular grain may
be much more important in promoting potentially active sites for S02
reaction than the fine (less than 0.1 micron) pores. If the retention
time in the reaction zone is short, the geometric surface area is even
more important than the total surface area. The micrographs also seem
to indicate that the sulfate reaction takes place along a reasonably
well-defined front so that diffusion models based on such a concept would
be realistic for long reaction times.
A study of a series of photographs, of which- the two presented
here are a part, shows that the reaction zone is about 70 microns vide
in this sample. The calculated volume of the reaction zone was compared
with that of the unreacted oxide. It was assumed that the particles
were spherical and that the bulk densities of the two zones were the
same as the densities of the pure material (i.e., no allowance was made
for voids or packing). The results of the calculations show the weight
ratio CaS04:Ca0 to be about 1:U, or that about 20$ of the weight is
calcium sulfate.
U8
These studies showed also that there is a general decrease
in'the crystallite size of the calcium sulfate as the reaction zone is
approached. The average size of the calcium sulfate crystals at the
outer edge of the sulfate reaction zone is about 1 micron, whereas the
average size of the crystallites in the reaction zone, as seen in
Figure 19, is 0.5 micron or less.
Properties of Calclte and Dolomite Calcines
In a comparison of the properties of calcines of a calcite
a dolomite as affected by the conditions of calcination and the
initial particle size of the stone, the effects of the variables
listed in the tabulation were stu'lied in a factorial manner.
Levels	
Calcination	1	2
Time, minutes	1	5
Temperature, °C.	1100	1225
Limestone
Particle size, microns	38b	136
Type	Calcite	Dolomite
Hie calcite, a very pure, coarse-grained limp stone,
designated Sample no. 89 Coarse, was the same stone that had been used
in the textural comparison with Iceland spar and in the dead-burning
tests. It contained only 0.h% accessory minerals (quartz and lllite
clay). Its mean crystallite size was 3875 A., and its lattice constant,
a, was U.987 A., which shoved that there had been no significant
substitution in its crystal lattice.

-------
*9
The dolaaite, designated the Janes River Dolomite, BCR no.
2069, vas supplied by a bidder for supplying limestone for the Shawnee
full-scale injection tests. It contained 1.9fp accessory minerals,
largely quartz and feldspar, and vas the purest dolomite available for
©
the tests. Its mean crystallite size vas 2075 A., and its lattice constant,
a, vas U.8O5 A.
The stones were hand ground and dry screened, arid the fractions
then were dispersed and vet screened. The vet-screen fractions vere
washed with acetone and again dry screened. The final products vere
free of aggregates and fairly uniform in size. The minus 3U- plus
U2-mesh and minus 100- plus 115-mesh fractions vere used in the tests;
the average particle size vas calculated as the arithmetic average of
the tvo screens that just passed and just retained each fraction.
The calcinations vere made in a b.5-cm. l.d. vertical tube
furnace. About 0.5 cram (essentially a monopartlcle layer) of the stone
fraction was placed in the bottom of a 5.5- x 5.5-cm. cylindrical holder
made of platinum and covered with a perforated lid. A 1^00-cc./minute
stream of a mixture of 15$ C02 and 85$ N2 vas used to flush the furnace
tube; the flov vas sufficient to change the furnace atmosphere about 2.5
times per minute. The preheated furnace vas raised abruptly around the
sample, held for the desired time, and then lowered. The temperature
vas measured vith a thermocouple with the bead Just belov the sample
holder. A total of 5 to 8 heatings vas made under each set of conditions,
and the products were composited to obtain sufficient calcines for
measurements of their properties. The properties of the calcines that
vere measured are listed in Table V.
p
c
tr- fr- M h
13 H
I k* H
|d a
SP-
x m-
h><3
£3
C+ or
(0 a
14 %
p
R
o
3
Nj>

o
B
3
vO
VH
o
4
to
hi M i-»
ro to q o to ro g g
h* t-» »— H V Vf V	OJOD&QD
Cnononovc- r p*r
CO
«?
§
a
&
00000000
CD Co ro ro ro ro
fO ©o -p" —. CD ON
fO H H H
OOOOOOOO
WIOHMMNHH
ro oj^o v-n o o
W H rm o h
OOOOHH-'MH

n0 NO ^OSO VO \£> vf> NO
0d (» Co CdvS Cb Co-J
vo ^"NJi -e- 00 On
MfOHPIOHHH
osrovo-J^*o—J-c*
vi ovsp o OU1VJ1
OOOOOOOO
»— h- y- t-> \-h
NjiV^ViiVjl 00
On C* <7n On -P"

ro
8.
OOOOOOOO
*C -t* VJ» ON -c- -C"VJi \J}
-£- vji vji n> vji -c 0>
O^—J-J—JO* On H"
OOOOOOOO
ss io ss
M H- o ^ »
.t#|
o 3
• *0
a> »u »
{J <
N

y +- iu r— vi 1— iu r—
CD Q3 —J N> fONOV* 00
88888888
3"
<9
fe'S
rn s?
1 is
-J V_N OWJ1	-J
-J SO
«8
0
0
p
0
*3

M- 0)

ft

O £

M
&

10
<+

0
3
c
a

I
oS

-------
51
Statistical analysis was made of the data in which nl 1 four
factors were tested—time and temperature of calcination, and particle
size and type of stone. Analyses were made for each measured property
and for Its logarithm, as shown In Table VI. The residual variance in
each property was considered to be the }- and lv-factor interactions;
these variances were recalculated in terns of residual error and are
shorn in Table VI in units of the property. Thus, the residual error
is U.8 mg- SO3/IOO ng. calcine for the analysis of the effects of the
factors on SO3 absorption; it is 22.of the absorption when the logarithm
Is considered. The residual error should approximate the experimental
error of replicates provided none of the 3- or It-factor interactions
are significant.
Absorption of SO*.: The absorption of sulfur oxides was
measured as the gain in weight of the calcine in 30 minutes at 930° C.
In an atmosphere of S02, 0S, and 92$ Ns. The capacity to absorb
SO3 was decreased by Increasing the calcination tine from 1 to U minutes,
by Increasing the calcination temperature from 1100" to 1225° C., and by
grinding the stones to 136-mlcron (100-mesh) particles rather than to
38^-micron (bO-mesh) particles when the stones were calcined for these
times at these temperatures. The dolomite was a much more effective
absorbent for SO3 than vas the calcite under these conditions of
calcination. These overall effects' of the independent variables (called
"main effects" in statistical language) were significant with respect
to the residual error, regardless of consideration of the numerical change
in the weight during absorption or of the percentage change in weight
(corresponding, respectively, to the analysis of j or log where j
is the weight gain).

-------
53
The type of Btone, calcite or dolomite, Interacted significantly
with both the time of calcination and the temperature of calcination in
affecting both the numerical and percentage change of absorption. These
simple 2-factor interactions can be demonstrated by the following
tabulation.
Absorption of SO^, mg./lOO tag.,
under indicated conditions of
	calcination	

Time.
min.
Tenro.
. 'c.

1
k
1100
1225
Calcite (C)
23-P
15-5
27-0
11.5
Dolomite (D)
52.0
fc9-0
60.0
Ul.O
Ratio D:C
2.30
3-15
2.2b
3-5U
In the tabulation the SO3 absorption is averaged for all particle sizes
and temperatures when considering the Interaction between type of stone
and time of calcination. Similarly, in considering the interaction
between type of stone and temperature of calcination, the averages for
all particle sizes and times are calculated. For example, for the four
calcite heatings for 1 minute It is calculated from the data In Table V:
23 = (36 + 13 + 31 + 11)A- The ratios of the absorption by dolomite
to that by calcite Indicate the Interactions. The dolomite absorbed
more than three times as much SO3 as the calcite for all calcination
times of U minutes, whereas it was only 2.3 times as effective for all
calcination times of 1 minute. Similarly, the dolomite was 3.; and 2.2
times as effective as the calcite at 1225° and 1100° C., respectively.
5*
These conclusions are based upon a standard error of 22.5^ in
SO3 absorption or a numerical error of h.8 mg. SOa/lOO mg. calcine.
This relatively high error suggests that one or more of the high-level
interactions used in computing the error may also be significant-
A better estimate of the error for SQa absorption was obtained
by running partial duplicates on the eight calcite samples. Canplete
duplicates were not made because this would entail duplication of the
screening and firing; the firing was done in small, monolayer batches and
composited as described earlier. Duplicate absorption measurements were
therefore made on the composited calcines. Each absorption of S03 was
made In the TGA apparatus and the recorded weight increase was checked
by weighing before and after absorption on an analytical balance and by
chemical analysis of the product after absorption was complete. The
recorded absorption was taken as the average of the three methods of
determination. Table VII shows the analysis of variance from which the
error of the partial duplicates is estimated as 1.2 mg. SO3/IOO mg.
calcine or 5.9$ of the average absorption by the calcite samples. The
error of true duplicates would be expected to be only slightly greater
than these values.
The data in Table VII show that there was a significant
difference among the absorptions of the eight samples, that there was
a difference among methods of determining the weight increase, and that
the duplicates agreed better for some samples than for others. The
TGA recorded weight increase was higher than that determined by the
analytical balance or by chemical analysis; this was due to a drift in

-------
55
TABLE VIX
Analysis of Variance for Partial Duplicates of Eight Absorption
Samples Using the Average of Three Methods of Analyses
Corrected
Soufce of
variation
sums of
squares
Degrees of
freedom
Mean
square
F-ratio
Samples (S)
3727.75
7
532. %
123 **
Methods (M)
42. $k
2
21.!»7
U.95**
Duplicates (D)
0.12
1
0.12
0.03
8 x M
57.52
lit
lt.ll
0.95
S x D
111.31
7
15.90
3.66*
M x D
0.83
2
0.U2
0.10
Residual
60.77
1U
i».3Ua
-
Total
U001.2U
*•7


Significance at Indicated confidence level: " 99$;» 95%.
a Error = V».3*»/3 • 1.2 mg. EO3/IOO mg. calcine.
56
the calibration during the absorption run and Is routinely corrected
for In precise TGA work. However, the Interaction between the duplicate
and samples suggests that the error might be proportional to the measure
meat. This was verified by considering the data for the three methods
of analysis for the eight dolomite samples. The residual error was
2.2 mg. SO3/IOO mg. calcine or 7-5/5 of the average absorption by the
dolomite samples. The error analyses indicate that a percentage error
is more appropriate for the absorption of SO3 and that the analysis in
Table VII of the logarithm of the absorptions is preferred for this
property of the calcines.
The residual error of 22.5$ In Table VI for the logarithm
of the absorption of SO3 Is considerably larger than the error (5.94 to
T»5$) estimated for the runs. This suggests that there are one or more
high-level Interactions that are significant for the absorption of SOa.
The sums of squares for the 3- and U-factor Interactions are
Sums of squares
t x T x PS	0.0022T
t x T x type	0.01875
t X F8 X type	0.00006
T x PS x type	0.00093
t x T x PS x type	0.000U7
t ¦» time, T = temperature, PS =
particle size.
The heterogeneity of these variances Is evident from the fact that the
sum of squares for the interaction of 'time,- temperature, and type is
nearly an order of magnitude greater than the next greatest sum of

-------
57
squares. When this Interaction Is ellnlnnted fraa the residuals, the
error Is calculated to be 7-3$ of the measurement, vhlch is in good
agreement with the estimated error fron the partial duplicates.
The interaction of tine of calcination, temperature of calcina-
tion, and type of limestone can be demonstrated by the following
tabulation which averages the results for the two particle sizes used.
Time, 	Calclte	Dolomite	
mln. 1100° C. 1225° C. 1100° C. 1225° C.
1	53.5 12.0	60.0 1A.5
It 20.0 11.0 6O.5 37.0
The effect of changing the time of calcination from 1 minute to U minutes
is insignificant for the calclte calcined at 1225° C. or for the dolomite
calcined at 1100° C.; however, and 2Oj decreases, respectively. In
the SO3 absorption are obtained by increasing the calcination time for
calclte at 1100° C. and for dolomite at 1225° C. An insignificant
3-factor interaction would have shown more nearly the same effects for
the two materials at the same temperature.
The interaction of time of calcination, temperature of calcina-
tion, and type of limestone on the capacity of the calcine to absorb SO3
is shown in Figure 20. The isopleths of SO3 absorption were constructed
fran an equation that contained only the significant effects shown in
Table VI plus the three-factor interaction, time x temperature x type.
In coded form the equation is
log y = 1.U6M - O.OltZjIfXi - 0.13055x2 - O.O2SHX3 +
0.2293Ux4.+ 0.02356x5^X4 + 0.01»U66x2X4 - 0.03U23x1x2x4 (20)
Figure 20. Isopleths of SO-, Ahsori>ticr. As functions of
the Time and Temperature of Calcination. [Numbers on
curves indicate weight gain (a3 SOaJ/lOO mg. calcine]

-------
59
where *1 refers to the time of calcination, xa to temperature of
calcination, *3 to particle size of the limestone, and *, to the type
of limestone. The coded variables, x^, have the values of -1 and +1,
respectively, for the low and the high levels of the real variables
shown In Table V. The variables , x2, and 313 are quantitative or
continuous variables, since numbers can be assigned to the levels;
variable X«, however, Is a qualitative or discrete variable and refers
to a type of material for which the assignment of numbers to inter-
mediate levels is questionable. The coefficients of equation 1 were
calculated from the data In Table V by analysis of the logarithms of
the SO^ absorptions.
The contours of absorption of 8O5 by the calclte shown in
Figure 20 are flatter with time when the temperature is high (1225° c-)
than when the teiperature is low (1100* C.); for the dolomite, the
contours are flatter at the lover teniperature. This results in a response
surface for the calclte absorption that is steeper at short calcination
tines than at long calcination times; for the dolomite, the response
surface Is steeper at h minutes calcination time +>«" at 1 minute
calcination time. This suggests that at 1225" C. the calclte has
already lost Its reactivity toward SO2 after 1 minute of calcination
whereas the dolomite still has capacity to react at this teinperature.
The optimum conditions for the calclte absorption would therefore be
at some calcination time shorter than 1 minute, whereas the temperature
of calcination would be more critical than the time of calcination In
obtaining the optimum conditions for the reactivity of the dolcmlte.
60
(In Figure 20, no units have been shown for intermediate times and
temperatures because it cannot be established from these tests whether
the scales should be linear, logarithmic, exponential, or of some other
form.)
Other Properties of the Calcines: The total pore volume (of
pores between 98 and O.O35 micron), the two density-measurements,-,and
the degree of calcination have residual errors either in y or log y
that approximate the errors of measurement of these properties. The
residual errors for the crystallite size of CaO are 920 A. in y and
555k In log y, and for the logarithm of the pore volume of pores between
IT*5 and 0.(9; micron the error is about 11$. These are considered
larger than the measuring errors and they therefore suggest that there
are significant J- or k-factor Intel actions of both of the variables
studied that affect these properties. These Interactions are Identified,
In coded form, as
X1Z2X4 for the pore volume (IT-5 to 0.035 micron)
both 11X2X3 and x^xaxg** for the crystallite size
of the CaO.
lbe 3-factor interactions can be illustrated In the same manner as was
tone above fbr the Interaction of"xIxsx4 In the absorption of SO3. The
factor Interaction, x1xsx3x^, is sore difficult to understand in
physical significance, but it can be readily demonstrated In mathematical
terns by considering the two 2-factor Interactions, xjx2 and 13X4, and
constructing the following tabulation.

-------
61
Crvatallltc Site of CaO for Different
Combinations of XiXp and x->xA Interactions
		
xi»?	-	+
25UO A.	2080 A.
+	2050 X.	2570 A.
The low value (-) for each 2-factor interaction Is taken as a combina-
tion in which one of the factors is at the lov level and the other is
at the high level; the high value (*) for the interactions indicate that
both factors are at the high lefel or both are at the lov level. The
fa..factor interaction for the CaO crystallite size is then demonstrated
by the decrease that occurs as one 2-factor interaction is changiu from
its low value to its high value while the other 2-factor interaction is
maintained at its low value, and by the increase that occurs at the high
value of the second interaction; In the absence of a factor interaction
these effects would be about the sane.
A very prominent interaction in the total pore volume, and the
density obtained therefrom, Is that of the particle size of the limestone
and the type of limestone. This is evident frao the 5th column of
Table V where the smaller particle size of the dolomite gives rise to
a much larger total pore volume over the coarser grind than that for the
calcite. This may explain why the total pore volume is not as good an
index for reactivity as is the pore volume which includes only those
pores 17.5 to 0.055 micron in diameter; the effect of particle size for
62
different types of stones is largely eliminated by se3Action of the
proper pore volume.
The pore volume which includes only pores of 17.5 to 0.035
micron, column 6 in Table V, has a significant interaction between the
type of stone and the temperature of calcination. Changing the tempera-
ture of calcination from 1100° to 1225° C. caused a decrease lit this
pore volume for the calcite and an increase for the dolomite. This
temperature change caused a decrease in the absorption of SO3 for both
types of stone, although the decrease was not as pronounced, percentage-
wise, for the dolomite as for the calcite.
The degree of calcination, column 9 of Table V, ranged from
about 96 to 100£ under the conditions of the tests. The dolomite was
slightly more nearly completely calcined at 1100° C. than was the cal-
cite, but it is doubtful that differences in the degree of calcination
can be related to the reactivity with SO3. The particle size of the
stone was not significant for the degree of calcination under these
conditions.
The absorption of SQa by a calcine is influenced greatly by
the conditions of calcination as well as by the type of stone and the
degree of grinding the stone prior to calcination. The significant
interactions of these factors on several properties of the calcine have
been identified. There is not a complete correlation of the effects of
the factors for SQ3 absorption with any single measurable property; it
is quite likely that the absorption of depends on several of the
other properties of the calcine.

-------
a
The presence of high-level interactions of factors far many
of the properties of calcines raises serious questions as to the
applicability of second-order designs to study of the system. In the
ranges of the variables studied, third-order designs appear to be
required for many of the properties, and the crystallite growth-of CaO
may require even higher-order design. It. should be noted, however, that
the order of the design required to represent a property adequately vlll
be decreased as the ranges of the factors are decreased. This nay permit
the use of second-order designs for injection conditions where the
experimental conditions vary less widely, particularly for the tine
of calcination. It Is probable, however, that this information can be
obtained only under actual Injection test conditions at a power plant.
6U
COMPARISON OF LMESTOITES
Introduction
The injection of limestone into a power plant combustion
chamber results in dispersion of the particles of the stone In the gas,
calcination, and reaction of the' calcine .with the S0a in the gas. The
system is polythermal in that the temperature of the injected material
rises from ambient to nearly that of the flame, and then cools at an
unknown and perhaps variable rate. The calcination reaction is endo-
thermic and serves to lower the temperature attained by the particles on
Injection; the sulfation reaction is exothermic and, if sufficiently
rapid, serves to raise the tempervture of the particles. The short
retention times of both calcination and sulfation, the high tempera-
tures attained, and the high velocity of the gas stream and Its
entrained solids make it difficult to reproduce the conditions In the
laboratory and measure the reactions with the precision desired.
Laboratory simulations of the injection system Include com-
promises of one kind or another. Fixed-bed tests with calcined stone
do not include the calcination reaction; dispersion of the particles in
the gas stream is not attained in fixed beds of either raw stone or
calcine, and the rate of mass transfer is much slower than is attained
In practice. Most laboratory tests have been capacity tests rather
than rate tests, and the correlation of capacity with rate has not been
demonstrated sufficiently to establish such confidence in capacity tests
alone. Results of Isothermal tests of capacity vary markedly with the

-------
65
tenperature and vith the stone (2). The dlspersed-phase reactor of
Battel1e (6) approaches the Injection system physically, but the
problems of sampling, freezing all reactions, and sensitivity to
chemical analytical errors preclude the precise determination of
mechanisms, rates, and efficacy of the process that is desired.
This section of the report describes tests of selected stones
in an attempt to relate their calcination and sulfation properties—both
capacities and rates—to the chemical and mineralogical properties of
the stones. Like other laboratory tests, these suffer from compromises
that prevent determination of the absolute correlations desired.
Several rate and capacity measurements were made, but it cannot be
determined at, present which, if any, of them apply closely to the
injection system.
It was hoped that activation energies and frequency factors
of the sulfation reaction could be obtained fran the polythermal TGft
studies, and that the resulting kinetic parameters could be used In the
equations derived In Appendix C. The weight of the sample in the TGA
apparatus is the combined result of the amount of calcination that has
occurred, the uptake of SO2, the amount of oxidation to sulfate, the
amount of decomposition of the sulfite as the temperature is raised,
and perhaps the substitution of S02 for C02 without calcination. It
became obvious that a determination of weight as a function of time
and temperature would hardly suffice to define the kinetic parameters
of sulfation, even though many of the kinetic parameters of the other
reaction were known. The TGA was not sufficiently sensitive to detect
a change in weight wben SOp was in the atmosphere until the temperature
was JO* to 100°C. above that at which decomposition began when SO^ was
absent. This implies either a retardation of the calcination reaction
by S02 or a substitution reaction of 80a for COs in which no appreciable
weight change occurred. In either case, the kinetics of calcination in
nitrogen cannot be used to predict calcination losses when S02 is present
The multiplicity- of simultaneous reactions, as described above, precludes
the accurate determination by TGA of the kinetic parameters for sulfation
The sulfation of the various stones, however, has pointed out distinctive
characteristics that nay be correlated with their chemical and mineralogl
cal properties.
Selection of L<""»gtcoes
A group of TO to 80 stones was supplied by the Air Pollution
Control Office (AHX>), and each atone was studied petrographically to
identify the rtimftlc carbonates and the accessory minerals, and to
determine the range of grain sizes in each stone. Prom this group 35
stones were selected for calcination and sulfation tests, and a detailed
ndneraloglcal characterisation was made of each stone.
The coiqposltlons of the 35 limestones are shown in Table VIII;
18 stones were primarily calcites, 13 were dolomites, 2 were mixtures
of calcites emu uulumites, and k were mixtures of magnesite and dolomite.
The chemical compositions were determined spectroscopically by Bituminous
Coal Research, Inc.; the Ignition losses agreed with those we obtained

-------
67
Ch'-alccl ind Hln
TAPI-E VIII
leaJcnl flvmU. b of Itcritunin
Ccnpo«ltion. wt. f*
BCD
l«n.




fe.
Ion
CnO
*b0 ftA
BftgO
'go





Calclt» •
13*
*3.?
53.*
1.5 0.?
0.0057
0 0#*
*3*3

5-.1
O.J 0.*
0.0115
0.0->/',
1350
*3.6
5*.7
0.5 0:?
0.0056
o.ccn?
1355
33.1
*1-5
1.2 ] .6
0.25t2
0.*)*9
1359
*).6
*.7
0.7 0.1
0.0056
0.03&2
1>63
3G.1
*2.2
0.9 J. 6
0.1*06
0.6MB
13&
to 1
*9-1
1-3 0.9
O.OP98
0.26)6
13*9
*3.*
>.3
0.0 0.1
O.OJ96
0.073C
1377
27.J
ae.*
a.e 3.1
0.2326
j.9985
1379
*1.3
*8.7
s.a 0.*
0.1526
0.2016
16TI
26.5
31.2
1.1 0.7
0.0735
0.2058
1679
*2.7
52.1
1.* 0.)
0.0*17
o.tfte
l€9l
37.7
ki>.9
? 2 0 6
0 C»2)
O.PO*
14£)
*3.9
VS.?
*.6 0.*
0.0505
O.CPfiO
1(67
kb*
33 *
1.2 0.1
0 0P&J.
Q.&jCA
1*9*
je.6

1.7 0 6
¦0 ce^o
0.3*38
169?
*2.9
5*i
0.3 0 1
0.0266
O.CU96
1693
0.7
St).)
0.6 0.2
O.OJ73
0.10)J
Oraln
•ue,
coded
Uult-ccll Cryntnl.
length, lit"
nlf, J
O.OOA?
1.6
3.0
*.9fl
0.1*9
0.03*5
0.7
1.0
* 909
O.Kf
0.0b 51
2.6
2.0
*.996
0.1//
0.0.^68
2*. 6
1.0
* 96?
0.0/.
0.01*1
0 9
1 0
*.965
0.12*
0.1278
lA.O
1 0
*.969
o.i-r«*
0.9165
a.e
?.6
* 979
0 097
0.0005
o.a
1.0
*.9^7
0.12/
0.0672
*3-8
1.0
*¦975
0.16?
O.C*» 11
6.6
3-0
*.5*0
0 0'jt
0.0110
*5.0
2.0
* 9&7
0.1/7
0.00V>
3.6
1 0
*.97r»
0.93
0.0033
13.8
1.0
* 9*>
0.11'.
0.^*17
3 0
.2.8
* 996
0 1 *'•
0 a-3V
1.5
2 f>
* 986
0.1//
0.0*50
O.OC66
12 *
1.0
*.96*
0.10/
0.6
3.0
*.9«5
0 153
0.00*.0
22.*
15.5
1.1
0.0395
1.1266
1695
*P.2
£9.9
23.1
0.6
0.0289
0.0299
1701
ka.5
3*.5
:?.5
k.)
0.0920
0.1W3
17®
u.6
33.2
ib.o
0.8
0.0277
0.2715
0.0079
1.0
2.0
*.811
0 J19
0.2690
2.0
1.5
* .808
0.1*1
0.0090
0.7
3.0
*.89*.
0.211
0.5610
7.0
1.0
*.016
0 ov.
O.OOdO
0 6
2.5
*.805
O.I89
0.0M5
35.0
1.5
*.808
0.1*1
O.O69I
9.2
i.O
*.806
0.171
O.OOQO
1.0
3.0
*.806
0.1*9
0.0)59
0.5
1.5
*.807
0.1Q?
0.0792
2*.?
3.0
*.809
0.098
0.00)7
*.8
).0
*.8c8
0.189
0.*5*2
8.6
1.3
* .822
0.090
0.008)
*.6
1.3
*.810
0.970
1*0
169^
1375
1)76
tf.o
*6.*
*8.8
*5.0
>7.?
ij.e
7.1
7.2
12.*
3*.2
*1.0
0.7
0.6
0 0277
0.0)10
O.0277
0.0)10
Btgwiltw
0 0322
O Qlji»
0.13*0
0.005*
0.0588
0.0073
0.0000
O.Ol^k
).*
12.6
10.0
96.0
2.5 *.910
3.0 *.886
2-5
9.5
*.7?0
*.720
0.1
0.1S9
0.1*7
0.220
*	C&lcultied froa ecaro«!tlc-i cf etlclar.
*	Accessory altterals.
e 1 • fine, < C) i( } ¦ ctdia, fj-250 u. 3 ¦ coerce, >>230 11.
68
in our calcination tests. The CaO, MgO, Fe^O3, Na^O, Kp>0, and MnO^
content'	determined on the calcines: they were recalculated to
reflect the composition:, of the original stones. The amounts of
acces:-'^, minerals, column 9 of Table VIII, were determined by decomposing
5 to 5 grams of each limestone in UOO ml. of 1$ acetic acid and main-
taining the pH at U by addition of acetic acid until all the rhombic
carbonates vere dissolved; this required about hours for the calcltes,
1 to 2 weeks for the dolomites, and several weeks for the stones that
contained considerable magnesite (MgCOs). The accessory minerals were
recovered unaltered, and their amounts are listed in Tabic IX.
The last three columns of Table VIII show the microscopic grain
texture; the x-ray unit cell length, a, of the major carbonate phase;
and the x-ray crystallite size within the grains. The grain texture,
column 10, was determined petrographically and coded to agree with the
scale used by /'.rvey (10); for the purpose of this tabulation the rather
broad groups of fine, medjwn, and coarse grain sizes are labeled 1, 2,
and 3, respectively, with an occasional variation from Integers as the
predominant grain si?e approached the arbitrary values of 65 or 25O
microns that separated the groups. The grains were Irregular in
shape, and there was a large range of grain sizes In each limestone;
the predominant grain size in each stone is therefore given a semi-
quantitative rating for correlation with its calcination and sulfation
parameters.

-------
69
TftSLE n
Accessorr Mineral Content of Llae»toces
Accessory	vt. ^
BCB


n-
ear-
ill-

treo*
Mpift-
Feld-
Musco-
Glass
go.
Total
(tovts
lite
onlte
Chert
Lite
alte
spar
vite
frit





C&lcltes



1JJ6
1.6
0.0
0.0
O.O
0.0
1.0
0.0
0.0
0.0
0.0
iy>5
0.7
0.3
0.1
0.0
0.2
0.0
0.1
0.0
0.0
0.0
1550
2.6
1.6
0.9
0.0
0.0
0.0
OJ.
0.0
0.0
0.0
1355
2b.6
16.0
8.6
0.0
0.0
0.0
0.0
0.0
0.0
0.0
1359
0-9
O.h
0.0
0.0
0.0
0.0
0.0
0.0
0.1
O.fc
1363
18.0
6.0
6.0
0.0
0.0
0.0
0.0
0.0
6.0
0.0
1368
8.8
fc.2
h.2
0.0
0.0
0.0
O.h
0.0
0.0
0.0
1369
0.8
O.h
O.h
0.0
0.0
0.0
0.0
0.0
0.0
0.0
1377
1*3.8
13.1
1M
0.0
0.0
0.0
*•5
0.0
13.1
0.0
1379
6.6
fc.6
0.0
0.0
0.0
0.0
0.0
1.0
0.0
0.0
16T7
fcs.o
lh.0
0.0
lb.0
0.0
0.0
3.0
0.0
0.0
lh.0
1679
3.6
1.7
l.T
0.0
0.0
0.0
0.2
0.0
0.0
0.0
1681
lj.8
b.6
h.6
0.0
b.6
0.0
0.0
0.0
0.0
0.0
168$
3-0
1.0
1.0
0.0
0.9
0.0
0.1
0.0
0.0
0.0
168?
1-5
0.5
0.5
0.0
O.h
0.0
0.0
0.0
0.0
0.1
1691
12.il
5-a
6.0
0.0
0.0
0.0
0.6
0.0
0.0
0.0
1692
0.6
0.3
0.2
0.0
0.0
0.0
0.1
0.0
0.0
0.0
1693
2.8
1.0
1.8
0.0
0.0
0.0
0.0
0.0
0.0
0.0
polcmltes
1337
1.0
0.7
0.0
0.2
0.0
0.0
0.1
0.0
0.0
0.0
13U0
2.0
2.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
13U1
0.7
0.0
0.0
0.0
0.0
0.5
0.0
0.1
0.0
0.0
1367
7.0
0-3
0.0
0.0
0.0
0.0
6.7
0.0
0.0
0.0
1380
0.6
0.1
0.1
0.0
0.0
0.0
0.0
0.1
0.0
0.3
1678
35.0
15.7
15-7
0.0
0.0
0.0
1.8
1.8
0.0
0.0
1680
9.2
3-1
0.0
0.0
0.0
0.0
0.0
3.1
0.0
3.0
1666
1.0
0.3
0.3
0.0
0.3
0.0
0.0
0.0
0.0
0.0
1688
0.5
0.2
0.2
0.0
0.0
0.0
0.0
0.0
0.0
0.1
1690
2h. 2
20.6
3-6
0.0
0.0
0.0
0.0
0.0
0.0
0.0
1695
k.Q
l.h
0.5
0.0
0.0
l.h
0.0
0.0
l.h
0.0
1701
8.6
3.0
0.0
0.0
0.0
0.0
5.6
0.0
0.0
0.0
1702
h.e
2.3
2-3
0.0
0.0
0.0
0.0
0.0
0.0
0.0





Mixtures





1360
}.h
0.0
O.li
0.0
1-3
0.0
1.3
0.0
0.0
O.b
169*
12.6
h.O
0.0
0.0
0.0
h.O
0.0
0.0
0.6
0.0





Kagnesites




1375
10.0
1.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
1376
28.0
0.7
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
70
The unit-cell length, a, coluzm 11 of Table VIII, vas
determined by measuring the d-spacings corresponding to various Miller
indices (hXi) in the x-ray diffraction pattern and using the method of
least squares (18) to obtain the unit-cell dimensions or lattice
constants. Goldsmith and Graf (£) shoved that both the a and c
dimensions of the unit cell were functions of the degree of substitu-
tion of magnesium for calcium in both calclte and dolomite and of
calcium for magnesium In dolonite. The larger ratio of the radius of
Ca to that of 0 (0.67) compared with that of Mg to 0 (0.U7) results
in a decrease in the cell dimension when magnesium substitutes for
calcium, and an increase when calcium substitutes for magnesium.
Thus, the length of the lattice constant, a, is an Indication of the
Kind and degree of substitution in the rhombic carbonates; the pure
constituents have ideal values of U.990 A. for calcite, U.808 A. for
dolomite, and ^.633 A* for magneslte.
The x-ray crystallite size, column 12 of Table VIII, represents
the mean size of the crystallites within the grains of the major
carbonate minerals--calcite, dolomite, or magnesite--ln each stone. This
property was measured by the x-ray line-broadening technique described
by Rau (22) using the value of the pure diffraction breadth due to
crystallites as defined by Azaroff (l). The (101*0 reflection was used
to calculate the mean crystallite size for the stones because It is the
strongest diffraction maximum for the rhombohedral carbonates and is
particularly useful when the samples contain mixtures of carbonate
species. The crystallite sizes are reported in microns rather than the
customary Angstrom units (1 A. = 10"V)- Both the crystallite size

-------
71
and the lattice dimension of a limestone are properties of the com-
position of the carbonate fraction of the stone, particularly the
distribution of calcium and magnesium atoms In the unit cell.
The accessory minerals of each stone, column 9 of Table VIII,
were further identified as to amount and type of Impurity by petrographic
methods. The results are presented In Table IX for ten of the more
cosmos Impurities found in the stones, Including an unidentified fritted
glass. Pyrite, mica, and some unidentified phases were found in a few
stones but in very smMI amounts. The total residue from the acetic
acid treatment ranged from 0-5 to of the stone. The 10 components
listed in Table IX account for essentially all the Impurities for all
the stones except the magnesites (BCR 1375 and 1376); for these two
stones the bulk phase of the acetic acid residue, after 3 months
treatment, was undissolved magnesite.
Calcination
The calcination tests were made in the TGA with 50 mg. of the
minus 2b- plus 28-raeah fraction of each stone. The sample was spread
as a single layer In the sample dish and covered with glass wool and a
perforated lid to prevent sample loss and allow escape of gases. The
equipment was flushed with nitrogen gas at the rate of 500 ml. per
minute, and the heating rate was approximately 5* C. per minute.
Typical smoothed curves of weight versus temperature are shown in
Figures 21 and 22.
72
tcmperature.'k.	TIME.MIN
Figure 21. Calcination and Sulfation of a Calcite, BCR 1679 -
Figure 22. Calcination and Sulfation of a Dolomite, BCR I69O.

-------
T>
The parameters for calcination are given in columns 2 through
5	of Table X. The inflection temperature is the point at which the
curve changes from concave dovnward to concave upward, and also is the
temperature at which the rate of calcination, column is greatest
for the particular sample size, particle size, and heating rate. The
activation energy and pre-exponential term of the Arrhenlus equation,
columns It and 5» were determined by least squares as described in
Appendix B; the relation between these parameters is shown in Figure 1.
Sulfation
Isothermal : Capacity measurements of the absorption of sulfur
oxides by 50 mg. of precalcined stone were made in the TGA at ?'*0' C.
Each stone was soft-calcined at 900° C. for 50 minutes in a nitrogen
atmosphere, and the minus 2k- plus 28-mesh fraction of the calcine
was subjected for JO minutes to a flow of 500 ml./min. of an atmosphere
of 1$ SO2, hit 02, and 92$ Ng. Curves similar to those of Figure 2
were drawn to obtain the first-stage and total weight gains. Columns
6	through 9 of Table X show the important characteristics of each stone
with respect to isothermal sulfation at 930° C.; these include the
total capacity, the maximum rate of weight gain, the first-stage
capacity, and the time to complete the first stage under these conditions.
Polythermal Calcination-Sulfation
The polythermal calcination-sulfation tests were made on
50 mg. of stone at a heating rate of about 5° C. per minute in a flow
of 500 ml./min. of an atmosphere of k S02, 02, and 92?' N2. The
7*
TABLiv I
r»irf»tn*tlon and Xso the real Sulfation of Llccstonea
Polythfnal calcination

Inflec-




I so the real
sulfation*


tion



Total

First
ftve
BCR
point
-(dw/dt)i.
E, kc&l./

vt. gain.
(dv/dt^,
vt. gain,
Tlse,
Ho.
ti, 'K.
ca./"K.
nole
lOR Ka
Bff.
ea./din
eg.
nln.




Calcltes




1536
1079
0.2t9
51->
955
lb.6b
529
lb. 10
3 1
w
1068
0-578
68.2
1315
25.76
11.1b
22.90
3-2
1)50
IO58
0.388
68.0
13-29
28-35
10.27
25.50
b.O
1355
1056
0.2UO
52.1
10.11
20.86
7.57
I6.5O
3.7
1359
1057
0.251
51.8
9.76
13-78
>¦99
12.10
36
1J6)
1067
0.520
52.0
9.71
2b.63
10.75
20.20
35
1568
1036
0.269
bl.b
7-60
2b. bb
7.93
19.00
53
1369
1087
0.299
>0.2
5-00
16. bl
7.58
13.80
b.O
1577
1076
0 217
73 7
1*1.06
16.21
521
11.00
*•5
1379
1078
0.297
69.O
13.28
16.30
7.96
lb 60
b.l
1677
1031
0.206
55-6
6.1b
12.72
6.17
9-60
53
1679
10t>8
O.35I
57.0
11.10
39.99
16.b7
35-00
3.1
1681
106;
0.366
517
9.67
27-91
10.80
21-30
3-2
166;
1052
o.zBo
b8.b
9.32
25-59
1J.17
22.50
2.8
1687
1029
0.2M)
b7-9
9.21
26.97
11.19
23 50
3.3
1691
1073
0.255
65.5
12.59
33.13
13.67
25.60

1692
1073
0.371)
72-5
15.01
23 79
9-11
17.50
3.*
1695
1057
0.358
61-9
10 Ob
26.bl
16.82
23-50
2.7




Dolomites




1537
1069
0.250
55-1
10 36
18.36
9.22
16.20
3 b
131)0
1060
0.398
70-7
13.85
15.60
8 71
13.60
3.5
13ltl
1060
0-390
55.O
10 U2
10.90
6.12
9.80
•3.2
1367
101)0
0.330
38.5
7.19
21.81
15.51
18.00
2.9
1580
1025
0.320
51 2
10.13
32. b7
9.00
29.20
3-9
1676
I0i>8
0.201)
63-2
12.25
20.88
10.71
lb .90
2.b
1680
loi)6
0-356
67.1
13-33
31.3 r
13.00
27.80
b.2
1686
103I)
0 589
55-7
30 98
3b 62
13.98
28.50
5-1
1688
ioi)5
0.332
75 5
15-02
38.90
15 05
32.20
b«5
1690
101)0
0.265
vr.6
9.10
1970
10.27
17.50
2-9
1695
10G5
0-355
56.5
10.72
12. b8
6.>b
10.00
3.6
1701
1025
0.289
51.0
10.02
3b. 13
lb.07
26.80
3.0
1700
lcM
0-557
53-7
10.28
26.03
9.b3
22.00
b.l




Kixtures




1560
101)9
O.58I
66.1
12.96
38.50
lb. lb
2b. 60
2.9
16911
101)9
0.365
28.3
b.76
12.b5
8.62
8.60
1.7




M&grte sites




1375
910
0.211)
25-5
5.22
13.70
5.B1
10.00
3.7
1376
922
0.327
50.9
11.21
10.22
1.87
5.60
b 8
a For y> minutes at 930* C.; based on 50 teg Initial calcine.

-------
75
heating rate was approximately the same as that for the polythermal
calcination tests, and the atmosphere was the same as for the isothermal
sulfation tests. Typical smoothed curves for the change in weight with
rising tecperature are shown in Figures 21 and 22 for stones BCR 1679,
a calcite; and BCR 1690, a dolomite. The temperature rise stopped at
9to° C. (1213° K.), the upper temperature limit of the TGA equipment;
the reactions were isothermal at 930° C. beyond this point and the
units on the abscissa of Figures 21 and 22 then are indicated by time
in minutes.
The behavior of a stone (BCR 1679) that had sulfated markedly
during calcination is shown in Figure 21, and that of a stone (BCR 1690)
In which little reaction had occurred during calcination is shown in
Figure 22; this effect is indicated by the height of the minimum of the
calcination-sulfation curve above the final calcination curve (yis).
A test tor carbonate at the completion of each run shoved that decatena-
tion had been completed. Two inflection points, II and 12, are
indicated by arrows on the calci nation-sulfation curves; these arc the
points of most rapid weight loss and weight gain, respectively. The
Inflection points are calculated by the computer in applying the
spline-curve fit to the data; the same is true of the minimum in the
curve. The nlnlran and the second inflection 12 frequently-occurred
while the systea was Isothermal at 9UO° C., as in Figure 22.
77
PBLH XI
ft>lrthcrcal Calcination nrrl Sulfation of L'Tftorw

Initial
Inflection
Wo

•0
vt. lo*af
point,"
vt.,

Wo.

TTi

Idv/tfTW1
15J6
U39
U8*
1239
0.57)
15*3
1103
1203
1266
0.176
1390
U09
lie?
1301
O.123
1555
1066
1191
1210
0.109
1399
1150
1190
12)1
0.325
1563
111ft
12Q5
1243
0.195
Utt
U38
1209
12*2
0.500
15^9
uv
12**
1261
0.1*7
13TT
lCfl*
U97
1219
0.221
1579
1190
1200
1230
0.270
U7T
1100
U05
1227
0.068
*8?
1036
iott
1133
O.O65
1681
1150
1212
1251
0.353
Ig*
1050
1079
1151
0.1*6
1657
1U6
U*9
W*
0.199
1091
nab
1210
1233
0.196
1692

U»
U»5
0.395
1695
1157
1200
1251
0*221
33
Vt.	gain, ag/K) as.
Initial old pi	Tin.	alnuf	Betting
farloe	AfUr	bur leg	Afttr	rata,
calci-	cald-	emicl-	calci-	dT/4t,
eatlwj	nation	TDtal	ration	mtlon	TC./oln.
Calclf
9.86	3.35	13.20	22.0	56.0	4.33
6.96	17.39	2k.33	33.1	83.9	i.68
13.18	».9k	18.62	J9.6	70.k	k.Sa
9.87	12.kl	22.27	Jl.8	68.3	4.77
U.jk	ia.sk	27.kfi	l6.k	93.5	k.9i
13.T7	12.83	36.60	22.9	77.1	5.k7
12.32	13.03	27.33	2o.k	ts.6	3.07
17.8k	9.311	27.39	3k.k	73.6	k.63
3.71	13.62	17.33	30.3	*9.7	k.79
12.78	k.k]	17.21	19.}	80.7	3.:8
12.2*	3.3*	13.78	26.3	6J.3	k.50
27.73	11.69	39-"	2k. 3	95.5	».jo
3.33	J3.k2	33.96	20.8	39.2	k.83
23-33	12.k8	)<.al	25.2	74.8	k.80
23.33			...	....
17.78	7.73	23.33	30.1	79.9	k.63
16.9*	17.98	>*.83	22.9	37.1	k.78
13.39	7.71	23.29	13.7	96.2	k.38
17.10	17.»3	3*.33	18.1	91.9	5.1a
Pallw
JJ2 "a "5 °-l<> 23.11	U.93	33 »	29.1	70.9	5.00
Hi? .23 uo7 °-323 12.27	0.7*	13.01	29.2	70.8	k.T7
^5 JZS °-367 l,jS	0.38	13.76	39.6	ao.k	5.33
HS f JS }2£ ?-i?2 a?«	"g	33.69	*9J	60.8	k.78
}J5 1283 0.218 li.82	8.6	23.6k	66.J	6J.3	k.87
f* 1») 1069 122} 0.217 k.62	U.J8	18.21	36.3	33 5	k «k
}£2 J£ 02 £2 J-S *•»	»-2	fi:r	sis
MSB 109 1066 L293 0.2J8 17.J3	9.98	27.31	Jl.i	k6.2	3 52
>2 tfS fS °-a> 6-6a	13.69	22.31	68.9	61.1	3.kj
%£ *5S "2 "JJ 0.171 a.ka	»o.91	23-33	k7.1	52.9	s.kk
JS? tS. M2J "2 °-y; 13-37	s.ka	17.99	78.k	n.6	5.1a
21 15* "8a 0.080 25J2	xo.96	36.17	55.3	kk.7	3.33
1T® 936 1053 1258 0.159 16.61	lfl.93	35.36	37.7	k7.3	3.2k
Wtotar—
Hff ^ #a2? 25.95	10.00	36.0k	k^.6	4.92
169* IOBO 10*7 1396 0.109 0.39	k.60	13.06	5.2	50.1	*.60
flBGSllttS
9* 1091 0.366 17.91	5.97	85.06	37.*	9*.*	*.93
9*3 975 1053 0.7*9 7.(9	U.26	21.29	22.0	100.0	5.00
* Mapvntur* at vblch raw of wt«ht low vm uiira.
to
rat* of v*t«ht loaa, at Tj4 .

-------
76
The Important characteristics of all the stones in the poly-
thermal calcination-sulfation tests axe given In Table XI. The three
temperatures are T0, the temperature at which weight loss was first
detected; Tji, the Inflection temperature at which the rate of weight
loss was greatest; and Tm
-------
79
Calcination end Sulfation Characteristics (cont'd)
Isotherml Sulfation at 9^0° C.
ys	Total weight Gain, rag./jO mg. calcine
y0	Kaxinun rate of weight Eain> i°S-/(5° <°S- calcine)(nin.)
y7	Weight gain during first stage, mg./50 mg. calcine
y0	Time for completion of first stage, min.
Polythermal Calcination-Sulfation
ye Temperature at which weight loss becan, °K.
yl0 Temperature of first inflection point "K.
yu Temperature at which weight was minimum, *K.
yl£ Kaxisua rate of loss of weight, mg./(50 mg. stone)("K.)
yia Weight gain during oalclnatian, mg.7$0 Eg. calcine
y14 Weight gain alter calcination, ng./^O ng. calcine
yxn Total weight gain, B&./yO ng. calcine
yio Tine to completion of calcination, nln.
y1T Tine after completion of calcination, nln.
yle Heating rate, *K./min.
In addition to the x variables of chemical and mlneraloglcal
composition and the y variables of calcination and sulfation parameters,
there vere derived free combinations of these variables a group of
parameters related to the capacity of the stone for sulfation and the
rate of sulfation. These values, presented in Table XII, Include the
weight gain per 50 og. of stone while calcination Is taking place, the
weight gain up to the Inflection point 12 where the gain Is most rapid,
and the total gain In weight of sulfation. It is assumed that calcination
vao completed at the point at which the weight was a nrfnlmmt, although
it probably was cornleted slightly later. The fact that the inflection
point 12 was reached close to the point of minimum weight In many of the
runs indicates that calcination was completed at about this point. It
80
TABLE ril
•Capacity of Llncatonco for SO-, and Rate of Sulfation
Capacity mg./50 mc-	Bole, ma./nln. with 50 r«.
Initial atone			Initial atone	
To	Av. to Max. after Cocbl-
BCR Tnln.f To 12, Total,	Tnln.,ft calcination, natlcn,c
No.
"1	 _wa	 »•>	r,	r,
Calcltco
1336
5.6
5-9
7-5
e 0
£3
13*3
b.o
5-5
lb.O
1350
8.7
8.9
IO.3
0.220
1355
6.6
7-9
lb .9
0.203
1359
8.2
10.1
15-5
0.500
1363
8.8
lb .2
17.0
0.38b
1368
7-5
10.5
16.5
0.368
1369
9-1
9-8
lb.5
0.265
1377
2-7
31
12.6
0.089
1319
7-5
11.9
10.1
0.5#
1677
9.0
10.5
11.6
O.JhO
1679
15-9
18.8
22.6
0.6b9
1681
8.2
10.2
22. b
0.106
1685

lb .3
20.2
0.52b
16»r
8.6
10.1
13.0
0.266
1691
10.l>
lb.6
21 .b
O.bjb
1698
8.9
12.5
13-3
O.65O
O.5S1
1693
9-8
12.8
19-8
7.6
9.b
13-9
0.261
0.209
6.6
6.6
7.0
0.826
0.015
9-9
10.0
10.1
O.25O
0.01b
1J.0
lb .2
18.9
0.265
0.819
7-9
13-9
12.6
0.119
0.107
3-8
6.1
12.6
0.088
0.215
9-3
13-2
12.3
0.135
O.125
9-2
lb .9
lb-s
O.I78
0.28b
5-5
7-9
11.8
0.051
0.168
1.6
7-3
15-b
0.03b
O.568
9-0
11.b
10.b
0.115
0.06b
lb.5
17-3
20.8
0.262
0.160
9-2
19-5
19-7
O.I59
0.301


Mixtures

11.1
12.7
16.7
0.578
0.837
5-8
8.7
8.1
0.136
O.OEL


Ka&neslt«a

9.6
15-7
12.8
0.857
0.061
3-6
2.6
IO.9
0.16b
0.078
0.059	0.2'
0.2a8
0.2ad	0.2K
O.O58	0.215
O.ObB	0.210
0.0T1	0.b6b
0.03b	0.556
0.65b	0.885
0.81b	0-579
O.boS	0.b08
0.1?6	0.41J
O.15I1	0.586
0.8b2	0.786
0.T5B	0.690
O.25T	0.678
0.17?	0.564
O.52I	0.791
0.065	0.6oe
0.1)10	0.761
1557	7.6	9.k	15-9	0.861	0.209	0.5(7
1340	6.6	6.6	7.0	0.226	0.015	0.216
15U	9-9	10.0	10.1	0.250	0.01k	0.216
1367	15-0	ll>.2	I8.9	0.265	0.219	0.399
1380	7-9	15-9	12-6	0.119	0.107	0.176
I678	3-8	6.1	12.6	0.088	0.215	0.216
1680	9-3	13.2	12-5	0.155	0.125	0.199
16B6	9-2	14-9	lb-5	O.I78	0.2ft	0.556
1688 3-5	7-9	11-8 0.051	0.168	0.167
1690	1.6	7.5	15-b	0.03b	0.5C8	0.520
1695	9.0	11.b	10.lt	0.115	0.06b	0.1J0
1701	lb.5	17.5	80.8	0.262	0.180	O.5&5
1702	9-2	19-5	19-7	O.159	0.301	0.366
1360 11.1 12.7 16.7 0.378	0.237	0.501
l69t> 5.2 8.7 8.1 0.156 0.0a	0.170
1375	9.6 15.7 12.8 0.857 0.061	0.291
1376	3-6 2.6 10.9 0.16b 0.078	0.229
Temperature at which weight Is elninun.
c Inflection point at which rate of velght ealn la aaxicum.
*3 ¦ (*1 ~ ra)X
-------
81
is possible that calcination was completed before the ninicu=! weight
was reached, and that the minlmrn weight resulted fran the decomposi-
tion of CoSOa; this possibility is unlikely, however, because many of
the ninima occurred in the isothermal region where the rate of oxidation
of sulfite to sulfate should be either constant or proportional to the
amount of sulfite present.
The last three columns of Table XII are rates of gain In weight
resulting fron sulfation. Column 5» ri, is the average rate of weight
gain, mg. per 50 of stone per minute, during calcination; it is the
quotient of the weight gain at the minimum point of the curve, Wj,
divided by the time required for completion of calcination, ylc„ from
Table XI. Colunn 6, rEl is the slope at the inflection point 12 and is
obtained from the computer printout of the spline-curve fit; it is the
maximum rate of weight gain In nig. per 50 mg. of stone per minute after
calcination was complete. The last column of Table XII, r3, is a
combination of the two rates with a consideration of the temperature,
Tmin °K- (yu)i at which the weight was minimum; the rate r3 is
arbitrarily defined as
r3 ¦= (»"i + r^lO3/^,,	(21)
This expression gives equal weight to the average rate of sulfation
during the calcination period and the maximum rate of sulfation after
calcination is complete; it also given higher values of r3 for stones
that calcined conipletely at lover temperatures in the polythermal tests.
8e
The reason behind this arbitrary use of T^in in equation 21 is that
stones that require higher temperatures than other stones for coznplete
calcination should have higher activation energies for reaction and
consequently be less reactive at any lover temperature; the reciprocal
of the absolute temperature was chosen to express this factor, and 103
was Introduced to keep the magnitude of the rates at about the same
order as those of ri and r^.
The correlations of the chemical and mineralogic&l properties,
x variables, of the limestones with each other are shown in Table XIII,
and the correlations of the calcination and sulfation parameters, y
variables, with each other are shown in Table XIV. These tables are
triangular because the lover left half of the matrix Js identical vlth
the upper right half, that is, the correlation of ^ with xj is identical
with that of xj with xi. The numbers in the tables are the correlation
coefficients of each pair of variables.
The correlation coefficients range from -1 to +1 and measure
the association betveen each pair of variables. For example, the correla-
tion of each variable vith itself is unity as seen from the values on the
diagonals of the matrices. The closer a correlation coefficient approaches
-1 or -*1 the more nearly perfect is the association betveen the two
variables.
A positive value of the correlation coefficient indicates that
as one variable increases the other variable also Increases; a negative
value Indicates that one of the variables of the pair decreases when
-------
IntercorrtUtloa of Olclnatloo and Sulfation Par—tin
0.06
1.00
Jfcu
0.87
a
35 0.2$
00 0.99
JCa— 7n /r JCa.	Jb.
-0.37 -0.)6 -0.58 0.08	2^
0.S9 0.35 0.30 -0.21	.0.32
0.32 0.20	-0.03
0.3* 0.22	-0.10
0.96 -0.(9	-0.20
0.79 -0.3a	-0.23
1.00 0.01	-0.22
1.00	-0.00
1.00
O.Jb
0.^7
1.00
0.19
0.18
o.co
1.00
JUa
0.S9
-o.jo
0.06
-0.01
-0.16
-0.27
-0.21
0.20
0.93
1.00
-0.03
0.08
•0*07
-0.08
0.0)
-0.0)
0.06
0.16
-0.12
-0.06
1.00
0.43
0.06
0.2V
0.23
-0.27
-0.»
*0.23
-O.ll
0-32
0.22
•0.21
1.00
-0.22
0.20
•0.12
•0.10
0.37
0 M
0.37
•0.12
-0.07
-0.13
-0.17
-0.43
1.00
•0.01
•0.12
0.02
-0.01
0-37
0.33
0.32
*0-13
0.19
0.8}
0.00
0.09
-0.27
1.00
•0.19
0.07
-0.03
•0.09
0.61
0.63
0.S7
-0.80
0.10
0.08
-0.14
-0.30
0.60
0.61
1.00
•0 .u
0.86
0.03
0.07
0.81
0.13
0.25
0.18
-0.8?
-0.73
0.62
-0.39
0.01
-0.15
-0.12
1.00
JUl
0.88
•0.16
0.00
•0.03
•0.01
•0.08
0.07
•0.(9
0.61
0.47
-0.36
O.32
0.06
0.08
0.12
•0.6b
1.00
•0.14
0.16
0.21
0.19
0.13
0.16
0.1b
0.0b
-0.31
-0.23
0.43
O.OOL
•0.10
0.17
* 0.06
0.43
-0.47
1.00
9
wu nn
jattwomUttwiot OmXc*). ttomlolcal Pwwtlw
38
H M .Fa.	Jl Jl	^
35 -0.6T -0.6*	0.00 0.89	-0.31
-0.31	.0.09	-o.ok	o.a
-0.12	-0.00 0.26'	-0.9T
0.ST	0JX	-O.J3	-0.19
Q.62 O.U> -0.31
1.00 0.01 -0.20
O.JT
0.09
1.00 -0.0* -0.08
1.00 >0.20
£ia.
0.12
-0.18
o.er
-0.29
-0.35
-o.oe
-0.3T
0.35
•0.16
1.00
-0.96
'-0.43
-0,18
.0 .44
m
0.01
•0.90
0..1
•0.06
1.00
¦o.er
•0.43
-O.U
0.90
O.60
0^
-0.01
•0.16
O.0B
•0.80
o.ae
1.00
-0.10
-0.9?
•0.80
0.33
0.(6
0J2
0.02
•0.32
0.13
•o.oo
0.70
o.er
1.00
•0.48
-0.1?
-0.13
-0.04
0.05
-0.03
-o.oe
•0.16
0.16
0.18
0.32
0.33
•0.10
1.00
•o.oh
0.15
•0.16
•0.09
.0.03
-0.07
-0.10
-0.06
0.19
-o.or
0.00
-0.09
0.05
-0.0?
1.00
-0.05
-o.or
o.ij
•0.10
-0.15
-o.u
-0.13
0.38
-0.11
0.28
-o.cc
•O.OT
-0.16
-0.03
-0.31 -o.u
-0.9 -0.30
0.03	o.ei
0.Q9	-o.Qi
0.23
0.37
0.03
O.OT
0.47	-0.06
•0.32	-O.U
-oa?	-0^7
-0.81	0.10
0.42
0.19
0.19
.0.49	-0.49
-0.18	-0.18
-0.19	-0.11
0.54	-0.09
0.S6	0.08
0.88	-0.03
-0.01	*0.09
0.1T
0.1T
0.18
0.24 -0.09
-0.09 -0.08
•0.13 -0.08
1.00 -0.0k
1.00
-0.21
0.18
0.10
0.52
0.30
0.52
-0.05
•0.21
O.U
o.ea
0^
0.31
-o.u
0.98
•0.07 -0.06
-0.00 -0.06
0.38 0.81
-0.08 O.U
1.00 -0.06
1.00
9
-------
85
the other increr ses. For example, the correlation between the ignition
loss, xx> and tha erarunt of accessory mineral in the lirestone, *X1,
is -0.96; as the ezount of impurities in era so s in the limestone the
ignition loss decreases because of the lover carbonate content. This
correlation is not perfect (-1) because some of the impurities may lose
weight on ignition, and because of different ratios Ca:I>tg In the stones.
The correlation between the contents of montmoriXlonite, x14, and the
fritted glass, x20, was O.98; these impurities occurred together in all
the stones In the test group.
Significance at the 0.01 (99^) confidence level was used in
all comparisons; Rqi = 0.^5. The significant correlations in Table XIII
are: xxxs, xl%e, xjx^, Xix12, xxxx3, x^x3> x^xq, x3xe, x4x€; x4x17, x4xi
*5*6; *5*11# *5*12# *5*13i *G*IP> *Cxlli *CX12» *0*1.3> *C*1S» *11*12>
*11*13> *n*i*j xii*i8* *ia*i3i *i3*io> and xi+x^o* Most of these
correlations are obvious from knowledge of the properties of the line-
atones and of the effect of one property upon another; other correlations
are less obvious and nay be characteristic only of the group of lime-
stones tested.
The significant correlations among the calcination and
sulfation characteristics, Table XJV, are: yiyg, yiy10, y3y<, y5ys,
ysy7» ysyis. yoyr.- y?yi5, yvyis, ysvio, yeyie, ynyie, yi3yis» ynyis,
anl yieyiT-
86
Among the calcination parameters, the association of the
activation energy, y3, with the frequency factor, y4l was 0.99 as was
seen in Figure 1; the only other significant correlations of a
calcination parameter were those of the inflection temperature, yA,
with yo and yIO, the temperatures of first weight loss and of the
first inflection point H in the polytherinal calcination-sulfation
tests.
The isothermal sulfation characteristics, y5, yG, and y,t all
correlated significantly among themselves; each of these values also
correlated at a value of about 0.6 with the corresponding value in the
polythennal calcination-sulfation test, yis«
Among the polytherraal calcination-sulfation tests, the time
to complete the calcination of the stone, y1G, was related inversely to
the temperature of first v/eight loss, y0, the first inflection temperature,
yJO, o-nd the time after calcination to cessation of reaction, yi7; it
was directly associated with the temperature at which calcination was
complete, y^. The total weight go-in, y\s» was correlated at about 0.6
with both the weight gain during calcination, yi3, and the weight gain
after calcination, yi4, as one would expect, since yi5 = yi3 ^ yi4; the
differences among stones in absorbing SO^ during calcination accounts
for the lack of perfect correlation.
The correlation copf'Ti ^	each y variable and each
x variable are shown in Table XV, those between each derived capacity,
w, and rate, r, property and each x variable are shown in Table XVI, and
those between each derived property end each y variable are shown in
Table XVII.
-------
o
O
b
O
O
o
4
o
*
3
S
M
8
S
ro
s
6
6
b
4
o
o
4
o
k
io
3
is

s

c
•©
i
o
o
o
b
o
4
o
h
w
3
5

o
v
>6
©
o
o
o
b
o
4
O
a
a
a
c
is
s

~4
©
p
o
&
&
o
4
4
<«
&
8
*8
is
s
S
o
o
t
o
o
o
o
4
4
&
a
s
~-
iu
V
8
8
&
o
o
&
4
o
4
4
8
¥
8
3
V
is
k
Va
IP
o
o
o
4
4
p
4
&
h
s
3
'*
?
•S
8
o
©
&
4
&
o
o
O
8
3
is
k
S
8
b

O
o
b
b
b
o
o
O
o
p
>4
is
s
cs
»-
IT
is
b
&
O
b
o
4
b
4
B
s
5
s

9
8

4
o
b
4
o
o
4
o

e
S>
V*
a
B
it
8
b
&
&
~
4
4
b
4
9
&
P
k
s
5
8
S
b
o
4
i
4
IP
b
o
M
8
M
W»
B
~-
»
8
is
4
4
&
o
b
o
4
o
io
v*
k
8
k
s
*
e
3
b
b
o
o
o
4
4
e
o
8
a
p

H
SO
o
k
b
&
b
b
&
4
4
4
&
P
o
»
r
k
8
i3
H
b
o
o
b
o
4
4
o
S
8
i-
-4

H
8
a
fe
.° .° p
a o 9
IK
c «
o o
^ 3
o o
S 5.
o o
8 9
4 &
k a
&
is
o o o
P s s
o o
8 «
O O l<
5 i If
4 b
3 S
p «
o
§ &
& o
k g
4 4 4
14 4
lif £
4	4
3	8
o	4	o
3	P	»
4	&	o
5	»	3
o	bo
3	3	8
6
3 £
& 4
a k
P b
C )?
s » e is
4 o o o
#855
o o &
g i: «
o fa o
5 fc 3
8 8 3
•° P k
£ e r
i£\t
s S |j
fo s* r
P B 8 B P
o o
s k
o 4
8 &
4 o o
3 9 8
p a c a s
o b
is »
b 4 o
P fc 3
o	o
s	s
o	o
s	s
l»
S 8
O &
8 8 3

8 5
b	14
k	lb
S	I*
0	M
1	o
b	s
t	&
B	P
4 o o
B '8 3
8 & 8
CSS
bob
a a c
4 o |p
abb
o 4 14
'8 3 I s
o o lo
e a te
ii 11>
$tt
i p |j
P s F
o o k
3 G F
set*
O O ^4
ft 8 F
0	O h
a h F
II 8 £
IS S|f
1	t k
v* -» r
1$
TABLE XVI
Corrojatjon or CupuciUcs und Rates of Reaction With
Chemical and Mtncrnlcclcnl l*ropc-rtlea of Llmcstonca
88

vt

va

0.J9
o.jO
• 0 01
*2
o.J>6
0.10
0.P2
>3
-O.C»
0.10
-0.26
*4
0.07
-0.03
0.16,
*3
-0.10
-0.0?
0.19
*C
-0.M
-O.hO
-0.03
*7
0.20
0.11
0.19
*e>
•0.06
-0.07
-O.hk
*u
0.0^
-0.11
0.20
*io
-0.25
-0.38
-0.6^
*11
-0.37
-0 3"»
-0.02
*ar
•O.fcY
-0.37
-0.0>
*\o
-0.J9
-0.30
oar
*14
0.0?
-0.0'»
-0.13
*1»
-0.P1
0.01
0.37
Ijo
-0.15
-0.16
-o.*»o
*»T
0.2J»
0.07
0.22
*io
-0.03
-0.01
-0.18

-0.25
-0.28
-0.06
*20
0.06
-0.01
-0.1k
TABLE XVII
Correlation of Cnr?cltlcn and Rntco of Reaction With
CalclnntJon and Sulfation Pn.r reactor 8 of Limestones

•
vg
va

-0.27
-0.35
-0.20
9p
0.19
0.2?
-0.0>
*3
-0.1 r
-0.15
-0.)l
94
-0.15
-0.12
-0.1>
yr.
0.33
O.SP
Q.f»S
7c
o.fci
JteL
0.S8
Tr
0.3*1
0.^3
0.X>

-0.07
-o.os
-tTTB
7o
'0.12
-0.P2
0.11
yio
-0.17
-0.?6
0.12
y»i
-0.23
-0.03
-0.21
yis
-0.*5
-0.33
-0.2C
yi3
Q.oJi

O.bO
yi4
-0.30
0.05
0.72
y>»
0.S2
0.67
0.92
yxo
0.00
0.17
-0.17
yiT
0.05
-0.07
0.09
yio
-0.11
0.J6
0.06
rg
ra
o.ofl
0.)'.
-0.11
-O.O0
-0.05
-0.02
-0.10
-0.06
0.30
0.31
0.2B
0.3h
0.2)
0.?6
-0.11
-0.16
0.27
o.po
0.30
oTCy
-0.02
-O.bh
0.10
0.12
-0.16
0.30
0.79
0.^9
0.52
0.66
-0.22
-0.62
-0.02
o.*»i
0.0^
-0.22
-------
89
The x-y correlations show that the ignition loss, xlt of a
linectone is correlated positively with the	calcination rate,
ys, and the weight gain during calcination, y13.
The CaO content of the stone, x2t is correlated positively with
the temperature of first weight loss, ys, with the temperature of the
inflection point H, yi0> and with the time after calcination is
complete to cessation of sulfation x2 is correlated negatively
with the time to complete calcination, yiQ. The magnesium content,
x3, has the opposite effect and is correlated negatively with y©, y10,
find yn, hut correlated positively with y16. This means that the
calcites start to lose weight at higher temperatures than the dolomites,
but are completely calcined sooner. The Ka20 content, vas cuirelated
negatively with the maximum rate of calcination, ys.
The lattice constant a, y9, was correlated positively with
the inflection temperature at H, yio> and with the reaction tijae
after calcination was completed, y17, but vas correlated negatively with
yio, the time during calcination. The crystallite sifce, xl0> varied
inversely with the total polythermal weight increase, yls, whereas
increasing the quartz content, x12, adversely affected the weight gain
during calcination, yi3.
Increase in the accessory mineral content, x^, decreased the
maximum rate of calcination, y2, whereas the weight gain after calcination,
y14, was proportional to the chert content, x15.
90
The derived capacity and rate propcrV us were correlated
significantly with a few x variables, Table XVI, but there were several
significant relations of the y variables with these properties, Table
XVII. Both Wj. and we, the capacities to the end of calcination and to
the inflection point 12, were not related significantly to any chezxical
or mineralogical property; w£, the total polythermal capacity was
related inversely to the crystallite size within the carbon&te grains,
xi0*
The average rate of SO? pickup during calcination, rx, vas
correlated positively with the CaO content, xs, and with the unit-cell
constant, xg; it was related inversely to the KgO content, X3. These
correlations Snply a higher initial rate of reaction with SO? for
calcitcs than for dolomites. This is illustrated in Table XVIII in which
the values of the average rate, Ti, are arranged from lowest to highest
for 33 stones (omitting the magnesitos), and'the properties of the
slowly reacting stones (upper 16) were compared with those of the
faster reactors (lower IT). The slo«Oy reacting stones included 10 of
the 13 dolomites, 5 o* the 1$ calcites, and 1 of the 2 mixtures. There
was a significant difference between the average unit-cell length, x?,
of the slowly reacting stcnos end thrt of the better performers-
-------
91
TABLE XVIII
Correlation of Initial Reaction Rate with Chemical
Properties and Texture of Mgestonea
(Initial rate, ij, la the average rate of weight
lose up to the point of alnltcum velght)
	 Initial
BCH rate| n,
Ho. mg./nln.
1690
1688
1678
1377
1681
131.3
1693
1380
16P0
1694
i7oe
1686
1J33
1350
1340
13<>1
0.034
O.OJl
0.088
O.089
0.106
0.114
0.113
0.119
0.133
0.1)6
0.159
0.178
o.aoe
0.220
0.826
0.250
Weight ration
Khaablc-
carbonatea
COajOO, to CaO. FejOj, Groin alte,
*U*r— (100-Xn i/xs. I,	codetl.
Uhlt-cell Crystal -
length, lite alic,
a' **»
Average 0.1J9
Std. da*. 0.016
1336
1337
1701
13«9
1367
1687
1677
1}68
1360
1363
1379
1691
1359
I6B3
1693
1679
1692
0.853
0.261
0 .262
0.263
0.263
0.266
0.340
0.363
0.378
0.384
0.389
0. (15b
0.300
0.52li
0.541
O.6I19
0.650
Average 0.398
Std. dev. 0.032
1.318
1.619
1.563
0.961
0.81(0
0.786
1.460
1.536
1.304
1.022
1.343
1.579
0.798
0.797
1.303
1.567
1.973
0.084
0.809
1.623
1.232
0.799
1.331
0.809
0.849
0.817
0.989
0.853
0.848
0.816
0.797
0.911
0.801
0.820
0.783
0.936
0.058
Slow-reacting etonee
3.384
3.419
3 299
1-979
1.920
1.835
3.291*
3.270
3.220
2.349
2.873
3.333
1.817
1.781
3.192
3.332
2.769
0.170
1.843
3.414
2.649
1.827
2.862
1.845
1.763
1.857
2.147
1.943
1.918
1.852
1.812
2.012
1.824
1.850
1.814
2.072
0.112
1.1
3.0
0.5
1.3
1.3
1.5
5.1
1.0
0.6
1.0
o.t
1.0
0.6
3.0
0O
2.3
0.9
1.0
0.6
3.0
0.6
1.3
0.2
3.0
1.6
1.0
0.2
2.0
l.k
1.5
0.5
3.0
1.0
1.9
0.3
0.2
ig stonea

O.S
3.0
0.2
2.0
*.3
1.3
0.1*
1.0
3.2
1.0
0.1
2.6
0.Y
1.0
0.9
2.6
0.7
2.5
1.6
1.0
O.k
3.0
0.6
1.0
0.1
1.0
O.k
2.8
0.2
1.0
0.3
1.0
0.1
30
0.9
1.8
O.k
0.2
4.809
4.807
4.808
*••975
<>.983
">•989
4.808
4.805
4.806
4.886
4.810
I1.806
4.932
4.986
4.806
4.6o6
4.667
0.021
4.978
4.811
4.622
k.987
4.616
4.966
<1.967
4-978.
4.910-
4.969
4.960
4.984
•¦.965
4.986
4.963
*>•973
4.983
fc.950
0.016
O.O98
0.182
0.141
0.162
0.113
0.167
O.I89
O.I89
0.171
0.159
0.070
0.189
0.096
0.177
0.141
0.211
0.154
0.010
0.149
0.119
0.090
0.127
O.O63
0.177
0.177
O.097
0.135
0.124
0.097
0.107
0.124
0.145
0.130
0.068
o.iw
0.124
0.007
92
A similar difference was noted for the ratios jxr and
(100 - *n)/x2, «n of which agree with the dolomite-caleite distribution.
The relationships between xs, x3, and x11 can be determined by
considering each limestone to consist only of CaC03, MgC03, and accessory
mineral, and the ignition loss to be due to C0a alone. Thece considera-
tions lead to the equations
*i + *2 + *3 + *11 = 100	(22)
*1 = 0.78VTx2 + 1.0911* x3	(23)
*u «= 100 - 1.781+7x2 - 2.091UX;,	(21*)
from which is derived
100 : x" -1 + a ~ §i	(25)
*2	*2 *S
so that any two of the ratios In equation 25 fixes the third ratio. In
the multiple regression that follows, only two of the ratios are used to
avoid redundance of major composition variables.
The stones that had the highest reaction rates, rr, also had
smaller average crystallite sizes, Xi0, than those of the slouer reactors.
There were no significant differences, however, between the Fe20s contents,
x4, or the microscopic grain sizes, x^, of the two groups of stones.
The maximum rote of absorption of SOe after calcination was
complete, rs, was correlated positively with the chert content of the
stone, x15, (Trble XVI), but the combined rate expression, r3, was
correlated positively with both the CaO content, Xp, and the unit-cell
-------
93
length, xg, and negatively with the ignition loss, xi. A summary
comparison of the 16 stones with the lowest values of r3 and the 17
stones with highest values of r3 is shown in Table XIX. Only 3 of the
13 dolomites were in the better performing group, and only 5 of the 18
cplcites were In the poorer performing group. This is reflected in the
average values of *x/*2 and X9 for the two groups. The effect of the
impurities is shown by the average values of (100 - Xu)/x-» for the two
groups. The x-ray crystallite size, x1D, was smaller for the better
performing group, but the FegOa content, x4, and the microscopic grain
size, xB, bore no relation to the rate of reaction, r3.
The many correlations in Table XVII between the capacity and
rate expressions and the y variables can be explained from the relations
Vi
= yis'UOO-XjJ/lOO
(26)
V3
= yis*(loo-Xi)/ioo
(27)
ri
= *i/yie
(28)
r3
= (ri + r^.ios/yn
(29)
and the correlations discussed previously. The characteristics and
r2, the capacity and rate at the inflection point 12, were obtained from
the computer printout and have no inherent correlations by derivation.
The weight gain after calcination, w2, was correlated significantly with
ys, y6, and y7 from the isothermal tests and with y13 and y15 from the
polythermal calcination-sulfation tests. The only significant correlation
of r2 among the y variables was a ueak proportionality with the total
capacity, yis, of the polythermal calcination-sulfation tests.
9U
ramj? nx
Correlation of C^Mpatlon Reaction Rotea vltfa
Chenlcal Properties and Textures of Lice a tone a
Catblna-
bCR 11 on rate, r3f8
Bo.	eg./aln.
1695	O.IJO
1683	0.167
169b	0.170
1J80	0.176
1680	0.199
1353	0.210
1JS0	0.21J
1J40	0.216
1341	0.218
1678	0.346
1JJ6	0.253
13l>3	0.286
1690	0.}20
1363	0.336
1701	0.345
1686	0.356
Average 0.240
8td. «n. 0.017
Velj&it ratios
Kb coble
carbonates
CO^CaO, to CaO.
(lQO-»l¦)/*2
Slov-reac
l.l>6o	3-294
1.619	3-419
1.022	2.349
1.536	3.270
I.5C1	3.220
0.798	1.817
0.797	1.781
1.505	3.192
1.567	3.332
1.563	3.299
0.809	1.8113
0.786	1.835
1.518	3.384
0.855	1-943
1.232	2.61)9
1.579	3.333
1.259 2.747
0.087 0.172
FCgOg, Grain size,
coaed. xa
stones
0.6	3.0
0.5	1.5
0.6	).0
0.3	2.5
0.9	1.0
1.6	1.0
0.2	2.0
1.h	1.5
0.3	>.0
1.3	1.3
0.2	3.0
0.4	1.0
1.1	3.0
1.6	1.0
U.3	1.5
0.2	3.0
1.0	2.0
0.3	0.2
Unit-cell Crystal-
length, lite alee,
»¦ ""	»¦ »¦"	
h.808	0.189
4.807	0.182
11.686	0.159
>1.805	0.189
4.806	0.171
4.982	0.096
4.966	0.177
4.808	0.141
4.806	0.211
4.808	0.141
4.978	0.149
4.989	U.167
4.609	0.098
4.989	0.124
4.822	0.090
4.806	O.189
4.868	0.155
0.021	0.009
1687	0.364
1702	0.366
1337	0.367
1369	0.379
1677	0.386
1367	0.399
1377	0.408
1379	0.413
1359	0.464
1360	0.501
1692	0.602
1685	0.678
1681	0.690
1693	0.7ol
1679	0.786
1691	0.791
1368	0.823
Average 0.540
8td. dev 0.0-'»}
Fast-reac
0.809	1.845
1.343	2.873
1.628	3-414
0.799	1.827
0.849	1.763
1.351	2.862
0.961	1.979
0.848	1.918
0.797	1-812
0.989	2.147
0.733	1.814
0.911	2.012
0.840	1.920
0.801	1.824
0.820	1.850
0.816	1.852
0.817	1.857
0.951	2.092
0.060	0.116
atonea
0.1	2.6
0.8	1.3
0.2	2.0
0.1	1.0
0.7	1.0
5.2	1.0
5.1	1.0
0.4	3.0
0.1	1.0
0.7	2.5
0.1	3.0
0.4	2.8
0.6	1.0
0.2	1.0
0.3	1.0
0.6	1.0
0.9	2-6
1.0	1.7
0.4	0.2
4.986	0.177
4.810	0.070
4.811	0.119
4.987	0.127
4.987	0.177
4.616	0.083
4.975	0.162
4.980	0.097
4.985	0.124
4.910	0.135
4.985	0.153
4.936	0.145
4.985	0.115
4.985	0.130
4.975	0.068
4.984	0.107
4.978	0.097
4.949	0.123
0.016	0.008
¦ r3 - 
-------
95
The }• variable, y10, which was included in the correlations to
determine whether there was a bias for any particular property of the
stone that was due to the heating rate, is a random variable with a mean
of It.95° ± 0.31' C./min. There was no correlation of this rate with
any calcination or sulfation characteristic, which indicates that slight
variations in the heating rates had no significant effect on the
properties neasured. There was a weak positive correlation, O.5O, with
the magnesium content of the stone, xr,, anci a similar weak negative
correlation, -O.5I, with the lattice constant, x0; since x3 and x0 are
highly correlated, -0-97, the correlations with yjB amount to a single
event. It is to be expected statistically that there will be, on the
average, one errcneous correlation out of 100 tests of correlation vhen
the confidence level is set at 0.01 (or 995). It is concluded that the
variation in the heating rate had no effect on the measurements.
Multiple Correlations
The si-^lc correlations discussed above indicate that no single
property of a limestone is sufficient to predict its capacity for, or rate
of absorption of, SO2. An effort was made to find the significant factors,
or combination of factors, that by linear recression might be useful in
predicting the effectiveness of a limestone for removal of S02 from the
96
stack gas of a power plant. The method used to identify the significant
factors in a group of 10 or less is shown in Appendix D. Two models
were used
« ¦= X Vl + b°	(3°)
and
Iog Q =	+ b0	(31)
vhere Q is or rj from Table XII and the ten z factors are defined in
terms of the x factors of Tables VIII and IX.
*1

*l/*2
ratio, Ignition loss:CaO
22

(100 - x31)/*a
ratio, rhombic carbonates:CaO
Z3

*4
Fep03 content, %
z«
-
*e
microscopic grain size, coded

a
X£.
unit-cell length, a, k.

B
Xio
x-ray crystallite size, ^

e
log X3
log $ NasO
ZU
=
log (*5 + *a H x7)
log $ (ttap0 + K^O + HnOp)
zo
Q
*11-(*12 + X,s + xso)
sum of illitc, montnorillonitc

s
*13 + *11 + *16 +
tremoltte, llmonite,


*1T + *10 + Xi9
feldspar, and rnuscovdtc
Zxo
a
*lv/*4
fraction of uncombined Fe203a
& In some of the stones the value of z1Q vas greater than unity because
of the rather Inexact determination of x17; for these stones the value
of zJO was taken as 1.0 in the regression.
-------
97
The coefficients of the significant factors for the linear
(equation 50) and the logarithmic (equation 31) models are listed in
Table XX for each of the capacity, v, and rate, r, characteristics. In
each regression, the two magnesites, BCR 1375 and 1376, were emitted from
consideration, so that the equations apply only to calcites, dolomites,
and their mixtures.
Of the ten z factors considered in the multiple regressions,
four to eight were eliminated as not contributing significantly to a
particular characteristic. Six factors vere significant in the regressions
for the capacity measures Vi and wd, and five factors vere significant for
the combined rate expression r3. The rate ts was poorly described by all
10 factors, and only 2 factors vere significant in predicting this property.
The unit-cell length, zS} was a significant factor in each linear model
and in most of the logarithmic models; increasing the unit-cell length,
which is equivalent to selecting limestones that are purer calcites,
caused most of the properties to increase (an exception is v2). The
clay mineral combination represented by factor affected adversely
each property except r2. The NasO content in its logarithmic form,
z7, was a significant factor In all capacity measures, and many of the
z factors had different effects on the linear and logarithmic models.
A preliminary study of principal component analyses for sane
of the y variables in relation to the x variables is given in Appendix
E, from vhich it is concluded that multivariate analysis of the present
data has little promise of producing a useful relation for evaluating
stones for the injection process.
98
Coefficients of Significant Factors in Multiple
Correlation Eouations
(Q
rj)
J*2_
Q = ^bjij + bQ
_£a_
(10)
*1
-
5k.k6
-
_
-
0.6592

-
-2k.89
-
-0.1259

-
b3
1.163
-
-
-
-
-
t>4
-
-
-1.571
-
-
-
*»S
2.278
-J0.k2
5.685
0.1179
O.O6258
-
be
-
-
-52.1k
-
-
-2.k05
bv
5 973
5-576
3.605
-
-
-
bB
-3.289
-
-
-
0.13U3
-
be
-O.3U89
-O.U755
-0 ^280
-0.01191
-
-0.009091
*>10
3 757

2.818
0.1651
-
0.1818
b0
-
589.51
-
-
-
-17.897
S.D.a
2.657
2.835
5.001
0.119
O.16I1
0.139
Rb
0.701
0.757
0.769
0-735
O.U15
0.776
bi
3-066
-
b2
-2.122
-0.5157

-
-
*5
-k-555
-2.8k5
be
-
-
by
0.259k
0.21H*
bB
-O.I838
-
bs
-0.01818
-0.01968
bio
-
-
bo
25.27k
16.1U9
S.D.
a 53
33
pb
0.697
n < toJ.
log Q	b0
0.05llk -0.2681
-0.06688
0.29U2
O.lSjl*
-O.OlUOj
0.1200
25
A Tfl
-0.02129
0.2755
66
'¦"TjO
-5.026
0.2787
lko
(11)
O.7816
k.218
-2.86k
-0.009271
0.2122
-21.65k
-
O.Ui-U
a Standard deviation.
b R = multiple correlation coefficient for equation containing only the
significant z^'s.
-------
99
Application to Stones Considered for Use In Shawnee Teats
Battelle Memorial Institute (6) used the dispersed-phase
reactor to study five stones that are to be used In the injection teats
at TVA's Shawnee steam plant and determined apparent reaction rates of
each stone at five temperatures. These stones were characterized
mineraloglcally and chemically to obtain the ten z factors that are
needed in equations 30 and for calculation from the coefficients
In Table XX of the six capacity and rate characteristics defined In
Table XII. The results are given In Table XXI.
The first row of reaction parameters for each stone listed
In Table XXI was calculated by the linear model 30, the second rcw by
the logarithmic model 31» and the third row Is the arithmetic average
of the two results. Taking the average of the two results decreased the
prediction error by a factor of Jz. There are, however, at least two
glaring discrepancies In the predicted properties: the weight gain
during calcination, Vj, for stone 2062 is predicted to be greater than
that at the inflection point 12, w2j and the value of w2 for stone 2069
Is predicted to be higher than the total capacity, w3. These predictions
are not in agreement with the shape of the calcination-sulfation curves,
and result from prediction errors for the different properties as listed
In Table XX.
The predictions of the reactivity parameters of Table XXI for
the various stones vere compared with the smoothed apparent reaction rates
reported by Battelle (Table F-3 of reference 6). The rates were smoothed
by deriving by least squares the equation of the linear relation between
100
TABLE XXI
Predicted Beactivity Parameters of Limestones
to be Tested at Shavnee
Calcd.
BCR by eq. 	Predicted parameters
Ho.
Ho.
*1
va
w»
n
r 0
r.i
2060
10
11
Av.
7-399
6.930
7.16U
10.905
10.188
10.547
17.072
15.176
16.12 If
0.298
0.250
0.27U
0.229
0.22U
0.227
0.566
0.550
0.558
2061
10
11
Av.
7.00U
6.802
6.903
7.326
7.773
7.5>»9
lit. 709
11.057
12.883
0.352
0.315
0.331*
0.127
0.122
0.12U.
0.572
0.552
O.562
2062
10
11
Av.
10.638
6.8U9
8.7U3
7.280
7.672
7.V76
11.758
15.539
13.6it8
0.511
0.579
0.5U5
0.126
0.030
0.078
0.268
0.238
0.253
206U
10.
11
Av.
6.9U5
6.648
6.797
7. Slit
7-951*
7.63I1
13.881
11.072
12.U77
0.35"»
0.318
0.336
0.128
0.101
0.115
3,3 £
-9
Odd
2069
10
11
Av.
6.870
6.602
6.736
13.161
lli. 191
13.676
11.055
13.219
12.137
0.160
0.137
0.IU9
0.229
0.091
0.160
0.03)
0.13i»
0.085
-------
101
the logarithm of the apparent rate and the reciprocal of the absolute
temperature as illustrated in Figures F-L through F-5 of reference 6,
and recalculating the apparent rates at the temperatures reported. (The
apparent rate at 1910° F. for stone 2060 was omitted in the regression.)
The comparison of the predicted parameters from the chemical and
mineralogical properties vlth the smoothed reaction rates at five
temperatures is shown in Table XXII In terms of the correlation coefficients-
Since there were five stones in the comparisons, each significance test is
based on only three degrees of freedom.
As shown in Table XXII, there is a significant correlation
between the total weight of SO3 absorbed, w3, in the polythermal
calcination-sulfation test and the apparent reaction rates at 1603° and
1681° F. No other calculated parameter is correlated significantly with
any of the isothermal reaction rates of Battelle, and the association of
the values of v3 with the apparent reaction rates decreases as the
temperature at which they were measured increases. Hie relationships
between the apparent reaction rates and two of the calculated parameters,
w3 and r2, are shown in Figure 2}. The correlation with w3 of the rates
at the two lowest temperatures Is rather good, but the rates at higher
temperatures have no consistent association with the predicted values
of w3. The calculated rate, ra, which Is the maximum rate of weight
gain after calcination Is complete, is slightly associated with the
apparent reaction rates of Battelle, although the relation is not signifi-
cant for any of the temperatures. The facts that r2 is predicted by only
two properties of the stone, and its prediction error (Table XX) is quite
102
TABi-g XXII
Correlation with TVA Results of Reaction
Parameters Calculated frttq Battelle Data
(Battelle data were snoothed by regression; TVA
results are averace values from Table III-XTV)
Correlation coefficient0 at


indicated
temperature,
*F. (*C.
)
Reaction
160}
1681
1807
1910
2000
parameter
(875)

(986)
(IOI13)
(1093)
Vl
0.29
0.13
-O.36
-0.'i2
-0.51

0.10
0.15
0.26
0.2U
0.21
w3
0.97
0.91
0.3U
0.18
-0.08
ri
0.10
-0.02
-O.36
-0.38
-O.I4I

0.63
O.69
0.55
0.1i5
0.28
r3
0.32
0.33
0.18
0.12
O.OU
a Ros = 0.83,
R01 =
O.96, R001
= 0.99.


-------
105
(i
g a 240
*§
OS
5^20°
2 co
160
i§
120
2
eo
m 1 »

—tO i	
0 0 _

O 0
0
0
~ °
TEMPERATURE
0
0 0 0
•Oh A -
•F. *C.
t 0 A
A 0
O- 2000 - 1093

A 0
&
0- 1910 - 1043
° °a
&
A - 1807 - 966
6

• - 1601 - 919


¦ - 1603 - 873
•





¦ _


1% D 1
Wj,UG./00 mo.
at a2 0.3
r2,M0 A*0MGJ(MIN )
Figure 25. Comparison of Calculated Reaction Parameters
With Battelle*s Apparent Reaction Rates.
Figure 2U. Relation Between Reduced Time and Degree of
Calcination of Minus 2k- Plus 28-Mesh Colbert Limestone.
10k
large, would indicate that the slight association of Tp with the reaction
rates for these five stones probably is a fortuitous rather than a
dependable relation; it is also possible that a better expression for
r2 may be found with different combinations of chemical and mineralocical
properties of the,stones.
It should be noted that the correlations of v3 with the
Battelle rates vere significant only in the temperature range in which
w3 was measured for determining the regression coefficients. The
calcination-absorption tests were polythermal only up to 9W# C. (1724° F.),
and the temperature was never higher than this because of equipment
limitations. The rates that were measured at higher temperatures by
Battelle—at 1807°, 1910°, .and 2000° F.; 986°, 10ky, and 1093* C.—
apparently are not associated with the same chemical and mineral ogical
properties that determine the rates of reaction at lower temperatures.
Two possible explanations of failure to extrapolate the low temperature
results to those at higher temperature are the increased fluxing of
some of the Impurities and the increase of the crystallite size of
CaO as the temperature rises; both of these factors are minimized below
1750® F., and their effects on reactivity were not available to be related
to the chemical and mineralogical composition.
The relative importance of the properties of limestones as
reflected in their reactivity with sulfur oxides is discussed in Appendix F.
-------
105
REFSlSiCES
1.	Azaroff, ii.-V., "X-Ray Diffraction Methods," s6th Borelco X-Rav
Analytical S;' ~ol. p. 3, Philips Electronics Instruments, Mt. Vernon,
N. Y., 196T.
2.	Bituminous Coal Research, Inc., Progress Report, "Research on
Methods for Control of Sulfur Dioxide Bn1salons From Coal-Burning
Power Boilers'- (BCR Report L-273) March 7, 1968.
3.	Borgvardt, R. H., "Kinetics of Reaction of S02 With Calcined
Limestone," U. S. Public Health Service Symposium on Limeatone-
Sulfur Dioxide Reaction Kinetics and Mechanisms. NAPCA, Cincinnati,
Ohio (1969). (Abstract)
U. Borgvardt, R~. H., Bivironnental Science and Technology. (1), 59 (1970).
5.	Coutant, R. H., Barrett, R. E., Simon, R., and Lougher, E. H.,
"Investigation of the Kinetics of Reaction of a Limestone With S02
in Flue Gas." Paper presented at the Dry Limestone Injection Process
Symposium, Gilbertsville, Ky., June 1970.
6.	Coutant, R- W., Campbell, B., Barrett, R. E., and Lougher, E. H.
"Summary Report on Investigation of the Reactivity of Limestone and
Dolomite for Capturing SO2 Fran Flue Gas." NAPCA Contract No.
PH-86-67-115. June 27, 1969.
7.	Fischer, H. C., J. Am. Ceram. 80c.. 58 (7), 2"»5-51 (1955)-
8.	Fischer, H. C., J. An. Ceram. Soc.¦ 58 (8), 28U-8 (1955)-
9.	Goldsmith. J. R.. and Graf, D. I.. Am. Mineralogist. fc3. 8^-101
(1958).
10.	Harvey, R. D., "The Mineralogy and Petrography of Carbonate Rocks
Related to Sulfur Dioxide Reactivity." Presented at the 3rd Limestone
Symposium, St. Petersburg, Florida. December l»-8, 1967.
11.	Hastings, C., "Approximations for Digital Computers," p. 189, Sheet
65, Princeton Univ. Press, N. J. (1955) -
12.	Hatfield, J. D., and Kim, 1. K., "Limestone Calcination and Sulfation
Investigations," U. S. Public Health Service Symposium on Limestone-
Sulfur Dioxide Reaction Kinetics and Mechanisms, NAPCA, Cincinnati,
Ohio (1969J• (Abstract)
l}. Hedin, R., "Investigation of the Lime-Burning Processes," Svenska
Forsknlngglnst¦ Cenent Betong Vld Kgl. Tek. Hogskol. Stockholm.
Hand!.. 32. Stockholm. Sweden. 123 PP. (1961)
106
11». Heek, K. H. van, and Jlintgen, H., "A Contribution Concerning
Kinetics and Mechanism of Calcination and Reaction of Carbonates
With S02 Under Hon-Isothermal Conditions." Paper presented at the
Limestone Symposium, St. Petersburg, Florida, Dec. 1967.
15.	Heek, K. H. van, and Jiintgen, H., "Kinetics of the Reactions of
Calcium Compounds With S02." Paper presented at the Dry Limestone
Injection Process Symposium, Gilbertsville, Ky., June 1970.
16.	Klaus, R. L., and Van Bess. H. C.. A.I.Ch.E. J.. 13 (6). 1152-6
(1967).	~~
17- Mayer, R. P., and Stove, R. A., "Physical Characterization of
Limestone and Lime," Azbe Award Ho. U, National Lime Association,
Washington, D. C., 26 pp. (I96U).
18. KcClellan, G. H., and Lehr, J. R., Am. Mineralogist. I37U-91
19.	Mezaki, R., and Kittrell, J. R., Ind. Eng. Chem.. 59, 65-9 (1967).
20.	P»!ache, C., Be man, H., and Frondel, C., Dana's System of Mineralogy.
Sj 7th ed., John Wiley "« Sons, Sew York, N. Y., p. 211 (1951).
21.	Potter, A. E., "Sulfur Oxide Capacity of Limestones," U. S. Public
Health Service Symposium on Lime stone-Sulfur Dioxide Reaction
Kinetics and Mechanisms. NAPCA. Cincinnati. Ohio (1Q6Q^. (Abstract)
22.	Rau, R. C., Advan. X-Ray Anal.. 5, 10U-16 (1962).
23.	Snedecor, G. W., Statistical Methods, p. 91, Iowa State College
Press, Ames, Iowa, 1956.
-------
107
APPHCDIX A: MECHANISM OF CALCINATION REACTION
A phase-boundary-controlled reaction is represented (1) by
the expression
dor/dt = k(l . a)1	(1A)
where or = the fraction reacted
1 = the order of the reaction (I.e., the exponent of the unreacted
portion, 1 - cr, at time t)
k = the rate constant
When the reaction is controlled by the external surface of the reactant,
the value of I depends upon the shape of the particles. For spherical
particles, the value of I is 2/5, since the surface varies as t.hg p/j
power of the volume; when the reaction progresses faster in tvo
dimensions than in the third dimension, however, as with cylinders or
discs, the value of t approaches 1/2.
The Integrated form of equation 1A is
1 - (1 - or)1 ^3 = 0.206J0 t/to.s	(2A)
for 1 = 2/5, and
1 . (1 - or)i/2 = 0.29809 t/to.s	(JA)
for X e 1/2. In equations 2A and JA, a is the fraction of reaction at
time t; tc.5 is the time for 50$ reaction; and the constants result from
geometrical considerations of the particle shape and the advancing front
of the reaction.
108
The degree of calcination, or, is plotted against the reduced
time coordinate, t/t0.5, in Figure 2k for all the Isothermal calcination
runs that have been made in this laboratory on minus 2k- plus 28-mesh
Colbert limestone for different partial pressures of C02. The theoretical
curves, equations 2A end }A, are shown for comparison with the experimental
points.
It is concluded that the phase-boundary-controlled model
represents the data adequately, and that for this particle size, minus
2k- plus 28-mesh, equation 2A represents the data better than equation
3A. This Implies that the reaction proceeds at the same rate in all
three dimensions of the particle. It Is possible that finer grinding of
the limestone will produce flatter particles that may follow equation 3A
more closely than equation 2A; the fact that the results at higner tempera-
tures, where spelling is more pronounced, approached the theoretical curve
JA supports this assumption.
-------
109
APPHIDIX B: CALCULATIOII OP KUTETIC PARAMETERS
FOB CALCIHATIOII
The kinetic equation for theroogravinetrie studies of
calcination is
dor/at ° X(1 - or)2/3	(IB)
Substituting the heating rate equation, dT = qdt, and the Arrhenius
equation, k = ko exp(-E/RT), Into equation IB and integrating, there
Is obtained
T
(1 - cr)l/3 = 1 - (Ito/Jq) f exp (-E/RT)dT	(2B)
J To
where
or c fraction of } lmestone calcined at time t or at
temperature I
q o heating rate, *K./min.
To, T = temperatures at which the fraction decomposed is 0
and or, respectively
ko# B *> Arrhenius parameters for specific rate constant, k
R o gas constant
The thermogravlmetrlc measurements are the weights, w, of the
undecomposed limestone at different times, t, or temperatures, T, and
the relation to the fraction decomposed is
1 - a = (w - vf)/(v0 - vf)	(3B)
where w0 and w^ are the original and final weights of the sample,
respectively, (that is, when o s 0 end a = 1, respectively).
110
The data (w, T) were smoothed by a spline-curve coaputer
program based upon the technique of Klaus and Van Hess (16), which
assures no discontinuities in the first two derivatives, dw/dT and
d^/dT2. The 6zaoothed data were used to evaluate the kinetic parameters,
ko and E, of equation 2B by a nonlinear least-squaring technique. The
integral in equation 2B was evaluated precisely by integrating once by
parts
T	T	ox
r exp (-E/RT)dT = [T exp (-E/RT)] - (E/R) exp (-X) ^ (UB)
J T0	T° JXo ^
where X = E/RT
Xq = value of X at To
evaluating the residual integral by a Hastings approximation (11) with
an error of less than 10~6.
The least*squares surface was made more regular by redefining
the pre-exponential parameter of the Arrhenius equation (19) by the
relation
k = kx exp (-E/RTo.s)	(5B)
to obtain
k = kt exp [-(E/R)(l/T - 1/T0.5)]	(6B)
where T0.s is the temperature, °K., at which the calcination is $0$
ccmplete (a = 0.5). The reparameterization changed the poorly conditioned
Bwo-of-squares surface into a veil-conditionsa surAace. The uuji>uiu» vf
the surface were more circular or elliptical with a well-defined minimum
rather than long and skewed with an ill-defined minimum, and convergence
was improved markedly. The size of the confidence regions of the para-
meters was unchanged because there was no change in the rate equation itself.
-------
Ill
APPENDIX C: KINETIC DITSRPiCTATIOH OF DIFFERENCES
AMONG STOHES
If ve assume thatj when limestone Is Injected into a combustion
chamber to decrease the emission of S02, the Important reactions are
CaCOa JSlo-CaO + C02	(1C)
CttO + S02 + O.SOo k?»CaSO„	(2C)
CaO ^ c (CaO)	(3C)
where each con be represented in Arrhenius form by
\ » klo e*P (-Ei/RT)	(UC)
as derived In equation IT for a particular stone, and that the injection
is made at sufficiently high temper*ture (above 900" C.) so that calcina-
tion is not rate limiting, the rate equation for reaction 2 in excess
oxygen is
dx/dt 1 k;?(a - x)O>0 - x)(l - c) - kaX	(5C)
where a • the initial concentration of S02 In the gas
b a the concentration of lime after injection, of which 0
is the effective fraction that reacts In the first stage
before a complete shell of product is formed
x - the amount of CaS0« formed at time t
c = the degree of dead-burning that occurs under the Injection
conditions.
If the injection temperature Is 1500° C. or belov, the last term in
equation 5C can be ignored, since the rate of decomposition of CaS0« vill
he less than l£ of its rate of formation. Assuming also that the particles
112
cool from their highest temperature at a constant rate, p, and that
the sulfation reaction stops when the temperature drops to 800" C.
there is obtained on integration
1 ~	= exp [-k20a(S - l)l/»]	(6C)
vhere S = the effective ratio cao:S02 In the flue gas,
S = (b9/a)(l-c)
9 = the degree of desulfUrizatlon of the flue gas
I = r 2 exp (-Es/RT)dT	(7C)
J Tl
b injection temperature, *K.
T2 ¦ temperature of cessation of sulfation, *K.
The value of the dead-burned fraction, c, in the definition of S must
also be evaluated from the Arrhenius parameters of reaction 3 by
1 - C O exp [K»o/p) J S exp (-Bj/KT)dT]	(8C)
The Important factors for desulfurisation are therefore the effective
ratio CaO:SOa, 8; the concentration of S02 in the flue gas, a; the
injection temperature, Ti; the cooling rate, p; the kinetic parameters
of the 8tone for sulfation; and dead-burning.
Assuming the dead-burning parameters of equation 17; an
effective fraction, 0, of lime to react to be OA; and a cooling rate
of 220* C./sec*> the relationship between the injection temperature,
-------
113
the ratio C&0:S02, the S02 concentration, and the sulfation paranetera
Is shovn In Figure 25 • In this figure the data are plotted to show the
beneficial effects of a high frequency factor,	in the Arrhenlus
equation be, and of a low activation energy, E2. The optimum Injection
temperature Is predicted to decrease with the activation energy, E2,
for sulfation and with Increasing S0a concentration In the flue gas;
these relations might he modified with stones of different dead-burning
characteristics and with different cooling rates. The effect of the
S02 concentration is more pronounced when both Arrhenlus parameters,
kso and Ea, are either low or high; when one Is low and the other
high, the effect of concentration of SOg on Its degree of removal Is
less pronounced. A large frequency factor, ka0j makes the temperature
range for good desulfurlzatlon less critical.
Ill*
100
12 13 14 15 16 13 14 15 16 13 14 15 16 13 14 15 16
TEMPERATURE .°C., *10
-2
Figure 25. Effects of Amount of Injected Limestone and
Kinetic Parameters on the Desulfurlzatlon of Stack Gas.
-------
115
APPENDIX D: DETERMINATION BY REGRESSION ANALYSIS OF
THE SIGNIFICANT FACTORS THAT AFFECT THE
SOa ABSORPTION PROPIBTIES OF LIMESTONE
Consider a dependent variable, y, that is a function of
several, n, independent variables: *t, xs, *3.. ,xn. It ia desired
to obtain the significant Independent variables that, In conjunction
vith each other, affect the value of y, when a number of measurements
are available that associate the value of y with a set of values of z's
The method is based on the statistical t test for the algnlfi
caace of the coefficients of the x variables in a linear model. The
process involves the successive e1 lmlnation of one x factor at a tine
until each of the remaining factors or independent variables
attained a t -value that Is statistically significant, at some arbitrary
confidence level, In the regression model.
Two regression models are used:
y = *1*1 + + 83X3 +...aDxn	(ID)
y = b^ + bax2 + >>3X3 +...bnx3 + b0	(2D)
where the a's and b's are the coefficients of the factors in the
respective models. Model 2 contains a constant term, b0, but model 1
contains only a linear combination of the independent variables.
The coefficients of aj in equation ID and in equation 2D
are determined by minimizing the sums of squares, E(y0bs - yic)s and
2(y0bs " yac)2» respectively; y^s is the set of observed values of the
116
dependent variable, y, and ylc and yzc are the corresponding sets
calculated by equations ID and 2D respectively. The corresponding t
statistic,
til = at/SEai, 1 = 1, n	(3D)
tai = bj/SEbi, 1 = 1, n + 1	(ltD)
for each coefficient is calculated by equations 3D and UD, where the
first subscript of t refers to the model number, the second subscript,
1, refers to the corresponding coefficient in the model, and the terms
SBa^ and SEbj are the standard errors of the respective coefficients
that are derived from the least-squares fit of the models and the lack
of fit, or residual variance, with respect to the observed values of y.
The least significant factor of the n factors is eliminated
by testing the value of t for the term b0 in model 2; if this particular
t value Is greater than the t statistic at the prescribed confidence
level, the remaining t values of model 2 are ranked from lowest to
highest In absolute magnitude, and the variable corresponding to the
least t value Is eliminated unless It, too, exceeds the t statistic at
the same confidence level. If the tenn b0 In model 2 is not significant,
the above procedure Is used to eliminate the least significant factor In
model 1. The remaining n - 1 factors are then treated by regression with
the same 2 models, equations ID and 2D, and the second least contributing
factor is eliminated. This process Is repeated until all the remaining
factors contribute significantly to the regression. (It should be noted
that the order of listing the factors in equations ID and 2D Is unimportant,
because the same least-squares solution is obtained in any order for a
given number of linear factors.)
-------
117
When the number of factors has been reduced to a set of
significant factors, the observed values of y are further treated by
regression to Include these factors and their interactions with each
other. The model is (for significant factors , xs, and *3)
y = *>!*! + + b3*3 +	+ bjjXjXs + b 23*2X3	(5D)
Again, the term b0 is added to a second model to correspond to equation
2D. The least significant of the terns in equation 50 is eliminated as
before, and the process Is repeated until only significant terns remain
in the expression.
The above procedure of selecting the most important factors
ftor a particular response, or dependent variable, is carried out on a
computer that makes the choice of model, eliminates the proper term,
tests the t statistics, regroups for the next regression, and finally
sets up and solves for the significant interaction terms.
The confidence level for the t statistics used in the multiple
correlations was 90$. The process of elimination was limited to a total
of 10 independent variables, n = 10, because the solution of a matrix of
larger size frequently resulted in unrealistic values of the constants.
The loss of significant figures in the matrix inversion of the sums-of-
squares equations when n was larger than 10 sometimes gave R values large
than 1, as well as erroneous standard errors for the coefficients. For
this reason, when the solution of equations ID and 2D gave five or more
significant factors, the interaction equation, 5D, was omitted. With
four or fewer significant factors, the elimination procedure on the inter
action equation usually returned to the original solution by eliminating
the Interaction terms.
118
APPRJDIX E- MULTIVARIATE AHALYSIS—A LOOK k?
PRIKCIPAL-COMPONENTS PREDICTION
Principal-component analysis is a mathematical technique for
simnmrlzlng a set of related measurements as a set of derived varlates,
frequently fewer in number, which are definable as Independent linear
functions of the original measurements. For example, if y, a dependent
variable, is a function of n independent variables, Xi, xa, Xg, ... in,
it is possible to express y as a function of m combinations of the n
variables by the equation
y = BjXi + 1)2X2 + B3X3 + ...	(IE)
where n £ n, and
n
Xj = Z ^ BijXj, i = 1, m	(2E)
The derived varlates, represented by capital X"s, are orthogonal to each
other, and the first derived varlates, X^ will account for the largest
possible proportion of the total variation In the original x variables,
the second will account for the largest possible proportion of the
remainder, and so on. The derived varlates are known as principal
components, and the corresponding variances as latent roots. When all
the principal components are taken, equation IE results in a regression
equation, when transformed to the original x variables, identical with
that obtained by linear regression. Since the derived varlates, X^, are
independent and arranged in the order of decreasing accountability of the
total x variance, it is frequently possible, by geometrical manipulation, to
obtain linear functions of the original variables that are chemically
meaningful or consistent with other observations.
-------
U.9
A total of thirteen x variables tr)
yia	s t.87	t-90 5.1#
H 0.813	O.ttLl 0.755
y5	S 8.27	9-01	9.02
R 0.656	0.569 0.568
yia/yis S 0.213 0.217 0.229
R 0.662	0.6M> 0.593
120
Fiaure 26. Number of Roots Required to Explain Variations
Among 13 Measured Properties of Limestones.
-------
121
The prediction of the polythermal uptake of S02 ng
calcination, y13, was affected greatly by dropping the four ssallest
roots, but was not significantly affected by dropping the two smallest
roots. The prediction of tile total isothexval weight gain at 930° C.,
ys, vas affected more vhen the first two roots vere dropped than when
the third and fourth roots vere dropped. The prediction of the fraction of
S02 uptake during calcination, yia/yis, was affected Increasingly as more
roots were dropped.
It appears likely that the multivariate approach to elucidate
the principal components of the chemical and mlneraloglcal properties of
llmeatonea in relation to the S02 absorption will not produce a meaningful
function that will have a practical value for the injection process based
on the data of this report. The derived functions can be no better than
the linear regression models, which In themselves leave much to be desired.
Possible explanations Include the Imprecise measures of sane of the x
variables or even the inadvertent omission of one or more properties of
the stones that may affect the sulfation characteristics markedly.
122
APPHTODC P: RELATIVE IMPOHEAHCE OF PROPERTIES 0?
T.TtreSTOHBS FOR REACT IDS WITH SULFUR OXIDES
The relative Importance of the significant variables that
affect the reactivity of limestones with S0a is shown in the attached
table. The results are shown as the changes in the factors required to
produce either an increase of 1 in the value of v3 or a 10% increase In
the value of w3, depending on whether the linear or logarithmic equation
is used to calculate the capacity v3 using the coefficients of Table XX.
An Increase of 0.81 in the ratio of rhombic carbonates to CaO
in the stone (in the range 1.76 to 3.U2) causes a 10$ increase in the
absorptive capacity of the stone; the same effect Is obtained by either
decreasing the microscopic grain size (code) by 0.62, increasing the lattice
constant by O.lU- A., increasing the HajO content by 68flt of its value,
decreasing the content of clay, feldspar, and limonite by 2-95^ in
absolute amount In the stone, or increasing the fraction or iron (if it
has to be present) in limonite form by a value of O.^U.
Similarly, an increase of 1 in absolute capacity (next to last
column in Table XXIII) requires either a decrease of 0.031 ^ (310 A.) In
the x-ray crystallite size, a decrease of 3.05$ in the clay, feldspar,
and limonite content, or an increase of 09-yf> in the HasO content. All
of these Increases or decreases must be within the ranges listed.
-------
J*
a
u
£
-i
0)1
s
•s
01
8
8
t
8
e
t
~>
9
U
-3 1
I
09
Cj
S
3
1).
4)
H
5
Q
a
6
«
4
3
O
4»
•H
5
3
U
i
St
11
i i
S 6
» 3
M %j
11
1 J?
o
f-(
.O
s
a
01
*
o
i!
fl
5
0
a
£
t>
a
¦&
r?
d
¦e
a
4>
3
e
•a
o
d
n
3 i
I 5
$
i
s
O
aJ
5
i
I :
*4
0
Q £ |
w4 B +»
•p 5 u
¦ 1
P 03
«H g O
* « &
8
•H «
I 8
I I«
& «
$ *
%« 5
* a
& -
i *
f %
£ *
6
d
d| o
5
3
%
O
H
N
O
O
3
3
3
Q
6
8
U
8
vH
•P
3
3
ns
a
o
3
TABLE mn
Changes In Limestone Properties Required to Produce
Measured Changes in Reactivity
Change In property
to produce

Property

1
SI
&l
|j
A 10$ Increase
In weight gain
Designation
Description
Raiwe
per SO (jwrnrn n-P ntnnn
za
i> rhombic carbonates/^ CoO
1.76s>M-
-
+0.81
z«
Microscoplo grain size, codeda
1-3
-0.6k
-0.62
Zs
Lattice constant, a, k.
U.805-^.989
+0.18
+0.1l»
Za
X-ray crystallite size,
0.0T0-0.211
-0.(01
-
ZT
Logarithm of Ha^ content
O.OO5-O.25
~89.5*"
+68.2*»
Z9
$ clays, feldspar, and llmonlte
0.1-31
-3.05
-2.95
*10
Fraction of Fa as llmonlte
0-1
+O.35
+0.3U
a
1 h <6J ji;
2 = 63- 250 p; 3 = >250 p.



b Percentage Increase in sodium content.



B
-------
Q
G. H. McClellan' and J. I.. Eades-
Authorlzcd Reprint
rrum Hpeolul Technical I'ulillcnlloti 472
Copyright
American Society for Testing and Materials
1310 Race Street. Philadelphia. Pa.. 19103
1970
The Textural Evolution of Limestone Calcines*
REFERENCE: McClellan, G. H. and Eades, J. L., "The Textural Evolu-
tion of Limestone Calcines,"The Reaction Parameters of Lime, AST\!
STP 472, American Society for Testing and Materials, 1970. pp 209-227.
ABSTRACTi A pure limestone and Iceland spar were shock calcined a!
several constant temperatures in the range 750 to 1300 C in a study of thr.
pore structure of the calcines. The calcines were examined with a scanning
electron microscope to follow the development of the pore structure and tc>
correlate changes in the pores and in the related surface area with the
properties of the lime and the temperature of calcination.
Limestone calcines prepared at temperatures below 1000 C had largei
pores and lower surface areas than Iceland spar calcines prepared at the
same temperature. Raising the temperature of calcination caused a decrease
in surface area and an increase in pore size of both materials The decrease
in surface area on calcination at 1000 C resulted from the growth of large
pores at the expense of smaller pores, but the physical characteristics ol
the lime were largely unchanged. At temperatures above 1000 C the calcium
oxide crystals sintered and grew in size as the pore size continued to increase
and the surface area decreased.
The larger initial crystallite size of the Iceland spar resulted in an
unfavorable pore size distribution in its calcined products prepared below
1000 C. The limestone calcines prepared at the same temperatures had
larger pores and smaller surface areas. Literature reports confirm that there
is an optimum temperature of calcination for each stone for producing the
proper surface area, pore distribution, and lime condition for maximum
solid-fluid reactivity.
KEY WORDS: calcite, limestone, roasting, calcium carbonate, calcium
oxide, electric power plants, flue gases, sulfur dioxide, air pollution, injec-
tion, scanning electron microscope, density (mass/volume), porosity, poroism-
eters. evaluation, tests
The injection of dry limestone into power plant furnaces is the most
economical means of decreasing sulfur dioxide concentrations in the
effluent stack gas. The calcium oxide produced from the stone reacts with
* Part of this work was performed under conlract TV-29232A with the National
Air Pollution Control Administration of the U S Public Health Service
1 Research chemist. Division of Chemical Development, Tennessee Valley Authority,
Muscle Shoals, Ala. 35660.
1 Research assistant professor. Department of Geology, University of Illinois,
Urbana, III. 61801. Personal member ASTM.
209
-------
210 THE REACTION PARAMETERS OF LIME
sulfur dioxide and oxygen to form calcium sulfate that can be removed by
conventional dust-collecting systems. The results reported here are a part
of a continuing laboratory study of the properties of calcines, picpared by
shock heating (calcinuiions curiiul out in an environment pivlu-atcd to a
selected temperature in contract to calcining at rising temperatures). The
materials examined were portions of the samples used by McClellan et ala
in reactivity tests, and the physical properties of the calcines were corre-
lated with their reactivities. The previous results showed that the nature of
the original stone influences the properties of its calcines, including their
reactivity with sulfur dioxide.
Limestones and their calcines have been evaluated for industrial use
largely on the basis of their chemical purity. Advances in industrial tech-
nology [/]4 have shown, however, that physical properties such as surface
area, porosity, pore size distribution, and crystallite size are important fac-
tors'in determining the rate and degree of the chemical reactions of the
calcines.
In this study, two high-calcium limestones were shock calcined at tem-
peratures between 750 and 1300 C. These two materials differ only in
texture, so that the effects of composition can be ignored. The calcines of
these stones were studied by scanning electron microscopy and by mercury
porosimetry, and the results were used to interpret the changes in texture
with rising temperature of calcination.
Experimental Procedure
The two limestones were a high-purity calcitic limestone (USDA Refer-
ence Sample 89) and Iceland spar (Ward's Scientific Company) in the
form of clear crystals. Both stones were crushed and screened to produce
a minus 20- plus 24-mesh fraction from which samples were hand picked
to minimize the content of accessory minerals.
X-ray and petrographic examinations showed the samples to be free of
accessory minerals. Their cell constants showed that both materials were
pure calcites as defined by Goldsmith and Graf [2] and were without
significant isomorphism.
.A 3-g sample of each calcite was calcined in a platinum boat for 16 h
at constant temperatures in the range 750 to 1300 C. The samples were
placed abruptly in a preheated furnace to effect initial shock calcination;
the long retention time was sufficient for complete calcination at each
temperature.
See p. 32.
The italic numbers in brackets refer to the list of references appended to this
paper
-------
McClELLAN AND EADES ON LIMESTONE CALCINES 21 1
Several particles of each calcine were stored under vacuum until they
were examined in the scanning electron microscope. Immediately prior to
the examination, each sample was coated in a vacuum evaporator with a
2:1 mixture of gold and palladium that was deposited from several direc-
tions to minimize charging of the rough, porous surfaces of the calcines
during examination.
The pore volumes and pore size distributions were measured with a mer-
cury porosimeter that had an upper pressure limit of 3,000 psia which
limited penetration to pores 0.03S fim in diameter or larger. Apparent
densities of the calcines were calculated from weights of the porosimeter
penetrometer containing the sample and mercury. Densities calculated from
measurements with only the pores larger than 17.S /im in diameter filled
with mercury are comparable to bulk densities determined by mercury
displacement.
Results
Raising the temperature of calcination decreased the total pore volume
and shifted the distribution toward larger pores in all calcines of sample 89
except that made at 1100 C (Fig. 1). This calcine showed a higher pore
volume and a more even distribution of pores in the range 17.5 to 0.175
/xm than the 1000 C calcine. Each Iceland spar calcine (Fig. 2) had less
than half the total pore volume of the corresponding limestone calcine,
and almost all the pores in the 750 and 850 C spar calcines were smaller
PORE DIAMETER, MICRONS
FIG. I—Pore volumes of I6-I1 limestone calcines.
-------
212 THE REACTION PARAMETERS OF IIME
PORE DIAMETER, MICRONS
FIG. 2—Pore volumes of 16-h Iceland spar calcines.
than 0.2 fim. The total pore volume of the spar calcines decreased, and the
pore distribution shifted toward larger pores with rising calcination tem-
perature except for the 1200 C calcine. The total pore volume of this cal-
cine was larger than that of the 1100 C spar calcine and may have re-
sulted from the mechanism that caused the larger pore volume in the
1100 C calcine of sample 89.
Murfay et al [3] showed that the reactivity of a lime is related directly
to its surface area. Rootare and Prenzlow [4] derived a method for cal-
culating surface areas from mercury porosimeter measurements that gives
results in good agreement with those of gas-absorption measurements, and
surface-areas calculated by this technique may be more applicable to solid-
fluid reactivities than those measured by some other techniques. This con-
clusion is based on the assumption that diffusion times for fluids in pores
smaller than 0.30 fim are longer than the contact times used in measuring
reactivity [J]. The surface areas calculated from the porosimeter data are
shown in Table 1.
The surface areas and pore sizes of the calcines determined from mer-
cury porosimetry data and from measurements of the scanning electron
micrographs are shown in Table 2. The individual surface areas and equiv-
alent mean spherical pore diameters were calculated by the method de-
scribed by Mayer and Stowe [/]. Surface areas and mean pore diameters
were calculated only for calcines prepared at 1100 C or below because the
-------
MtCLEllAN AND EADES ON LIMESTONE CALCINES 21 3
number of poies decreases rapidly at the higher temperatures as sintering
becomes pronounced.
There is good agreement between the surface areas calculated from the
porosimeter and from the electron microscopic data. The distribution of the
surface area within a sample also is important. The data in Table 1 show
that most of the surface area of samples calcined below 1000 C is asso-
ciated with small pores, those smaller than 0.2 /xm. If Potter's data [5] on
the sulfur dioxide capacity of limestones are correct, diffusion into poies
this small and smaller is too slow to be effective in short-time reactivity
tests. The formation of films of calcium sulfate on the surface could block
areas containing many of these pores, further decreasing their availability
for reaction. The pores provide the fluids access to the surface area of the
solids, and materials with mean pore diameters larger than 0.2 /xm ere
more responsive in short-term reactivity tests because of the shorter times
required for diffusion of sulfur dioxide into these larger pores. As shown in
Table 2, calcines of sample 89 had mean pore diameters larger than these
of Iceland spar calcines prepared at the same temperature, and the pore
diameters of both materials increased with rising temperature of calcinatio n.
Raising the temperature of calcination changes the nature of the solid
lime as well as its pore size distribution. McClellan et al3 showed that the
crystallite size of the limes of both materials slowly increases with rising
temperature of calcination below 950 C and decreases slightly at 1000 C.
At temperatures above 1000 C, the rate of crystallite growth increases with
rising temperature. These changes in the physical properties of the lirne
decrease the potential reactivity of the lime surface as crystal perfection
increases and crystal growth influences the texture of the surface itself.
Scanning electron microscopy has previously been used to study the sur-
face of limes prepared in commercial kilns.8 The work reported here is the
first laboratory study of shock-treated limestone calcines prepared under
laboratory control at several different temperatures.
The main advantage of scanning electron microscopy over optical and
transmission electron microscopy is its greater depth of focus, which is as
much as 100 times that of the other techniques. This advantage, combined
with the simple sample preparation, allows the investigator to examine the
unaltered surface of materials and to obtain information that was pre-
viously unavailable, or only to be guessed at, on surface texture and pore
size, shape, and distribution. In the study of limes, scanning electron
microscopy presents clear evidence on surface areas and the properties of
pores.
See p. 3.
-------
214 THE REACTION PARAMETERS OF UME
I'AHI 1. I-- Suijucv areas oj calcines calculated fttmi /mrosimeier mcti\uieinems.

Surface Area"
Surface Area"
Surface Area"
Pore Size. pm
M'/h
% ofTotiil
M'/g
% of Total
M'/t!
of I'otiil


l.tMIMONI 89




750 C calcine
850 C calcine
950 C calcine
17.5
0.12
0.3
0.08
0.04
0.07
1.5
5.0
0.17
0.5
0.12
0.6
0.11
2.4
1.75
0.22
0.6
0.16
0.8
0.19
4.0
1.0
0.36
1.0
0.26
1.3
0.28
5.9
0.5
0.54
1.5
0,43
2.1
1.15
24.5
0.4
0.71
2.0
0.84
4.1
1.87
39.8
0.2
3.72
10.3
8.62
42.4
3.76
80.0
0.1
26.26
72.9
20.17
99.3
4.70
100.0
0.05
35.68
99.0
20.31
100.0
4 70
100.0
0.035
36.03
100.0
20.31
100.0
4.70
1000.

1000 C calcine
1100 C calcine
1200 C calcine
17.5
0.04
0.6
0.06
5.2
0.03
5.9
5.0
0.08
1.1
0.09
7.3
0.05
9.7
1.75
0.15
2.2
0.17
13.7
0.09
16.1
1.0
0.32
4.7
0.37
29.2
0.31
57.6
0.5
0.67
9.9
0.82
65.4
0.50
92.3
0.4
0.96
14.1
0.92
73.2
0.54
100.0
0.2
4.88
71.8
1.06
84.5
0.54
100.0
0.1
6.79
100.0
1.12
89.6
0.54
100.0
0.05
6.79
100.0
1.25
100.0
0.54
100.0
0.035
6.79
100.0
1.25
100.0
0.54
100.0
Iceland Spar

750 C calcine
850 C calcine
1000 C calcine
17.5.
0.01
0.0
0.01
0.0
0.01
0.1
5.0 .
0 03
0.0
0.01
0.0
0.01
0.1
1.75
0.09
0.1
0 04
0.1
0.02
0.3
1.0
0 19
0.3
006
0.2
0.05
0.7
05
0 25
0.4
006
0.2
0.05
0.7
0.4
0.28
0.4
006
0.2
0.05
07
0.2
0.38
06
0.06
02
0.08
1.2
0 1
0.38
0.6
4 26
13.2
1 02
14.5
0.05
47 38
75.3
29 86
92.7
5 95
84.3
0.035.
62.93
100.0
32 20
100.0
7.0ft
1000
-------
McCLELLAN AND EADES ON LIMESTONE CALCINES 215
TABI.F. I (continued).

Surface Area"
Surface Area"
Surface Aiea"
I'orc Si/e, /im
M'/g
% of Total
M'/g
of Total
M'/g
"o oTTota

1100 C calcine
1200 C calcine
1300 C cak me
17.5. . ..
0.00
0.0
0.00
0.1
000
2.0
5.0	
0.01
0.7
0 01
0.4
0 01
5.0
1.75	
0.01
0.7
0.01
0.6
0.02
11.6
1.0 . ...
0.03
2.5
0.02
1.0
0.03
15.6
0.5	
0.04
3.4
0.02
1.0
0.06
19 1
0.4	
0.05
5.0
004
1.9
0.10
50.8
0.2	
0.08
7.6
1.14
56.6
0.20
nw.o
0.1.
0.40
37.9
1.89
93.8
0 20
11)0 0
0.05	
0.69
65.2
2.02
100.0
0.20
1000
0.035	
1.06
100.0
2 02
100.0
0 20
100 0
0 Area contributed by pores of indicated and all larger sizes.
TABLE 2—Correlation of porosimeter and scanning electron microscope measuremtnts.
Porosimeter	Scanning Electron Microsocpe
Calcination	Surface	Pore	Surface	Pore	No of
Temperature, deg C Area, M'/g Size, Mm° Area, M'/8 Size,	Pores6
Limestone 89
750 .
36.0
0.05
27.4
0.06
481
850	
20.3
0.09
14.6
0.12
514
950..
4.7
0.37
42
0.42
471
1 000 ..
6.8
0.26
6.7
0.26
50.!
1 100 . . .
1.2
I 44
1.4
1.27
181!
Iceland Spar
750 	
62.9
0.03
66.5
0.03
477
850	
32.2
005
46.4
004
. 470
950


46.0
004
348
1000 	
7.0
0.25
108
0.16
368
1 100	
I.I
1.58
0.9
1.92
75
0 Calculated from the relation A = 6/Op, where A = surface area, M'/g; D = mean
diameter, ^m, of pores assumed to be spherical, p = absolute density, g/cm'.
6 Number of pores measured for calculation of pore size.
-------
216 THE REACTION PARAMETERS OF LIME
In calcines prepared at temperatures below 1000 C, the most diagnostic
features are the pores and their growth by coalescing. A representative
area of the 750 C limestone (sample 89) calcine is shown in Fig. 3 in
which the evolving pore structure of the lime is superimposed on a residual
rhombohedral cleavage fragment of the stone. There is a wide range of
pore sizes, even in this soft-burned stone. The solid lime is generally fea-
tureless, and no significant points of contact have developed between the
individual lime units within each particle. Electron micrographs of 750 C
Iceland spar calcine show essentially the same evolution of pores, but the
pores are smaller and there are more per unit area.
The 850 C calcines (Fig. 4) show continuing coalescing of the pores,
although residual rhombohedral features of the original stone persist. This
figure shows that the larger pores grow at the expense of the smaller ones,
as observed in the porosimeter data. There is no evidence of individual
calcium oxide units within the larger particles.
The 950 C calcines (Fig. 5) show additional coalescing of pores super-
imposed on detectable cleavage features of the original stone. The pores
FIG. 3—Limestone calcined at 750 C showing pore evolution superimposed on
cleavage fragments 1X6300).
-------
McClEUAN AND EADES ON LIMESTONE CALCINES 217
FIG. 4—Limestone calcined at S50 C showing coalescing of pores (X 2640).
seem to be oriented randomly, and there was no detectable crystallographic
contiol of pore evolution in any of the calcines studied. Discrete subunits
of calcium oxide, which appear for the first time in calcines prepared at
950 C, can be seen throughout Fig. 5.
The 1000 C calcines show final pore growth by coalescing, and the cal-
cine of sample 89 (Fig. 6) is representative of this stage. Pores 0.5 /am or
larger are abundant, and the subunits of lime appear relatively unchanged
from those in the 950 C calcines. The open textural appearance and an
essentially unchanged lime appear to indicate optimum surface availability
for solid-fluid reaction.
The Iceland spar 1000 C calcine (Fig. 7) shows only a few pores as
large as 0.05 /im, and most of the pores are much smaller. The distinguish-
able lime subunits seem smaller (about 0.1 /xm) than the lime in the
limestone calcine (about 0.5 fim) and seem to have proportionally larger
points of contact.
Calcines prepared at 1100 C or above show sintering and shrinking of
the oxide that is not seen in the lower-temperature calcines. The limestone
-------
218 THE REACTION PARAMETERS OF LIME
calcine, shown in Fig. 8, has a more compact texture with extensive con-
tacts developed between the lime units, and a pronounced decrease in
porosity over that shown in Fig. 6. Growth loops or spirals are present on
the individual lime units in the right half of the figure, indicating that
crystal growth and development of single-crystal morphology has begun,
even though residual rock cleavage features are still present.
FIG. 5—Limestone calcined aI 950 C showing continuing coalescence of pores
and the development of individual lime units (X 2665).
-------
McCLELLAN AND EAOES ON LIMESTONE CALCINES 219
The I MX) (' Iceland spar calcine in Fig. l) also shows the increased com-
pactness of the lime and the resultant decrease in porosity. Althoagh
growth spirals are not well developed, many of the lime units arc develop-
ing llat faces and straight edges between crystal faces. The pores in the
1100 C calcines of both materials arc beginning to show straight sides as
crystal growth begins.
FIG. 6—Limestone calcined at 1000 C showing completion of pore growth by
coalescence (X2888).
-------
222 the reaction parameters of iime
FIG. 9—Iceland spar calcined at 1100 C showing developing crystal faces and
geometric pores (X 9720).
or larger. Because of this restriction there was no plateau in porosimeter
pore volume as in the other calcines, indicating that this calcine had pores
smaller than 0.035 /tm and would have a surface area larger than 63 m2,/g.
Potter [5] concluded from his studies of calcines sulfated in fixed-bed
tests that pores in the range 17.5 to 0.3 /xm are most important in the sulfa-
tion reaction and that smaller pores require diffusion times too long to be
effective when dry limestone is injected into a power plant furnace. The
porosimeter data show that 4 m-/g of the 36 m-/g of the surface area of
the 750 C calcine and 9 m2/g of the 20 m2/g of the surface area of the
850 C calcine of sample 89 are accounted for by pores in the effective
range, and that rising temperature increases the fraction of pores in this
size range. In contrast, the Iceland spar calcines have only 0.1 m-/g of the
1.1 m2/g surface area contributed by pores in this size range at 1100 C
and 1.1 m- g of the 2.0 m-/g surface area at 1200 C. This study has
shown that both the porosity and the physical condition of the lime in the
calcines prepared at temperatures up to 1000 C were influenced strongly
by the nature of the starting materials. The Iceland spar calcines had
-------
McClEUAN AND EAOES ON LIMESTONE CALCINES 221
FIG. 8—Limestone calcined at 1100 C showing sintering texture and growth
spiral (X1350).
Discussion of Results
The application of scanning electron microscopy to the study of lime-
stone calcines demonstrates the meaning of pore volume, pore distrib ution,
and surface area more clearly than other techniques. The evolution of both
solids and pores as the temperature of calcination is raised can be observed
directly rather than imagined by conceptual models as required in the past.
These studies illustrate the access that pores provide reactant gases to the
surface area of the calcines.
The calculation of surface areas from porosimeter measurements pro-
vides an insight into the physical characteristics of the calcines that is not
provided by inspection of the pore-volume data alone. The calculated areas
were essentially the surface areas generated by the pores observed in the
scanning electron microscope studies. The surface area calculated for the
750 C Iceland spar calcine should be considered a low approximation
because the data were calculated only for pores approximately 0.035 /im
-------
220 THE REACTION PARAMETERS OF LIME
FIG. 7—Iceland spar calcined at 1000 C showing the pores and individual lime
units (X 36,450).
As shown in Fig. 10, the 1200 C limestone calcine is quite compact,
and the porosity is confined to a few large pores. The figure contains a
good example of a residual cleavage fragment on which typical features of
the calcine are clearly apparent. Epitaxial growth of a potential crystal face
is noticeable on six or seven oxide units, and the few remaining pores are
large and contribute little to the surface area of the lime. The 1200 C Ice-
land spar calcine in Fig. 11 also shows well-developed and developing
crystal faces and has the compact, nearly pore-free appearance of extended
points of contact.
The 1300 C Iceland spar calcine (Fig. 12) has a very compact and
practically pore-free surface texture with abundant growth loops. A very
well-developed oxide crystal appears at the right of the figure. This photo-
graph shows why the calculated surface area of this sample is only 0.2
m-'/g and why the surface would be considered unreactive in a solid-fluid
system.
-------
McCLELLAN AND EADES ON LIMESTONE CALCINES «!23
FIG. 10—Limestone calcined at 1200 C showing epitaxial growth of crystal feces
on cleavage fragment (X2J83).
smaller mean pore diameters and larger surface areas than the correspond-
ing calcines of sample 89. Calcination at higher temperatures decreases
the porosity and sinters the oxide of both materials so that the reactivity is
decreased.
The Iceland spar calcines have a distribution of pores that is unfavorable
for sulfation reactions, and sintering has occurred at temperatures at which
suitable pores begin to develop, with a resultant decrease of surface area
and potential reactivity.
This interpretation is consistent with the observations of McClellan
et al3 on the relative reactivities of limestone and Iceland spar calcines
with sulfur dioxide in fixed-bed tests. They found that the limestone cal-
cines were superior reactants under all the experimental conditions tested
and concluded that unfavorable pore distribution and poor properties of
the lime accounted for the relative inactivity of the Iceland spar calcines.
Eades and Sandburg" drew similar conclusions from a series of limes on
which they measured hydration reactivity. Their data showed that a de-
veloped pore structure of the limes was necessary for satisfactory hydra-
-------
224 THE REACTION PARAMETERS OF IIME
FIG. 11—Iceland spar calcined at 1200 C showing well-developed crystal growth
(X93I5).
tion of the surface, and that softer burned stones without suitable pore
structure but with larger surface areas were less reactive.
Murray et al [J] have shown that the time required for a coarsely crystal-
line limestone to reach 920 C during its calcination has a significant effect
on the porosity and activity of the lime produced. If the time is too short,
the porosity and activity are quite low. If the time is too long, porosity
develops satisfactorily, but the activity decreases slowly with prolonged
exposure. Fischer [6] found similar results for the densities of a stone he
shock calcined at different temperatures. McClellan et al3 measured bulk
densities of shock-heated limestone and Iceland spar calcines prepared in
the temperature range studied by Fischer. The bulk densities of sample 89
shock-heated calcines were consistently lower than those of the correspond-
ing Iceland spar calcines. The nature of the starting material again appears
to influence the physical properties of the calcine, even though the rate of
heating prior to and during calcination, as well as the maximum tempera-
tures, were the same within experimental error for corresponding samples
of both materials.
-------
McClEllAN AND EADES ON LIMESTONE CALCINES 225
The differences in physical properties of equis alen: calcines of Ice
spar find limrMoiw <<9 ma\ reflect differences in the difficulty of nnuump
carbon (Ini*tilt from ihe caltite structure of the materials |7) The cry,tal-
lilc t>\/jc of the Iceland spar measured by X-ray line broadening was 0.71
/xin, whereas Ihe ciystallilc size of sample K9 was 0.39 /xm. This is the
only difference in physical properties measured on the starting materials.
FIG. 12—Iceland spar calcined at 1300 C showing a sintered surface with growth
loops and a well-developed single crystal (arrow) (X4042).
-------
226 THE REACTION PARAMETERS OF IIME
Perhaps the larger crystallite size of the Iceland spar inhibits the migration
of curium dioxide from the crystal structure.
The results of this study show that the texture of limestone calcines
evolves in two distinct steps. Pores form initially as carbon dioxide is re-
leased during calcination. Raising the calcination temperature causes the
pores to grow by coalescing (that is, big pores grow at the expense of small
pores). This tcxtural feature predominates in materials calcined below
1100 C. At about 1100 C, the change in the physical characteristics of the
calcium oxide predominates as the limes become compact as a result of
sintering and crystal growth of the calcium oxide. These changes are
accompanied by a marked decrease in porosity as the calcination tempera-
ture is raised. Both stages of textural development may be present in a
single particle if the external surface and edges are calcined before the
center, as reported in the literature [6,7].
Summary and Conclusions
Results of scanning electron microscope and mercury porosimeter studies
of the calcines confirm several general observations, among which is the
fact that the nature of the uncalcined carbonates influences the properties
of their calcines. The larger crystallite size of the Iceland spar confers
undesirable physical characteristics on its calcines.
Raising the temperature of calcination causes a systematic evolution of
texture in the calcines that is characterized by stages dominated by pore
growth and later by sintering and calcium oxide crystal growth. The rate
and nature of these changes also arc influenced by the characteristics of
the starting material. The undesirable pore size distribution of the Iceland
spar calcines indicated that its properties were similar to those of some
shock-calcincd, coarsely crystalline materials that are described in the
literature. Propeities of calcines prepared from limestone 89 compare
favorably with those of some of the reactive limes described [<5]\
It can be inferred from these results, that solid-fluid reactivity is corre-
lated with pore distribution and surface area.,-r' The pores provide the
reactant fluid with access to the surface area. If diffusion of the gas in small
pores is slow f J], then each stone has some optimum time and temperature
of calcination for development of a favorable pore size distribution and
surface aiea that results in a reactive lime.
Acknowledgment
R. C. Mullins contributed to this paper through discussions of the pore
sizes and pore volumes, all of which were measured under his supervision.
-------
McCtELLAN AND EADES ON LIMESTONE CALCINES 227
References
[/] Mayer, R. P. and Stowe, R. A., "Physical Characterization of Limestone and
Lime," Azbe Award No. 4, National Lime Association, Washington. D. C.. 1964
[2] Goldsmith, J. R. and Graf, D. I., "Relation Between Lattice Constants and
Composition of the Ca-Mg Carbonates," American Mineralogist. Vol 43. 1958.
pp. 84-101.
13]	Murray, J. A., Fischer, H. C., and Sabean, D. W„ "The Effect of Tine and
Temperature on the Properties of Quicklime Prepared from Calcite," Proceedings,
American Society for Testing and Materials, Vol. SO, 1950, pp. 1263-12*7.
14]	Rootare, H. M. and Prentzlow, C. F., "Surface Areas from Mercury Poroi.imeter
Measurements," Journal of Physical Chemistry, Vol. 71, No. 8, 1967, pp 2733—
2736.
[5]	Potter, A. E., "Sulfur Oxide Capacity of Limestones," U. S. Public Health
Service Symposium on Limestone-Sulfur Dioxide Reaction Kinetics and Mecha-
nisms, National Air Pollution Control Authority, Cincinnati, Ohio, 1969 (abstract).
[6]	Fischer, H. C., "Calcination of Calcite: I. Effect of Heating Rate and Temperature
on Bulk Density of Calcium Oxide," Journal of the American Ceramic Society,
Vol. 38, 1955, pp. 245-251.
[7]	Hyatt, E. P., Cutler, I. B., and Wadsworth, M. E., "Calcium Carbonate Decompo-
sition in Carbon Dioxide Atmosphere," Journal of the American Ceramic Society,
Vol. 41, 1958, pp. 70-74.
-------