JUiL
mm

             SELECTED  STUDIES
                    ON
           ALKALINE ADDITIVES
          SULFUR DIOXIDE CONTROL
             U. S. ENVIRONMENTAL PROTECTION AGENCY

-------
SELECTED STUDIES ON ALKALINE ADDITIVES

      FOR SULFUR DIOXIDE CONTROL
                  by

    R. H. Borgwardt, D. C. Drehmel,
    T. A. Kittleman, D. R. Mayfield,
             and J. S. Bowen
    ENVIRONMENTAL PROTECTION AGENCY
National Environmental Research Center
Research Triangle Park, North Carolina
             December 1971

-------
The APTD  (Air Pollution Technical Data) series of reports is
issued by the Office of Air Programs, Environmental Protection
Agency, to report technical data of interest to a limited num-
ber of readers.  Copies of APTD reports are available free of
charge to Federal employees, current contractors and grantees,
and nonprofit organizations - as supplies permit - from the
Office of Technical Information and Publications, Environmental
Protection Agency, Research Triangle Park, North Carolina 27711
or from the National Technical Information Services, 5285 Port
Royal Road, Springfield, Virginia 22151.
          Office of Air Programs Publication No. APTD-0737
                                    ii

-------
                TABLE OF CONTENTS



A.  Summary                                  1

B.  General Introduction                     5

C.  Discussion of Projects                   9

    1.  Kinetics of the Reaction of SO-
        with Calcined Limestone

        a.  Introduction                     10
        b.  Experimental                     11
        c.  Discussion of Results            15
        d.  Conclusions                      28
        e.  Nomenclature                     29

    2.  Properties of Carbonate Rocks
        Related to S02 Reactivity

        a.  Introduction                     31
        b.  Experimental                     33
        c.  Results                          37
        d.  Discussion of Results            45
        e.  Conclusions                      64
        f.  Nomenclature                     67

    3.  A field Study of the Role of
        Overburning

        a.  Introduction                     69
        b.  Experimental                     71
        c.  Discussion of Results            76
        d t  Conclusions                      87

    4.  Methods for Testing the  Degree
        of Overturning

        a.  Introduction                     89
        b.  Experimental                     92
        c.  Discussion of Results            95
        d.  Conclusions                      112
                      iii

-------
    5.  Capacity of Limestones for
        Sorption of SO-

        a.  Introduction
        b.  Experimental                    115
        c.  Discussion of Results           119
        d.  Conclusions                     130
D.  General Conclusions                     132

E.  Application of Results to Limestone
    Processes                               135

F.  Recommendations                         138

G.  References                              140

Appendix A   Kinetic Data of Carbonate
             Rock Types                     143

Appendix Bl  Procedure for Dead-Burning
             Tests                          158

Appendix B2  Statistical Analysis of
             Dead-Burning Data              162

Appendix B3  Crossplots of Data from
             the Dead-Burning Study         184

Appendix Cl  Calcined and Uncalcined
             Limestones Ranked According
             to Bed Weight Gain             197

Appendix C2  EPA-APCO Limestone Inventory   205

Appendix D   Sorption of S02 by Waste Kiln
             Dust from Portland Cement
             Manufacturing Operations       218

Appendix E   Copper Oxide Sorption of SO-   224

Appendix F   Char Sorption of S0» and
             Regeneration                   228
                      iv

-------
                 TABLE OF FIGURES
1-1   Differential Reactor                   13

  2   Sorption of S02 by 1351 at Various
      Reactor Temperatures                   17

  3   Sorption of S02 by Different
      Particle Sizes of 1351                 17

  4   Log r Versus Log C to Estimate "m"     19

  5   Arrhenius Plots for the Reaction of
      S02 with Four Limestones               19

  6   Reaction Rate Versus Particle
      Diameter                               24

  7   Sorption of SO, by Calcined
      Limestones                             24

  8   Relation of Reaction Rate and
      Sulfate Loading                        26
2-1   Comparison of Reactivity of
      Calcines with S02 at 980°C             38

  2   Sorption of S02 by Different
      Particle Sizes                         40

  3   Estimation of Initial Rate and
      Effectiveness Factor                   41

  4   Comparison of Reactivity with
      B.E.T. Surface Area                    49

  5   Reaction Rate at Constant Sulfation
      as a Function of Particle Diameter     52

  6   Reaction Rate Versus Surface Area      55

  7   Effect of Mode of Calcination
      on Reactivity                          60

  8   Relationship between Effectiveness
      Factor and Pore Size                   65

-------
3-1   Boiler Plan and Sample Locations       75
  2   Reaction of SO, with Calcined
      Limestone                              78
  3   Rate of Hydration of Calcined
      Dolomite                               80
4-1   Typical Firing Cycle                   94
  2   Acid Titration of #2061                97
  3   Surface Area Versus Flue Gas
      Sorption                               99
  4   Pore Volume Versus Flue Gas
      Sorption                              100
  5   Small Pore Volume Versus Flue
      Gas Sorption                          101
  6   Small Pore Volume Versus S02
      Capacity                              102
  7   Flue Gas Sorption Versus S02
      Capacity                              103
  8   C02 Test:  Effect of Reaction
      Temperature                           105
  9   S02 Test:  Effect of Reaction
      Temperature                           106
  10  C02 Test:  Effect of Reaction
      Time                                  107
  11  S02 Test:  Effect of Reaction
      Time                                  108
  12  Slaking Test Main Effects             109
  13  COn Test:  Comparison of Measured
      Reactivities for Sulfated and
      Unsulfated Limes                      111
  14  Reactivities of Sulfated and
      Unsulfated Limes Measured by
      Hydration Weight Gain                 113
                        Vi

-------
5-1   Flow Diagram for Fixed Bed
      Reactor System                        117

  2   Reaction Temperature and Form
      of the Reactant                       120

  3   Frequency Distribution of
      Capacities of 86 Stones Tested        120
                      vii

-------
                 LIST OF TABLES
1-1   Properties of Calcines                 12
  2   Kinetic Parameters for Sorption
      by Calcined Limestones                 22
2-1   Stone Identification and
      Description                            32

  2   Chemical Analyses                      34

  3   Kinetic Parameters for S0£
      Sorption by Calcines                   39

  4   Parameters for SC^ Sorption by
      Calcines as a Function of Particle
      Size                                   43

  5   Summary of Physical Properties
      of Calcines                            46

  6   Physical Properties of Raw
      Limestones                             50

  7   Effect of Calcination Temperature
      on Physical Properties of Calcite
      Spar                                   57

  8   Comparison of Reactivities of
      150/170 Mesh Stones at 980°C           59
3-1   Composition of Additives               73
  2   Sulfate and Carbonate Content of
      Boiler Samples                         81
  3   Specific Volume of Pores and
      Degree of Sulfation                    84
  4   Physical Properties of Boiler
      Samples as Determined by Oil
      Absorption                             86
4-1   Composition of Limestones              93
  2   Correlation Coefficients               98
                       viil

-------
5-1   Effect of Calcination                 122

  2   Correlations between Physical
      Properties and 980°C Capacities       125
  3   Capacity of Various Classes of
      Limestone                             128
                      IX

-------
                         A.  SUMMARY









Concentrated  on limestone-sulfur dioxide (S07) reactions,




alkaline additives research has defined the kinetics and capacity




as well as the effects of everburning.  The kinetics of S09




sorption by limestone calcines has been determined with a high




temperature differential bed gas solids contactor.  It has been




shown that for small particles the reaction is first order and




chemical rate controlled with activation energies in the range




of 8.1 to 18.1 k cal/g. mole.  In general, reaction occurs




initially throughout the particle volume and the internal dif-




fusion resistances become limiting only after conversion of at




least 20% calcium oxide.  The reaction rate was weakly dependent




on particle size and strongly depressed by sulfate loading.




For  average particle diameters less  than 500 microns, the rate




was essentially independent of particle size.  The effect of




increasing conversion on rate is explained as an exponential




relationship between the frequency factor and sulfate loading.









Experiments on a wide range of carbonate rocks using the dif-




ferential reactor confirm £he importance of chemical reaction




rate controlling for S02 sorption.  Furthermore, the strong




influence of physical properties on limestone reactivity has




been delineated.  Small pores and large surface areas lead to




high reaction rates while large pores lead to high capacity.

-------
Pores smaller  than 0.1 micron plug rapidly with, reaction products




and are apparently responsible for shell formation on certain




stones.  The results- confirm the exponential decay of reaction




 rate during the  chemical reaction-controlled phase, which




generally extends from 50 to 90 percent conversion of CaO in




95 micron particles  at 980°C.  Stones of varying geological




type yield calcines  of distinctly different physical structures




which show correspondingly large differences in both rate of




reaction and total capacity for S0_ sorption.









The loss of reactivity due to everburning has been studied




using an oil fired boiler belonging to the Florida Power Cor-




poration (in St. Petersburg),  During these field tests, samples




of lime were collected from the flue gas at various points in




the boiler and analyzed to determine the degree of calcination,




extent of sulfation,  and changes in physical properties related




to chemical reactivity.  The boiler samples were found to be




considerably less reactive with S0_ than stone calcined in the




laboratory.  The low porosity and high density of the injected




lime indicates that  everburning is at least partly responsible




for the low boiler desulfurization achieved when additives are




injected with fuel.









The effects of everburning for calcination temperatures ranging




from 1700 to 3200°F  were determined by measuring more than a




dozen chemical and physical properties of these calcines.  While

-------
density increased, other reactivities (e.g., SO- sorption,  CO-




sorption, extent of slaking), pore volume, and surface area




decreased markedly with calcination temperature.  Loss of




reactivity is as high as 80% for sorption of SO- from flue gas




in 120 seconds when the calcination temperature is raised from




1700°F to 2600°F.  Moreover, a strong intercorrelation was




noted between S0? reactivities, surface area and pore volume.




This supports the conclusions that (1) reactivity is a function




of surface area and pore volume and (2) loss of reactivity (or




dead-burning) is attributable to growth of CaO crystals which




decreases the surface area and pore volume.  The best test for




determining the loss of reactivity of partially sulfated lime




involves measuring the weight gain of hydrated samples.









Eighty-six carbonate rock samples were tested in a fixed-bed




reactor to determine their capacity to react with flue gas




containing sulfur dioxide.  Although most of the work was




performed with the carbonate and the oxide at standard test




conditions, supplementary tests were made on hydrates, oxides




and carbonates over a wide range of reaction temperatures and




calcination conditions.  At 980°C the average sulfur loading




of the carbonate feed was -practically equivalent to the loading




for the precalcined samples.  Differences among the samples




were only slightly related to chemical composition; porosity.




as measured by mercury pore volume, best explained variations




in capacities of the samples.  Chalks and oolitic samples were

-------
 the most efficient absorbents, and magnesite and Iceland spar,




 the least efficient of the stones tested.








 In addition to limestone as a material for sorbing SO., other




materials have also been considered.  Interest had been shown




in kiln dust from Portland cement manufacturing as a SO. sorbent.




Tests conducted with kiln dust samples exposed to flue gas in a




differential reactor and an aqueous batch scrubber indicated




that kiln dust has absorptivity and reactivity comparable to




limestone, but no better.  Metal oxides such as cupric oxide




have also been suggested as potential SO. sorbents.  Thermo-




gravimetric analysis (T6A) of a sample of cupric oxide exposed




to pure SO. had a maximum weight gain of 48 weight percent at




738°C.  However, at the temperature of interest (315°C) the




cupric oxide had only a 2-3 percent weight gain.  Finally,




the reactivity of char with sulfur oxides within the Westvaco




sorption-regeneration scheme has been studied.  At the time




ten percent of the feed SO. concentration broke through the




bed of char, loading on the char was 15.8 wt. percent SO..




Loaded S0« in the form of H.SO, was regenerable with H_S to




form sulfur.

-------
                  B.  GENERAL INTRODUCTION







Limestone processes for controlling sulfur oxides, especially



the dry limestone injection process, have long been of great



interest because of their simplicity and low cost.  The Division



of Process Control Engineering at an early date selected the



dry limestone injection process for development and large scale



demonstration of optimum performance on a power boiler unit of



the 150 MW class.  This is in accord with a report of the


                               (26)*
National Academy of Engineering     , which recommended the



development of throw-away processes for S0~ control on the



highest priority basis.  In support of this goal, the activities



of the Process Research Section have been directed toward specific


problem areas relating to the limestone processes, with major



emphasis over the past two and one-half years on acquiring



information on the dry injection process which will be required



to apply the process on a wide scale.  Little was known about



the mechanism and rate of reaction between limestone and sulfur



oxides, nor had the differences between limestones been related



to their potential reactivity.  That this information was vital



for optimization of the process can easily be seen.  The lime-



stone injected into a power boiler must calcine Cevolve CO_ to



become lime) and react with most of the sulfur oxides present -



all within two seconds or less.  If the limestone is injected
*  Refers to bibliography listed under G._ REFERENCES.

-------
too close to  the fireball, the lime produced will be dead-burned




and become unreactive.  If the limestone is injected higher or




farther from  the fireball, the residence time for the lime




particles in  the zone of the boiler favoring reaction with




sulfur oxides will be seriously reduced.  Hence, an understanding




of the kinetics of lime-SO_ reactions and the characteristics




of dead-burning were necessary.  In addition, data on the




capacity of limestones was of interest to evaluate the maximum




potential of limestones to absorb sulfur oxides, and to identify




those unsuitable for injection due to markedly low capacity.








From the general objectives of defining the capacity and kinetics




of limestone-SO_ reactions as well as determining the nature of




everburning, there evolved five discrete investigations:




     1.  Kinetics of the Reactions of SO- with Calcined Limestone




     2.  Physical Properties of Carbonate Rocks Related to SO-




         Reactivity




     3.  A Field Study of the Role of Overburning




     4.  Methods for Testing the Degree of Overburning of




         Calcined Limestones




     5.  Capacity of Limestones for Sorption of S0~








The first of these was undertaken to determine the mechanism




and rate of reaction of limestones after calcination under




standardized conditions.  Effects of reaction temperature,

-------
particle size, and extent of reaction on the reaction rate were




investigated.  The second study characterized ten diverse types




of naturally occurring carbonate rocks for their S09 reactivity.




It was the goal of this study to identify those physical or




petrographic properties (such as pore size, surface area and




crystal structure) which contribute to greatest S0« reactivity,




and to identify those types of limestones which perform most




effectively as sorbents for sulfur dioxide.








For the field study of everburning, limestone was injected with




the fuel in an oil-fired furnace.  In so doing, the limestone




was given the maximum solids residence time and high tempera-




tures for fast calcination and fast reaction.  If SO- removal




efficiency were to be poor, it would be the result of calcination




at excessively high temperatures leading to dead-burning.  Other




laboratory tests were made to confirm everburning of the lime-




stones.  To develop a test for dead-burning in the presence of




fly ash as well as partial sulfation, which could be used to




quantify the influence of the effect of dead-burning in full




scale evaluation of the process, a fourth study was initiated.




Not only was a dead-burning test desired, but the nature of




dead-burning itself in terms of physical and chemical property




changes needed to be defined.

-------
The last Of the above listed investigations was responsible for




determining the ability of a large number of limestones to




achieve a high absorption efficiency.  Thus, a comprehensive




limestone inventory was established and a ranking within this




inventory as to SO. reaction capacity was identified so that




selective recommendation of limestones would be possible.








Other materials besides limestone have been studied as sorbents




for controlling sulfur oxides.  Among them are kiln dust, copper




oxides, and char.  In general, these investigations were designed




to test the feasibility of proposed processes and were not intended




to be in depth experimental programs.  Brief discussions of the




results may be found in the appendices.  The kiln dust tests




are discussed in Appendix D; the copper oxide study, in Appendix E,




and the char, in Appendix F.

-------
                 C.  DISCUSSION OF PROJECTS
In this section are given individual detailed discussions of




each of five project areas which have been investigated during




the past two and one-half years (July 1968 to December 1970)




by the Process Research Section.  The intent of each of these




projects as discussed fully in the general introduction was




to delineate some aspect of limestone reactivity with respect




to SO- sorption.  Each of the five parts of this section




identifies its own' objectives and conclusions as well as the




details of the experimental procedure and results.  General




conclusions drawn from the whole of the experimental effort




may be found in the section immediately following.

-------
                              10
1.  Kinetics of the Reaction of SO, with Calcined Limestone
a.  Introduc tion



Processes in which limestone and dolomite are used to desulfurize



flue gas are being intensively investigated under the sponsor-



ship of the Air Pollution Control Office.  Such processes include



the dry injection of pulverized stone into boiler furnaces, the



wet scrubbing of flue gases and combinations of dry injection



and wet scrubbing, and the use of fluidized bed contactors,



fluid bed combustion, and thin fixed beds.  An understanding of



the chemical reactions associated with these processes is desirable



to obtain most efficient performance when such process is applied.







It is generally assumed that limestone absorbs SO~ by a mechanism



involving two consecutive steps - dissociation of the calcium



carbonate, followed by reaction of CaO with sulfur dioxide.  It



is expected that the rate of the second step will be important



in any of the proposed pollution control processes, and especially



in the dry injection process    .  Several investigators   '   '



have determined the saturation capacities of a large number of



naturally occurring limestones and dolomites under various



conditions of reaction with SO^.  Other studies under way will



define the rate of reaction of uncalcined limestones in the


              (8)
disperse phase   .  The purpose of the investigation reported



here was to determine the rate of reaction of limestones after



calcination under standardized conditions.

-------
                             11


b.  Experimental

Bench scale experiments were carried out to measure the rates

of reaction for several selected limestone calcines when ex-

posed to a fly ash-free flue gas of controlled composition.

Table 1-1 shows the chemical composition and primary physical

characteristics of the calcined stones used in this work.

Complete geological descriptions of the stones are given in a

separate report    .  Calcination was carried out in 180 g.

batches in an Inconel kiln 12.5 cm. long and 8 cm. in diameter,

rotated at 1 r.p.m.  It was heated to 980°C. in an electric

furnace and then charged with 10/28 mesh stone.  The kiln was

maintained at 980°C. and purged with air for 2 hours to remove

C02 during calcination.  Conversion to the oxide was complete

under these conditions (C0_ < 0.5%).  The calcined stone was

cooled, crushed, and screened into size ranges of 12/16, 42/65,

and 150/170 mesh (Tyler).  The calcined samples were stored

in airtight containers until used.




The rate of reaction with SO- was determined in a differential
                           I
reactor (Figure 1-1) constructed of Inconel alloy.  In this

type of reactor the thin layer of solid and high gas flow

prevent gas-phase concentration gradients around the reacting

solid.  The gases, preheated in an outer annular section, pass

downward through the inner reactor tube about 73 cm. long and

3.42 cm. in inside diameter.  The Sample is supported on a

30-mesh Inconel screen in a removable carrier.  The carrier

-------
                                  12
                  Table 1-1.  Properties of Calcines


                                                     Bulk  .            b
                                                   density,   Porosity,
Stone    LOI %a  %CaO    %MgO    %Fe2°3    %Si02     g./cc.    cc./cc.

1337     47.4     55      43      0.33     0.92      1.41       0.58

1351     42.4     54      28.5    7.0      8.2       1.59       0.56

1343     42.8     94       0.8    0.66     2.98      1.88       0.45

1360     43.8     81      13.0    1.25     3.65      1.51       0.56



     a  Weight loss on calcination at 980°C.

        150/170 mesh particle size.

-------
                           13
      FLUE GAS
                                                 :-—N2 GAS
       TO INFRARED


          TO ORSAT
     Figure  1-1.  Differential Reactor    .

1. Teflon solenoid valve 2. Heating tapes 3. Thermometer 4. Preheat
lection 5. Thermocouple 6. Sample 7. Reactor tube 8. Sample carrier 9.
Heating furnace 10. N» Purge exhaust 11. Alundum filter 12. Constant-
temperature oven 13. Orifice 14. Manometer 15. Flow control valve

-------
                             14
 is  sealed against a flange in the center of the reactor tube
 so  that the entire gas flow passes through the solid during
 exposure.


 In  this investigation a sample consisted of 30 mg. of calcined
 stone, distributed uniformly over the 2.65 cm. diameter screen.
 For small particle sizes a disk of woven refractory fabric was
 placed on the screen and a 1 cm. thickness of refractory
 (fused quartz) gauze was placed on the fabric.  The lime
 particles were dispersed into the gauze.


 The mass flow rate of gas through the screen was maintained
 constant at 0.075 g./(sq. cm.)(sec.), which at 870°C. corre-
 sponds to a superficial velocity of 240 cm. per sec.  A high
 gas velocity reduced gas film resistance to a negligible
value, so that mass transfer to the particle surface did not
affect rate measurements.  The gas fed to the reactor was a
flue gas generated by combustion of fuel oil containing carbon
disulfide.  The composition of the flue gas was 10.5% CO.,
3.4% 02, 9.9% H20, 0.27% S02> 0.003% S03, and 75.9% N£ by
volume.  The sulfur dioxide concentration was monitored con-
 tinuously with an infrared analyzer after removal of the
water vapor from the sample stream.


The reactor was mounted in an electric furnace containing
 three heating sections.  The center section was energized

-------
                             15
by a proportional controller acting on a thermocouple located




3.4 cm. above the screen supporting the lime sample.  The other




two sections were equipped with variable transformers set by




thermocouples in the top and bottom of the reactor tube to




maintain a uniform temperature over the full length of the




reactor and preheater assembly.  The thermocouples were cali-




brated in situ against a multiple shielded, high-velocity




thermocouple.  A multipoint recorder continuously monitored




reactor temperatures.









Before a run was started, the carrier and sample were allowed




to heat up for 5 minutes to the reactor temperature.  The




time of exposure of the solid to the gas stream was controlled




by solenoid valves that started the gas flow at the beginning




of the run and purged the reactor with nitrogen at the end of




the run.  The sample was removed from the carrier after expo-




sure (along with the refractory gauze, if used) and analyzed




for sulfate.  The exposed sample was dissolved in water by




soaking with ion exchange resin, filtered, and the filtrate




titrated in 80% isopropyl alcohol with barium perchlorate




using thorin indicator.









c.  Discussion of Results




The chemical reaction between limestone and sulfur dioxide at




high temperature in the presence of excess oxygen is:




                CaO 4- S02 + 1/2 02  ^ CaS(>4              (1)

-------
                             16
Equilibrium data for the reaction have been summarized by


                                          (34)
researchers at Battelle Memorial Institutev   .   The reaction




proceeds to the right at temperatures up to 1230°C. at partial



pressures of SO- corresponding to flue gas concentrations of



about 3000 p.p.m.  Equilibrium would also permit the MgO com-




ponent of dolomite to react at temperatures below 840°C., and




it has been reported to participate in the reaction with SO^



                                        (35)
in the presence of iron oxide impurities    .  More recent


              (4)
investigations    have shown that SO- reacts only slightly




with MgO at 430° to 700°C. in a fluidized bed, reaction




occurring preferentially with calcium even when Fe2°3 content



is as high as 7%.  Calcined magnesite (MgCCL)  and calcined bru-




cite [Mg(OH)_] were shown to have low capacities for sorption


                                                           (2)
of sulfur dioxide.  Dolomites injected into a pilot furnace




have been examined by X-ray diffraction analyses, but showed




no MgSO, as a reaction product, although small amounts were



detectable by DTA methods.  No distinction is made in this




study between CaO and MgO, although the data support the con-




clusion that only CaO has a significant reaction rate under the



conditions investigated.
When reaction 1 takes place in flue gas containing high concen-



trations of carbon dioxide, equilibrium also favors a competing



reaction(34) below 770°C.:



                     CaO + C02 ,=&: CaC03                    (2)



Typical experimental results for  the sorption of S02 are shown



in Figure 1-2; the mg. of S03 found in 150/170-mesh particles

-------
                             17

                        Figure  1-2
                            60       80

                           TIME, seconds
                                           100     120
          Sorption of sulfur dioxide by dolomite 1351  at various
reactor temperatures

SO» concentration - 3000 p.p.m. (dry basis), particle size  = 150/170
                       Figure  1-3
   10
                            60      80

                            TIME, seconds
           Sorption of sulfur dioxide by different particle sizes of
 dolomite 1351

   SO> concentration - 3000 p.p.m. (dry basis), temperature = 870° C.

-------
                             18
 (D  = 0.0096 cm.) after reaction is plotted against exposure
 time at various reaction temperatures.  Figure 1-3 shows a simi-
 lar plot for different particle sizes at a reaction temperature
 of 870°C.  Total conversion of the CaO in this dolomite would
 correspond to an ordinate value of 23.2 mg.  These figures
 illustrate the strong sensitivity of the reaction to tempera-
 ture and the surprisingly low sensitivity to particle size,
which were characteristic of all the stones examined.


The rate of sorption was measured as the tangent to the smooth
curve drawn through the data and is defined as:
                         r -
                         r ~ w  dt
where W is the grams of calcined stone exposed in the reactor,
and n' is the gram moles of SO- in the stone at time t.  (Note:
Nomenclature is summarized under Subsection e. at the end of
this section) .
The data were correlated according to the rate expression for
                                   (31)
chemical reaction in a porous solid     :
since dn'/dt = -dn/dt

                        W  dt = p~ kvC
The effect of SO- concentration, C, on the rate of sorption is
shown in Figure 1-4.  The SO- concentration was varied between

-------
    0.03
                       Figure 1-4
                                       SLOPE, m  1.088
      0.02          0.1                  1

                    S02 CONCENTRATION 
-------
                              20
 58 and 6000 p.p.m. by changing the carbon disulfide content



 of the fuel oil burned in the furnace.  The rate was measured



 at a conversion of 10.5% of the CaO in 12/16-mesh stone reacted



 at 870°C.  The line fitted to these data by the method of least



 mean squares has a slope of 1.008 or m = 1, indicating that



 the reaction is first order with respect to the concentration



 of SO- in the gas phase.  For the remainder of the experimental



 work, which is reported below, the SO^ concentration was fixed



 at 3000 p.p.m. (dry basis) or 2.9 X 10   g. moles/cc. (wet



 basis, 870°C., 1 atm.).







 The rate constant k  is a function of temperature and also some



 function of nf/W, the sulfate loading; it decreases as the reaction



 progresses and the solid reactant is consumed.  The temperature



 dependency was correlated by the Arrhenius equation:



                         ,     . -E/RT                      ,,,
                         kv = Ae                           (6)



An Arrhenius plot for each of the four calcined stones is shown



 in Figure 1-5 for reactions with S02 at temperatures between



 650° and 980°C.  The rates were measured at a sulfate loading


         _q

 of 1 X 10   g. mole/g. of 150/170-mesh particle size sample.



 This loading corresponds to approximately 10% conversion of the



 CaO.  The data show a linear correlation between log r and 1/T,



 as specified by Equations 5 and 6.  The apparent activation



 energy determined from the slope of these plots was distinctly



 different for each stone, ranging from 8.1 to 18.1 k cal/g. mole.

-------
                             21
When rates were measured at higher CaO conversions - up to 20% -




the plots shifted toward the abscissa, but remained parallel




to the lines shown in Figure 1-5, thus indicating no significant




change in the activation energy.









The high sensitivity of the rate to temperature suggests chemical




reaction to be the predominant rate-controlling resistance during




the initial period of S0_ sorption by small particles.  The




apparent activation energy for sorption, controlled solely by




bulk diffusion, would be only 3.4 k cal./g. mole.  A summary




of the empirical kinetic parameters estimated for the four stones




is given in Table 1-2.









The data in Figure 1-5 further indicate that reduction in rate




of reaction with S02 as a result of competition with C02 was




not important at 650°C.  At 540°C., however, there was evidence




that reaction 2 was significant and the Arrhenius plots could




not be extrapolated to that temperature.  The plots also failed




when the reaction temperature was raised from 980° to 1100°C.,




sorption rates decreasing at the higher temperature.  Subsequent




experiments in which the calcine was heated for 10 minutes at




1100°C. and then reacted at 870°C. showed the same difference




in rate when compared to a sample which was not exposed to the




high temperature.  It was concluded that the loss of reactivity




was due to the changes in porosity and bulk density which occur




when lime is "hard-burned"^  .

-------
                                  22
             Table 1-2.  Kinetic Parameters  for Sorption
                     by Calcined Limestones

                                                          *
            Activation        Reaction rate        Frequency
             energy, E,        constant, k ,         factor.  A,
Stone      cal./g. mole          sec.~^   V          sec.~l


1337        10,000            4.8 X 103            2.4 X 105

1351        18,100            7.2 X 103            9.0 X 106

1343        14,200            4.0 X 103            1.1 X 106

1360         8,100            2.3 X 1Q3            5.5 X 10^
     a                                        —3
        Evaluated at sulfate loading of 1 X 10   g.  moles/g.,
        150/170 mesh particle size; 980°C.

-------
                             23
To ensure that gas film diffusion was not influencing the rate




measurements, a supplementary experiment was made with a 0.95-cm.




diameter sample carrier which permitted exposure of the solid




at high gas velocity.  Samples of dolomite 1337 (D  = 0.13 cm.)




were exposed for 2 minutes at 980°C.  The amount of sulfate




found in the stone was only 5.3% greater at 2400 cm./sec. gas




velocity than found at 225 cm./sec.  The negligible effect of




gas velocity verified that the rate of sorption was not limited




by mass transfer to the solid surface.









The effect of particle size on reaction rate is shown in Figure 1-6,




in which the value of r at a sulfate loading of 2 X 10   g. mole/g.




is plotted against the inverse of particle diameter.  If the




particles are assumed to be spherical, the total exterior surface




of a given mass of stone (specific surface) would increase with




1/D  when the particle size is reduced.  If the reaction were




occurring only at the outer surface, the plot shown in Figure 1-6




would be expected to be a straight line through the origin.  It




is clear that the rate was not proportional to specific surface




and for fine particles was essentially independent of it.  The




results suggest that some reaction takes place within the interior




structure of the solid and that the relative importance of the




internal reaction becomes greater as the particle size decreases.




These observations are similar to the effects associated with




highly porous catalysts and are consistent with the fact that




the pore space in calcined limestones usually accounts for 50%

-------
 s
 5:   3
 o
 <
 ut
                               24
                         Figure  1-6
                             MESH SIZE
        12/16        42/65
                                               1337
              20        40        60        80       100       120
                          Reaction rate vs. l/BP
              Sulfate loading = 2 X 10-' g. moles/g.
   SO
Ul
3
8
&
   0.5
                        Figure  1-7
           1351
          Dp * 0.0096 cm
     S   10
                         100              1.000
                           TIME, seconds
10,000
             Sorption of sulfur dioxide by calcined limestones

-------
                             25
or more of the total volume of the particles.  The porosities



(or fraction of particle volume that is pore space) for the



four calcined stones are given in Table 1-1.







Figure 1-7 shows the conversion versus time response for dif-



ferent particle sizes of stones 1351 and 1343 over long periods



of exposure at 870°C.  An analysis of the response according to


                            (32)
the method of Shen and Smith     was made to test for intra-



particle diffusion.  This model is based on diffusion through



the product crust as the rate limiting mechanism and a nonporous



solid reactant in which reaction occurs only at the interface



of the unreacted core.  The data from this investigation could



not be correlated by the shell diffusion model given by Shen



and Smith's Equation 31.  Figure 1-6 shows that the model for



shell diffusion, which predicts that the rate of sorption at


                                             _ £
a given sulfate loading will increase with 1/D  , is clearly



inconsistent with data on particle size versus reaction rate.







The data could be correlated empirically by a plot of log r



against sulfate loading, n'/W, as shown in Figure 1-8.  The



data were linear for all particle sizes of each of the four



stones examined at the reaction temperature of 870°C.  The



observed response can be interpreted in terms of a change in



the frequency factor, A, of Equation 5.  The frequency factor,



which relates the reaction rate to the number of molecular

-------
                            26
                        Figure 1-8
 1C
 a
 I
                     Dp =0.13
                                              Op = 0.0096
                                    Op = 0.051


      0   12     34     5
                                         7    8    9    10
1337
        Relation of reaction rate and sulfate loading for dolomite

-------
                             27
collisions occurring per unit volume per unit time,  is



dependent upon the amount of SO- and the amount of unreacted



CaO present (as well as other factors such as surface effects



and probability of reaction of the molecules) .   As the reaction



progresses and CaO is consumed, the frequency factor will be



expected to decrease in some manner related to the amount of



sulfate formed.  If r  and A  are the rate and frequency factor
                     o      o                     -i    ./



at zero sulfate loading, then, from Equations 3, 5,  and 6,




                     nAoC   -E/RT

                     ~-
designating the slope of the straight line fitted to the data



in Figure 1-8 as 3, the equation for the rate of reaction can



be written as a function of sulfate loading:




              ,   n AC     E     „ n1 , r                  /QN

              lo* -^ --- RT = ^ W  + 10g ro             (8)




Equations 7 and 8 reduce to





                         A = A e-en'/W                     (9)
                              o




The observed relationship between rate and sulfate loading thus



indicates an exponential decrease in the frequency factor.  The



data of Figure 1-8 show that coefficient 3 is a function of



particle size, the rate of reaction of large particles being



more sensitive to sulfate loading than small particles.  Esti-



mates of the effectiveness factor, obtained by extrapolation



of the straight lines of Figure 1-8 to zero sulfation, indicate

-------
                             28
n - 1 for particles smaller than 0.05 cm. as shown by the


approximately equal intercept of the two lines on the ordinate.


The lower intercept for 5  =0.13 cm. indicates n < 1> implying
                         P

the presence of pore diffusion resistance of significant magnitude


for the larger size particles.





The results of this study show that the rate of sorption of S02


by calcined limestones is dependent to a very large extent upon


the kinetics of the chemical reaction, particularly at small


particle sizes, and that the rate of reaction predominates as


the overall rate-controlling resistance for conversion of at


least the first 20% of the CaO.





d.  Conclusions


Differential reactor techniques were used to measure the rate


of reaction of sulfur dioxide with four natural specimens of


limestone, after calcination at standardized conditions.  The


rates were measured as a function of S0« concentration, particle


size, and CaO conversion between 540° and 1100°C.  A first-


order chemical reaction was the predominant resistance limiting


the rate of sorption of SO^ by small particles.  The activation


energy was dependent upon the type of stone, and ranged between


8.1 and 18.1 k cal./g. mole.  The rate was essentially independent


of particle size for D  < 0.05 cm.  Reaction occurs initially


throughout the particle volume under the isothermal reaction

-------
                             29
conditions studied and the internal diffusion resistances become

limiting only after conversion of at least 20% CaO.  The reaction

rate decreased rapidly with increasing conversion, explained by

an exponential relationship between the frequency factor and the

sulfate loading.



e.  Nomenclature

    A     = frequency factor, sec.  (g. moles/cc.)

    A     = frequency factor at zero solid conversion,
            sec.~-'-(g. moles/cc.)m~

    C     = gas phase concentration of sulfur dioxide, g. moles/cc.

    D     = mean particle diameter, cm.

    E     = activation energy, cal./g. mole

    k     = reaction rate constant per unit volume of solid,
     V      sec. ~1(g. moles/cc.)m~1
    m
= order of reaction with respect to sulfur dioxide
    dn/dt = rate of change of S0_ in the gas phase, g. moles/sec.

    nf    = sulfate in solid as S03, g. moles

    R     = gas constant, 1.987 (cal./g. mole°K.)

    r     = rate of formation of SO, in solid, g. moles/(g. sec.)

    r     = rate at zero solid conversion, g. moles/(g. sec.)
     o

    V     = volume of solid, including intraparticle pores, cc.

    W     = weight of solid sample, g.

    T     = temperature, °K.

    t     = time, sec.

    g     = empirical correlation coefficient defined by Equation 9

-------
                   30
= effectiveness factor, ratio of reaction rate to
  the rate that would be obtained if entire volume
  of particle participated equally in reaction

= bulk (particle) density of solid, g./cc.

-------
                             31
2.  Properties of Carbonate Rocks Related to SO,, Reactivity
                                 __—          Z








a.  Introduction




This report summarizes an experimental study of the S0_ sorption




characteristics of ten diverse types of carbonate rocks (lime-




stones, dolomites, magnesites, etc.).  Specimens representative




of a broad spectrum of naturally occurring stones were selected




for detailed petrographic, mineralogical and chemical examin-




ations under Contract CPA 22-69-65 with the Illinois State Geo-




logical Survey (ISGS).  It is the goal of that contract to




establish criteria, based on petrographic and mineralogical




properties, for selecting carbonate rocks of greatest potential




for the desulfurization of stack gases.  Tests conducted by




APCO, to evaluate the isothermal rates of reaction of calcines




prepared from these stones are reported here.  The rate of




reaction of the raw stone under non-isothermal conditions will




be determined by Battelle Memorial Institute under Contract




PH 86-67-115 using a dispersed phase reactor.









A brief description of the rock types selected for this study




is shown in Table 2-1.  This was taken from the ISGS Quarterly




Report, dated December 4, 1969.  Two additional stones were




also tested which are not included in that report:  Stone No.




2129 (ISGS Type 11), a Michigan marl, which is an unconsolidated




form of CaCO,, and Stone No. 1336 (ISGS Type 10), which is a

-------
                              32

                         Table 2-1


            Stone Identification and Description
Type 1.  Calcite, variety Iceland spar, transparent, colorless
         and optically near perfect; cleavage rhombs up to 1 inch.

Type 2.  Calcite, very coarse calcite spar, translucent milky in
         color due to abundance of crystalline imperfections;
         cleavage rhombs up to 1 inch.

Type 3.  Calcite limestone, with few scattered fine-grained
         dolomite rhombs; coarse unequant granular and porous;
         consists of recrystallized crinoid fragments and a few
         finely recrystallized bryozoans.  Gray.

Type 4.  Calcitic limestone, very fine, equant granular, and
         dense; light brownish gray.

Type:5.  High purity dolomite, medium-grained, granular and
         porous; gray reef type of dolomite.

Type 6.  81 percent dolomite, contains clay and fine quartz silt
         impurities; medium-grained, granular and microporous;
         buff color, non reef type of dolomite.

Type 7.  Magnesite, high purity, very fine, equant granular and
         microporous; near white.

Type 8.  Aragonite sand, contains traces of magnesium calcite;
         consists of oolitic and concentrically banded fossiliferous
         sand-sized particles.  The particles have a banded
         structure and consist mostly of fibers of aragonite.

Type 9.  Calcitic dolomite (70% dolomite, 18% calcite) with
         limonite along grain boundaries and a few scattered
         particles of chert; fine-grained, moderately equigranular,
         and microporous; mixture of brown and gray colored stone.

-------
                             33
marble.  These two stones were tested because of the unusually



high reactivity found for the marl in pilot scale coal-fired



furnace tests conducted by Babcock & Wilcox (Contract PH 86-67-127)



and the contrastingly low reactivity found for the marble in



similar tests.  Chemical compositions of all stones are reproduced



in Table 2-2.






b.  Experimental



The reactivity of the eleven types of limestone was measured



and correlated by the methods outlined in the previous section



(C 1 b.) and in the paper, "Kinetics of the Reaction of SO-



with Calcined Limestone"   .  Calcines prepared in a small



rotary kiln  (2 hours at 980°C.) were crushed and screened



into 3 particle size ranges: 12/16 mesh (5  = 0.13 cm.),



42/65 mesh (D  = 0.025 cm.) and 150/170 mesh (D  = 0.0096 cm.).
             P                                 P

Thirty-milligram samples of the calcines were exposed for varying



lengths of time to flue gas containing 3000 p.p.m. SO-, dry


                 8            3
basis (2.63 X 10   g. mole/cm. , wet basis, 980°C.).  Using



the smallest particle size, reaction rates were determined at



reactor temperatures of 650°C. (1200°F), 760°C. (1400°F), and



870°C. (1800°F).  The rates were evaluated and compared at a

                                                 _o

sulfate loading of 2.5 mg. SO-/30 mg. calcine (10   g. mole/g.).



The sorption curves and Arrhenius plots for these data are given



in Appendix A.

-------
                                                      Table 2-2
                                         Chemical Analyses in Weight Percent
ISGS Stone Type
APCO Stone No.
Mineralogical
Type
SiO,
Ti02
A1203
Fe203
FeO
MnO
MgO
CaO
Na20
K20
P 0
C02
S03
SrO
Cl
C (organic)
H20 (100°C)
Ign. Loss
1 2
2201 2202
Iceland calcite
spar spar
ND
ND
ND
ND
.13
ND
ND
55.3
0.03
0.02
b
trace
43.95
0.01
0.14
ND
_
-
43.49
ND
ND
0.01
0,19°
—
0.06
ND
55.5
0.07
0.02
^ ,
43.35
0.17
0.002
0.04
_
-
43.15
3
2203
calcite
ND
ND
ND
0,20C
—
0.10
1.86
53.4
0.08
0.02
—
43.75
0.20
0.009
0.03
—
-
43.67
4
2204
calcite
1.53
ND
0.01
0.31C
—
0.09
0.00
54.8
0.06
0.04
w
43.35
0.15
0.019
ND
-
-
43.15
5
2205
dolomite
0.03
ND
0.02
0.34C
—
0.02
21.40
30.30
0.10
0.03
fm
47.30
0.13
0.019
0.09
—
-
47.24
6
2206
dolomite
11.8
0.02
1.77
0.13
0.41
0.02
17.4
26.5
0.06
0.90
0.02
40.27
0.03
0.04
trace
—
j
40.46d
7
2207
magnesite
0.47
ND
0.08
ND
0.07
ND
44.2
2.93
0.03
0.03
trace
50.96
0.01
0.01
ND
—
-
51.56s
8 9 10 11
2208 1701f 1336 2129
calcitic
aragonite dolomite marble marl
0.19
ND
0.27
ND
0.01
ND
ND
55.2
0.29
0.03
0.01
42.10
0.37
0.10
0.24
-
-
43.33
5.88
0.15
0.69
2.82
1.75
0.21
15.33
30.82
0.28
0.22
0.10
40.68
0.42
0.04
ND
-
-
41.85
0.85 3.63
0.05
0.20 0.95
0.15 ND
-
0.01
1.4 ND
53.7 46.6
0.14
0.18
0.67
43.4 37!l8 £
0.50
-
- -
5.60
4.04
43.4 47.06
  ND - not detected:  Limits of detection for Si02, 0.03; Ti02 and MnO, 0.01; A1203, 0.05; Fe203, 0.01;
                      MgO, 0.10; Cl, 0.02.
b Trace of P20tj is approximately 0.005%.
c Percentage of total iron expressed as Fe203.
d Includes 0.30% H20+.
e Includes 0.61% H20+.
^ Stone No. 1701 came from the same source as Stone No. 1351, studied by various investigators.

-------
                             35




Previous data were correlated by the expression



                     1  dn'   kv „
Because of inaccuracies involved in measuring particle densities,


P , of finely divided calcines, the rate data are correlated in


this report on the basis of total surface area, S , which is
                                                 O

more easily and accurately determined.  Since k  = k  p  S ,
                                               v    s Hp  g'
where k  is the reaction rate constant per unit surface with
       S

units of cm. /sec.  Values of S  were obtained from B.E.T.
                              g
                                     2
measurements and are expressed in cm. /gm.  Equation (2) is


the simplified model used for data correlation in this report.
The effectiveness factor is used here to represent the degree


to which reaction occurs within the internal structure of the


solid.  The maximum value of n = 1 indicates that reaction occurs


equally throughout the internal pore structure.  Low effectiveness


factors, n « 1 are associated with strong pore diffusion resis-


tances and indicate that internal structure does not participate


in the reaction, i.e., that reaction takes place primarily on


the outside surface of individual particles.  By analogy with


catalytic reactions, n would be expected to be a function of


particle size, and the ratio of reaction rate/pore diffusion

-------
                              36
 rate (and  hence,  pore  size).  All other factors being equal,




 n  decreases with  (1) increasing particle size, (2) increasing




 temperature or  reaction rate and (3) small pore size.









 Three different particle sizes of each calcine were exposed at




 980°C. for periods of  5 seconds to 2 hours to obtain additional




 parameters which have  not been measured previously.  These include




 effectiveness factors, n; reaction rates at zero sulfation, r ;




 the  $ coefficient (sensitivity of S0~ sorption rate to sulfate




 loading); and total CaO utilization.  The estimation of these




parameters is based on plots of log reaction rate versus sulfate




 loading, which have been shown to give a linear correlation over




a wide range of solid  conversions   .









Total CaO utilization was determined for each particle size




by exposure for 2 hours at 980°C.









 Samples of calcines were submitted to International Minerals




 and  Chemical Company for complete physical characterization




 (pore size distribution, pore volume and B.E.T. surface) and




 these data are  summarized in the discussion of results.  Calcines




were also submitted to Illinois State Geological Survey for




 study by scanning electorn microscope.

-------
                              37
c.  Results




Large differences were found in SO- sorption rates for the calcines




prepared from the various types of limestone examined in this study.




Figure 2-1 compares the reactivities of 150/170 mesh particles




at 980°C.  In general, the marble and spars showed the lowest




rates of sorption, whereas the dolomites and marl showed the




highest.









The kinetic parameters for the reaction are summarized in Tahle 2-3.




The activation energies, determined from the Arrhenius plots




given in Appendix A, fall within the range previously reported




for the reaction of calcined stone with S0~, and again showed




distinct differences between stones.  The lowest activation




energies were found for Iceland spar (Type 1), and an impure




dolomite (Type 6).  The Iceland spar showed other clear effects




of strong diffusional resistance.  The high impurity content




of the dolomite may account for its kinetic behavior.









Particle size had a marked effect on sorption rate for nearly




all stones at 980°C.  Figure 2-2 illustrates typical sorption




curves for three different particle sizes of one calcine.




Rate measurements obtained from such plots were correlated in




the manner shown in Figure 2-3.  Extrapolation of the rate data




to zero sulfation permitted the initial rate, r , to be evaluated




independent of the effects of sulfation.  The reaction rate

-------
                        38
    2O
Figure  2-1.
60       80
 TIME (SEC)
100
12O
                                                         I4O
Comparison of Reactivity  of  Calcines with SCK at 980°C.
Particle size = 150/170 mesh.

-------
                             39
Table 2-3.  Kinetic Parameters for S09 Sorption by Calcines
              (150/170 mesh particles, 980°C.)
ISGS
Stone
Type
1
2
3
4
5
6
8
9
11
APCO
Stone
No.
2201
2202
2203
2204
2205
2206
2208
1701
2129
Activation
Energy, E,
cal./g. mole
9,500
26,500
15,500
12,500
19,500
9,200
14,400
18,100
_
Reaction rate
constant, k ,
cm. /sec.
0.186
0.737
0.228
0.219
0.186
0.068
0.152
0.147
0.16
Frequency
factor, A
, 0
cm. /sec.
8.50
3.18 X
1.19 X
31.2
4.84 X
2.78
47.3
2.20 X
_
>

io4
io2

io2


io2


-------
                             40
30
27
                                               150/170 MESH
                                                     MESH
          100
200
500
600
700
                    300     400
                 TIME, SECONDS
Figure 2-2.
Sorption of SO- by Different Particle Sizes of Type 4  Calcine
at 980°C.

-------
                                   41
 40
 30
 20
         T	1	1	  I
  10
JB
o>
UJ

*
tr
UJ
oc
 0.2
150/170

 mesh
                                                       _l	I
               24           6          8           10



                   SULFATE LOADING,  (gm moles/flm )  * I03
                 12
       Figure 2-3.


       Estimation of  Initial  Rate, r  , and Effectiveness Factor, n, for the
                                   o


       Sorption of  S02 by Type  4  Calcine,

-------
                             42
constants given In Table 2-3 were computed from r  for the




150/170 mesh particles, based on the surface area of the unreacted




calcine.  The effectiveness factors given in Table 2-4 were




estimated as the ratio of r  for the particle size in question




to the value of r  for 150/170 mesh particles.  This assumes




n - 1 for the smallest particle size.  This assumption appears




to be well justified by the experimental evidence for all stones




except the Iceland spar (Type 1) and the marble (Type 10), as




will be discussed later.








The empirical relationship shown in Figure 2-3, which was




previously reported to correlate reactivity at 870°C., was




found in this study to be also valid at 980*C.  Effectiveness




factors, however, which had been evaluated at n - 1 for




particles of 42/65 mesh at 870°C., were lower at the higher




temperature used in this evaluation.  At 980°C. the value of




n was generally in the range of 0.7 for 42/65 mesh and about




0.3 for 12/16 mesh particles.  Log r versus sulfate loading




was linear for small particles to at least 50% CaO conversion




at 980°C.; for two stones, marl (Type 11) and aragonite  (Type 8)




results were linear over the full range of conversion.








The capacities of Iceland spar (Type 1) and marble (Type 10)




were especially sensitive to particle size, utilization increasing




by a factor of 2 when particle size was reduced from 12/16 mesh




to 42/65 mesh, and increasing again by a factor of 3 when particle




size was reduced further from 42/65 to 150/170 mesh.

-------
          Table 2-4.Parameters for SO  Sorption by Calcines as a Function of Particle Size (980°C.)
Stone   Particle
Initial rate, r ,
Effectiveness  Coefficient 6,   Total
Total CaO
Type   size, mesh  gm.moles/gm.sec.( x 10 )   factor, n
                                        ( x 10~ )    Capacity^3'   Utilization^

1


2


3


4

12/16
42/65
150/170
12/16
42/65
150/170
12/16
42/65
150/170
12/16
42/65
150/170
0.0090
1.04
5.00
0.274
1.41
1.61
0.673
0.849
1.08
0.649
1.02
1.39
< 0.0018
< 0.21

0.17
0.85

0.62
0.79

0.47
0.74

25.4
30.4
13.4
12.3
10.5
7.75
6.62
3.83
2.77
12,5
4.80
2.89
2.6
6.0
18
17
18
21
20
25
33
6.8
18
33
6.2
13.6
43
40
43
50
49
61
80
16
43
80
                                                                                                         U)

-------
Stone    Particle
       Initial rate,  r ,
Effectiveness  Coefficient g,   Total
Total CaO
type

5


6

7
8


Marl


1336

size, mesh
12/16
42/65
150/170
12/16
42/65
150/170
150/170
42/65
150/170
12/16
42/65
150/170
12/16
42/65
150/170
4
g.moles/g.sec. ( x 10 )
0.610
0.935
1.91
0.195
0.457
0.728
—
0.316
0.316
0.894
0.894
0.894


0.50
_2
factor, n ( x. 10 )
0.32 17.9
0.49 3.93
3.03
0.27 4.58
0.63 3.00
1.92
—
1 2.38
2.38
1 2.14
1 2.14
2.14



, (a)
capacity
9.0
22
22
19
19
19
1.3
36
39
35
35
35
2.0
4.8
14.3
utilization
37
90
90
100
100
100
50
86
93
100
100
100
4.9
11.7
35
             (a)
             (b)
Milligrams S0_ absorbed by 30 mg. calcine, 2 hr. exposure.
Percent conversion of CaO to sulfate, 2 hr. .exposure.

-------
                             45
Type 7, a magnesite, showed no SO^ sorption beyond that attribu-



table to CaO impurities for any reaction temperature from 540°C.



to 980°C.  To ensure that poor performance was not a result of



dead-burning at the 980°C. calcination temperature, additional



tests were made in which the raw stone was calcined in-situ at



reaction temperature.  The maximum reactivity was found at



650°C., at which temperature 6.5 mg. of S0_ was absorbed by



30 mg. calcine in 100 seconds.  At 760°C. sorption dropped to



2.6 mg. per 100 seconds exposure.  At 540°C., as at 980°C.,



pickup was only about 1 mg. of SO..
                                 •J






d.  Discussion of Results



    (1)  Pore Structure



Since 50 percent or more of the total particle volume of calcined



limestone consists of pore space, it would be expected that the



characteristics of the internal particle structure, such as pore



volume, pore diameter and surface area, would significantly



influence reactivity with S0«.  This would be especially anti-



cipated where large particles are concerned.  The primary physical



properties of the calcines, determined by American Instrument



Company and International Minerals & Chemicals Company, are



summarized in Table 2-5.  These measurements show marked differ-



ences between the properties of the calcines prepared from the



different types of stone.  Mean pore size, for example, varied



from 0.07 micron (Type 1) to 4 microns (Type 8): a factor of 60.

-------
                             46





                         Table 2-5




         Summary of Physical Properties of Calcines






Stone    Particle    Mean, pere diam.,   '  Pore volume,      B.E.TA  Surface,
Type
1
2
3
4
5
6
7
8
9
10
(marble)
11
(marl)
size, mesh microns
12/16 0.075
42/65
150/170
12/16 2.1
42/65
150/170
12/16 0.60
42/65
150/170
12/16 0.44
42/65
150/170
12/16 0.27
42/65
150/170
12/16 0.30
42/65
150/170
12/16 0.01
42/65
150/170
12/16
42/65 4.0
150/170
12/16 0.42
42/65
150/170
12/16 0.065
12/16 1.6
42/65
150/170
cc./g.
0.26
0.23
0.26
0.34(C)
0.27
0.28
0.31
0.29
0.26
0.32
0.31
0.32
0.39
0.37
0.50
0.40
0.40
0.40
0.34
0.39(C)
0.34
0.37
0.39
0.31
0.032
1.19
1.28
m /g.
9.8
9.9
10.2
0.83
0.7
0.7
1.5
1.6
1.8
1.9
2.1
2.4
3.6
3.4
3.9
3.6
3.9
4.1
22.7
29.0
37.8
0.65
0.79
3.4
3.7
3.1
0.63
3.5
2.2
3.5
   by mercury intrusion     •'excluding voids > ly     excluding voids > lOy

-------
                             47
Surface areas were also quite different,  ranging from 0.6 to
     2
10 m. /g. for different limestones.  Complete pore spectra were

obtained for each calcine; however, since the distribution of

pore sizes in a given calcine covered a narrow range (indicating

uniform-sized pores), only the mean pore diameters listed in

Table 2-5 are used in making comparisons.
For a given calcines pore structures showed no apparent dependence

on particle size.  The data of Table 2-5, for example, show that

the total pore volume of smaller particles was not consistently

or appreciably different from that of the larger particles.

Mean pore size also showed no trend with particle size.  A small
                                      2
increase in surface area, about 0.4 m. /gm., is evident when

particles are reduced from 12/16 to 150/170 mesh.  This is

attributed to the new surface exposed at the points of fracture

when the solid is crushed to make smaller particles.


Previous work has indicated that SO- sorption by small particles

is controlled by chemical reaction rate.  If diffusion into the

internal pore structure is fast relative to the reaction rate,

the entire B.E.T. surface area will participate in the reaction.

Under such conditions n = 1 and, from equation (2) the rate of

reaction of different calcines (compared at a given sulfate

loading, temperature, and SO- concentration) would be expected

to be directly proportional to B.E.T. surface area, S .  The
                                                     O

-------
                             48
 rate curves  of  Figure  2-1 are reproduced in Figure 2-4 with




 the  B.E.T. surface  of  each calcine  (determined prior to




 reaction) shown in  parenthesis.  With three exceptions the




 reactivity does increase with surface as expected.









 Before pursuing this analysis further, some discussion of these




 three exceptions, Type 6, Type 1 and Type 10, is in order.  In




 Type 6, considerable amounts of impurities are present, the




 calcine containing about 23 percent inerts.  These inerts, in




 addition to reducing the amount of CaO per unit weight of calcine,




 can presumably  block S02 from access to the reactive CaO in a




manner similar  to CaSO,, thus affecting the initial reaction




 rate as though  the stone were partially sulfated.  Furthermore,




 physical measurements of the raw stones (Table 2-6) show an




 unusually high  surface area for Type 6, indicating that much of




 the pore structure is not attributable to CaO, but rather to




 the impurities.  As a result of these considerations Type 6




 is not further  considered in this analysis.








 The two other stones, Type 1 and Type 10, have no chemical




 impurities which could account for their low sorption rates.




 Examination of  the data in Table 2-5 reveals that these two




 calcines are physically similar with regard to their small




 pore size.  Both have mean pore diameters less than 0.1 micron.




 These small pores, which account for nearly all of the surface




 area, are apparently the cause of the slow SO. sorption rates.

-------
                                  49
    20
     15k
 ui
 z
 o
i
    10k
                                                MARL(in-situ)

                                                   (3.5?)
                                                               T5
                                                               (3.9)
                                                               1336

                                                               (0.63)J
     3         20       40       60       80       100      120       140



                                  TIME (SEC)

      Figure  2-4.


      Comparison of Reactivity with B.E.T. Surface Area (in parentheses,m /gm.)

-------
                         50
                     Table 2-6

 Physical Properties of Raw Limestones, 10/28 Mesh
Stone
Type
1
2
3
4
5
6
7
8
10
11
Mean Pore Diam.
Microns
40
35
5
2
20
0.3
0.15
0.2
20
1
Pore Volume,
cc^/gm.
0.040
0.019
0.041
0.041
0.016
0.051
0.054
0.069Ca)
0.016
0.89
B.E.T.2Surface
m /gm.
0.19
0.17
0.32
0.46
0.18
2.87
1.54
3.26
0.09
4.48
-42 mesh particle size, voids larger than 20 microns
 omitted.

-------
                             51
Several different sources of evidence show that these small




pores lead to high diffusion resistance and plug easily with




reaction products.  For example, the data of Table 2-5 show




an unusually strong sensitivity of capacity to particle size




for these two stones.  In no other case were such large increases




in utilization found when particle size was reduced.  A test




of sorption capacity with -325 mesh Stone Type 10 (D  = < 44y)




showed that 83 percent CaO utilization can be achieved with




very fine particles.  This compares to 4.8 percent utilization




with 12/16 mesh stone, which indicates clearly that reaction




occurs with this stone only at the outside surface of the particles.









Comparison of reaction rates (at a given sulfate loading) is made




for several stones in Figure 2-5.  Most stones had responses




characteristic of pore diffusion resistance for particles down




to 150/170 mesh (pore diffusion corresponds to a straight line




through the origin).  Marl (Type 11) and aragonite (Type 8)




showed complete absence of diffusional effects (horizontal




line).  Iceland spar (Type 1} and marble (Type 10), however,




showed the concave-upward response which is characteristic of




shell diffusion on a non-porous solid.  These two stones are




the only specimens of approximately 15 examined at APCO in which




shell formation could be confirmed by S0_ sorption characteristics.




This result is logically attributable to quick pluggage of the




small pores with reaction products.  The conclusion that SO^




sorption by Type 1 involves shell formation is in accord with

-------
                                     52
                                 MESH SIZE
        12/16
42/65
150/170
                 STONE   SULRftTE LOADING
                  TYPE   (gm.motes/ gm.)x^O3
-   4
X

u
o>
«n
*>
O   Tt
E   3

E
o«
UJ


tr

z   9
O   2
111

-------
                             53






the results reported by TVA for Iceland spar calcite as




observed by direct microscopic examination of reacted particles




(March 1968 progress report to APCO).









From the experimental evidence discussed above, it is concluded




that the Iceland spar (Type 1) and marble (Type 10) are, in




effect, non-porous solids.  This results in the case of  the




marble from a combination of low porosity and small pore size.




Although the Iceland spar has sufficient porosity, the mouths




of the pores are of uniform small size which plug quickly with




reaction product.  In both cases the reaction with SCL is




confined to the outside particle surface.  It should be  noted




that these two stones are not typical.  Mercury intrusion tests




on calcines prepared from 85 different limestones generally




available in the United States showed average mean pore sizes




of about 0.3 micron.  Calcines with mean pore sizes of 0.1




micron or less are rare, when prepared under these conditions.









Referring again to Figure 2-2, the correlation of reaction rate




data with B.E.T. surface was made by comparison of rates at a




sulfate loading of 7 mg. S0» per 30 mg. calcine.  Comparison




at a constant degree of sulfation is necessary to ensure that




the effect of sulfation does not mask out the effect of surface




area, since it has been established that rate decays exponen-




tially with sulfation.  The rates evaluated at constant sulfation




are plotted against B.E.T. surface  (of 150/170 mesh particles

-------
                             54
prior  to  reaction) in Figure 2-6, for reaction at 980°C.


Additional data which is available on other stones are also


included.  It  is apparent from Figure 2-6 that the reaction


rate does increase with B.E.T. surface area in a manner which


is approximately linear.  This result confirms again that


chemical reaction controls the rate of sorption of S0» by


small particles and indicates that n - 1 for D  < 0.01 cm.
                                              P

at 980°C.





The presumption that small pores plug more rapidly than large


pores (made with regard to the discussion of Type 1 and Type 10)


was confirmed by an experiment in which samples of a calcine


were reacted with SO- for varying lengths of time, and the pore


spectra determined by mercury intrusion measurements on each


reacted sample.  Comparison of the changes in spectra showed


considerable difference in the rate of decrease in pore volume


for pores of different size.  The volume of small pores


(between 1 and O.ly) decreased with sulfation at a rate about


2.3 times greater than the rate of decrease in volume of pores


larger than 1 micron.





Surface Biffusion. Ionic Diffusion


A short experiment was made to test for the presence of surface


diffusion resistances and also to test for effects of S0~


diffusion through the CaO lattice as possible rate-limiting steps.


For this purpose, Type 2 calcine (which has large pores and


therefore would be expected to show such effects most strongly)

-------
                                   55
O
UJ
UJ
c»
E
O
10
ow
(O

9
E
 A
UJ
    .20
     .18
     .16
     .14
     .12
     .10
    .08
2   .06
h-
o
UJ
(T
    .04
    .02 -
         T2bB
                                  T3
                                 1343
                                                                 T5
                                                       T9
                       1234

                   B.E.T.  SURFACE AREA OF CALCINE, m^/gm.

        Figure 2-6.

        Reaction Rate vs.  Surface Area of 150/170 mesh Particles.  Rates


        Evaluated at 980°C.  and Sulfate Loading of 29.1 x 10~  gm.moles/gm

-------
                              56






 was  exposed  in the reactor  to SO.  for short periods followed




 by long  periods of soaking  at reaction  temperature but without




 SO-  present.   For  example,  a  sample was exposed for 35 seconds




 to flue  gas, purged and held  at  980°C.  for 30 minutes and then




 exposed  again  to SO- for 35 seconds.  The amount of sulfate in




 the  sample was  then compared  to  that absorbed by another sample




 exposed  continuously for 70 seconds.  The results showed no




 difference between the two  modes of exposure, indicating that




 diffusion of sulfate into the solid during long periods of




 soaking does not increase the reactivity and therefore solid




 diffusion would not appear  to be a significant factor.  This




 experiment was repeated several times with different exposure




 times and multiples of exposures with the same result.








 Calcination Temperature




Four samples of Type 2 calcite spar CIO/28 mesh) were calcined




 in the rotary kiln for 2 hours at temperatures of 930°C. (1700°F),




 980°C. (1800°F), 1040°C. (1900°F) and 1115°C. (2040°F).  The




 physical properties of the  calcines thus prepared are shown in




 Table 2-7.  The maximum surface area was obtained at 980°C. and




 the  surface area decreased  sharply at temperatures above 980°C.




 The  SO- reactivity shown in the table was determined at 980°C.




with 150/170 mesh particles,  by exposing five samples for 300




 seconds.  The average amount  of S02 absorbed again correlated well




 with the B.E.T. surface area.  The observed reduction in SO,,

-------
                             57


 reactivity  is  clearly a case of loss of surface area (due to

 everburning).   Since this stone contains no silica impurities,

 chemical  side  reactions are obviously not responsible for loss

 of reactivity.



                         Table 2-7

            Effect of Calcination Temperature on
            Physical Properties of Type 2 Calcite
                      Spar, 12/16 Mesh
Gale.
Temp.
1700°F
1800°F
1900°F
2040°F
Mean Pore
Diara. Microns
1.5
2.0
2.2
3.0
True Density
gm./cm.
3.46
3.43
3.67
3.46
Pore Vol.
cc/gm.
0.35
0.40
0.38
0.26
BET Surface
m2/gm.
0.65
0.83
0.59
0.32
S02 , s
Reactivity
10.7
12.5
10.1
6.2
     (a)
        mg. 863 absorbed by 30 mg. 150/170 mesh particles in
        300 seconds at 980°C.
In-Situ Calcination

The kinetic parameters presented in Tables 2-3 and 2-4 were

determined with stone calcined in a rotary kiln at 980°C. for two

hours.  In order to ensure that the (reactivity of such samples

was not altered as a result of their being coaled and stored

for some time prior to reaction, a series of tests was made

in which the stone was calcined in the reactor immediately

preceding exposure to SO-.  This was done by placing 150/170

mesh particles of raw stone in the reactor carrier and inserting

the carrier into the heated reactor at 980°C.  The amount of

raw stone placed in the carrier was equivalent to 30 mg. calcine

-------
                             58
 (about 52 mg. uncalcined).  The sample was thus subjected to shock




 calcination when the carrier was inserted into the reactor.  After




 5 minutes flue gas was admitted to the reactor and the sample




 exposed for 100 seconds.









 Ten such runs were made with each stone; the average S02 sorption




 is compared in Table 2-8 with the sorption found for kiln cal-




 cined stone.  Although several stones showed significant changes




 in reactivity, depending on the mode of calcination, the relative




 ranking of the reactivity was generally the same as that obtained




 for kiln-calcined material.  Several samples, Type 8, Type 10




 and Type 11, were more reactive when tested in this manner, the




marl (Type 11) being the most striking example.  Several tests




were made with marl in which the sample was left in the reactor




 for 2 hours after shock calcination prior to S0? exposure.




 These samples showed no loss of reactivity compared to samples




 heated for only 5 minutes, thus indicating that the 2 hour cal-




 cination used in preparing kiln calcines was not responsible




 for the loss of reactivity.  Figure 2-7 compares the reactivity




 of marl for the two modes of calcination.  No other stone




showed such an extreme enhancement of reactivity when calcined




 in-site.

-------
                              59
                         Table 2-8

                           (a.)
Comparison of Reactivitiesv ' of 150/170 mesh Stones at 980°C.
Stone
Type
1
2
3
4
5
6
7
8
10
11
2061
Raw Stone
Calcined in- situ
8.8
10.3
11.3
14.4
10.0
10.9
1.1
8.5
5.5
20.4
11.8
Standard Deviation
of 10 in-situ tests
± 0.8
± 0.9
± 0.9
± 0.7
± 0.7
± 0.8
± 0.4
± 0.7
± 0.4
± 2.3
± 0.8
Kiln Calcined
for 2 hours at 980°C
7.7
9.0
13.0
13.7
15.6
10.9
1.3
6.5
2.3
12.5
-
   	                      !


    Milligrams SO- absorbed by 30 mg. calcine  in  100  seconds.

-------
                                 60
   20
    15
     L
o
O»
E
   I0
IT
o
                  CALCINED IN-SITU
CALCINED IN ROTARY
      KILN
                               IN-SITU 650° C. REACTION
                                     26% CALCINED
             20      4O      60       80       100      120      140
                              TIME (SEC)
        Figure  2-7.
        Effect of Mode of Calcination on Reactivity of  Michigan Marl.

-------
                             61
Reaction Model



It is concluded from the above discussion that a simple model



which assumes chemical reaction to be the rate-limiting resistance



will adequately correlate the data for pure calcines which have



pore diameters larger than 0.2 microns.  The expression given



by equation  (2) defines the rate in terms of the initial



(unsulfated) characteristics of the calcine:




                      1  dn'     _  „
                  r=—  -rr=kSC    n
                      W  dt    s  g  so2




Previous data have indicated, and the present work has confirmed,



that reaction rate decays exponentially as sulfation progresses.



The relationship between rate and sulfate loading, nT/W, can



be described empirically by:





                         r = e-Bn'/W                      (3)





This change has been interpreted in terms of a decrease in the



frequency factor A , since



                        .     .  -E/RT
                        k  = Ae
                         s


the experimental results can be expressed by:




                       A=A  e-*n'/W
                            o


An alternative interpretation of the observed relationship



between rate and sulfation could be made on the basis of changes



in k  S  n as reaction progresses.  It is known that the surface
    s  g


area and pore volume of the calcine/reaction-product mixture



decreases as sulfate loading increases.  This loss of surface,

-------
                             62
however, is not necessarily due to changes in CaO surface, but




more probably results from the reduced pore size brought about




by the accumulation of reaction products on the pore walls.




It may be argued that the intrinsic surface of CaO is not in




fact altered at all by the presence of CaSO, product.  Note




that the linear relationship between rate and surface area at




constant sulfate loading (shown in Figure 2-6) will not




account for the exponential decay of reaction  rate.








A comparison of the experimental 3 values also suggests that




changes in the effectiveness factor, due to pore plugging,




likewise will not account for the observed rate decay.  The




data for marl and aragonite, both of which have effectiveness




factors of unity, show equivalent slopes of $ ~ 2.2.  These




two stones have pores considerably larger than the other stones,




but 3 values of similar magnitude.  The fact that the sensitivity




of rate to sulfation is similar despite the difference in pore




sizes, indicates that the accumulation of products in the pores




is not limiting the rate of S02 diffusion into the solid as




long as the pores do not plug.  The fact that aragonite and




marl are free of pore diffusion is confirmed by the fact that




particle size does not affect the rate at any degree of sul-




fation up to about 80 percent.  It is concluded from this




reasoning that the observed decay in rate of reaction does




not result from pore plugging.  A more likely mechanism which




would more closely agree with the observed responses would be

-------
                             63
a blockage of individual CaO sites by larger CaSO,  molecules.
It is concluded from these considerations that the  change of
rate with sulfation is best described in terms of the reaction
rate constant and the preferred expression of the overall model
would be in the form:
             1  dnf  .    -en'/W  -E/RT _                   ,..
             W  dF = Ao e       e      CS02 n              (4)

Values of A , 3 and n determined experimentally for the stones
tested in this study are given in Table 2-4.  This model
adequately describes the sorption of SO- by calcines under
isothermal conditions up to about 50 percent conversion of the
CaO.  The limits of applicability of the model, presumed to be
the point at which pore plugging begins (and diffusional effects
become rate limiting rather than chemical reaction), are:

                              Maximum Percent CaO
          Stone               Conversion to which
          Type                Equation (4) applies
            1                         29 (n < 1)
            2                         50
            3                         48
            4                         50
            5                         62
            6                         63
            8                         77
           10                         19 (n < 1)
           11                         86

Note that the 5 purest stones (Tl, T3, T4, T5 and T8) show
fairly close agreement on the value of k .  The average "intrinsic

-------
                             6'4
rate constant" of these five stones has a value:



                k  = 0.194 ± 0.029 cm./sec.
                 s


The effectiveness factor, n> is strongly dependent upon pore



size, as would be expected.  Figure 2-8 shows the relationship



between the effectiveness factor  (evaluated at zero sulfation)



and pore size for 42/65 mesh calcines at 980°C.  Pores larger



than about ly correspond to n - 1 for this particle size; when



particle size was increased to 12/16 mesh, TI dropped to about



0.3.  Estimates based on these data using the Thiele diffusion



modulus indicate that n values in the region of 1.0 should be



expected for 150/170 mesh particles.








e.  Conclusions



The isothermal rates of reaction of S0? differ widely among



calcines prepared from carbonate rocks of varying geological



type.  Physical properties of calcines prepared from different



geological types of stones also differ widely.








Many of the observed differences in S0_ reaction rate can be



logically interpreted in terms of the physical characteristics



of the calcine, particularly the size of the pores.








The total S09-sorption capacity of the limestone calcines



studied increased with pore size.

-------
                                                                               MARL   T8
                                                                                      (4,1)
                             T3
                                                                                                       Ui
O.I    0.2   0.3   0.4   0.5    0.6   0.7   0.8    0.9    1.0    I.I    1.2   1.3
                           MEAN PORE  DIAMETER,MICRONS
.4    1.5    1.6   1.7

-------
                             66
The rate of reaction increases with decreasing pore size until


a critical pore diameter of about 0.1 micron is reached.


Presumably pores smaller than 0.1 micron are rapidly blocked


by reaction products.  These blocked pores inhibit diffusion


of SO- into the solid interior, resulting in a lower rate.


Maximum rate results when B.E.T. surface area is in the region

         2
of 3.5 m. Ig. (corresponding to pore diameters of 0.2 - 0.3y)


under isothermal reaction conditions.





Chemical reaction is the sole limiting resistance for particles


smaller than 0.01 cm. at temperatures up to 980*0. when pores


are larger than 0.2jj.  The isothermal reaction of pure limestones


with SO- show an intrinsic rate constant (per unit of surface)


on the order of 0.2 cm./sec.





At conversions of about 50 percent the rate of sorption of S0~


by 90p particles of normal pore size changes from chemical


reaction to a combination of chemical reaction and diffusion.





Both rate and capacity of S0_ sorption are highly dependent upon


particle size.  The effect of particle size is not the same for


all stones but is determined primarily by the size of the pores.


Small pores lead to the highest sensitivity between the reac-


tivity of calcines and particle size.  Calcines with very large


pores may show no dependence of reactivity upon particle size.

-------
                             67





No evidence of surface diffusion or ionic diffusion could be


established as a major limitation on SO- sorption.






The relative ranking of calcines prepared from the eleven types


of stones listed in Table 2-2 was established in the following


order with respect to isothermal rate of reaction with SO- at


980°C.:  Til (marl) > T5 > T9 > T4 > T3 > T6 > T2 > Tl > T8 >


T10 (marble).  In this rating marl was calcined in the reactor.


T7 (magnesite) reacts only slowly with S0_ at any temperature


between 540° and 980°C., under isothermal conditions, and


consequently, T7 was omitted from this rating.






f.  Nomenclature


    A      frequency factor, cm./sec.


    A      frequency factor at zero solid conversion, cm./sec.


    C      gas phase concentration of sulfur dioxide, g. moles/cc.
     DU —


    D      mean particle diameter, cm.
     P

    E      activation energy, cal./g. mole


    k      reaction rate cpnstant per unit surface, cm./sec.
    k      reaction rate constant per unit volume of solid,


           sec.   (g. moles/cc.)


    dn/dt  rate of change of SO- in the gas phase, g. moles/sec.


    n'     sulfate in solid as SO-, g. moles


    R      gas constant, 1.987 (cal./g. mole °K.)


    r      rate of formation of SO- in solid, g. mole/g.(sec.)

-------
                         68
r      rate at zero solid conversion, g. moles/g. (sec.)


                                         2
S      B.E.T. surface area of calcine, cm /g.
 O

W      weight-of calcine sample, g.



T      temperature, °K.



t      time, sec.



3      empirical correlation coefficient defined by Equation 3



n      effectiveness factor, ratio of reaction rate to the rate

       that would be obtained if the entire volume of the particle

       participated equally in reaction



p      bulk density of calcine  (particle density), g./cc.

-------
                             69
3.  AField Study of the Role of Overturning of Limestone








a.  Introduction




Full-scale boiler-injection tests to evaluate the dry injection




process for removing sulfur  dioxide (SCL) from the flue gases




of power plants were begun in late 1969.  These tests called




for the injection of additives into the boiler at several points




downstream from the combustion zone.  However, a. comprehensive


          (33)
TVA report     pointed out the desirability of injection directly




into the burners by mixing the additives with the fuel.  This




approach was reported to have several advantages over separate




injection.  (1) It would allow maximum residence time for the




solid to mix and react with the gases.   (2) It would provide




optimum distribution of additive in the high-temperature zone.




(3) It would reduce maintenance problems by eliminating the




need for much of the injection equipment.  (4) It would allow




easy control of the additive-to-fuel ratio, and would auto-




matically adjust the ratio to changes in boiler load.








The decision to inject separately during the full-scale dry




injection process evaluation tests, instead of mixing with the



fuel, was based on the results of several experiments in which




additives fed to coal-fired furnaces with the pulverized coal


                                                          (13)
produced little desulfurization.  For example, Goldschmidt



found that dolomite hydrate removed only 8.9 percent of the S02




when added in stoichiometric proportion to the sulfur.  Finely

-------
                              70
 pulverized dolomite injected with coal removed 12 percent of



 the  SO- when  fed  to a pilot furnace compared to 25 percent


                                            (2)
 removal when  the  dolomite was fed separately   .  Limestone



 injected through  the upper row of coal burners (2700°F), but



 separately from the fuel, removed only 20 percent of the S0_


                                 (19)
 at twice stoichiometric feed rate    .  However, the same author



 reported desulfurization efficiencies of 50 to 60 percent when



 the  stone was injected at 2100°F.
It is generally believed that the low efficiency of desulfuri-



zation achieved in the high temperature injection tests was a



result of the phenomenon of everburning.  Overturning is the



process whereby a relatively unreactive lime is produced when



calcination occurs at high temperature and is known to involve



both physical and chemical changes in the stone.  Overburned



lime can range from the qualitative extremes of hard-burned



to dead-burned depending upon the severity of calcination



conditions.  Among the physical changes associated with over-



burned lime, as compared to lime calcined under moderate (or



soft) conditions, are loss of porosity, increased density   ,



loss.of surface area, growth of crystallites, and larger mean


             (23)
pore diameter    .  The extreme condition of everburning results


                                                           !   :

in a lime that has lost all porosity and the bulk density of



which approaches the absolute density of calcium oxi4e, 3.40



grams per cubic centimeter.  This dead-burned lime, which is



chemically inert, is produced at temperatures of 3QOO°F or




more' '

-------
                             71
In addition to physical changes, chemical changes can also occur


in limestone when it is overburned.  These changes result from


side reactions in which CaO unites with impurities in the stone


to form silicates, aluminates, and ferrite, which make calcium


unavailable for reaction with S0_.  These chemical changes are


of particular concern where limestone injection is contemplated

                      (34)
for coal-fired boilers     because of the possibility of reaction


of the additives with coal ash, which is composed primarily of


the oxides of silicon, iron, and aluminum.





The purpose of this field study was to determine whether additives


are overburned when injected with fuel.  This information is


needed because no examination of the stone was made in previous


tests that definitely established that the materials were hard-


burned or dead-burned.  Our object was to take the simplest case,


in which the chemical effects of everburning would not be a factor


(by using pure additive and an oil-fired furnace), and determine


whether the physical properties and chemical reactivity of the


lime produced during boiler calcination is significantly different


from that of stone calcined under controlled conditions in the


laboratory.





b.  Experimental


Two separate series of injection tests were made using four


different additives (two limestones and two dolomites).  The


test boiler was a 300,000 lb/hr., 900 psig, 900°F., Babcock &

-------
                              72
Wilcox* type FH integral boiler, located at the Bayboro Station

of the Florida Power Corporation in St. Petersburg, Florida.

The boiler fired-No. 6 fuel oil containing 2.3 percent sulfur

at a rate of 10,000 pounds of oil per hour at an operating

load of 150,000 pounds of steam per hour.



The compositions of the stone, determined spectrographically

by Bituminous Coal Research, Inc., are given in Table 3-1.

The possibility of tying up calcium in side reactions was mini-

mized by using pure forms of limestone and dolomite.  Because

the ash content of the fuel oil was very low (0.08 percent),

the reaction of the additives with ash was considered an insignif-

icant factor.  By using an oil-fired boiler, it was possible to

approach the problem specifically in terms of the physical changes

in the lime that might be related to its reactivity with S0».



During the first series of tests, the effect of boiler load on

the degree of burning of additives injected with the fuel was

investigated.  The dry, pulverized additives were mixed with

fuel oil and fed by a constant-displacement pump into the fuel

line that leads to the burners.  The feed rate was stoichio-

metrically equal to the sulfur content of the oil on the basis

of the CaO and MgO content of the lime.  Each additive was fed

under conditions of both high boiler load (300,00 Ib. steam

per hour) and low boiler load (150,000 Ib. steam per hour).


*  Mention of company or product name does not constitute
   endorsement by the Air Pollution Control Office.

-------
                    73
Table 3-1. COMPOSITION OF ADDITIVES,  IGNITED BASIS
Component
CaO
MgO
Si02
A100_
2 3
Fe-0,
2 3
Ti°2
Na20
K20
Test series 1
Dolomite
56.0
37.0
4.55
0.60

0.86

0.03
0.25
0.1
Limestone
96.0
1.0
1.13
0.41

0.20

0.03
0.02
0.1
Test series 2
Dolomite j Limestone
57.0
39.0
1.0
0.2

0.02

0.03
0.02
0.1
95.0
1.9
2.15
0.2

0.02

0.03
0.02
0.1

-------
                             74
At the boiler outlet ahead of the air preheater particulate




samples were drawn by means of stainless steel probes that




fed into miniature cyclones.  The temperature of the cyclone




was 450°F.  The sample collector was purged with dry nitrogen




to prevent contact of solids with the flue gas during sampling.




Each run consisted of a sample period of about 4 hours.








During the second series of tests, the effects of particle size,




iron content, residence time, and injection temperature on the




amount of S02 absorbed by the additives were investigated.




Stabilized dispersions of limestone and dolomite in light fuel




oil were prepared by Basic Chemicals in two particles size dis-




tributions.  Most of the work was done with the finer dispersion,




in which 50 percent (by weight) of the particles were smaller than




2.4 microns and 10 percent larger than 5 microns.  The other dis-




persion, which will be referred to in the following discussion as




"coarse", contained particles 50 percent of which were smaller




than 8 microns and 20 percent of which were larger that 30 microns.




A special preparation of fine dolomite was also made containing




4 percent mill scale as an articicial source of Fe»0-.  The




additive dispersions were injected at two points:  at the burners




with the fuel, and downstream from the combustion zone at a




temperature of 2430°F.  The sampling points were located around




the boiler as shown in Figure 3-1 (during separate injection



tests the first sample point was used as the injection site




and samples were taken at the fourth sample position only).

-------
                                75
                            2ND SAMPLE
                           SCREEN TUBES
1ST SAMPLE
    3RD SAMPLE
SUPERHEATER TUBES
     4TH SAMPLE
     BOILER EXIT
     (4 PROBES)
                                       rMBUSTION
                                        ZONE
                                         BURNERS
              BOILER PLAN  AND SAMPLE  LOCATIONS, TOP VIEW
                                Figure 3-1.

-------
                                76
Water-cooled sampling probes were used at the first and second




positions, but otherwise the sampling equipment was the same




as discussed earlier.  The residence time of the particulate




in the probe during sampling was less than 0.1 second.








The temperatures shown in Figure 3-1 were measured at high boiler




load with a water-cooled, high-velocity thermocouple.  At low




load the temperatures at sample positions 1 and 2 were 2270°




and 1970°F., respectively.









c.  Discussion of Results




Preliminary observations based on the first series of tests




showed that only a small amount of the lime had reacted with




SO- in the boiler.  Samples collected at the boiler outlet




during high-load conditions contained 7 percent SO-.  The amount




of sulfate found in the stone was consistently greater for




samples collected at low boiler load than for samples collected




at high load, but it never exceeded 14 percent utilization of




the CaO and MgO.  Carbonate analyses (evolution method) showed




that calcination of the stone was complete when fed to the




burners during either mode of boiler operation.









The chemical reactivity of the boiler samples was determined




in the laboratory in two ways:  as the rate of reaction with




SO- and as the rate of hydration.  The S02 reactivity was




measured in a differential reactor    by exposing the samples

-------
                                77
to flue gas containing 0.27 percent S0?.  Figure 3-2 shows the

amount of sulfate formed in 30-milligram samples as a function

of exposure time when reacted at 1600°F.  These data are for the

limestone fed during the first test series.  Data for the same

stone calcined under soft burn conditions (1700°F., 30 minutes)

are shown for comparison.  The rate of reaction, measured as

the slope of the curve at a given value of sulfation, was in

all cases greater for the laboratory calcine than for the limes

calcined in the boiler.  For the examples shown, the rate of

reaction of SO- with the boiler-calcined limes were one-fifth

the rate of reaction with the soft calcined lime when compared

at a sulfation of 4 mg SO-/30 mg calcine.  On the basis of

this comparison, it may be concluded that calcination in the

furnace flame resulted in a lime that was considerably less reac-

tive than stone calcined under less severe conditions.



The total S0_-sorption capacity of stones calcined in the boiler

was measured by exposing samples in the laboratory reactor for
                              \
2 hours at a temperature of 1800°F.  Under these conditions 75

percent of the theoretical conversion was obtained for the

dolomite samples (CaO component), and 65 percent for the lime-

stone.  These utilization values were not significantly different

from those obtained with stones calcined under soft conditions.

Because the primary difference was in the rate of SO- sorption,

rather than capacity, it was concluded that the limes were

hard-burned but not dead-burned.

-------
                          78
   14
  12
  10
  '8
 d°
 CO
   6
        I    I   I    I    I   I

        BOILER SAMPLE
        TEST SERIES
                             I   I    I   I    I    T
        I   I    I    I    I   I    I    t   I   I    I    I   I
           20     4O      6O     SO     100
                    EXPOSURE TIME, SECONDS
                                            120
  14
  12
  10
 o
 m

 £
 1

  *
               I    I   \    II\    \
    LABORATORY CALCINE
_     (30 MIN., 1700°F)
           I    I    I    I   I    I    I   I    I    I   I
140
  0        2O      4O     60      8O     IOO     120     140
                    EXPOSURE TIME,  SECONDS
Figure 3-2.  Reaction of S02 with Calcined Limestone at
             1600°F in Flue Gas, 0.27 percent  S02.

-------
                             79
The rate of hydration test, a standard technique used in the




lime industry for evaluating the degree of burning,  was performed




for four boiler samples by G. & W. H. Corson, Inc.  A series of




standard laboratory calcines was prepared from the raw stone at



temperatures of 1800°, 2100°, 2400°, and 2700°F.  The percent




of hydration versus time was measured for each calcine and compared




to that of the boiler samples.  Figure 3-3 summarizes the results




of tests with the dolomite run in the first test series.  The




laboratory calcines show decreasing reactivity as calcination




temperature was raised, in accordance with the effects of hard-




burning.  The rate of hydration of the boiler limes was lower




than the hardest-burned laboratory calcine prepared at 2700°F.




Limestone samples showed similar results  (e.g., 50 percent




hydration of the boiler calcined limes required 202 minutes,




compared to 21 minutes for the 2700°F. laboratory calcine and




1.2 minutes for the 1800°F. laboratory calcine).  Although the




accuracy of this test is confounded by the presence of sulfate,




which may also influence the hydration rate, the results




qualitatively confirmed that the boiler-calcined limes were




overburned.








Table 3-2 is a compilation of the chemical analysis data for




samples obtained from the second series of tests.  Comparison of




the carbonate and sulfate analyses of samples collected at




sample position 1 with those of samples collected at position 4




(runs 3, 4 and 5) shows that both calcination and sulfation of




the lime were completed in the furnace section before it reached

-------
                                    80
  100
   80
H

yj
o
   60
   40
o
o

0  20
        I8OO°F
          A BOILER SAMPLES


          0 LABORATORY  CALCINATION
                       I
10
100
1000
                                     TIME, MINUTES



        Figure  3-3.  Rate of Hydration of  Calcined Dolomite:  Test Series 1

-------
                                     81
Table 3-2.  TEST SERIES 2:  SULFATE AND CARBONATE CONTENT OF BOILER SAMPLES
Run
1
2
3


4


5



6
7
8
9
10
Boiler load, Sample
Additive 000 Ibs. stearo/hr position
i
Coarse dolomite
Fine dolomite
Fine dolomite


Fine dolomite


Fine limestone



Coarse limestone
Fine limestone
Separate injection-
Fine dolomite
Separate injection-
Fine dolomite + Fe_0
Separate injection-
Calcined limestone
305
275
250


150


250



250
150
250
250
250
4
4
4
2
1
4
3
1
4
3
2
1
4
4
4
4
4
Weight percent
Sulf ate ' Carbonate
As SO.^Y As C02
6.5
3.6
3.7
3.2
4.7
13.2
13.6
12.5
4.3
4.2
3.8
5.1
7.7
13.7
13.3
17.0
6.4
1.3
2.0
1.8
0.4
0.4
0.7
1.7
0.3
1.3
0.7
0.04
0.9
2.0
2.1
10.6
30.3
2.4
                 (a) Calcined basis

-------
                             82
 the first sample position, at which the temperature was 2430°F.



 No significant amount of S02 reacted with the lime while it



 passed through the boiler section, even though the lime was



 completely calcined.  These facts show that the rate of calci-



 nation was not limiting S0_ sorption, (i.e., the calcination



 step was completed prior to entry into the boiler, but there



was no further sulfation even though the S0» capacity of CaO



 is greatest in that temperature zone).  It was concluded that



 the deactivation occurred in the highest temperature region of



 the furnace.  The additional residence time was not the cause



of higher sulfation at low boiler load.







The coarse additives fed during runs 1 and 6 (Table 3-2) absorbed



more sulfate than did the finer additives.  This effect was also



evident in comparing the results of series 1 and 2, greater S02



 sorption being attained in the former case where coarser feed



was used.  This might be a result of less rapid calcination of



 the larger particles, a factor known to affect the density of



lime:  higher densities being associated with high rates of


           (12)
calcinationv   .







The physical properties of boiler samples obtained from both



 the first and second series of tests were examined and compared



 to those of soft-calcined stones.  Mercury porosimetry measure-



ments on 80 limestones and dolomites calcined in the APCO



 laboratory show that most of the pore volume is in a range of

-------
                             83
pore size between 1.7 and 0.1 micron.  Pores in this size range




typically comprise about 85 percent of the total pore volume in




soft-calcined limes.  Furthermore, for a given stone, the greatest




mercury intrusion occurs over a narrow range of pressure, indi-




cating that the pores are of uniform size.  The mean diameter




of these uniform pores can vary from stone to stone but is




usually about 0.3 micron.  Porosimetry measurements of the boiler




samples ±n no case showed the large volume of pores of uniform




size found in the laboratory calcines.  The specific volume of




pores in the range of 1 to 0.1 micron is compared in Table 3-3




for limes calcined in the boiler during test series 1 and labor-




atory calcines prepared from the same stones.  Pores larger




than 1 micron were necessarily excluded from this comparison




because of the fact that interparticle voids cannot be differen-




tiated from intraparticle pores by mercury porosimetry measure-




ments, and the former represent most of the total void volume




in the finely powdered boiler samples.  The data of Table 3-3




show that the boiler samples had markedly fewer pores in the




particular size range characteristic of soft calcined lime.








The bulk density and porosity of limes obtained from the second




series of tests were estimated by oil sorption tests.  This




test gives a good indication of the total volume of intraparticle




pores when the amount of oil sorbed by the interior of the




particles is large compared to the amount retained on the

-------
                            84
     Table 3-3.  SPECIFIC VOLUME OF PORES I  TO 0.1 MICRON

                DIAMETER (BY MERCURY INTRUSION) AND DEGREE

                OF SULFATION OF ADDITIVES:  TEST SERIES 1
Dolomite

   Laboratory calcine

   Boiler low load

   Boiler high load
                             Specific pore
                             Volume, cc/g
0.37

0.078

0.063
                 Sulfation.
                   % SO
none

16.6

 7.3
                                                        •3—
Limestone

   Laboratory calcine

   Boiler low load

   Boiler high load
0.31

0.16

0.15
none

11.5

 7.1

-------
                             85
particle surface.  The powder is immersed in No.  2 fuel oil,




filtered and dried on absorbent paper.  The volume of oil




contained in the pores is determined by multiplying the weight




gain by the specific volume of the oil.  Laboratory calcines




prepared from the same stone were also carried through the




procedure as a basis for comparison.  As shown in Table 3-4,




the boiler samples had high densities and low porosities.




These effects are both associated with hard-burning during



calcination   .  Comparison of the high- and low-load samples




showed that the greater sulfation achieved at low load could




not be explained on the basis of differences in porosity.




Tliis is confirmed by the data in Figure 3-2, which shows that




samples collected at different boiler loads did not have sign-




ificantly different rates of reaction with S0_.  It was concluded




that the degree of everburning was not related to boiler load




and that improved sulfation at low load was the result of some




other variable - possibly differences in oxygen content of the




flue gases - associated with the higher excess air fed at low




boiler load to improve heat transfer.  The adverse effect of  low




oxygen on the desulfurization reaction has been demonstrated  in




pilot-scale experiments on coal-fired furnaces for both separate




injection  ' and addition with the fuel^   .








Several supplementary tests were made injecting additives




separately from the fuel by spraying the suspension into the

-------
                             86
Table 3-4.  PHYSICAL PROPERTIES OF BOILER SAMPLES AS DETERMINED




           BY OIL SORPTION:   TEST SERIES 2,  SAMPLE POSITION 4
Dolomite, laboratory
Fine dolomite, run 2,
Fine dolomite, run 4,
Coarse dolomite, run
Limestone, laboratory
Fine limestone, run 5
Fine limestone, run 7
Coarse limestone, run
calcination^3'
high load
low load
1, high load
calcination
, high load
, low load
6, high load
Bulk density,
gm/cc
1.62
2.18
2.24
2.05
1.52
2.07
2.00
2.19
True density,
gm/cc
3.36
3.32
3.37
3.30
3.35
3.27
3.07
3.17
Porosity,
cc/cc
0.52
0.35
0.34
0.39
0.55
0.37
0.34
0.33
 (a)   Calcined 2 hours at 1700°F

-------
                             87
furnace at position 1 at high boiler load.   A comparison of  runs  3



and 8 in Table 3-2 shows that fine dolomite absorbed 3 times as



much sulfate under these conditions as when fed to the burners.



This was true even though the residence time was shorter and



calcination was not complete.  The addition of 4 percent iron



oxide to the dolomite that was injected separately improved  the



S02 sorption slightly, as shown by run 9.  A test in which



precalcined lime was injected separately gave poor results;



only half as much SO. was absorbed by this material as was absorbed



by uncalcined dolomite.  This observation verifies the findings


          (2  8)
of others   '    that no advantage is gained by precalcination



before injection.







The predominant form of sulfur in the lime was sulfate whether



injection was made with the fuel or separately.  Measureable



amounts of sulfide (averaging 0.27 percent by weight as S) were



found in the 13 samples fed with the fuel.  Limes from the



separate injection tests contained about 0.7 percent sulfite
                         I


(as S02) but no sulfide.







d.  Conclusions



The results of these tests have shown that the dry-limestone



process should not be applied by injection of the sorbent with



the fuel.  Instead, additives must be injected separately to

-------
                             88
achieve most efficient utilization of the limestone.  Overturning




is at least partly responsible for the low efficiencies found




when additives are fed to the burners.  The lime produced by




injection with the fuel is much less reactive with S0_ than




lime that is not calcined in the combustion zone.  This loss of




reactivity is associated with changes in the physical structure




of the stone during calcination in the combustion zone.  These




physical changes are sufficient to deactivate the lime, regardless




of further side reactions with impurities or ash.  Additives




injected into coal-fired furnaces with the fuel would therefore




be expected to yield only less-satisfactory results than those




reported here.








Boiler load was found to be an important variable affecting




desulfurization when additives were fed with the fuel.  This




is apparently due to the higher excess air used during low load




and is not a result of differences in flame temperature or




residence time.








These field tests indicate that there is an optimum particle




size as well as an optimum injection temperature.  The results




under the conditions of these tests show that injection




temperatures somewhat  higher than 2400°F. would be best for




2-micron particles.

-------
                             89
     4.  Methods for Testing the Degree of  Overburning



                     of Calcined Limestones







a.  Introduction



Currently, the dry injection process is being evaluated in a



full scale demonstration of performance at  the TVA Shawnee Steam



Power Plant, Paducah, Kentucky.  Optimum conditions for sulfur



dioxide abatement will be determined as well as the effect on



operation of the power plant and the economics of the process.


                                           (22  32)
Under the process evaluation planned by TVA   '    , carbonate



rock (limestone or dolomite) is ground to 70% less than 200 mesh



and is injected into the furnace above the fireball at tempera-



tures ranging from 2400°F to 2900°F.  The carbonate rock calcines



(evolves CCO and the resultant lime reacts with the sulfur



oxides in the presence of oxygen to form calcium sulfate.  The



calcium sulfate is removed with the fly ash in the standard dust



collection systems.  It has been supposed that the petrographic



nature of the lime accounts for its level of chemical reactivity,



and moreover, that the reactivity of the lime is dependent in



part on its calcination conditions.  An otherwise reactive



calcitic or dolomitic lime may become unreactive due to dead-



burning or everburning of the limestone.  Causes of this loss



of reactivity may be physical changes or side reactions with



impurities in the carbonate rock or both.

-------
                             90
As a limestone is calcined at higher temperatures, it loses



reactivity as well as surface area and porosity   .  Volumetric


                                         t-j  23)
shrinkage as high as 52.6% has been noted  '     for an oolitic



stone calcined at 2450°F   .  Loss of surface area with increased



calcination temperature has been reported   *   '   *     such


                        2             2
as a decline from 62.9 m /g. to 0.20 m /g. in specific surface



area for a corresponding increase from 750°C to 1300°C in ealc-


                                            (24)
ination temperature for Iceland spar calcite    .  Corresponding



to the volumetric shrinkage there is an increase in particle or



apparent density  '  '    from a low of 1»5 gm/cc for lime pre-



pared at low temperatures up to 3.0 gm/cc for limes prepared at



very high temperatures.
Changes in porosity (i.e., fraction of lime particle which is



pore space) with increased calcination temperature include loss



of porosity(3' 6> 7> 21> 24' 29J and increase in the mean pore


        (24  29)
diameter   '    .  Thus, the large pores grow at the expense



of the small pores which contribute greater porosity.  In con-



junction with the growth of pores there is a loss of porosity



attributable to the growth of large crystals by assimilating



small crystals   '    .  As a result of increased calcination



temperature, the reaction rate with sulfur dioxide or hydro^-



chloric acid is retarded  '    , the C02 absorption is lessened^   ,



the slaking rate is slower   , and the extent of sulfation


.  , .   (20, 25)
declines   '

-------
                             91
Several tests for the reactivity of calcium oxide have been



widely used already.  Two of these tests, the ASTM slaking



rate test and the G. and W. H. Corson X-ray hydration test rely



on decreased reactivity of overburned lime with water.  The



former test follows the reaction rate from the temperature rise



during slaking in an insulating bottle.  Another rate test,



the coarse grain titration test, depends on titration at one-



minute intervals with hydrochloric acid to construct characteristic



titration curves.  The titration curve for a lime calcined at



high temperature will rise far more slowly than that for a lime


                                                               (27)
calcined at low temperature.  A fourth test proposed by Y. Ohno



would expose the test sample to CO- and judge the degree of



everburning from the extent of recarbonization.








In order to determine if the limestone injection at the Shawnee



Plant led to optimum calcination of the limestone, it was necessary



to identify a test for lime reactivity using small samples of



limes diluted with fly ash.  In addition to considering the tests



listed above, many other possible tests were evaluated.  The



procedure was to determine various chemical and physical pro-



perties of different limestones calcined at a range of tempera-



tures.  The more promising tests were then applied to samples



from tests at Babcock & Wilcox and from preliminary injection




runs at the Shawnee Plant.

-------
                             92
b.  Experimental




The carbonate rocks tested were two calcites and a dolomite which




have been identified as Stone Nos. 2061, 2062, and 2069, respec-




tively.  An analysis of these is listed in Table 4-1.  Two sets




of these stones were sent to G.. and W. H. Corson, Inc., Plymouth




Meeting, Pa., where they were batch calcined at 1700, 2000, 2300,




2600, and 3200°F in a rotary kiln.  The stones were calcined




for two hours at temperature after the initial heat-up period




(see Figure 4-1).  The first set (Set I) was returned as cal-




cined; the second set (Set II) was ground to minus 170 mesh at




Corson1s.  All grinding operations were done under nitrogen and




samples were shipped in cans sealed under nitrogen.









Some of the tests were performed by contractors.  General




Technologies Corporation was sent samples of the first set for




infrared (IR) analysis.  To American Instrument Corporation




(AMINCO) went minus 70, plus 140 mesh (70/140) samples of the




first set and 170/270 samples of the second set.  Test results




reported by AMINCO were surface area by the B.E.T. (Brunauer,




Emmett, Teller) method and mercury penetration porosimetry




data.   The latter yielded information such as density, porosity,




distribution of pore size, and median pore size.









The tests conducted in the APCO laboratory may be divided into




two groups - chemical reactivity tests and physical property




tests.  The chemical reactivity experiments were as follows:

-------
                  93
Table 4-1.  Composition of Limestones
                   Analysis, weight per cent
Stone Number
Component
Si°2
Al_0_
2 3
Fe00,
2 3
MgO
CaO
Ti°2
Na2°
K20
MnO
2061
1.40
<0.2

0.27

1.77
95
0.03
<0.02
<0.1
<0.03
2062
2.25
<0.2

0.31

2.42
94
0.03
<0.02
<0.1
<0.03
2069
2.63
0.72

0.70

36.0
58.0
0.05
0.06
0.47
<0.03

-------
   4.0
   3.5
   3.0
   2.5
                         Figure 4-1
              TYPICAL   FIRING   CYCLE
tn
  2.0
ui
   1.0
  0.5
                                                   i    .     i
             1000      1500      2000
                     TEMPERATURE °T.
2500      3000     3500

-------
                              95
           1.   Absorption of sulfur dioxide from flue gas




           2.   Absorption of sulfur dioxide from pure SO



           3.   Absorption of carbon dioxide




           4.   Absorption of steam




           5.   Modified coarse-grain acid  titration




           6.   Hydration-weight gain.




 The physical  property tests were  density  determinations made




 in one  case by using an air  pycnometer and in  the second, by




 immersion  in  oil.  A  description  of the procedure for these



 experiments is given  in Appendix  B.








 Preliminary to the final results  all proposed overburning or




 dead-burning  tests were checked for their feasibility in the




 present TVA Shawnee injection  test application.  The most




 severe limitation was. that  the sample size could not be larger




 than several  grams.  Both the  slaking rate test   ' and the




 coarse grain  titration     called for sample sizes an order of




magnitude  larger.  In  the latter  case, it was found that the




 test could be modified and  still  give reproducible results.




Unfortunately,  this was  not  the case for  the slaking rate test




which had  to  be eliminated  for  consideration in this application.








c.  Discussion  of Results




     (1)  Effect of Calcination Temperature




The results of the various physical and chemical tests show




that apparent densities, median pore diameter, and average IR

-------
                             96
band shift increased with increasing calcination temperature



while all other properties decreased.







Results of the acid titration test for Stone No. 2061 are shown



in Figure 4-2.  All of the chemical and physical properties were



significantly affected by calcination temperature; however, the



most dramatic changes were in SO- absorption from flue gas, pure



S02 absorption, pure CO- absorption, hydration-weight gain,



B.E.T. surface area, and pore volume.  All of these properties



decrease in about the same way except for CO- absorption which



drops abruptly at 2000°F.  This behavior has been noted before


          (27)
by Y. Ohno     and has been attributed to the "shrinking,



melting, and fusing of the crystals'     , thus, obstructing



carbon dioxide penetration into the lime.






Statistical analysis of the data shows that such properties as



S0_ absorption from flue gas, pure SO- absorption, B.E.T. surface



area, and pore volume change significantly with calcination



temperature but not with stone sample.  Hence, the data of all



three stones of a set of calcines were taken together to calc-



ulate correlation coefficients between different properties.



The results shown in Table 4-2 demonstrate the high degree to



which sulfur dioxide reactivities, surface area, and pore volume



are intercorrelated (also refer to Figures 4-3 to 4-7 which are



the crossplots for the data from Set I).  This evidence supports



the conclusion that loss in surface area and pore volume at



higher calcination temperatures account for loss in reactivity.

-------
o:

.3
".{
i
Z
9
o

o
o
                           Figure 4-2
              ACID  TITRATION   OF  ^2061
                                              CALCINE  TEMPERATURE
                                               +  1700° F
                                               D  2000° F
                                               o  2300° F
                                               A  2600° F.
                                               X  3200° F
1 t 1 1 1 1 1 1 1 1 1 1
1 1 1
                68   10  12   14  16   18  20 22  24  26 28  3O
                         TIME  IN MINUTES

-------
                         98
          Table 4-2:   Correlation Coefficients














Flue Gas
Absorption
mg./30 mg.
Pure SO,, % gain
Absorption
B.E.T. Surface
Area, M /g.
Pore Volume
cc/g.



 0
0
-.
3 «0 •
t-l rf» 00
P»« < B




0.807

0.863

0.858
Set I
       Set II

-------
CT1
p^
en
              *
    M    K
                    Q£flD  BURN  STUDY/CROSS  PLOT NO  31
                       B.E.T. SURFACE  AREA  VS.  FLUE  GAS  CAPACITY
                                 *

(\J


                           K



  -,                     *


LJ


(X
L_



cn"""".




 n

UJ
 ••o
QQO
                           |        ^        |        |         f
  1.00      .3.00     5.00      7.00     9.00      11.00     13.00     15.00     37.00
           FLUE GRS CflPRClTY CMG/30MG3--X

-------
  u?_
  ir\ —

UJ
cc
o
o
                     [OD  BURN  STUDY/CROSS  PLOT  NO  35
                        HG PORE VOLUME  VS. FLUE GAS CAPACITY
                          *
                                      m
                                                                        §
     ~   •  —)•     T—"   •  i	1	1	• • i -  •	r—•	]
   1.00      3.00      5.00      7,00      9.00      11.00     33-00     J5.00     17.00
            FLUE GflS  CflPflClTY CMG/30MG]=X

-------
  r-

  •T -,
r~\

O
\
u
C_J
21 '•
ID
_J
O




i
Q_

_J
_J
a:

So
                   4-5= DERD  BURN  STUDY/CROSS  PLOT  NO  36
                          SMALL PORE* VOLUME VS.  FLUE  GAS CAPACITY
)K
     *                                      *SMALL PORE - PORE  HAVING

                                                        A DIAMETER LESS

 *                                                       THAN TWO  MICRONS



K
   j'.OO      3.00      S.OO     '/.DO     9.00      11.00     13.00     15,00     J7.00

             FLUE GflS CflPflCITY CMG/30MG)=X

-------
LD
\
LJ
C_J
Z)
_1
O
  "°-
_ I
COO

                   4-6: DEflD  BURN  STUDY/CROSS  PLOT  NO  47
                          SMALL PORE* VOLUME VS.   S02 CAPACITY
                   JK
K                              TSMALL PORE- PORE HAVING
                                        A DIAMETER LESS
                                        THAN TWO MICRONS
  •          I         I        I        I         I        I         I        I
   7.00     15.00     21.00     31.00     39.00     47.00     55.00     63.00     71.00
             502  CflPflCITY CPERCENTD-X

-------
O
S-i
c— 1
o
o
>-
I!
<—>
l~
•ZL
LJ
LJ
OC
 .
to
            Hwr.4-7,  CERD  BURN  STUDY/CROSS  PLOT  NO 27
                       PURE S0t  CAPACITY  VS.  FLUE  GAS  CAPACITY
o
u>
                      X.
                  r     	r~      ~r~     "  i ""~      r        \~      i
  .00      3.00      c>.00      7.0(1     9.00     11.00     13.00     IV 00    .17.00
           FLUE GflS  CnPnCITY  IMG/30MGD-X

-------
                              104
     (2)  Effect of Test Conditions

Upon review of the data on the effect of calcination temperature,

three reactivity tests were selected for further study; they

were pure C0_ absorption and pure S0» absorption and hydration-

weight gain.  Possible parameters for the first two included

reaction time, reaction temperature, gas flow rate, and particle

size.  The effect of particle size was studied first by finding

the absorption for samples of 70/140, 140/200, 200/270, and

-270 mesh ranges.




Carbon dioxide absorption was almost independent of particle

size but sulfur dioxide absorption was slightly lower (less than

10 per cent) for the largest size ranges.  On the other hand,

carbon dioxide absorption was affected drastically by the temp-

erature of the absorption test whereas sulfur dioxide absorption

was not (see Figures 4-8 and 4-9).  Both reactivity tests were

independent of gas flow rate and weakly dependent on reaction

time after ten minutes exposure (refer to Figures 4-10 and 4-11).

Possible parameters for the hydration-weight gain test were slaking

time, slaking volume, drying time, and drying temperature.  Effects

of the last two are not significant while the other effects are
                                                          i   '
illustrated in Figure 4-12.  Increasing either slaking time or

slaking volume of distilled water increases the hydration-weight

gain even for extremely hard-burned stones.

-------
  4O
                   105


                Figure 4-8
               C02  TEST
                                    CALCINED AT 1700° F.
CO
   30

h-
Z
UJ
u

cc
UJ
(L
   10
                                  x CALCINED AT 2000° R
                                  I
       500
    600          TOO


REACTION  TEMPERATURE °C.
800

-------
                               106
                             Figure 4-9
                    S02  COMBUSTION BOAT

                             TEST
   80
   70
<  60
CD

h-

CD
         CALCINED AT  1700 °R
o

QC
UJ
Q.
40
   30
         CALCINED AT  2000° F.
   20
   10
          1400   1500   1600   1700   1800   I90O   2000


                   REACTION  TEMPERATURE  °F.

-------
     107
  Figure 4-10


  C02  TEST
               CALCINED AT 1700° F.
                              o
               CALCINED AT  20OO° F.
                              X
                             120
REACTION  TIME-MINUTES

-------
   80
   70
§  60
   50
8  40
tr
u
CL
   30
   20
   10
                         108
                      Figure 4-11
              S02  COMBUSTION BOAT

                        TEST
                     CALCINED AT  1700° F.
CALCINED  AT 2000° F



   X.
           J_
   _L
J.
           10    20     3O    40    50


              REACTION  TIME- MINUTES
                      60
                   70

-------
               109
            Figure 4-12
SLAKING  TEST  MAIN  EFFECTS
   30
   28
   26
   24
   22
   20
   18
1  l6
I  l4
i  12
g  10
o
er  8
    6 -
    4-
    2 -
          I
                                 2000° F
                                 CALCINE
                                 3200° F
                              / CALCINE
                               X - I ML SOAK
                               o - 5 ML SOAK
            0.5       1.0      1.5      2.0
               SOAKING TIME - HOURS
                                  2.5

-------
                              110
     (3)   Effect of Partial Sulfation and Fly Ash




 Besides  the samples already listed above, certain samples exposed




 to SO -containing flue gases were received from Babcock and Wilcox
     X



 (B&W).  Under a contract to APCO, B&W injected the Stones No. 2061,




 No. 2062, and No. 2069 into their pilot scale coal-fired furnace.




 Samples of lime-fly ash mixtures from the cyclone downstream of




 the furnace were tested by all of the methods mentioned above.




 For the tests based on steam absorption and acid titration, the




 measured reactivity was very low or there was none at all.




 Results for physical property tests of these samples were dif-




 ficult to interpret in comparison to results for pure limes.









 To verify the exact relationship between the extent of sulfation




 and the degree of dead-burning measurements, the following experi-




ment was conducted.  Samples of Stone No. 2061 were exposed to




 either air or flue gas at 1800°F for varying lengths of time.




 The reactivity of the samples was determined using the C0_




absorption test and the hydration weight gain test.  The measured




reactivity for samples exposed to flue gas was corrected for




 the calcium sulfate content,  i.e., the calcium oxide content




which has already reacted and is unavailable for the dead-burning




 test reactivity measurement.   After this correction, the degree




 of dead-burning as measured for the samples exposed to air and




 for those exposed to flue gas should be the same.  This is the




 case for the hydration weight gain test but not for the C09




 absorption test.  As shown in Figure 4-13 there is a large

-------
                           lii
iu  80
   70
   60
!«
   40
LU
O
X 30
O
5


1  20
cc
    10
               Figure 4-13:   COMPARISON  OF MEASURED


              REACTIVITIES  FOR SULFATED AND

              UNSULFATED LIMES
                          AIR EXPOSURE-

                            UNSULFATED  SAMPLES
                           F.LUE GAS EXPOSURE-

                                 SULFATED SAMPLES
          \5    30    45    60    75    90


            TIME  OF  EXPOSURE,  MINUTES
                                              IO5   I2O

-------
                             112
deviation between the measured reactivity for the air exposed




samples and that for the flue gas exposed samples when reactivity




is determined by- CCL absorption.  As shown in Figure 4-14 there




is no such deviation for the hydration weight gain test; thus,




this test is independent of the extent of sulfation.









d.  Conclusions




Four reactivity tests - namely, flue gas absorption, SO,, absorption,




C02 absorption, and hydration weight gain are suitable as dead-




burning tests for samples of limited size diluted with fly ash.




The later three are sufficiently uncomplicated to be used as




field tests.  For sulfated samples, the hydration weight gain




test is recommended since it is independent of the degree of




sulfation.  Excellent correlation between SO- reactivities,




B.E.T. surface area, and pore volume support the conclusion




that dead-burning can be accounted for by loss of surface area




and pore volume.

-------
                       113

                     Figure 4-14
  40
  36
  32
 . 24
X 20
o
   16
Si
g  12

X
   8
REACTIVITIES OF SULFATED AND
UNSULFATED LIMES MEASURED BY
    HYDRATION WEIGHT GAIN
          REACTIVITY  BY HYDRATION
                 EXTENT OF
                      SULFATION
                              X-AIR
                              0-FLUE GAS
                                EXPOSURE

                              A-PER CENT S02
20
                                     18
                                     16
                                              14
                                     12
10
                                     8
          15     30   45     60    75
           TIME OF EXPOSURE, MINUTES
                              90

-------
                             114
       5.  Capacity of Limestone for Sorption of SO,.
a.  Introduction



The low cost and widespread availability of naturally occurring



carbonate rocks, particularly limestones*, has led to interest



in their use as reactants for desulfurization of combustion gases.


                              (2B)
Potter, Harrington, and Spaite     have presented some of the



background and engineering considerations for development of a



limestone desulfurization process.  In the development of the



process, it became apparent that some limestones are better



absorbents than others and that selective recommendation of



limestones would be necessary to optimize full scale field trials


                         (29)
in power boilers.  Potter     has presented results of the



present study which was designed to determine (1) the differences



in the sulfur dioxide reaction characteristics of a large number



of limestones, and (2) the physical and chemical properties



responsible for these differences.







A fixed-bed reaction test was selected because of its relative



simplicity.  However, the fixed-bed approach is not easily



suited to reaction rate studies because of the changes in



conversion and S0_ partial pressure throughout the bed:  There-



fore, the work has been limited to determining the ability of



a sample to achieve a high absorption efficiency.  In this
* In this report "limestone" includes all carbonate rocks


  containing magnesium and calcium.  High purity CaCO  is

  termed "calcite".                                  3

-------
                             115
report, absorption is measured as bed weight gain and is



generally expressed as capacity.








Using the results of preliminary tests of ten samples    ,




standard calcination and reaction conditions for comparing




capacities were selected.  Since the early work showed little




correlation between sulfur oxide capacity and chemical com-




position of the limestone, in the work reported here characteri-




zation of physical properties and classification of samples




was emphasized.








b.  Exp er imental




    (1)  Sample Preparation and Analyses




Crushed samples of limestone were screened to obtain a -18 +20




U. S. mesh cut (mean opening size 0.92 mm).  After a crushed




stone was thoroughly blended, a representative sample was




analyzed spectrochemically.  The carbonate content of the rock




was determined by dissolving a portion of the raw stone in




IN HC1.








The calcined material for the standard screening studies was




prepared in a bench scale calciner consisting of an Inconel




pipe rotating in a muffle furnace.  A 180 g. charge was fed to




the preheated kiln and held at 980°C for 4 hours.  To form




hydrates, hot water was slowly added to the calcined stone.




After the cake was dried to constant weight at 110°C, it was

-------
                             116
crushed and scteened to obtain -d8 +20 U.S. mesh hydrate pellets.


Thirty-nine of the calcined stones were tested for B.E.T. surface


area and mercury pore distribution.  The bulk specific gravity


was determined by an oil sorption test (ASTM test C-127-54)


modified for use with No. 2 fuel oil.






    (2)  Capacity Determination


The fixed bed reactor system used to determine sulfur oxide


capacity is shown in Figure 5-1.  No. 2 fuel oil (C:H = 6.8 by


weight) blended with controlled amounts of carbon disulfide (CS»)


was burned at 17% excess air to give a flue gas with the following


average composition by volume:  10.5% CO-, 3.4% 0-, 9.9% H20,


0.27% S02, 0.003% S03, and 75.9% N2>






The 35-mm diameter Inconel reactor was preheated to the desired


temperature, 980°C for the standard test condition, before the


sample was dumped from the top of the reactor onto a 30 mesh


screen in the center of the reactor.  A 20-gram charge of cal-


cined stone is stable at 980°C.  For uncalcined and hydrated


samples a weight equivalent to 20 grams of calcined stone was


used.  The flue gas was passed through the bed at a rate of 425
                                                        i   :

standard liters per hour for 3-1/2 hours.  In an oxidizing I


atmosphere at 980°C the net reaction is


              CaO, N + 1/2 0.    + SO-    + CaSO.

                            2(g)     2(g)       4Cs)

-------
                     117
                Figure  5-1
                                     TBIP ERA TUBE CONTROL
                                      CONDENSATE
                                   DIFFERENTIAL
                                   PRESSURE VALVE
Flow diagram  for fixed-bed reactor system

-------
                              118
 The S0_  content  of  the effluent was monitored continuously




 with an  infrared analyzer.  At the end of the exposure period,




 air was  admitted to the reactor to inhibit side reactions.




 The final weight of the bed was then determined.  The sample




 size, flow rate,  and contact time were selected so that even the




 most reactive stones had nearly ceased reaction.  Therefore,




 the data give a  comparison of the saturation conversions of




 the samples.








 The bed weight gain of a 20-gram calcined sample  (expressed as




 g.  SO-/100 g. of  calcined stone) was selected as  the index of




 sulfur oxide capacity.  The bed weight gain, also referred to




 as  "loading" or "capacity", is the weight of SO- removed from




 the flue gas.  No side reactions with carbon dioxide or water




were observed during the cooling and handling of the reacted




 sample.  In order to determine the reproducibility of the test,




 single replicate  tests were made on 22 different samples.  The




 standard deviation was found to be 2.98 g./lOO g.  Since loadings




varied between zero and 85 g./lOO g., the test is considered to




be  sufficiently sensitive and precise.

-------
                             119
c.  Discussion of Results




    (1)  Form of Sorbent and Reaction Temperature




The carbonate, oxide, and hydroxide forms of two samples were




tested at 430, 705, 980, and 1095°C.  The results are given in




Figure 5-2 for a high-calcium stone and a dolomite,  designated




A and B, respectively.  This work shows the relative performance




and temperature sensitivity of the different forms of the stone




and also the relation between optimum capacity and capacity at




980°C, the temperature at which the bulk of the experimental




work was done.









The hydrates have greater capacity  than the oxides and carbonates




throughout the temperature region examined.  In an air pollution




control process, the additional capacity of hydrates has the




greatest relative advantage at low  temperatures where reduced




gas volumes, lower viscosity, and cheaper materials of construction




might offset the cost of sorbent preparation.  Below 900°C the




capacity of the calcined stone is greater than that of  the carbo-




nate.  Since the low  temperature  (400  to 800°C)  loadings are




substantially different from  the high  temperature loadings,




conclusions presented here are most valid for temperatures in




the region of 1000°C.








Below 1000°C the capacity of  the  carbonates falls off rapidly




with  decreasing temperature.  Above 1000°C  the  loading  generally




decreased with increasing temperature.   The reduced capacity at

-------
                  120
              Figure 5-2
                REACTION I(«»l**IUH[, '
Reaction  temperature and form of reactant
             Figure 5-3

11

•« 1
5
I
*" *
I
8 ,
* '

__

-

_

-
1


















0 1 10 IS J
REACTION TEMPERATURE. MO'C
—








. js





AVERAGE CAPACITY OF U SAilPL ES.






\\
T|1-T_,




5




4
"V •





2.25 9 SPs p* 100 g
psa PTisa

	 ^


;•




	 i


•:;:



	

—
— ]
•- t
0 H » 35 01 *S SO SS W 41 70 75 ID 1
               AVERAGE CAPACITY, ff 10, ,., IOC ,
 Frequency distribution of  capacities
  of 86 stones tested

-------
                             121
higher  temperatures may result from changes in physical properties.



For  example,  significant crystallite size growth has been reported



above 1000°C     .  Reduced capacity may also be due to thermo-



dynamic limitations:  little CaSO, formation is expected above


                               f O / \

1230°C  in combustion flue gases    .  Disregarding these reactant



considerations,  there is little incentive for higher temperature



desulfurization  because increased gas volumes and viscosity



coupled with  more costly materials of construction would result



in higher operating costs.  The 980°C standard test temperature



seems to be near the most practical temperature for a high-



temperature,  SCL control process based on the use of uncalcined



limestone.








     (2)  Calcination



A series of fixed-bed calcinations were made at 1316°C for 16 hours



to evaluate the  effect of high-intensity calcination on the SO
                                                              X.


capacity and  physical properties of calcined stones.  The results



are given in  Table 5-1.  There was in all cases a loss in surface



area and an increase in the mean mercury intrusion pore size



when the calcination temperature was increased from 980 to 1316°C.



Loss in pore volume and surface area probably resulted from shrin-


                           (23)
kage and crystallite growth    .   The variations in SO, loading



that resulted from changes in calcination procedure are well



explained by  the volume of pores larger than 0.3y.  The minimum



pore diameter measured by the mercury porosimeter is 0.017u.

-------
Table 5-1.  EFFECT OF CALCINATION


                                          Volume of pores
mple

C
D
E
F
G
Hg pore volume
(cm3/g)
985°C
0.31
0.24
0,34
0.38
0.43
1316°C
0.21
0.25
0.07
0.32
0.14
B.E.T. surface area
(m2/g)
985°C
9.2
8.4
4.1
1.7
2.0
1316°C
1.1
1.1
1.0
0.1
0.6
Mean Hg pore size
(microns)
985°C
0.06
0,10
0.78
0.92
0.62
1316°C
0.35
1.3
6.0
18.0
3.5
larger than 0.3y
(cm3/g)
985°C
0.03
0.10
0.33
0.35
0.43
1316°C
0.18
0.18
0.07
0.32
0.14
SOX capacity
(g S03/100 g
985'C
13.6
26.8
29.0
47.0
63.0
1316°'
20.1
31.5
9.15
18.1
24.6
                                                                              to
                                                                              fo

-------
                             123
Those samples whose volume of large pores was reduced had a

corresponding loss of capacity, whereas the two samples whose

S0~ loading was increased as a result of high-temperature calc-

ination had an increased volume of pores larger than 0.3p.



Each of the 81 samples was tested in the fixed-bed reactor in

both the uncalcined and calcined conditions.  The correlation

coefficients* between capacities of uncalcined samples and

capacities of calcined samples was +0.734.  The mean loadings

of uncalcined and calcined samples were 41.7 and 43.1 g. SCL/

100 g. of sample, respectively.  These values suggest that at

980°C the relative capacity of a sample is, on the average,

the same in either the uncalcined or calcined state.  Analysis

of variance tests on 33 combined samples and on 8 individual

classes did not show statistically significant differences

between uncalcined and calcined samples.  Because of the equi-

valence of the capacities of uncalcined and calcined stones,

averages of tests on uncalcined and calcined samples are used

for each stone in the balance of this discussion.
*  The correlation coefficient, r, is an index of the strength
   of the linear relationship between the changes of two variables.
   For example, r   = 0.65 means that 65% of the time a positive
   change in y is accompanied by a positive change in x.

-------
                             124
     (3)  Capacity Correlations

The average SO, loadings of 86 samples reacted at 980°C are

summarized in a frequency distribution chart (Figure 5-3).

It is apparent that the capacities of various limestones for

sulfur oxides differ widely.  The average loading of 42.3 g./

100 g. for the -18 +20 mesh particles corresponds to 30%

utilization for pure CaO.  However, since the average CaO

content of the* samples was only 65%, the actual average utili-

zation of the CaO was 45%.  Incomplete conversion of the

reactant suggests diffusion limitation by the surface layer

of reaction products.  Electron microprobe analysis has verified
                             /Q\
the existence of such a layer   .  However, our studies with

different particle sizes suggested that the depth of penetration

of sulfur is not constant.  Further work is necessary to define

the limiting mechanism.




The factors that best explain the wide variations in SO- loading

at the standard test condition are summarized in Table 5-2.

With the exception of ignition loss and amount soluble in HC1,

the physical and chemical analyses were performed on the calcined

material.  Since a large number of samples was tested and

analyzed, it was found that the values of the correlation coef-

ficient did not change greatly with further increases in the

number of samples considered.  The absolute values of r are not

as important as their relative magnitudes.

-------
                           125
       Table 5-2.   CORRELATIONS BETWEEN PHYSICAL
            PROPERTIES AND 980°C CAPACITIES
Independent
variable
% CaO
% MgO
% Loss on
Ignition
% of Raw stone
soluble in HC1
B.E.T. surface
area (m /g)
Mean Hg pore
size (micron)
Area of pores
>0.3y (m2/g)
% Fe203
Bulk specific
gravity
Volume of pores
>0.3n (cm3/g)
Total Hg pore
volume (cm^/g)
% Oil
absorption
Number of
samples analyzed
59
59
39
39
39
39
39
59
39
39
39
39
Correlation
coefficient
-0.018
-0.043
-0.035
-0.202
-0.246
+0.398a
+0.400a
+0.429b
-0.507b
+0.620b
+0.635b
+0.643b
Significant at 95% confidence level
Significant at 99% confidence level

-------
                             126
Surprisingly, the chemical composition and carbonate content




(i.e., loss on ignition and amount soluble in HC1) explain




little of the variation in performance of the samples.  Of




the chemical components only the iron is significantly related




to capacity at or above the 90% confidence level.  Whether the




contribution of the iron is chemical or physical is not clear




at the present time.









The fact that the nitrogen adsorption surface area does not




correlate with capacity indicates that pores in the angstrom




size range do not significantly contribute to the fixed-bed




reaction.  On the other hand, a number of physical tests




show that pores in the micron range account for most of the




reactivity.  The correlation of the mean mercury pore size




indicates the relative importance of larger pores.  The area




of pores larger than 0.3p was calculated by integration of




mercury intrusion data    .  A trial-and-error calculation with




12 of the samples showed that the arbitrary 0.3u integration




limit gave the best correlation with loading.  Clearly, this




partial surface area is more directly related to capacity than




is the B.E.T. area.









Oil immersion led to significant correlations with capacity




when expressed either as bulk specific gravity or as percent




absorption.  The total mercury pore volume was the most useful

-------
                             127
parameter for explaining differences among samples because

(1) unlike oil absorption, mercury intrusion is not limited

to use on large particles and (2) correlation calculations have

shown that the bulk specific gravity, oil absorption, and volume

and area of larger pores are significantly related to the total

mercury pore volume.  Therefore, the iron content, mean mercury

intrusion pore size, and the total mercury pore volume are

considered to be the most fundamental variables for explaining

differences in the capacities of samples.  When these three

variables were used in a stepwise regression analysis, the

combined correlation coefficient was + 0.775.



    (4)  Classification of Samples

The carbonate form of limestone is the most likely feed

material in an SO,, removal process using limestone at temper-

atures higher than the calcination temperature.  Since most

of the physical properties in the preceding section were

determined on precalcined material, data on eight classes of
                            t
limestone are given in Table 5-3 to show the capacities of

various types of carbonate rock at the standard test condition.

The classes were  selected to give the greatest possible

variety of crystal structures, chemical compositions, and

geological origins.



Iceland spar was included because this high-purity calcite

is composed of very large crystallites.  A taagnesite was

-------
                             128
    Table 5-3.  CAPACITY OF VARIOUS CLASSES OF LIMESTONE
Classification



Iceland spar

Magnesite

Marble

Calcite

Dolomite

Aragonite

Oolite

Chalk
   Number of
samples in class
Variance
Mean capacity1
(g S03/100 g)
1
1
5
7
7
3
4
5
2.90
0.0545
6.55
2.14
10.32
6.45
3.97
6.54
17.3
19.7
32.0
32.7
43.8
52.1
57.1
66.4
  At 980°C

-------
                             129







tested  to determine  the reactivity of MgCCL.  Replicate runs



were made on  single  samples of Iceland spar and magnesite to




establish their capacities under the same conditions.  Hard,



highly  crystalline marble samples provide interesting compari-



sons with the soft,  fine-grained chalks.  The round-grained,



precipitated  oolitic samples represent another extreme in




crystal structure.   The orthorhombic aragonites were included




because most  limestones are rhombohedral.  Galcites were




defined as those having at least 95% CaO, and the dolomites




were defined  as samples with 40% to 44% MgCCL and 54% to




CaCCL   .  The results of these classifications are prest




in Table 5-3  in order of increasing mean capacity.
Iceland spar ±s apparently an inferior form of carbonate rock




for absorption of sulfur oxides.  A t-test on the various




classes showed that we can be 95% confident that Iceland spar




has lower capacity than the average calcite.  The low capacity




of magnesites was also shown by the t-test.  Thermodynamically


                      (34)
MgO can react with SO,,    , however, the rate of reaction is




apparently very slow.








It was determined, with 90% confidence, that for the test




conditions used the mean capacities of calcite and dolomite




differ.  The CaO utilization of the dolomites is about twice




that of the calcites.  Oolites and chalks appear to have




desirable properties for reaction with sulfur oxides.

-------
                             130
Although the five chalks had a higher mean capacity than the




oolites, a t-test showed that differences between these two




groups were not statistically significant.









Further refinement in the classification of samples could be




achieved if mineralogical and petrographic characterizations




were made.  For example, crystallite sizes may be important




in predicting the capacity of a stone.  This idea is supported




by the data in Table 5-3; the highly reactive chalks are




typically fine materials, whereas Iceland spar and marble are




relatively coarse materials.









d.  Conclusions




    1.  Between 900 and 1000°C raw and precalcined samples




        have the same sulfur oxide absorption efficiency.




        At these temperatures calcined stone does not lose




        capacity through the changes in physical properties




        that accompany overcalcination at higher temperatures.








    2.  The wide differences in capacity are best explained




        by reaction in pores of the size that are measured




        by mercury intrusion.  Chemical composition is of




        only secondary importance.









    3.  Under the chosen test conditions chalk and oolitic




        samples have the greatest sulfur oxide saturation

-------
                             131
        capacity.  Iceland spar, magnesite,  and marble are




        less reactive forms of limestone.








In addition to the data reported above a large amount of data




has been accumulated for other limestones and dolomites.




Approximately 150 limestones and dolomites have been inventoried




by the Process Research Section.  Data are available for most




of these sorbents which include chemical compositions, physical




properties (B.E.T. surface area, pore volume and distribution,




etc.), and S09 sorption characterization (bed weight gain, CaO




utilization, etc.).  All such data is being compiled by the




Process Research Section and will be available to interested




groups.

-------
                             132
                  D.  GENERAL CONCLUSIONS





1.  Chemical reaction is the sole limiting resistance for S0_


sorption when particles are smaller than 100 microns at tem-


peratures up to 980°C if pores are larger than 0.2 microns.


The isothermal reaction of pure limestones with S0~ show an


intrinsic rate constant (per unit of surface) on the order of


0.2 cm/sec.





2.  The rate of reaction increases with decreasing pore size


until a critical pore diameter of about 0.1 micron is reached.


Presumably pores smaller than 0.1 micron are rapidly blocked


by reaction products.  Maximum rate results when B.E.T. surface

                              2
area is in the region of 3.5 m /g (corresponding to pore


diameters of 0.2 - 0.3y) under isothermal reaction conditions.





3.  The total S02-sorption capacity increases with increasing


pore size, Furthermore, the capacity of limestones is correlated


to the area of pores greater than 0.3 microns but not to B.E.T.


surface area.





4.  Both rate and capacity of SO- sorption are highly dependent


upon particle size.  The effect of particle size is not the same


for all stones, but is determined by the size of the pores.

-------
                             133
Small pores lead to the highest sensitivity between the react-




ivity of calcines and particle size.  Calcines with very large




pores may show no dependence of reactivity upon particle size.









5.  The relative ranking of calcines with respect to isothermal




rate of reaction with S02 at 980°C is as follows:  marl > high




purity limestone > Iceland spar > aragonite > marble.  Magne-




site reacts only slowly with S02 at any temperature  between




540°C and 980°C.









6.  The relative ranking of limestones with respect to fixed




bed capacity at 980°C with -18 +20 mesh particles is as follows:




chalk > oolite > dolomite > calcite > marble > magnesite > Iceland spar.









7.  Loss of reactivity because of excessively high calcination




temperatures (known as dead-burning) may be attributed to the




growth of CaO crystals and the subsequent loss of surface area




pore volume.  Since loss of surface area and loss of pore volume




are similar functions of increasing calcination temperature,




both reaction rate and capacity of overburned limes are lost




proportionally.








8.  Injection of limestone with the fuel into a boiler leads




to dead-burning and poor S0_ removal efficiency.  Samples from




field tests where limestone was injected with the fuel in an

-------
                              134
oil-fired boiler demonstrated the high densities and low



porosities characteristic of overburned limes.








9.  Among many physical and chemical properties which are good



indications of the degree of dead-burning, CO, weight gain, SO
                                             fL


weight gain and hydration weight gain are suitable for samples



of limited size diluted with fly ash.  For partially sulfated



samples, the hydration weight gain test is recommended since



it is independent of sulfate loading up to 20% S0« in the test



sample.

-------
                              135
     E.  APPLICATION OF RESULTS TO LIMESTONE PROCESSES



In the dry limestone injection process, sorption of S0« will

be required of initially uncalcined limestone undergoing cal-

cination and reaction in a nonisothermal environment.  In this

regard, the conditions under which the process is to be run

and conditions under which experimental kinetic data of this

project were collected are quite different.  For instance, the
                       i
kinetic data were obtained for precalcined limestone reacted

under isothermal conditions.  This does not mean that the data

are not applicable.  Polythermal reactions may be regarded as

a series of isothermal reactions over short time increments

provided that the rate of temperature change with time may be

predicted and that there is no change in mechanism of the

reaction.  Rate of temperature change may be predicted with

heat transfer correlations and a heat balance.  While a multi-

tude of reaction mechanisms are possible, the high injection

temperature of limestone in the dry injection process would

preclude the sulfite reactions and suggest the dominant reaction

is direct conversion to calcium sulfate.  Thus, the dominant

reaction mechanism would not change.



The second difference between experimental conditions and boiler

injection conditions was precalcined  versus uncalcined limestone.

Once again, the reaction mechanism could be significantly different

-------
                             136
except if most of the limestone first calcines and then reacts




to form the sulfate.  In this case, the difference is one of




calcination conditions, i.e., calcination at a low temperature




for two hours versus calcination at a high temperature for a




few tenths of a second.  Effects of varying calcination condi-




tions are reflected in the surface area of the calcine and,




subsequently, in its reactivity.









The importance and functional relationship of the surface area




to the sulfur oxide reactivity of limes has been shown in the




investigations discussed above (for example, refer to conclusions




Subsection e, of Section C2).  It remains to describe the effect




of limestone type and injection conditions on the development




of surface area during calcination.  To this end work at Battelle




Memorial Institute has been committed using a dispersed phase




reactor.









There is little doubt that dead-burning is an important phenomenon




in limestone processes and in particular the dry injection process.




The above studies have shown that calcination at high tempera-




tures yields a dense lime with little surface area and little




sulfur dioxide reactivity.  Not only is injection of lime'stone




with the fuel into a boiler to be avoided, b"t also injection




should be sufficiently far from the fireball that particle




temperatures never exceed 2000°F.  Most tests for dead-burning

-------
                             137
are confounded by the presence of fly ash or  of  partial  sul-




fation.  Although modification of procedures  may be required,




a simple determination of the weight gain from hydration shows




the most promise as a dead-burning test for the  dry injection




process.









Selection of carbonate rock type for limestone processes depends




on the nature of the process.  For fluid bed combustion where




limestone remains as a chemically active fluid bed to react




completely, the proper measure of its reactivity is its capacity.




Rankings of limestone and limestone types for this application




are given in Table 5-3 of Section C5 and in Appendix Cl.  For




the dry injection process, limestone may react only during very




short residence times, and the proper ranking is derived from




reaction rate measurements as given in Figure 2-4^ Section C2.

-------
                             138
                    F.  RECOMMENDATIONS









1.  Sulfur oxide abatement processes involving injection of




limestone into a power boiler should be significantly enhanced




when highly reactive types of limestones are used;  especially,




aragonite and marl.  Testing of these materials in demonstration




scale applications is recommended.









2.  It has been shown that dead-burning is an important factor




in the dry limestone injection process.  Not only must injection




 of limestone with the fuel be avoided but also injection should




be into a temperature zone allowing greatest residence time




without everburning.








3.  Two fine-grained limestone materials have shown outstanding




S0» reactivity; specifically of the stones tested, marl with the




highest rate of reaction and chalk with the largest capacity for




absorption of S0_.  These materials should be studied in depth




to delineate those properties contributing to high reactivity.









4.  Much of the background information on limestone-SO_ reactivity




may be applicable to limestone dissolution rates in water and




subsequently to limestone slurry-S02 reactivity.  Those prop-




erties responsible for high dissolution rates of limestone in




water should be identified as well as the relative dissolution




rates of characteristic limestone types.

-------
                             139
5.  Limestone may well be an efficient sorbent for sulfur




emissions found in reducing conditions.  The kinetics of




limestone - hydrogen sulfide reactions should be determined




besides screening limestone types for reactivity.

-------
                               140
                         G.   REFERENCES
 1.  ASTM Standard C110-67

 2.  Attig, R. C., "Dispersed-Phase Additive Tests for SO- Control",
     Interim Report LR:68:4078-01:9 by Babcock and Wilcox Co. for
     contract PH 86-67-127 (Dec. 10, 1968)

 3.  Azbe, V. J.t "Fundamental Mechanics of Calcination, Hydration",
     Lime Manufacture May 1939

 4.  Bertrand, R. 'R. et al, "Fluid Bed Studies of the Limestone
     Based Flue Gas Desulfurization Process", Progress Reports 9
     and 10 by Esso Research and Engineering Co., for contract
     PH 86-67-130 (1968)

 5.  Borgwardt, R. H., Environmental Science and Technology, 4_ (1)
     59 (1970)

 6.  Borgwardt, R. H. et al, "The Dry Limestone Process for SO-
     Control: A Field Study of the Role of Overburning", APCA,
     New York, N. Y. (June 22-26, 1969)

 7.  Boynton, R. S., "Chemistry and Technology of Lime and Limestones",
     Wilev, New York 1966

 8.  Coutant, R. W. et al, "Investigation of the Reactivity of
     Limestone and Dolomite for Capturing S0? from Flue Gas", Reports
     by Battelle Memorial Institute for contract PH 86-67-115
     (Aug. 30, 1968 and Feb. 12, 1969)

 9.  Cunningham, W. A., I and EC, 43 (3) 635-8

10.  Eades, J. L. et al, "Scanning Electron Microscope Study of
     Development and Distribution of Pore Spaces in Calcium Oxide",
     2nd Annual SEM Symp., IITRI (April 1969)

11.  Eckhard, S., Z. Analyt Chem. 209:156 (1965)

12.  Fischer, H. C., J. Amer. Ceramic Soc., 38:7, 245-51  (1955)

13.  Goldschmidt, K., VDI Berichte. 6:21, 84 (1968)

14.  Harrington, R. E. et al, Am. Ind. Hyg. Assoc. J., 29, 52-8  (1968)

15.  Harvey, R. D., Environmental Geology Notes, 21, Illinois State
     Geological Survey (1968)

-------
                              141
16.  Hatfield, J. D. et al, Monthly Report for contract TVA-29232A
     (October 1968)

17.  Haynes, W., "Pilot Injection Studies at the Bureau of Mines",
     report presented at PHS Limestone Conference (Dec. 4-8, 1967)

18.  Hedin, R., "Structural Processes in the Dissociation of Calcium
     Carbonate", National Lime Assoc., Azbe Award #2 (1961)

19.  Ishihara, Y., "Removal of S02 from Flue Gases by Lime Injection
     Method"., report of Central Research Institute of Electric Power
     Industry, Komae, Kaitatama, Tokyo, Japan presented at PHS
     Limestone Conference (Dec. 4-8, 1967)

20.  Kim, Y. K., "Sulfation of Limestone Calcines", PHS Symp.,
     Cincinnati, Ohio (Feb. 5-6, 1969)

21.  Lougher, E. H., "Identification of Test Methods for Determining
     the Degree of Burning of Limestones", Final Report (Nov. 13, 1968)

22.  Ludwig, J. H. et al, Chem. Engr. Progress, 63_ (6), 82-4  (1967)

23.  Mayer, R. P. et al, "Physical Characterization of Limestone
     and Lime", National Lime Association, Washington, D. C.  (1964)

24.  McClellan, G. H. et al, "Scanning Electron Microscope Study
     of the Textural Evolution of Limestone Calcines", ASTM

25.  Mullins, R. C. et al, "Effects of Calcination Conditions on
     the Properties of Lime", ASTM

26.  National Research Council (National Academy of Engineering),
     "Abatement of Sulfur Oxide Emissions from Stationary Combustion
     Sources", 1970

27.  Ohno, Y., Gypsum and Lime, 28_, 22-28 (1957)

28.  Potter, A. E. et al, Air Engineering, 22-6, (April 1968)

29.  Potter, A. E., Ceramic Bulletin, Am. Ceramic Soc., 48  (9),
     855-8  (1969)

30.  Rootare, H. M. et al, J. Phys. Chem., H., 2733 (1967)

31.  Satterfield, C. N. and Sherwood, T. K., "The Role of Diffusion
     in Catalysis", Addison-Wesley, Reading, Mass. (1963)

32.  Shen, J. and Smith, J. M., Ind. Eng. Chem. Fundam., 4_, 293-301
     (1965)

-------
                              142
33.  Tennessee Valley Authority, "Sulfur Oxide Removal from Power
     Plant Stack Gas Conceptual Design and Cost Study" (1968)

34.  Ward, J. 0. et al, "Fundamental Study of the Fixation of Lime
     and Magnesia", Report by Battelle Memorial Institute for
     Contract PH 86-66-108 (June 30, 1966)

35.  Wicker, K., Mitt. Ver. Grosskesselbes. 83, 74-82 (1963)

36.  Wuhner, J., "On the Reactivity of Lime from Different Kiln
     Systems", National Lime Association, Azbe Award #5 (1965)

-------
                 143
              APPENDIX A




Kinetic Data of Carbonate Rock Types

-------
                             144
            KINETIC DATA OP CARBONATE ROCK TYPES









The graphs appearing in this appendix present the kinetic data




obtained in differential reactor test runs in accord with the




discussions of Section C2 of the report.  The carbonate rock




types indicated on the graphs are identified in Table 2-1 of




Section C2.

-------
                                                       145
     20


     18


     16


1
3
S?   12
O
fO
JT
•K   10
 CO
8    6
              I   I
I   I   I   I
        Type 1.  150/170 mesh calcined 925"C _
1   I   I   I    I   I   1    I   I   I   I   I   1
             20     40     60     80     100    120    140
              1   l   T  T  T  I   I   I   I   II   I

                 Type 1. 150/170 mesh calcined 2040"F _
20

18

16

14

12

10

 8

 6

 4
I   I   I   I   I   I   I   I   I   I   I   I   I

        Typel. 150/170 mesh calcined 325'C
                                                                                                     870 8C_
                                                                 1   l   I   I   I   I   I    I   l   I   I   l
                                       100    120    140    '  0
                                                                       60     80     100    120    140
                                                                           I   l  l   I   l  I   I

                                                                            Type 2.  150/170 mesh  _

-------
                                                  146



g
3
3
a
S
1
S
-s
a
S



20
U
16
14

12

10

S

6

4
2
I 1 I 1 1 1 1 1 1 1 I 1 1
Type 2. 150/170 mesh calcined 2040 "F _
-
-
— —
_ _
_ _
- —
— ~
-
_ —
-
1800 °F""
n 	 n ir~
-.V^***"^ 1200 'Fj
                                                                                  i   i  i   I   I   I   I   i   r
                                                                                    Type 2. 650 °C 150/170 mesh
                                                                                     I   I   l   l   l   I   I   1
                                                                100    200   300    400    500    600    700   800
    20
    II
t»
3
    12

    10

     8

     6

     4
                          I  I   I   l   I

                               Type 3.  150/170 mesh _
                    I   I   I   I  I   I

                  Type 4. 150/170 mesh _


                               S80°C
                I   1   1   I  I   I   I   I   I   1   I
                         Tint, IK
20    40     60    80     100    120    140
             Tim, sec

-------
                                        147
      20


      18


      16


 |    14
JG
 3
 r    12
S
J-
1

1    8
10
      18


      16


|    "
3
£    12
O
n


I    "
3     8
&e      i

^5     o


       4


       2
1   I   i   I   I   I
  TypeS. ISO/170 mesh
                   nc'
        I   I   I   I   I  I   I   I   I   I  I
                     GO     80    100   120    140
                                                        I   I   I   I   I   I  T   I   T  I   I   I   I
                                                                             Type6.  150/170 mesh
                                                                                                 -i
                                                                                           lit!'
                                                                 20
                                         40     60     80    100    120    140
              i   r T
                            i   riii   r
                          TypeS.  150/170 mesh
                                       870 °C
                                       650 °C
     i   I   I   I   I  II  JII   I   '  I  '
        20
                    40    GO

                          TiM,SK
                                 100   120    140
                                                                  120   14C

-------
                                              148
    20


    18


    IS
          I   I  I  I   T
s
    10
I   .
  I   I   I   I   I  T


Types. 150/170 mesh -
            20
                         60    80

                         TiM, MC
      100    120    140
20


II


16


14


12


10
I   I
                                                              Calcined in-situ
                                                           ""    to S02 exposure
                                                                     980 °C
                           I   I   1  T
                           •
                        MICHIGAN MARL-
                         150 170 mesh
                                                                                                      1
                                                                                 Calcined in rotary kiln
                                                                                      at 980'C
                                                                    I   1   I   I  I   1   I   i   i  I   I   i
20    40
            60     80

            HM.MC
                                 100    120    HO

-------
  100 c=r
   10 =
  0.1
    ^ i  ii Him   111 HUH   !  r
                     150 170 mesh
         11
                                          mi  inn
                                                         Type 1. 980 *C
                                                      T r nmirTTTmnr
                                                            42 65 mesh
rn=
                                             i  i m.iii   i  11 Him  i  11 nun  i  n
                                                                                        i  11 nun   i  11 imii   i  i nun   i  11
i         10         IN        1000      10.000  1
                                                        10         100
                                                                TIME,sec
                                                                       1000      10.000  1          10        100        1000      10.000
                                                                                                                                    •P-
                                                                                                                                    Vf)
  100
   10
N'
   0.1
         I  mill!   I I  I Illlll  I  11 mill  I  11113
                      150 170 mesh
                                                         Type 2. 980 °C
                                               "mwnrm
                                                                 42 65 mesh
1          10        100
                                          L_LLIHJJL_JL
                             1000      10,000  1         10
                                                               mini   !  n nun   i  ii
                                                                                                  I  111(1111   I  1 Illlltl
                                                                                                      II16 mesh
                                                                                          II HUH   I  I I
                                                                  100        1000
                                                                TIME, we
  10,000 1         10         100        1000      10.000
                                 N'  = mg.  SO   absorbed by  30 rag. calcine

-------
  100
  io
N'
  0.1
                       ISO 170 mesh
              Type 3. 980 "C
nriTiniir 7T
                                                                       42 65 mesh
                                                                 immn  i
                                                                   12 IS mtsh
    1          10          100         1000       10.000  1
         10          100
                  TIME, set
                                1000       10,000   ]
                                                                                                                                     1000       10000
                                                                    Type 4.  980 *C
   10
  0.1
     E  I  I ITTI1IT  I  I  I mill   I  I  Illllll    I
                       150 170 mesh
~i  iiiiiiii   TTimiir  TTTwiiii  11113
                   42 65 mesh
        I  II Ililll    II Illllll   I  I  Illllll    1  I Illllll  L_l_illili   I  I I HUH   I  ill.11 HI   I II
                                                                                                                        12 IGmesfi
                                                                 11 nun   i  11 nun   i  ii
     1          10         100         1000       10.000  1
                      100         1000       10,000  1
                   TIME. sec.
                                                                                                               10          100         1000      10,000
                                              N   =• mg.  SO   absorbed  b
                                                                            y  30  mg'.  calcine

-------
  100 =r
  10 =
N'
  0.1
                     150 170 mesh
             Type 5. 980 °C
7  j inrT'r
                                                                                n~Trn=3

                                               F
    1         10         tOO        1000      10.000  1
                                                                 42 65 i>esh
                               TmTn  r1^^   <
                                U 16 "esh
                  100
                TIME.sec
1000       10,000  1          10         100        1000      10.000
  100
N1
  0.)
                  111 nun   i  i iimn   nr
                     150/170 mesl)

                                                               Type 6. 980 °C
                42 65 mesh
                                                      ii inn   i  11 nun ..MI nun   i  i
                            !  I IIIUI   TTTTl
                                12 16 roesti
                                                                            I  I 1112
    1         10         100        1000      10,000  1
        10         100
                TIKE, sec
1000       10,000  1
10
    I  Illlllll.J.IlJJilil
too        loon       10.000
                                           N1 =  nij.',.  SO   absorbed  by 30 nip,,  calcine

-------
  100
   10
N'
  0.1
TypeS. 980 °C
       ETT
                        150 170 mesh
                                                                      TTTfflU   I  !  I Illlll   T
                                                                          42 65 mesh
     1          10          100         1000       10,000   1           10
                              100
                           TIME, set
1000       10,000
                                                                  MICHIGAN MARL  980 °C
  100
   It
N'
  0.1
                   TT
                          •65 mesh
                                                                            Mllll!   I  I III!!'
                                                                              1Z 16 mtsh
     1           10          100         1000       10.000   1
                  10          100
                           TIME, sec
                                                                                                                                                        Ul
                                                                                                           i  MiiiiH    i  muni   i	i    i  i
1000       10,000  1           10          100         1000       10.000
                                          N'  = rag.  S0~ absorbed  by  30  mg.  calcine

-------
                                                       153
    30





    27





    24





.1  21
O
T5


<»0  ID
E  18

o
CO



£  15

.g
1	T
                                 T	1	1	

                                      Typel. 980 °C
    12
 E


 CO
55'   9
                                       150/170 mesh
                                        42/65 mesh  -
            100    200     300    400    500    600    700
30




27



24




21



18




15
                                                                                  I      I

                                                                                Type 2.  980  C
                                                                                            150 170 mesh
            100    200   300    400    50u    600    700

                         TIME, sec
       100    200    300    400

                     TIME, sec
                                                                                 500    600    700

-------
                                                            I      I       1      I       t
                                                                             Type 6. 980 C
                                                      i      l      I      I
                                                  MICHIGAN MARL 980'C
100   200    300    400
             TIME, sec
500   600   700
100    200    300    400    500    600   700
             TIME, sec

-------
                                                       155
I   I    I   I   1   I    I   I   I   I   I -
          Type 2.  980  C 42 65 mesh -
I
                                                        ..  0.01
                                                                           I      I
                                           I      I       I
                                     Type 2.  980  C 12  16 mesh

                                                         0.0001
                                                               —  I   I   I   I   I   I   I   I   I   I    I
                                                           0.01
   0.001
                           15     20
                           S03l rag
30    35
                                                         0.001
                                                                                            150.170 mesh
                                                               12/16 mesh
                                                                            42/65 mesh
                      5     10      S03,mj     25     30

-------
                                                     156
              III\IIIIIIIF;
                       Typel. 980 C    150/17a mesh Z
                    I   I   I   I	1
            tilt
                                                     f     I     1      I    :
                                                      Typel. 980 C 42/65 mesh:
                                                        0.001 =
                   10     15
     20
25    30
35
                                                       0.0001
                                                                       I
4
5
7
      I            I    _
Typel. 980°C  12/lfrmeshI
o
5
fa
   0.001
OS
O
O
S5
OS
  0.0001
                                                                               r  \   T^   i   i   i   j
                                                                              Type 2. 980 C 150 170 mesh -
                                2
                           $03, mg
                             0.001

-------
                                     157
             I   I   I   I   I   I   I
                   Types. 980"C
12.16 mesh
   I   I   }   I   I   I   I   I    I   I
                                1.0 cr
                                      i   I   I  I   I   I   I
                                                      I   I   I   TJ
                                                   Types.  980'C Z!
                                                                    150 170 mesn
                                               I  t   I   I   I   1   1   1
                                       0.001
                                            0
                                            5
                                         10
                15
                   20
                      I   I  \   I —
                    TypeS. 980 °CZ
                                   FlIIIIiI  1I   T  T  T  T
                                                        MICHIGAN MARL 980  C
                                         0.1
   t  I	Ill
                                        0.01
   10
15    20
 SOs, mg
25    30
                                  35
                              0.001

                                     1   I   t  ..j_..J_
10
15    20
 SOa, mg
25    30
                                                                              35

-------
             158
          APPENDIX B
Bl.  Procedure for Dead-Burning
     Tests

B2.  Statistical Analysis of
     Dead-Burning Data

B3.  Crossplots of Data from the
     Dead-Burning Study

-------
                            159
Bl.  PROCEDURE FOR EXPERIMENTAL DETERMINATION OR COMPUTATION
     OF REPORTED DATA

     1.  Flue Gas Absorption - Method of  R.  H.  Borgwardt

         a.  Thirty milligrams of screened material are placed
             on a rock wool mat and into  a differential reactor.

         b.  Exposure is for 120 seconds  at  1800°F.

         c.  Recovered sample is analyzed for sulfate and results
             are reported as MgS03/30 mg.


     2.  Pure S02 Absorption

         a.  One gram of screened material is put in a combustion
             boat and into a horizontal tube furnace with quartz
             combustion tube.  (See Figure Bl-1).

         b.  Exposure is for 30 minutes at 1800°F with a gas
             flow of 2 SCFH.  The sample  is  in a complete
             atmosphere of S0?.

         c.  Results are reported as per  cent weight gain.
     3.  Pure CO  Absorption

         a.  One gram of screened material is put in a combustion
             boat and into a horizontal tube furnace with alumina
             combustion tube.

         b.  Exposure is for 60 minutes at 1400°F with a gas
             flow of 2 SCFH.  The sample is in a complete
             atmosphere of C02.

         c.  Results are reported as per cent weight gain.
     4.  Steam Absorption

         a.  One gram of screened material is put in a combustion
             boat and into a horizontal tube furnace.

         b.  Steam from a steam generator is fed to a preheat
             coil and into the stainless steel combustion tube.

         c.  Exposure is for 5 minutes at 500°C with results
             reported as per cent weight gain.

-------
                         160
 5.  Acid Titration

     a.  One gram of screened material is put in a 400 ml
         beaker and pH probes positioned in place.

     b.  At time zero, 200 ml of distilled water is added
         with vigorous agitation.

     c.  Titrations to pH 8 at minute intervals with IN HC1
         give results of cumulative acid versus time.
 6.  B.E.T. Surface Area - as reported by AMINCO.


 7.  Total Pore Volume - as reported by AMINCO.
 8.  Small Pore Volume - using the mercury penetration
     porosimetry curve, volume of pores with diameter less
     than 2 microns is calculated.
 9.  Mercury Intrusion Density - as reported by AMINCO.


10.  Average IR Band Shift - as reported by GTC    ; using
     infrared analysis of samples prepared by the KBr pellet
     technique, the average shift of three bands is calculated.
11.  Median Pore Diameter - using the mercury penetration
     porosimetry curve, the diameter corresponding to intrusion
     of 50% of the total volume is reported.
12.  Air PycnometerDensity - standard technique.
13.  Oil Absorption Density - No. 2 fuel oil is displaced
     in a graduated cylinder.
14.  Hydration-Weight Gain

     a.  One gram of material is slaked wxth one milliliter
         of distilled water for one hour.

     b.  The sample is dried for one hour at 260°C.

     c.  The percent weight gain is reported.

-------
           161
["I .X-ROTAMETER'
             EXPOSURE GAS
                SOURCE
           y    /Tr0.0!*805.!1-0/*. -B0.AT- jn-_  .V   	1
    HORIZONTAL TUBE FURNACE
                TO VENT
                                 SCRUBBING !
                                  SYSTEM  I
         Figure Bl-1

   COMBUSTION  BOAT
    TEST APPARATUS

-------
                            162
B2.  SOME STATISTICAL ANALYSES OF "LIMESTONE DEAD-BURNING" DATA









I.   SUMMARY









     1.  Data




As given in Tables B2-1 and B2-2 the data are in two batches




on limestones from three sources.  A measurement was taken for




each source at five preset calcination temperatures^  The




measurements for Batch II (B2) have eight components besides




temperature, and these eight will be termed the "basic set"




of variables.  The measurements for Batch I (Bl) have five




additional components.  This set of thirteen variables will




be called the "extended set" of variables.









     2.  Purpose




The amount of SO? absorbed from the flue gas is a decreasing




function of temperature in the range of temperatures for this




study.  Obviously, a number of other components are also highly




correlated to the flue gas measurement and hence they are all




controlled by temperature.  The purpose of this investigation




generally was to search for other, not so obvious, relationships




among the variables which may shed light on the dead-burning




effect.

-------
                               163


                           TABLE B2-la

                   Properties of Fredonia White

                     Set I, 70/140 mesh
                                       Calcine Temperature
                          1700    2000     2300
                         2600
                   3200
Flue Gas Absorption
1800°F, 2 min, mg/30 mg

Pure SO- Absorption*
1800°F. 30 min, % gain

Pure CO- Absorption*
1400 F, 60 min, % gain

Steam Absorption
500 C, 5 min, % gain

B.E.T. Surface Area
m2/g

Total Pore Volume
cc/g

Small Pore Volume
cc/g

Density by Mercury
Intrusion g/cc

Average IR Band
Shift, %
8.6
66.0
37.4
29.7
3.1
5.9
40.7
8.7
22.5
1.6
2,5
21.9
5.8
12.5
1.0
1.6
10.4
2.6
10.8
0.8
1.0
8.0
1.7
7.2
0.3
 .61     .51
 .28     .14
.33
.09
2.82    2.44    3.23
         62
80
.25
.04
          3.34
95
 .33
 .04
         3.30
100
* Note:  Listed values for 140/200 mesh

-------
                                164


             TABLE B2-lb:  Properties of Cedar Bluff
                           Set I, 70/140 mesh
                                      Calcine Temperature
Flue Gas Absorption
1800°F, 2 min, mg/30 mg

Pure S0_ Absorption*
1800°F, 30 min, % gain

Pure C0_ Absorption*
1400°F, 60 min, % gain

Steam Absorption
500°C, 5 min, % gain
B.E.T. Surface Area
m2/g

Total Pore Volume
cc/g

Small Pore Volume
cc/g

Density by Mercury
Intrusion, g/cc

Average IR Band
Shift, %
                           1700
 6.70
62.2
25.8
26.6
 1.9
  .70
  .28
 2.81
2000
3.76
27.3
8.7
17.0
1.2
.37
2300
3.81
26.7
6.6
15.5
0.9
.49
2600
2.60
18.8
5.6
12.3
0.8
.49
3200
2.51
13.2
3.6
10.2
0.5-
.37
 .10
.08
2.95     3.32
            25
          44
.05
          3.30
           67
.02
          3.19
           100
*Note:  Listed values for 140/200 mesh

-------
                                 165


                TABLE B2-lc:  Properties of James River
                               Set I, 70/140 mesh


                                          Calcine Temperature
                             1700
2000
2300
2600
3200
Flue Gas Absorption           9.56      9.04      6.60      2.28      1.80
1800 F, 2 min, mg/30 mg

Pure SO, Absorption*         67.4      62.1      46.2      13.2       7.6
1800 F,"30 min, % gain

Pure CO,, Absorption*         22.6      12.6       8.3       2.0       1.5
1400 F,"60 min, % gain

Steam Absorption             16.8      13.7      10.1       4.6       2.8
500 C, 5 min, % gain

B2E.T. Surface Area           4.1       2.2       1.5       —        —
m /g

Total Pore Volume              .69       .66       .55       .46       .37
cc/g

Small Pore Volume              .35       .31       .21       .01       .01
cc/g

Density by Mercury            2.84      2.74      3.39      3.44      3.31
Intrusion, g/cc

Average IR Band                 0        26        57        83        100
Shift, %
*Note:  Listed values for 140/200 mesh

-------
                  166
TABLE B2-2a:  Properties of Fredonia White
          Set II, 170/270 mesh
                         Calcine Temperature

Flue Gas Absorption
1800°F, 2 min, mg SO.,/30 mg
«3
Flue Gas Absorption
1800°F, 2 hrs, mg SO.,/30 mg
Pure S07 Absorption
1800°F, 30 min, % gain
Pure C02 Absorption
1400*F, 60 min, % gain
B.E.T. Surface Area
m^/g
Total Pore Volume
cc/g
Small Pore Volume
cc/g
Density by Mercury
Intrusion g/cc
Extent of Slaking
i _ t n •»_. 	 at 	 * __
1700
7.04
34.3
90.0
54.0
2.2
.77
.31
3.47
33.6
2000
3.52
22.6
16.8
9.8
1.1
.37
.07
2.91
28.1
2300
2.48
13.7
10.1
2.5
0.9
.30
.05
3.11
23.1
2600
2.08
11.2
6.0
3.9
1.3
.31
.08
2.92
20.2
3200
1.92
5.6
3.2
0.7
0.7
..22
.02
3.09
11.7

-------
                                      167


                 TABLE B2-2b: Properties of Cedar Bluff
                               Set II, 170/270 mesh
                                        Calcine Temperature

                               1700     2000     2300     2600      3200

Flue Gas Absorption             4.72     6.04     3.88     3.36      2.64
1800 F, 2 min, mg/30 mg

Pure SO- Abosrption            31.0     39.4     25.5     21.5       9.8
1800 F, 30 min, % gain

Pure CO- Absorption            11.0      6.7      5.4      6.0       2.9
1400°F, 30 min, % gain

B2E.T. Surface Area             1.4      1.2      0.9      1.1       0.5
m /g

Total Pore Volume                .59      .48      .36      .34       .33
cc/g

Small Pore Volume                .09      .05      .05      .06       .02
cc/g

Density by Mercury              3.27     3.15     3.17     3.16      3.21
Intrusion, g/cc

Extent of Slaking              32.1     30.4     26.9     27.0      10.4
1 ml, 1 hour, % gain

-------
                                      168


               TABLE B2-2c: Properties of James River
                             Set II, 120/270 mesh
                                      Calcine Temperature
Flue Gas Absorption
1800°F, 2 min, mg/30 mg

Pure S0~ Absorption
1800°F, 30 min, % gain

Pure C0? Absorption
1400°F, 30 min, % gain

B.E.T. Surface Area
m2/g

Total Pore Volume
cc/g

Small Pore Volume
cc/g

Density by Mercury
Intrusion, g/cc

Extent of Slaking
1 ml, 1 hour, % gain
1700
8.92
69.9
12.5
2.4
.62
.30
3.09
17.3
2000
4.92
49.4
6.7
1.3
.37
.15
3.16
10.4
2300
4.72
34.4
4.3
1.0
.37
.11
3.16
13.2
2600
2.92
10.3
2.8
1.0
.35
.07
3.25
2.6
                                                                      3200
1.80
8.0
1.7
0.7
 .26
 .05
3.24
5.2

-------
                            169







     3.  Analyses




Two computational methods of inspecting the data were tried:




1) principal components, and 2) regression with the flue gas




measurement as the dependent variable.








II.  GRAPHS AND CORRELATIONS









Correlation coefficient matrices are given for the measurements




in Bl and B2 on the computer printout as marked by the tabs.




The most notable difference between Bl and B2 is that mercury




intrusion density (MERDEN) in Bl is negatively correlated with




the temperature related variables whereas in B2 it is positively




correlated.  An inspection of the graphs of MERDEN vs. temperature




indicated this pictorially.  The B2 measurement from source 1,




Fredonia, at 2000°F is a major contributor, but even without




it the slope would be negative.








Some anomalies are evident in the ACID measurement, Bl, where




the three high temperature measurements for source 3, James River,




are much lower than the others.








III.  PRINCIPAL COMPONENTS








Temperature is a controlled variable, and variables in both the




basic set and the additional set are dependent upon temperature,

-------
                            170
hence correlational techniques such as principal components




analyses must be used and interpreted with appropriate caution.




The magnitude of any "correlation" cited in this report is not




meaningful itself, since its size is controlled by the selection




of temperatures, but it may be meaningful relative to another




such correlation from other variables, batches, etc.








For this situation we expect one large temperature component




with heavy loadings on those variables most related to temperature.




The size of this component is not of itself meaningful for reasons




mentioned above.  The second and higher components, since they




are orthogonal to this first major temperature component, hopefully




will exhibit factors perhaps not controlled by temperature and




yield useful information on the variable relationships.









     1.  Batch I, Basic Set of Variables




In the computer printouts as marked by tabs are the results of




the principal components computation on the Bl data for the




basic set of variables.  The.first component accounts for 86%




of the total variation and all variables are weighted more or




less equally with MERDEN, here coded as (MIDEN), and median pore




diameter (MDPDIA).  MDPDIA  have  negative weights as expected




because of their negative correlation with the other variables.




An inspection of case ranking by that component in Table B2-3




shows this to be a good, but not perfect temperature component.

-------
                                                         WT2- 3
flat oh No. *
Variables: 8 BASIC SET
Component No. 1
Rank

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Rank

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Source
123
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Source
123
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Temperature
17 20 23 26 32
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Temperature
17 20 23 26 32
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
                                                        Component No. 2
Rank

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Rank

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Source
123
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Source
123
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Temperature
17 20 23 26 32
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Temperature
17 20 23 26 32
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Component No. 3
Component No. 4

-------
                            172
The second component shows heavy weight, (-), for MERDEN with some




weight, (-), for mercury pore volume (MERPVO).   There is no clear




linear relationship to temperature or source.









Component three is largely a source component with CO- and BET




carrying positive weight and MERPVO a negative weight.  It is




doubtful whether the fourth component is meaningful.









     2.  Batch II, Basic Set of Variables




The first component for the B2 data is similar to that for the




Bl data except that MERDEN enters with the same sign as the




temperature related variables instead of the reversed sign as




for Bl.  This is expected because of the change in sign of the




correlation for MERDEN from Bl to B2.  This component is a strong




temperature component (Table B2-4).  Note that the direction




of the eigenvector is reversed from that of Bl, hence temperatures




are ordered from high to low.









The second component again has MERDEN as the major contributor




but MERPVO is no longer important.  Instead the MDPDIA seems to




be the contributor in its place in both the 2nd and 3rd component.




The third component is clearly the source component.








     3.  Batch I, Extended Set of Variables




The same computation was done on Bl with the extended set  of




variables less flue gas.  The first component is a pretty  good

-------
Batch No.   II
                                                    Tables J32-
Variables:  8  BASIC SET
Component No. 1
Rank

1
2
3
4
5
6
7
8
9
10
13
12
13
14
15
Rank

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Source
123
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Source
123
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Temperature
17 20 23 26 32
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Temperature
17 20 23 26 32
X
X
x
X
X
X
X
X
X
X
X
X
X
X
X
Component No.  2
Rank
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Rank
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Source
123
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Source
123
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Temperature
17 20 23 26 32
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Temperature
17 20 23 26 32
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Component No.  3
Component No. 4
                                                                                                               u>

-------
                            174
 temperature component as it was for the basic set, but the addi-




 tional variables change the later components.  In Table B2-5




 the strong source difference for ACID shows itself in component




 two where ACID has the heaviest weight and easily separates




James River from Fredonia and Cedar Bluff.  Notice the large




gap in component values between case no. 1 and case no. 11.  The




generally low values for AIKPY for James River influence this




also.  OILAB in component three aoes for Cedar Bluff what ACID




did for James River in component 2.  MERDEN (MINT) here tries




to influence the components as it did for the basic set, but




the source effect on the additional variables dilutes its efforts.








     4.  Summary of Principal Components Analysis




The temperature component appeared as expected in the first




component.  Source, another controlled variable just as tem-




perature, appeared strongly in the second or third component,




and may have obscured some of the variable relationships.  The




one variable that does seem to operate somewhat independently




of the temperature component is MERDEN.  MERPVO does this also




for Bl, but not for B2 where MDPDIA appears in a similar role.

-------
Batch No.
Variables:  12    EXTENDED  SET

Component No. 1
Rank

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Rank

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Source
123
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Source
123
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Temperature
17 20 23 26 32
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Temperature
17 20 23 26 32
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Component No.
Rank

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Rank

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Source
123
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Source
123
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Temperature
17 20 23 26 32
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Temperature
17 20 23 26 32
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
                                                                                                               -vl
                                                                                                               Ui
 Component No.  3
Component No. 4

-------
                            176
IV.  REGRESSION






Linear least square regressions were run on each batch with




flue gas as the dependent variable.  Only constant and first




degree terms were used in the model.  Batch I was analyzed first




using the basic set of variables (seven independent),  and then




using the extended set (eleven independent).  The stepwise re-




gression program in the UCLA Biomedical Series, BMD02R, and




the Linear Least Squares program of Wood and Toman were used.




Printouts are marked by tabs with comments below.









     1.   Stepwise Progress Tables




Tables B2-6, B2-7, B2-8 trace the progress of the stepwise re-




gressions for Bl basic set, for B2 basic set, and for  Bl extended




set respectively.  In these tables the independent variables




are listed in the leftmost column.  Across the top are the step




numbers and underneath them the estimated root mean square for




error at that step.  In the table for each step the independent




variables are ranked by their importance as predictor variables




as indicated by their F values.(F to enter for variables not in




the equation and F to remove for variables already in.  Variables




already in the equation are ranked ahead of those not yet'in,




and they have their ranks underlined.)








Generally the relative predictive value of an independent variable




as given by its rank is lowered at a particular step if the

-------
            177

         TABLE B2-6
 Trace of Stepwise Regression
      Batch I, Basic Set

             Step

so2
co2
BET
MINT
MPVOL
SFVOL
MDPV
a
0
a.
2
3
7
5
2
3

1
i
' 2
3
7
6
5
*
.61
2
1,
2
3.
-
-
-
4
.56
3
I
2
3
-
6
*
5
•53
4
1
2_
1
7
6
4
5
•52
5
1
2
1
6
5
i
-
.51
6
l
i
1
6
7
i
2
.47
7
JL
2
3
6
7
i
2
.50
         TABLE B2-7
Trace of Stepwise Regression
    Batch II, Basic Set

so2
co2
BET
MINT
MPVOL
SPVOL
MDPD
A
a
0
1
6
2
7
3
4
5

1
1
2
^
3
6
-
5
.82
2
1
i
4
5
3
-
-
.63
3
1
2
5
*
3.
-
-
A3
4
i
2
-
4
3
-
5
.38
5
1
i
7
i
i
6
2
.38
6
i
i
7
i
3
6
2
•38

-------
           178
        TABLE B2-8
Trace of Stepwise Regression
     Batch T, Extended Set
            Step
0
so2 i
COO 7
STEAM 9
ACID 12
BET k
A T"D"CV T ^
*i_L£\jri XU
OILAB n
MINT 8
MPVOL 5
SMPV 2
MDPV 3
I.R. 6
/\
cr
1
1
3
2
6
7
11
*
12
10
9
8
5
.61
2
1
9
£
12
5
6
7
3
11
*
8
10
.53
3
l
-
2
-
5
7
6
1
8
4
-
-
.50
if
1
-
2
7
5
-
9
1
6
i
8
-
.W
5
1
8
2
7
i
9
-
2
6
1
-
10
•35
6
1
11
2.
7
i
9
8
5
6
3
10
-
.34
7
1
-
£
7
i
8
-
5
6
1
9
-
•33
8
1
10
2
6
i
8
-
5
1
3
-
9
.29
9
l
10
2
6
i
8
-
5
7
1
-
9
•30
10
i
10
3
8
if
7
11
2
6
2
12
2
.32
11
1
10
3
1
i
6
11
5
8
2.
12
2
.35
12
3.
5
-
9
2
i
l
10
-
1
8
6
.11

-------
                             179
reduction in  the error sum of squares due to that variable




(or potential reduction for a variable not yet entered) is




made much smaller or "stolen" by the variable just entered,




i.e., the two variables yield redundant information.  The




value will be raised if this (potential) reduction is not lowered




very much while the error sum of square is considerably reduced




by the entering variable, i.e., the information for the two




variables is not very redundant.  If two independent variables




are completely uncorrelated the reduction due to one variable




will not be lowered at all by the other one's entry, hence its




rank will probably rise.









     2.  Residual Error




In Bl the additional variables in the extended step contribute




significantly to the regression equation as indicated by the




lowering of 6*, the root mean square for error, from about .50




to .30.  The drastic reduction of & to .11 in the 12th step




of the Bl, extended set, regression is probably specious due




to the very low degree of freedom for error.








     3.  Action of the Independent Variables




Several things should be noticed from this regression.




         a. SO, enters first and is consistently the most




            important predictor even though its importance




            is reduced little by little as correlated variables

-------
                   180
   enter the equation.  It is not as important in




   B2 as in Bl; CO- and MPVOL seem to contribute




   more in B2 to compensate.




b. CO- is worth more (relatively, compared to the




   other variables) in the presence of SO- than it




   is by itself.  In all of the regressions it comes




   from a low ranking to second or third with the




   entry af SO-.  Its information is apparently less




   redundant and more complementary to S0_, than




   others such as SMPV, MINT, MPVOL, MDDIAM.




c. SMPV is largely redundant with SO  since it loses




   heavily with the introduction of SO,,.  For Bl,




   extended set, it finally works up to have first




   ranking at the last step.  The drastic shuffling




   of rankings at this last step, just as the lowering




   of o, are suspect because of the very low degrees




   of freedom for error.




d. BET's influence is reasonably stable in Bl.  It is




   probably a good predictor variable, but of secondary




   importance to SO-.  In B2 it yields to the mercury




   penetration variables.




e. MINT and to a lesser extent SMPVOL have an interesting




   action in Bl.  With the basic set of variables they




   have little importance after S02 and (X>2 are intro-




   duced.  However, with the extended set, wherein

-------
                            181
            STEAM enters at the second step in place of CO ,




            we see that MINT assumes a high rank.   This suggests




            that the MINT, STEAM variables must be used together




            as a somewhat equivalent to,  but better predictor




            than CO..




         f. MPVOL apparently works with CO- in B2, but no such




            action is seen in Bl where it is constantly low




            in rank.








     4.  Other Notes




         a. In Bl the Cedar Bluff 2300°F measurement shows a




            rather large residual, but when it was eliminated




            there was no significant change  in the regression




            results.




         b. In B2 the Fredonia 1700° measurement has a large




            WSSD (Wood-Toman) indicating that the location of




            the independent variable is far removed from the




            other 14 measurements.  Elimination of this measure-




            ment has little effect on the regression equation




            indicating that the high value of FLUE GAS at that




            point is well predicated by the high values of the




            other variable.








     5.  Conclusions from the Regression Analysis




It appears obvious that the capacity measurements are  the strongest




predictors.  There are at least two useful degrees of  freedom

-------
                             182
 among  them which here seem to be given by either the S0_ and




 STEAM  combination or the SO- and C0_ combination.
After that the mercury penetration measurements add some




additional information with probably two or more degrees of




freedom.  MEKDEN AND SMPVOL (together with STEAM) or MPVOL seem




most promising.









BET stands as a useful variable in Bl somewhat independent of




the rest, but in B2 it is redundant to the mercury penetration




measurements.









V.   COMPARISON OF REGRESSION AND PRINCIPAL COMPONENTS ANALYSES






Generally, those variables which appeared with some force in the




non-temperature components proved to be rather poor predictors




in the regression equations.  For example, in Bl, basic set,




MERDEN and MERPVO were strong in the second component but law




in predictor value.  Similarly in B2, basic set, MERDEN and




MDPDIAM act similarly whereas MERPVO does not appear in the




second component, but it is instead a good predictor variable




for flue gas.  This suggests that such variables have a variance




component in this process which is unrelated to the basic




capacity measurements.







In Bl extended set, the additional variables AIRPY, OILAB, and




ACID displayed the same characteristic by carrying heavy weight

-------
                            183
in later components but having little predictor value.   These




variance components seem to be caused by source differences.

-------
                             184
B3.  CROSSPLOTS OF DATA FROM THE DEAD-BURNING STUDY









As outlined in Section C4, three limestones numbered 2061,




2062, and 2069 were calcined at a range of temperatures and




tested for chemical and physical properties.  The response




for each of these properties to calcination temperature is




given in the following crossplots.

-------
 o
 o
 <*"»-,
 o
 o
'"•o
CDO
CO
_

-------
  a
  a
  r-
  CD-
 CD
 o
 03
>-

II,
O
DC
a:
ex.
CEo
O
en

UJ
cc
Q_0

 r-*"
                     DEflD  BURN  STUDY/CROSS  PLOT  NO  02
                                                                      co
               *           *

                           *
                                                               *
   37.00     39.00     21.00     23.00     25,00     27.00     23.00     31.00     31.00
            CflLCINE TEMPERRTURE CHUNDREDS OF DEGREES

-------
 o
 O
 O
 O
>-
1!
 o
 o
 in.
CJ>
cc
UJ
o_
o
CE
0_
(X
o
CJ
cc
ID
Q_0
                     DEflD  BURN STUDY/CROSS  PLOT  NO  03
       oo
X
   17.00     19.00     21.00     23 00    25,00    27.00    .29.00    31.00    33.00
            CflLCINE TEMPERRTURE CHUNDREDS OF DEGREES

-------
 o
 o
 o
 o
 o
 o

               *
 O
 O
LJ

CJ
t—
h—I

u
                    DERD BURN STUDY/CROSS  PLOT  NO  04
               *
               x
oo
                                      K

                          K
5^'

21                                     w
cc
u

COO
   17.00    19.00    21.00     23 00    25 00    27.00     29-00     31.00    33.00

            CflLCINE TEMPERATURE (HUNDREDS OF DEGREES F5=X

-------
 en
 en .,,
 W I
 en
a
cr
u.
QC,
LJ

 •Q
                     OERD  BURN  STUDY/CROSS PLOT NO  06
31*
               X
               K
               )K
                                                                     vo
                                                              Mf-
         ^i        I       n        i        i       r     ^^

   17.00     19,00    21,00    23.00    25,00     27.00     29.00     31.00     33-00

            CRLCINE TEMPERflTURE CHUNDREDS OF DEGREES F)=X

-------
  en'
 en
 03

 cn'
 to
110-1
CJ
CJ
 in
en
~ZL
UJ
cn
•—foi
                      DEflD  BURN  STUDY/CROSS PLOT  NO  07
               *
               *
                                                                      VO
   17.00     19.00     21.00     23 00     25.00     27.00.    29.00    31.00    33.00
            CflLCINE TEMPERflTURE  (HUNDREDS OF DEGREES F:3=X

-------
 O1
 en
 en
 C71
 in
II
r->



O
CO
2
LJ
to
CD
cr
                     DERD*BURN STINGY/CROSS  PLOT NO*U8
               X
   17,00     J9.00    21.00    23.00     25,00    27.00    29,00     31.00     33.00
            tfiLCINE TEMPERflTURE CHUNDREDS OF DEGREES H=X

-------
en
en
in
in'
                   DEflD BURN STUDY/CROSS  PLOT  NO 09
                         X
                         X
                         X
                                                                  vo
             JK
 17.00    39.00     21.00     23.00    25.00    27.00     29.00     31.00    33.00
          CflLCINE TEMPERflTURE CHUNDREDS OF DEGREES

-------
 in_

II
f—>
CD
LJ
21
ID
_J
O
UJ
oc
                     DEflD  BURN  STUDY/CROSS  PLOT  NO  10
                                     *                              5
                         *
en
           I        I	1	1	\	\	1	1
   17.00     19.00    21.00    23.00     25.00     27.00     29.00    31.00    33.00
           CflLCINE TEMPERRTURE  CHUNDREDS OF DEGREES H=X

-------
 r-
 3- .
              K




   H
LJ

CJ
O
§°
Q_
                    DERD  BURN  STUDY/CROSS PLOT NO  11
X

           X

           K
__                                                           *
cos
   17.00    19.00     21.00    23.00    25,00     27.00    29,00    31.00     31 00

            ""      TEMPERflTURE [HUNDREDS OF DEGREES "  "

-------
 D
 O
 o
 o
 D

               ft
A
CD

0
I—/
cc
CE
i — i

Q


LdO
CE
»— «
CD
LJ
                    OEflD  BURN  STUDY/CROSS  PLOT  NO  12
                          K
                                                                    vo
                                                                    Ui
               *
           i        i        i       i        r       i        i       i
   17.00     19.00    21.00    23.00     25.00     27.00    29.00    31.00     33 00
            CflLCINE TEMPERflTURE CHUNDREDS OF DEGREES H=X

-------
  o
  o


  8.
 o
 o
  I
 o
 o.
 o

 03
CJ

DC

LJ
CO


Q-8
UJ

CD
CES
                     DERD  BURN  STUDY/CROSS  PLOT  NO 13

             *



fsi
   37,00     19.00    21.00    23.PO    25,00     27.00     29.00     31.00     33 00
            CRLCINE: TEMPERRTURE (HUNDREDS or DEGREES FD=X

-------
                  197
               APPENDIX C
Cl.  Calcined and Uncalcined Limestones
     Ranked According to Bed Weight Gain

C2.  EPA-APCO Limestone Inventory

-------
                            198
 Cl.   Calcined and Uncalcined Limestones Ranked Accordingto


      Bed Weight Gain




 The following limestones were tested in a fixed-bed reactor


 to determine their capacity to react with flue gas containing


 sulfur dioxide.  The equipment and experimental procedure has

                           (29)
 been  documented by Potter      and is also described in Section C5


 earlier in this report.  Essentially, a fixed bed of 20 grams


 of calcine is exposed to flue gas at 980°C for 3 1/2 hours.


 The flue gas, having an average composition of 10.5% C0», 3.4%


 02, 9.9% H20, 0.27% S02, 0.003% S03, and 75i9% N2, flows through


 the bed at a rate of 425 standard liters per hour.






 The bed weight gain of a 20 gram, -18 +20 U.S. mesh, calcined


 sample (expressed as g S0~/100 g of calcined stone) was selected


 as the index of sulfur oxide capacity.  For uncalcined samples


 a weight equivalent to 20 g of calcined stone was used.  The bed


weight gain, also referred to as "loading" or "capacity", is the


weight of SO- removed from the flue gas.  Data for samples pre-


 calcined at 980°C for 4 hours are presented in Table Cl-1.  Data


 for uncalcined stones are shown in Table Cl-2.






The time to 20% breakthrough is the time required for the S0'2


concentration of the gas effluent from the fixed bed of limestone


 to reach 20% of the inlet SO™ concentration.

-------
                            199
CaO utilization is the percent CaO which reacted based upon




the bed weight gain and percent CaO in the limestone sample.









Spectrochemical analyses were provided by Bituminous Coal Research,




Inc. (BCR).  The limestone numbers in this table were originally




assigned by BCR and limestones were referenced by a "BCR Number".




However, all limestones in the EPA-APCO limestone inventory are




now referenced by a "Lime&tone Number".

-------
                   TABLE  Cl-1.   CALCINED LIMESTONES RANKED ACCORDING TO BED WEIGHT GAIN

                        Composition, weight  %
Limes tone
Number

1371
1376
1336
1682
1341
1381
1379
1375
1382
1340
1677
1353
1354
1699
134
1342
1684
1696
1695
1343
1384
1694
1369
1363
1697

CaO

55.00
13.90
95.00
96.00
56.00
88.00
83.00
29.50
95.00
57.00
42.50
55.00
57.00
96.00
57.00
63.00
57.00
57.00
58.00
94.00
97.00
56.00
96.00
66.00
92.00

MgO
11.80
80.00
2.55
0.50
42.00
7.80
3.80
60.00
1.61
37.00
1.50
42.00
37.00
1.10
37.00
26.00
39.00
35.50
31.50
0.85
1.10
25.00
1.40
1.42
3.70

Fe 0
/ J
1.65
1.63
0.26
0.13
0.46
0.58
0.66
1.65
0.23
2.63
0.95
1.00
0.71
0.20
2.63
0.23
0.84
1.22
0.93
0.66
0.20
1.10
0.20
2.50
0.45

^3
4.65
0.86
0.36
0.20
0.12
0.46
1.27
1.12
0.20
0.27
1.35
0.19
0.43
0.30
0.27
0.30
0.30
0.60
1.32
0.73
0.20
1.40
0.30
6.60
0.66

SiO
*
21.20
3.20
1.49
1.40
0.91
2.35
6.60
4.55
1.12
1.95
53.00
0.96
5.30
1.26
9.5
9.50
2.53
2.00
4.50
2.98
1.00
13.30
1.16
21.00
2.40

CaO+MgO
66.80
93.90
97.55
96.50
98.00
95.80
86.80
89,50
96.61
94.00
44.00
97.00
94.00
97.10
89.00
89.00
96.00
92.50
89.50
94.85
98.10
81.00
97.40
67.42
95.70
CaO Utili-
zation %

00.00
4.92
6.60
9.40
17.00
13.60
17.10
48.65
15.20
27.20
36.60
30.70
31.70
18.60
30.00
29.80
32.80
33.30
33.50
21.60
21.10
37.50
22.70
32.20
24.90
Time to 20%
B . T. , min.

00.00
15.00
5.75
15.00
7.50
36.00
12.00
27.00
21.00
30.00
32.00
26.00
25.00
41.00
17.00
33.00
30.00
40.00
28.00
54.00
28.00
30.00
49.00
60.00
47.00
                                                                                          Bed Wt. Gain,'
                                                                                             grams	

                                                                                             00.00
                                                                                              1.37
                                                                                              1.82
                                                                                              2.52
                                                                                              2.71

                                                                                              3.41
                                                                                              4.08
                                                                                              4:11
                                                                                              4.17
                                                                                              4.41

                                                                                              4.44
                                                                                              4.85
                                                                                              5.12
                                                                                              5.14
                                                                                              5.34
                                                                                              5,
                                                                                              5,
                                                                                              5.
35
39
47
                                                                                              5.55
                                                                                              5.79

                                                                                              5.84
                                                                                              6.01
                                                                                              6.30
                                                                                              6.38
                                                                                              6.55
              S3
              o
              o
.  Breakthrough time is time in minutes required for S02 concentration of effluent gas to reach 20%
  of inlet S02 concentration.
  Gain in weight of bed having initial weight equivalent to 20 -grams of calcined stone.

-------
TABLE Cl-1.  CALCINED LIMESTONES RANKED ACCORDING TO BED WEIGHT GAIN  (continued)
   Composition, weight %
                                                                  1   Bed Wt. Gain,2
                                                                         grams
Limestone
Number
1702
1362
1358
1352
1368
1383
1364
1678
1356
1380
1382
1691
1690
1355
1686
1361
1357
1346
1367
1373
1680
1693
1343
1345
1688

CaO
«""•«••"'«•
60.00
56.00
59,00
56.00
82.00
76.00
57.00
28.50
53.00
57.00
95.00
77.00
34.00
.62.00
56.50
75.00
77.00
46.00
58.00
85,00
49.00
93.00
56.00
63.00
55.00

MgO
32.50
3.50
10.00
31.00
2.17
1.29
3.50
21.00
36.00
41.00
1.61
2.77
23.50
1.82
41.50
7.25
3.35
28.00
28.20
5.30
33.50
0.96
42.00
17.00
42.00

Fe^O
2 3
1.35
18.90
4.00
1.75
1.54
3,65
2.90
1.90
7.50
0.56
0.23
0.86
1.59
2.36
0.41
5.10
1.85
2.55
9.30
1.06
1.48
0.40
0.38
1.54
0.86

Al«0,
	 STT3
0.79
3.65
4.20
2.48
2.20
3.10
8.50
3.15
0.42
0.30
0.20
2.03
2.30
5.60
0.30
2.37
2.80
3.40
0.35
0,95
0.94
0.82
0.22
2.48
0.30

SiO
/
4.10
15.50
20.40
7.40
9.70
13.10
25.60
41.00
1.50
1.53
1.12
15.30
34.60
26.70
1.32
8.20
14.00
17.60
3.85
4.20
13.20
4.00
0.89
13.60
1.00

CaO+MeO
92.50
59.50
69.00
87.00
84.17
77.29
60.50
49.50
89.00
98.00
96.61
79.77
57.50
63.82
98.00
82.15
80.35
74.00
86.20
90.30
82.50
93.96
98.00
80.00
97.00
CaO Utili-
zation %
45.80
44.50
43.70
46.70
32.80
30.40
49.80
107.00
56.80
53.50
32.00
39.70
90.00
50.80
57.50
43.80
42.80
75.50
61.00
40.50
76.90
40.60
69.50
62.90
71.50
Time to 20%
B. T. ? min.
35.00
68.00
58.00
60.00
60.00
68.00
57.00
67.50
62.00
60.00
80.00
66.00
67.00
68.00
52.00
82.00
72.00
96.00
82.50
107.00
95.00
75.00
104.00
79.00
87.00
                                                                          7.08
                                                                          7.12
                                                                          7.42
                                                                          7.49
                                                                          7.76

                                                                          8.02
                                                                          8.08
                                                                          8.66
                                                                          8.66
                                                                          8.70

                                                                          8.70
                                                                          8.74
                                                                          8.75
                                                                          8.97
                                                                          9.21

                                                                          9.39
                                                                          9.41
                                                                          9.99
                                                                         10.14
                                                                         10.16

                                                                         10.76
                                                                         10.82
                                                                         11.17
                                                                         11.29
                                                                         11.35

-------
1701
                   TABLE Cl-1.  CALCINED LIMESTONES RANKED ACCORDING  TO BED WEIGHT  GAIN  (continued)

                       Composition, weight  %
Limes tone
Nuirih er
H \.iui^ c *•
1374
1378
1930
1698
1365
1370
1366
1351
±-J <~t _L.
1685
1679

CaO
65.00
56.00
55.00
72.00
52.50
53.00
66.00
54.00
86.00
91.00

MgO
**a~
18.50
41.00
43.00
19.50
22.00
36.30
16.50
28.50
8.20
2.40

Fe-0-
— 2—3
1.70
0.37
0.36
2.35
15.50
0.79
7.10
7.00
0.69
0.59

^•£0
— 5P3
1.28
0.23
0.14
0.39
1.81
1.21
1.27
1.55
0.25
1.00

SiO
12.20
2.42
0.78
5.85
7.25
7.50
7.50
8.20
2.72
4.10

CaO+MgO
89.50
97.00
98.00
91.50
74.50
89.30
82.50
82.50
94.50
93.40
CaO Utili-
zation %
60.10
74.30
75.90
58.40
84.20
84.80
68.60
86.00
55.00
53.20
Time
B. T.
107
87
132
110
114
100
115
117
138
118
to 20%"
, min.
.00
.00
.00
.00
.00
.00
.00
.00
.50
.00
Bed Wt. Gain,
grams
11
11
12
12
12
12
12
13
13
13
.47
.90
.14
.59
.63
.72
.92
.07
.51
.84
60.00   23.50
                           7.40
0.73
                                  6.80
83,50
                                                                81.00
                                         112.00
                                                                                   13.93
                                                                                                              NJ
                                                                                                              O

-------
TABLE Cl-2.  UNCALCINED LIMESTONES RANKED  ACCORDING TO BED WEIGHT GAIN

    Composition, weight %
Limes tone
Number

1371
1376
1341
1368
1355
1375
1379
1336
1369
1699
1695
1694
1697
1363
1702
1354
1686
1343
1362
1696
1358
1684
1381
1688
1352
CaO Utili- Time to 20%
CaO
55.00
13.90
56.00
82.00
62.00
29.50
83.00
95.00
96.00
96.00
58.00
56.00
92.00
66.00
60.00
57.00
56.50
94.00
56.00
57.00
59.00
57.00
88.00
55.00
56.00
MgO
11.80
80.00
42.00
2.17
1.82
60.00
3.80
2.55
1.40
1.10
31.50
25.00
3.70
1.42
32.50
37.00
41.50
0.85
3.50
35.50
10.00
39.00
7.80
42.00
31.00
Fe 0
• — «£— J
1.65
1.63
0.46
1.54
2.36
1.65
0.66
0.26
0.20
0.20
0.93
1.10
0.45
2.50
1.35
0.71
0.41
0.66
18.90
1.22
4.00
0.84
0.58
0.86
1.75
A1~P_
	 2 o
4.65
0.86
0.12
2.20
5.60
1.12
1.27
0.36
0.30
0.30
1.32
1.40
0.66
6.60
0.79
0.43
0.30
0.73
3.65
0.60
4.20
0.30
0.46
0.30
2.48
SiO.
L
21.20
3.20
0.91
9.70
26.70
4.55
6.60
1.49
1.16
1.26
4.50
13.30
2.40
21.00
4.10
5.30
1.32
2.98
15.50
2.00
20.40
2.53
2.35
1.00
7.40
CaO+MgO
66.80
93.90
98.00
84.17
63.82
89.50
86.80
97.55
97.40
97.10
89.50
81.00
95.70
67.42
92.50
94.00
98.00
94.85
59.50
92.50
69.00
96.00
95.80
97.00
87.00
zation %

0.67
67.20
18.00
14.30
19.00
41.00
14.30
15.33
16.20
17.20
32.20
33.70
20.80-
29.20
42.90
37.20
40.60
22.70
39.80
39.20
39.20
41.20
27.40
44.00
43.96
B. T. , min.

00.00
21.00
7.50
19.00
52.00
16.00
19.00
23.30
35.00
31.00
21.00
34.00
32.00
31.00
27.00
36.00
67.00
41.00
45.00
12.00
45.00
35.00
45.00
40.00
26.00
                                                                       Bed Wt. Gain,2
                                                                          grams
                                                                          00.00
                                                                           2.67
                                                                           2.89
                                                                           3.36
                                                                           3.41

                                                                           3.47
                                                                           3.54
                                                                           4.18
                                                                           4.49
                                                                           5.00
                                                                                           Ni
                                                                           5.35             8
                                                                           5.40
                                                                           5.47
                                                                           5.49
                                                                           6.03

                                                                           6.03
                                                                           6.05
                                                                           6.09
                                                                           6.37
                                                                           6.39

                                                                           6.68
                                                                           6.73
                                                                           6.92
                                                                           6.96
                                                                           7.03

-------
TABLE Cl-2.  UNCALCINED LIMESTONES RANKED ACCORDING TO BED WEIGHT GAIN,  (continued)
    Composition,  weight %
•Limestone
Number
1342
1373
1353
1677
1693
1680
1357
1678
1681
1380
1356
1361
1682
1384
1691
1690
1930
1345
1346
1366
1701
1367
1698
1364
1365
1351
1679
1374

CaO
63.00
85.00
55.00
42.50
93.00
49.00
77.00
28.50
72.00
57.00
53.00
75.00
96.00
97.00
77.00
34.00
55 00
63.00
46.00
66.00
60.00
58.00
72.00
57.00
52.50
54.00
91.00
65.00

MgO
26.00
5.30
42.00
1.50
0.96
33.50
3.35
21.00
3.45
41.00
36.00
7.25
0.50
1.10
2.77
23.50
43.00
17.00
28.00
16.50
23.50
28.20
19.50
3.50
22.00
28.50
2,40
18.50

FeJD
0.23
1.06
1.00
0.95
0.40
1.48
1.85
1.90
0.96
0.56
7.50
5.10
0.13
0.20
0.86
1.59
0.36
1.54
2.55
7.10
7.40
9.30
2.35
2.90
15.50
7.00
0.59
1.70

—2^3
0.30
0.95
0.19
1.35
0.82
0.94
2.80
3.15
1.20
0.30
0.42
2.37
0.20
0.20
2.03
2.30
0.14
2.48
3.40
1.27
0.73
0.35
0.39
8.50
1.81
1.55
1.00
1.28

SiO
9.50
4.20
0.96
53.00
4.00
13.20
14.00
41.00
16.90
1.53
1.50
8.20
1.40
1.00
15.30
34.60
0.78
13.60
17.60
7.50
6.80
3.85
5'. 8 5
25.60
7.25
8.20
4.10
12.20

CaO+MgO
89.00
90.30
97.00
44.00
93.96
82.50
80.35
49.50
75.45
98.00
89.00
82.15
96.50
98.10
79.77
57.50
98.00
80.00
74.00
82.50
83.50
86.20
91.50
60.50
74.50
82.50
93.40
89.50
CaO Utili-
zation %
39.80
29.20
47.20
62.00
28.40
55.50
35.80
100.00
39.60
51.20
54.90
40.00
32.40
32.70
43.80
101.80
66.30
58.30
81.50
58.20
60.80
70.00
55.20
85.30
90.20
93.60
55.00
.80.20
Time
B. T.
70
64
31
45
64
48
42
45
62
49
137
75
62
43
87
52
99
79
112
113
105
82
102
62
165
150
167
115
to 20%
, min.
.00
.00
.00
.00
.00
.00
.00
.00
.00
.00
.00
.00
.00
.00
.00
.00
.50
.00
.00
.00
.00
.00
.00
.00
.00
.00
.00
.00
Bed Wt. Gain,
grams
7
7
7
7
7
7
7
8
8
8
8
8
8
9
9
10
10
10
10
11
11
11
11
13
13
14
14
15
.17
.32
.45
.55
.58
.80
.91
.12
.14
.27
.32
.59
.96
.08
.63
.04
.46
.49
.78
.00
.34
.60
.85
.90
.93
.22
.42
.04

-------
                            205
C2.  EPA-APCO Limestone Inventory






The following is a listing of the current limestone inventory




maintained by EPA-APCO.  These samples are located in Cincinnati,




Ohio, at two main locations: (1) Research Laboratory Branch,




Fairfax Facility, 3914 Virginia Avenue, and (2) U.S. Post Office,




tyain Building, Room 37C.








The limestone samples are listed according to limestone number,




source, identification of material, and the amount of material




on hand as of August 1970.

-------
                            206
         EPA-APCO LIMESTONE INVENTORY - August,  1970
Limestone
 Number

  1335
  1336
  1337
(reordered
 as 1930)
  1338


  1339

  1340



  1341
  1342
  1343
(reordered
 as 1700)

  1344

  1345
  1346
         Source

 Hills Material Co.
 Rapid City, South Dakota
 Supt. of Quarry -
  John Holmes

 Georgia Marble Co.
 Tate, Georgia
  Mr. Hall

 Min., Pigm., & Metals,
 Chas. Pfizer Co.
 Gibsonburg, Ohio
  Mr. Plantz

 Number Not Used
 (changed to 1693)

 Number Not Used

 Nantahala Talc & Limestone
 Lexington, Kentucky
  Mr. Ferebee

 New England Lime Co.
 Div. of Chas. Pfizer
  & Co., Inc.
 Canaan, Connecticut
  Mr. Ingram

 Conklin Limestone Co.
 Lincoln, Rhode Island
  C. E. Conklin

 Hooper Bros. Quarries
 Weeping Water, Nebraska
  Mr. Hooper

 Number Not Used

 New Point Stone Company
 Batesville, Indiana
  Norman A. Wanstrath

Mankato Aglime & Rock Co.
Mankota, Minnesota
                 Amount of
Identification   Material on
  of Material    Hand. (Ibs.)

Limestone Dust       3
//2 White crushed
     marble
#10 glass house
     stone
Aglime
Limestone
Lime
132
38
83
59
67
1/8" down aglime   101
                   258
                   137

-------
                            207
         EPA-APCO LIMESTONE INVENTORY - August, 1970
Limestone
 Number

  1347

  1348


  1349



  1350
  1351
 (reordered
 as 1701)

  1352
 (reordered
 as 1702)

  1353
  1354


  1355
  1356
   1357
   1358
        Source

Number Not Used

J. E, Baker Co.
Millersville, Ohio

TVA Paradise Steam Plant
Drakesboro, Kentucky
  Owen C. Janow

Batesville White Lime
Batesville, Arkansas
General Manager - Mr. Cobb

Jeffrey Limestone Co.
Parma, Michigan
  John C. Jeffrey

Millard Limestone Co.
Annville, Pennsylvania
J. E. Baker Co.
York, Pennsylvania
  Mr. Paul

Montevallo Limestone Co.
Montevallo, Alabama

Elkins Limestone Co.
P. 0. Box 1228
Elkins, West Virginia
  Darrel Hankey

Valley Dolomite
Bonne Terre, Missouri
  Louis Huber

TVA Paradise Steam- Plant
Drakesboro, Kentucky
  Owen C. Janow

Hanna Coal Company
Adena, Ohio
Identification
  of Material
                Amount of
                Material on
                Hand, (Ibs.)
 Ground Raw Stone
 Greenville Quarry
       I '
                    205
                    244
                     147
                     130
                     210
Ground Stone
Dolomite stone
 Ag Lime
 Granular limestone   67
 8 mesh sand
                    232
 Ag Lime
                                                                 83
 Greenville Quarry   176
  II, Limestone gravel
                                            Ag Lime
                     133

-------
                            208
         EPA-APCO LIMESTONE INVENTORY - August. 1970
Limestone
 Number	

  1359
(reordered
 as 1699)

  1360
(reordered
 as 1698)

  1372
  1373
  1374
  1375


  1376


  1377


  1378


  1379
 1380
         Source

Grove Lime Company
Stephens City, Virginia
Monmouth Stone Co.
Monmouth, Illinois
  Dan Kistler

L. F. Rooney
Dept. of Geology
Indiana University
Bloomington, Indiana

L. F. Rooney
Dept. of Geology
Indiana University
Bloomington, Indiana

L. F. Rooney
Dept. of Geology
Indiana University
Bloomington, Indiana

Basic Chemicals
Cleveland, Ohio

Basic Chemicals
Cleveland, Ohio

TVA Russfield
Nashville, Tennessee

Verplanks Coal & Dock
Fettysburg, Michigan

Limestone Products of Amer.
P. 0. Box 490
Newtown, New Jersey 07860
  Mr. Thompson

Rockwell Lime Co.
Manitowac, Wisconsin
Pres. - Michael Brisch
                                              Amount of
                             Identification   Material on
                               of Material    Hand, (Ibs.)
Ag Lime
Sample B
Rockford, Indiana
Sample C
North Vernon,
Indiana
Sample D
Position 8
140
                     29
 72
 84
180
Fluxing Fines


Crestite #3
119


108


 78


136


190
Dolomitic Lime-     148
 stone 1/4" to 1/8"

-------
                            209
         EPA-APCO LIMESTONE  INVENTORY - August, 1970
Limestone
 Number

  1381


  1382


  1383


  1384
  1677
  1678
(reordered
 as 1931)

  1679
  1680
  1681
  1682
  1683


  1684
         Source

Teeter Stone, Inc.
Get tysburg, Pennsylvania

Hemphill Bsos., Inc.
Seattle, Washington

Colarusso & Sons
Hudson, New York

Southern Materials
P. 0. Box 218
Ocala, Florida
V. Pres. - H. B. Roberts, Jr,

Lone Star Materials
P. 0. Box 918
Austin, Texas
  Noble W. Prentice

Dolomite Products
Buffalo fie Howard Road
Rochester, New York

Indus Limestone
Immokalee, Florida

John S. Lane & Son, Inc.
Meriden, Connecticut

Countyline Stone Co.
County Line Road
Akron, New York 14001
Pres. - John W. Buyers

California Rock & Gravel
1800 Hobart Building
582 Market Street
San Francisco, California
Pres. - F. N. Woods III

Union Carbide Corp.
New York, New York

G. & W. H. Corson, Inc.
Plymouth Meeting, Pa. 19462
  Wm. Webster
                 Amount of
Identification   Material on
  of Material    Hand, (Ibs.)

                    158
                    116
                    120
Coarse Granular     200
 CaCO,
Limestone
Screenings
201
Crushed Limestone
5/32" + #35
3/4" limestone
                    114
                     88
                    172
249
                                                                127
                                                                 12
                    170

-------
                            210
         EPA-APCO LIMESTONE INVENTORY - August, 1970
Limestone
 Number

  1685
  1686
  1687
  1688
  1689
(reorder
 of 1361)
  1690


  1691


  1692


  1693


  1694



  1695
         Source

Raid Quarries, Inc.
F. & M. Bank Building
Budington, Iowa
  Paul R. Orr

Mayville White Limeworks
Mayville, Wisconsin 53050
  James 0. Smith
Identification
  of Material

Aglime
Amount of
Material on
Hand. (Ibs.)

   155
Fertilizer Grit
      Company
Frount & 8th Sts.
Quincy, Illinois 62301
Dir. of R&D - Vernon R. Heaton
Quincy Granular
Limestone
   234
   217
Marblehead Lime Co.
300 W. Washington St.
Chicago, Illinois 60606
  John F. Romanyak

Tennessee Valley Authority
Paradise Steam Plant
Drakesboro, Kentucky
  Gene Farmer

McConville, Inc.
Ogdensburg, New York

Warren Brothers Roads Co.
Syracuse, New York

Partin Limestone Products
San Bernadino, California

Pete Lien & Sons
Rapid City, South Dakota

Willingham - Little
Ga. Marble Co., Stone Div.
Lohitestone, Georgia

Willingham - Little
Ga. Marble Co. Stone Div.
Jasper, Georgia
Dolomitic  lime-
   185
                     158
Very Fine Ground
Pool Covering

Gravel
Georgia White
Grandlux
Willingham White
Size #10
   210


   260


   234


     76


   214



   166

-------
                            211
         EPA-APCO LIMESTONE INVENTORY - August. 1970
Limestone
 Number

  1696
  1697
  1698
(reorder
 of 1360)

  1699
(reorder
 of 1359)

  1700
(reorder
 of 1343)

  1701
(reorder
 of 1351)

  1702
(reorder
 of 1352)

  1921
  1930
(reorder
 of 1337)

  1931
(reorder
 of 1678)

  2057
  2058
        Source

James River Hydrate
  Supply Co.
Buchanah, Virginia

Marble Cliff Quarries Co.
Carntown, Kentucky

Monmouth Stone Co.
Monmouth, Illinois
Grove Lime Co.
Stephens City, Virginia
Hooper Bros. Quarries
Weeping Water, Nebraska
Jeffrey Limestone
Parma, Michigan
  John C. Jeffrey

A. E. Millard Limestone Co.
Annville, Pa. 17003
  Mr. Schredder

Chas. Pfizer & Co., Inc.
260 Columbia Street
Adams, Massachusetts
Chas. Pfizer & Co.,
Gibsonburg, Ohio
                                   Inc.
Dolomite Products
Buffalo & Howard Road
Rochester, New York

TVA Prototype Sample
Kentucky Stone
Irvington, Kentucky

TVA Prototype Sample
Kentucky Stone
Russellville, Kentucky
Identification
  of Material

 #3
 Aglime
 Aglime
 #2 Dolomite
 Limestone
Amount of
Material on
Hand, (Ibs.)

    196
                     249
                      40
    158
    205
                     138
    139
 Calcite crystals
 95 brand
 Limestone
                                            Limestone
     98
                      53
                                                 138
     16
                      16

-------
                          212
        EPA-APCO LIMESTONE INVENTORY - August.  1970
                                          Limestone
                                          Blue Limestone
                                          White Limestone
Limestone
 Number                Source

  2059         TVA Prototype Sample
              Road Material, Inc.

  2060         TVA Prototype Sample
              Fredonia Valley

  2061         TVA Prototype Sample
              Fredonia Valley

  2062         TVA Prototype Sample
              Cedar Bluff Limestone
2063         TVA Prototype Sample         Limestone
             National Gypsum

2064         TVA Prototype Sample
             Alabaster Limestone

2065         TVA Prototype Sample
             Longview Limestone

2066         TVA Prototype Sample
             Hoover Limestone

2067         TVA Prototype Sample
             Rigsby and Bernard Limestone

2068         TVA Prototype Sample
             Williams Limestone

2069         TVA Prototype Sample
             James River Limestone

2070         TVA Prototype Sample
             Greenville Limestone

2071         TVA Prototype Sample
             Marble Cliff Limestone
                                                          Amount  of
                                          Identification  Material on
                                            of Material   Hand,  (Ibs.)
                                           Sample 1697 from
                                           Marble Cliff,
                                           Caiutown, Kentucky
16


16


50


16


16


16


20
                                                                15
                                           Sample 1696 from     16
                                           James River Hydrate
2072         TVA Prototype Sample
             Lambert & Lambert Limestone
16
                                                                21

-------
                            213
         EPA-APCO LIMESTONE INVENTORY - August. 1970
Limestone
 Number

  2073

  2074

  2075

  2076

  2077
  2078
  2079
  2080



  2081



  2082



  2083


  2084


  2085
         Source

          TVA

          TVA

          TVA

          TVA

Longhorn Portland Cement Div.
Kaiser Cement and Gypsum
San Antonio, Texas 78218
  Mr. Art Knutson

Indiana Limestone Co.
P. 0. Box 72
Bedford, Indiana
  Mr. Gary (G. W.) Gaiser

Indiana Limestone Co.
P. 0. Box 72
Bedford, Indiana
  Mr. Gaiser

Ohio Geological Survey
Columbus, Ohio
  Mr. Horace Collins

State Geological of Kansas
Lamerence, Kansas 66045
  Mr. R. G. Hardy

Ward's Natural Science
  Establishment
Rochester, New York

Ward's
Rochester, New York

Ward's
Identification
  of Material
Ward's
                                              Austin
                                              Chalk
Standard Bluff
Rustic Gram
Amount of
Material on
Hand. (Ibs.)
Marl (Tufa)
Northeastern Ohio
Kansas Chalk
Iceland spar
Second grade
Iceland spar
First grade

Dolomite
Selasvann, Norway

White Chalk
Dover Cliffs
England
                     83
    58
    70
    98
    55

-------
                            214
          EPA-APCO LIMESTONE INVENTORY - August. 1970
 Limestone
  Number

  2086


  2087
  2105



  2106

  2107



  2108
  2109
  2110


  2111


  2112
2113
         Source

TVA
Spring Valley Limestone

Lone Star Cement
P. 0. Box 839
Demopolis, Alabama
  Mr. Mike Ried

Ward's
Ward *s

Filer's
P. 0. Box 995
Loma Linda, Calif. 92354
Identification
  of Material
Amount of
Material on
Hand, (Ibs.)
Colbert County
Limestone
Selma Chalk
100

94
Sharp pseudo-     20 pieces
hexagonal aragonite
crystals from Spain

New York marl        1

California, tan
banded aragonite
Eckert Educational Mineral
  Research Co.
1244 East Colfax
Denver, Colorado 80218
  Lois Hurianek (303) 272-8943
Massive Colorado
Aragonite
    12
Soil Conservation Service
U.S. Dept. of Agriculture
Midtown Plaza
Syracuse, N. Y.
  Mr. Bernard Ellis

Soil Conservation Service
U.S.D.A.

Soil Conservation Service
U.S.D.A.

Soil Conservation Service
U.S.D.A.
Marl, Canastota,
Madison Cok, N.Y.
    10
Marl, Canastota
Madison Co., N.Y.

Marl, Prattsburg,
Stuben Co., N.Y.

Marl, Gorham,
Ontario Co., N.Y.
    35
    36
    27
Missouri Lead Operating Co.  Lead doped Dolomite, 40
               Boss,  Missouri 65440
                 Mr.  J.  H.  Davis
                             5/8"

-------
                   215
EPA-AKO LIMESTONE INVENTOR* - August. 1970
Limestone
Number
2114
2115
2116
2117
2118
2119
2120
2121
2122
2123
2124
2125
2126
2127
2128
Source
Ward's
Ward's
Ward's
Ward's
Ward's
Ward's
Ward's
Ward's
Ward's
Ward's
Ward's
Ward's
Ward's
Ward's
Ward's
                                                   Amount of
                                  Identification   Material on
                                    of Material    Hand, (Ibs.)

                                  English Oolitic
                                  Limestone

                                  Indiana, Bull,
                                  Oolitic limestone

                                  Pennsylvania,dark
                                  gray, oolitic
                                  limestone

                                  Chocolate,            1
                                  Tennessee, Marble

                                  Black, Vermont,       2
                                  Marble

                                  Gray-White
                                  Mississippi Chalk

                                  Fine  crystalline      3
                                  White Vermont Marble

                                  Coarse crystalline    5
                                  White, Georgia Marble

                                  Coarse crystalline    2
                                  Pink, Georgia

                                  Coarse, White         2
                                  dolomitic, New York
                                  Marble

                                  Fine white dolomitic  1
                                  Mass. Marble

                                  Travertine,
                                  Tivoli,  Italy

                                  Calcareous, Tufa      2
                                  Mumford, N. Y.

                                  Cherty Limestone      2
                                  Leroy, New York

                                  Oolitic  Limestone
                                  Bedford,  Indiana

-------
                             216
          EPA-APCO LIMESTONE INVENTORY - August.  1970
 Limestone
  Number

   2129
   2201



   2202


   2203



   2204



   2205



   2206


   2207



   2208



  (1351)
        Source

Babcock & Wilcox
(from S. K. Vorres
R. R. #2
Fremont, Michigan 47912)

Ecjunga, Durango, Mexico
Supplier: Filer's
Redlands, California (1)

Hillside Mine Dump
Rosiclare, Illinois (1)

Columbia Quarry
  Company Mine
Valmeyer, Illinois  (1)

Allied Stone Co. Quarry
Milan, Illinois  (1)
Midway Stone Quarry
Osborne, Illinois  (1)
Abandoned Quarry
Bourbonnais, Illinois  (1)

Red Mountain District
Santa Clara County,
California   (i)
Identification
  of Material

 Newaygo County
 Michigan marl
Amount of
Material on
Hand. (Ibs.)

     20
 Calcite,  Iceland      1
 spar,  IGS Type 1
 (2)

 Calcite,              20
 IGS  Type  2

 Limestone, coarse-   30
 grained,  high
 purity, ISG Type  3

 Limestone, fine-      30
 grained,  high
 purity, IGS Type  4

 Dolomite, Reef type,  30
 high purity, IGS
 Type 5
 Dolomite,
 IGS  Type 6

 Magnesite,  fine-
 grained, high
 purity,  IGS Type 7
North Cat Cay, Bahama Islands  Aragonite,
Supplier: Ocean Industries,    IGS Type 8
Ft. Lauderdale, Florida
     30


     10



     18
Jeffery Limestone Company    Dolomite,
  Quarry                     IGS Type 9
Parma, Michigan  (1)
                 Same, as Lime-
                 stone 1351
(1)   Supplier:  R.  D.  Harvey,  Illinois State Geological Survey
(2)   ISG - Illinois  State Geological Survey

-------
                            217
         EPA-APCO LIMESTONE INVENTORY - August. 1970
Limestone
 Number

  2210
  2211
        Source

Chrzanow, Poland
Supplier: Power Metrology
  Research Organization
Energopomiar,
  R. T. Chrusciel, Glivice,
  Poland

Nova Huta Plant
Krakow, Poland
Supplier: Power Metrology
  Research Organization
Energopomiar,
  R. T. Chrusciel, Glivice,
  Poland
Identification
  of Material

 Dolomite for
 Nova Huta Iron
   & Steel Plant,
 uncalcined
 Limestone No.  2210
 Calcined in Nova
 Huta Plant, iron
 and steel process
Amount of
Material on
Hand. (Ibs.)

-------
                  218
               APPENDIX D








Sorption of S0_ by Waste Kiln Dust from




Portland Cement Manufacturing Operations

-------
                            219
          Sorption of SO^ by Waste Kiln Dust  from

          Portland Cement Manufacturing Operations
In addition to limestone as a material for sorbing S0_  from

power plant flue gases by dry-injection processes, other  materials

have also been considered.  Some interest has been shown  in kiln

dust as a S0? sorbent.  Kiln dust consists of the fines and con-

densed volatiles blown from the raw mix as it passes through a

rotary kiln in portland cement manufacturing processes.  This

dust is collected from kiln exhaust systems in cyclones and

baghouses.  In order to maintain an acceptable low alkali content

in the cement product, the kiln dust cannot be recombined with

the product.


Table D-l lists chemical analyses for three kiln dust samples.

Table D-2 lists the reactivity of these kiln dust samples to flue

gas in a differential reactor.  Details of the design and operation

of this reactor -are given in  the paper, "Kinetics of the Reaction

of SO  with Limestone"1.  Results for a 100 second exposure to

flue gas are reported in milligrams S03 per 30 milligrams calcined.

When as-received or uncalcined material is tested, the charge

weight used is equivalent to  30 milligrams calcined.  Absorptivity

of the kiln dust is comparable to but no better than Stone No. 2061,

an oolitic calcite.  Moreover, the Na20 content of kiln dust may

be as high as 2.5%,2 whereas, it is only a trace  quantity  (<0.02%)

1 Environmental Science & Technology, Jan. 1970,  p. 59-63.
2 Nebgen, J. W., et al., Midwest Research  Institute Prospectus,
  No. RC-199, May 5,  1970,  p. 2

-------
                            220
                        Table D-l

                Analysis o£ Kiln Dust^ ^
% Loss on Ignition @>1800°F
% Silica (Si02)
% Iron Oxide (Fe,03)
% Aluminum Oxide (Al^O.,)
% Calcium Oxide (CaO)
% Magnesium Oxide (MgO)
North Western
States* '
23.6
15.5
2.6
6.6
49,5
0.9
Medusa
17.3
15.1
3.1
7.3
52.9
3.1
                                                                  (4)
                                                         American

                                                            27.2

                                                            15.9

                                                            3.0

                                                            4.0

                                                            47.5

                                                            2.2
^ '  Analysis by Gilbert Associates,  Inc.,  2249 Fairview  St.,
    Reading, Pennsylvania 19606.  Analyses for alkali content
    not requested.
(2)
    Sample provided by North Western States Portland Cement Co.,
    Mason City, Iowa.

C3)
    Sample provided by Medusa Portland  Cement Co.,  Box 5668,
    Cleveland,  Ohio 44101.

(4)
    Sample provided by American Cement  Corp.

-------
                            221
                        Table D-2
     Reactivity of Kiln Dust in a Differential Reactor1
Temperature
°F
soa
1000
1200
1400
1600
1800
2
Stone No.
2061
(1.5)*
(3.1)*
4.6
6.4
8.0
9.4
mg 50^/30 mg calcine
Samples from Cement Companies
North Western
States Medusa
2.3
2.1
2.9
6.2
6.5
7.2
3.7
3.1
4.4
6.2
6.4
5.9
(uncalcined)
American
—
0.5
1.8
3.0
2.7
4.8
* extrapolated
Note: 1.  100 second exposure to flue gas of following compositions
               Component       Volume %
                 C02            10.5
                 °2              3'4
                 H20             9.9
                 S02             O-27
                 S03             0.003
                 N2             75.9
    2,  Sample supplied by Fredonia Valley Quarries, Kentucky
        (White limestone).

-------
                            222
 in most limes.  Unfortunately, the presence of sodium in fly

 ash decreases fusion temperature of fly ash which leads to

 increased quantity and strength of deposits encountered around

 the superheater and reheater tubes in a power boiler.  Thus,

 injection of kiln dust to remove S0_ before these tubes would

 likely lead to unacceptable periods of "downtime" for tube

 cleaning and loss of heat transfer efficiency.



Table D-3 indicates the reactivity of several kiln dust samples

in an aqueous batch scrubber.  A description of this apparatus

is given by Potter .  The procedure is to bubble flue gas

through a stirred slurry of the tested material and to monitor

the pH of the solution and the S0? content of the scrubbed gas.

An important datum is the breakthrough time, i.e., the length

of time during which the slurry successfully removes all of

the sulfur oxides.  The pH of the solution at breakthrough as

well as the breakthrough time in minutes is given.  Reactivity

of the kiln dust is comparable to finely ground uncalcined

limestone.
   International Symposium for Lime and Limestone Scrubbing,
   Pensacola, Florida, March 16-20, 1970.

-------
                        223
                    Table D-3

   Reactivity of Kiln Dust in anAqueous  Scrubber
North Western
States Kiln Dust
Uncalcined
Calcined
Medusa Kiln Dust
Uncalcined
Calcined
Stone No. 1700^
Dnealcined
Calcined
Breakthrough
Time, min.
44
41
41
39
45
61
pH at
Breakthrough
4.3
4.5
4.4
4.3
5.3
5.5
A calcitic aglime supplied by Hooper Brothers Quarries,
Weeping Water, Nebraska.

-------
            224
         APPENDIX E
Copper Oxide Sorption of SO-

-------
                            225
               Sorption of SO^ by Copper Oxides
The Air Preheater Company and the Owens-Corning Fiberglas




Corporation have established the fabric filterhouse to be an




effective chemical contactor for removal of sulfur dioxide from




flue gas streams.  It has been suggested that one potential



application for the filterhouse would be to remove S09 from




the flue gases from the reverbatory furnace in the copper




smelting industry.  The idea would be to inject either cupric




oxide or the solid effluent from the roaster or reverberatory




furnace into a filterhouse along with the effluent gases from




the reverberatory furnace.  After sorbing S0« the solid reaction




products would then be fed back into the roaster feed stream




where the sorbed S0? would be driven off in the roaster to




enrich the roaster effluent gases which contain five to ten




percent S0~ by volume.  The roaster effluent gases containing




SO  are normally sent to a sulfuric acid manufacturing process.








An exploratory investigation was made to show the reactivity




of cupric oxide to sulfur dioxide.  Chemical grade cupric oxide




(Baker A.C.S. grade) was exposed to pure S02 in a thermogravi-




metric analyzer (an Aminco Modular Thermo-Grav).  In the




thermogravimetric analysis (TGA) mode it automatically records




sample weight change as a function of sample temperature at

-------
                            226
pre-selected heating rates.  In this case the TGA was performed
with 100 milligrams of cupric oxide exposed to a 100 cc/min SO-
purge and heated at a rate of 6°C/min.  Figure E-l shows the
reactivity of cupric oxide with SO-.


As shown in Figure E-l the cupric oxide reached a maximum weight
gain of 50% at a temperature of about 720°C, after which the
reacted sample underwent decomposition.  The significance of
the break in the graph between 560 - 695°C is not known.


The significant information to be obtained from Figure E-l is
that at 3l5°C (600°F) there is only a 2-3 percent weight gain
by the cupric oxide.  Since commercially available fabric
filters have a maximum operating temperature of 600°F, cupric
oxide would not be art effective sorbent for a fabric filterhouse
application unless fabric filters with higher operating
temperatures can be developed.

-------
50
                                                                                           10
                                                                                           N>
    50     150    240     340     445  500       620   695  750    870   950

                                TEMPERATURE, °C.

        Figure E-l.   Thermogravimetrlc Analysis of Reactivity of Cupric Oxide  to SO-

-------
                 228
              APPENDIX F




Char Sorption of SO  and Regeneration

-------
                            229
CHAR SORPTION OF S02 AND REGENERATION
A number of processes under development for the control of




sulfur oxides emoloy the catalytic properties of activated




carbon to oxidize sulfur dioxide to sulfur trioxide on the




carbon surface and to retain the sulfuric acid formed from the




sulfur trioxide in the pores of the activated carbon.  Thermal




regeneration of the carbon results in significant consumption




of the rather expensive activated carbon.  In order to avoid




this problem associated with the regeneration of the spent




carbon, the West Virginia Pulp and. Paper Company (Westvaco)




has suggested a reductive type regeneration in which hydrogen




sulfide is used*  The hydrogen sulfide is reacted first with




a portion of the sulfuric acid to yield elemental sulfur.  A




part of the sulfur then reacts with the remainder of the




sulfurtc acid, presumably preventing reaction of the sulfuric




acid with the carbon and thereby averting consumption of the




carbon-








Several experiments to check the char sorption-regeneration




scheme proposed by Westvaco were run.  The following aspects




of this scheme were tested: (1> S02 sorption capacity of the




char at 25Q°F, (2) the reduction of H2S04-loaded char by ^S




at 300°F, 03) the reduction of H2S04 with sulfur at 600°F.

-------
                             230
 The  attached data summarize the results obtained in these



 experiments.  The following additional comments are offered:





 SO   Sorption
  £.
                                        3
 A fixed bed containing 9.38 grams (15 cm ) of char (supplied by



Westvaco) was exposed to flue gas containing 2800 ppm SO- (dry



basis) at 250°F.  The gas flow was 505 ml. per minute (21°C)



 through the bed, for a space velocity of 2000 hr. .  Figure F-l



 gives the'breakthrough curve for the run, as measured with an



 infrared analyzer continuously recording the SO- concentration



 in the reactor effluent gas stream.   S07 concentration in the



 effluent gas stream reached 10 per cent of the inlet concen-



 tration (10% breakthrough) in 6.65 hours, at which time the



char loading was 15.8 wt. percent SO-.  Exposure to flue gas



was ended at 8.75 hours.  The weight gained by the char during



 the total exposure period was 2.80 grams.  Integration of the



IR curve indicated an SO- sorption equivalent to 2.83 grams ELSO,.





         Reduction
Three samples of char from the experiment described above were



exposed to pure H_S at 300 °F using the apparatus shown in Figure F-2.



The gas volumes were measured in 100 ml burettes using saturated



salt solution (acidified with HC1) as the displacement fluid.



Different volumes of H-S were fed to each char sample, equivalent



to 33%, 88% and 100% stoichiometric to H-SO, based on the assumed



reaction

-------
   3000  -
                                                                    Space Velocity »  2000 hr
                                                                                           -1
   2500
                                                           S02 Feed Level, 2800 ppm

-------
                        232
                                     DRYING
                                     TUBE
                            I
                CHAR
               SAMPLE
                                         300° F.
                                       CHL BATH
                           HOT PLATE
    FEED BURETTE
GAS  COLLECTION
   BURETTE
Figure F-2.  Apparatus used  for H SO,/H2S Reduction Experiments

-------
                             233
The data  obtained  are  summarized in Table F-l, along with  three
additional  runs made with as-received char which had not been
exposed to  flue gas.   Figures F-3, 4 and 5 show the relationship
between the amount of  H S fed to the char (at constant rate)
and the amount of  gas  which passed through the char bed.   The
gas collected includes the air displaced from the reactor  tube
which had a volume of  28 ml.

H0SO,/S Reaction
""""".£ •~~Xjr~"i—•• •••    • ' •
The char  samples obtained from the above H.S treatment were
each heated to 600°F under flowing nitrogen using the apparatus
shown in Figure F-6.   The presumed reaction for this stage is
                H2S04  + 1/2 S •»• 3/2 S02 + H20
The sulfur dioxide collected in the peroxide absorbers was
analyzed  to determine  the total S0_ evolved.  When S0_ evolution
was completed, the char was removed and extracted in a Soxlet
apparatus with carbon  disulfide to recover elemental sulfur.
Results of these tests are listed in Tables F-2 and 3.  Finally,
the three samples  of as-received char which were exposed to H-S
(runs A, B, C), were composited and extracted with carbon  disul-
fide.  Sulfur recovery was 0.288 grams for the composite of nine
grams of unloaded  char for a baseline loading of 0.032 grams of
sulfur per gram of unloaded char.  This may be compared with
the results for runs 1, 2, and 3 in Table F-3 where the maximum
sulfur recovery was 1.52 grams of sulfur on 3.56 grams of  loaded
char or 0.745 grains of sulfur per gram of unloaded char (i.e.,

-------
                            234
        Table F-l: H2S04/H2S Reaction Stage at 300°F





Weight H0SO,-loaded char put in reactor tube » 3.82 grams each run
       ~~VL  *f


Feed rate of H S to char during reaction =1.6 ml/min  (20°C, 1 atm)





                               Total gas
Total H2& passed thru
added, ml. char, ml.
Run (Stoich.) (20°C, 1 atm.) (20°C, 1 atm.)
1 (33%)
2 (88%)
3 (100%)
Weight as-received
Feed rate of H-S to
A
B
C
219
575
655
char
char
219
575
655
28
34
76
put in reactor tube =
during reaction » 1.
80
262
381
Wt. gain Wt. gain
by char, by drying
gm. tube, gnu
0.187
0.576
0.713
2.92 grams
6 ml/min.
0.052
0.069
0.046
0.107
0.042
0.035

0.035
0.058
0.052

-------
                            Total H2S Fed  « 33% Stoichiometric
                                            H-SO, Loaded Char
30
60         90        120       150
            H2S Fed, ml. (20°C,  1 atm.)
       Figure F-3.   300°F H2S Reduction
180
210
240

-------
360 -
     Total H2S Fed - 88% Stolchiometric
                                         Char as received
                                                   HnSO,  Loaded Char
                                                                                                 10
           60
120
180     240     300     360     420
       H2S Fed, ml.  (20°C, 1 ata.)

    Figure F-4.  300°F ttJS Reduction
460    540
                                                                                        600

-------
360 -
                      Total H2S Fed « 100% Stoichlattetric
                                                          H-SO,  Loaded Char
                                                                                                         10
                                       H2S Fed, ml. '(20eC,  1 atm.)



                                    Figure F-5.  300°F H.S  Reduction
                                                                                                660

-------
                       238
           CHAR
          SAMPLE
DRYING
 TUBE
               ELECTRIC  FURNACE

                      600°F
                                         H202  ABSORBERS
Figure F-6.  Apparatus used for H-SO, Reaction Experiments

-------
                             239
       Table  F-2:  H2SC>4/S Reaction Stage at 600°F
Char from
Run No.
1 (33%)
2 (88%)
3 (100%)
Wt. char to
reactor, gm.
3.85
4.21
4.30
SO- recovered,
gm.
0.568
0.227
0.169
Wt. loss of
char during
reaction, gm
0.900
0.789
0.769
                Table F-3:  Sulfur Extraction
   Char from
Run No. (Table 2)
  Wt. char to
Extractor, gm.
   Sulfur
recovered, gm.
1 (33%)
2 (88%)
3 (100%)
2.96
3.42
3.56
0.183
0.449
1.52

-------
                            240
23 times the baseline loading).  The work showed conclusively




that ELSO, can be reduced to elemental sulfur in yields exceeding




80 percent at a constant temperature of 300°F, and that the pro-




duct of that reduction can be recovered by solvent extraction




methods.

-------