PROCEEDINGS OF SECOND INTERNATIONAL
            LIME/LIMESTONE
       WET-SCRUBBING SYMPOSIUM
                             VOLUME I

U     E N V I R 0 N M E N T A L  P R

-------
             PROCEEDINGS


        SECOND INTERNATIONAL


           LIME/LIMESTONE


       WET-SCRUBBING SYMPOSIUM
              VOLUME I
         November 8-12, 1971
       Sheraton-Charles Hotel
       New Orleans, Louisiana
   ENVIRONMENTAL PROTECTION AGENCY
       Office of Administration
Research Triangle Park, North Carolina
               June 1972

-------
The APTD (Air Pollution Technical Data) series of reports is
issued by the Environmental Protection Agency to report tech-
nical data of interest to a limited number of readers.  Copies
of APTD reports are available free of charge to Federal employ-
ees, current contractors and grantees, and nonprofit organiza-
tions - as supplies permit - from the Air Pollution Technical
Information Center, Environmental Protection Agency, Research
Triangle Park, North Carolina 27711 or from the National Tech-
nical Information Service, 5285 Port Royal Road, Springfield,
Virginia 22151.
                    EPA REVIEW NOTICE

These proceedings have been reviewed by the Environmental
Protection Agency and approved for publication.  The contents
of this report are reproduced herein as received from the
authors.  Approval does not signify that the contents neces-
sarily reflect the views and policies of the Environmental
Protection Agency, nor does mention of trade names or commer-
cial products constitute endorsement or recommendation for
use.
                   Publication Number APTD-1161
                                 11

-------
                       PREFACE
     The Second International Lime/Limestone Wet Scrubbing
Symposium was held November 8-12, 1971, in the Claiborne
Room of the Sheraton-Charles Hotel, New Orleans, Louisiana,
sponsored by the Environmental Protection Agency, Office
of Air Programs, Control Systems Division.

     The Symposium, under the chairmanship and vice-chair-
manship of Messrs. E.L. Plyler and D.R. Mayfield, began
Monday morning with the official welcome and opening remarks
by OAP's Sheldon Meyers, followed by an introduction by
Mr. Plyler.

     The Symposium consisted of nine sessions, divided into
five different areas: fundamental research, pilot scale
research and development, prototype and full scale tests,
panel discussion on scaling, sampling and analytical methods.

   Sessions 1 and 2, chaired by Philip S. Lowell, were
concerned with fundamental research.  Frank T. Princiotta
of OAP was the chairman for Sessions 3 and 4 and the first
three presentations of Session 5, dealing with pilot scale
research and development.

     The last two presentations of Session 5 and Sessions 6
and 7 were chaired by H.W. Elder.  The 17 papers presented in
these sessions were concerned with prototype and full scale tests

     The panel discussion on problems related to scaling in
lime/limestone wet scrubbing,  chaired by A.V. Slack, consisted
of the five papers of Session 8.  OAP's J.A. Dorsey chaired
Session 9, sampling and analytical methods, which concluded the
symposium early Friday afternoon.

     All papers presented during the symposium are included
in these proceedings except those which were given by notes
and for which there exists no written text.
                            111

-------
                              CONTENTS

Title                                                            Pacfe
                          VOLUME I
     Preface ..................................................  iii

     Philip S. Lowell
     SUMMARY: FUNDAMENTAL RESEARCH - PARTS I AND II  , ..........    1
                  FUNDAMENTAL RESEARCH - PART I

     C.Y. Wen and S. Uchida
     Simulation of SC>2 Absorption in a Venturi Scrubber
        by Alkaline Solutions
     M. Epstein, C.C. Leivo, and C.H. Rowland
     Mathematical Models for Pressure Drop, Particulate
        Removal and 803 Removal in Venturi, TCA, and
        Hydro-filter Scrubbers
     Delbert M. Ottmers, Jr.
     A Model for the Limestone  Injection-Wet Scrubbing
        Process for Sulfur Dioxide Removal  from Power Plant
        Flue Gas

     James L. Phillips
     Precipitation Kinetics of  CaSO4.2H20
     D.C. Drehmel
     Limestone Types for Flue Gas  Scrubbing  ...................   167
                  FUNDAMENTAL RESEARCH  -  PART  II

     J.M. Potts, A.V.  Slack, and  J.D. Hatfield
     Removal  of  Sulfur Dioxide  from Stack Gases by
        Scrubbing with Limestone  Slurry:  Small-Scale
        Studie s  at TVA 	   195

     L.H. Garcia
     Absorption  Studies of  Equimolar Concentrations of NO
        and NO2  in Alkaline Solutions 	   233

     J.D. Hatfield and J.M; Potts
     Removal  of  Sulfur Dioxide  from Stack Gases by  Scrubbing
        with  Limestone Slurry:  Use  of Organic  Acids 	   263

-------
                        CONTENTS  (continued)

Title                                                             Page

     J. S. Morr i s
     Potential Water Quality Problems Associated with
        Limestone Wet Scrubbing for SC>2 Removal from
        Stack Gases 	   285

     Linda Z. Condry,  Richard B. Muter and William P. Lawrence
     Potential Utilization of Solid Waste from Lime/Limestone
        Wet Scrubbing of Flue Gases 	   301
     Prank T. Princiotta
     SUMMARY: PILOT SCALE RESEARCH AND DEVELOPMENT - PARTS  I,
        II,  AND III 	   315

             PILOT SCALE RESEARCH AND DEVELOPMENT - PART  I

     B.N. Murthy, D.B. Harris and J.L. Phillips
     Sulfur Dioxide Adsorption Studies with EPA  In-House
        Pilot Scale Venturi Scrubber 	   325

     I.S. Shah
     SO2 Removal Using Calcium Based Alkalies Pilot Plant
        Experience 	   345

     R.A. Person, C.R. Allenbach, I.S.  Shah, and S.J. Sawyer
     A Pilot Plant Test Program for Sulfur Dioxide Removal
        from Boiler Flue Gases Using Limestone and Hydrated
        Lime 	   373

     R.J. Gleason
     Limestone Scrubbing Efficiency of Sulfur Dioxide in
        a Wetted Film Packed Tower in Series with a Venturi
        Scrubber 	   391

     T.M. Kelso, P.C.  Williamson, and J.J. Schultz
     Removal of Sulfur Dioxide from Stack Gases by Scrubbing
        with Limestone Slurry: TVA Pilot Plant Tests.
        Part I - Scrubber-Type Comparison 	   437

     N.D. Moore
        Part II - Experimental Design and Data Analysis for
        Spray and Mobile-Bed Scrubbers 	   462
                                   VI

-------
                         CONTENTS  (continued)

Title^                                                             Page

            PILOT SCALE RESEARCH AND  DEVELOPMENT  -  PART II

     A. Saleem, D. Harrison, and N. Sekhar
     Sulfur Dioxide Removal by Limestone  Slurry in  a  Spray
        Tower  	    481

     J.L. Shapiro and W.L. Kuo
     The Mohave/Navajo Pilot Facility for Sulfur  Dioxide
        Removal 	    507

     J.H. McCarthy and J.J. Roosen
     Detroit Edison Pilot Plant and Full-Scale Development
        Program for Alkali Scrubbing  Systems—A Progress
        Report 	    527

     A.L. Plumley and M.R. Gogineni
     Research and Development in Wet  Scrubber Systems 	    541

     D.E» Reedy
     Lime Scrubbing of Simulated Roaster  Off-Gas  	    561

                           VOLUME  II
           PILOT SCALE RESEARCH AND DEVELOPMENT - PART III

     John M. Craig, Burke Bell, and J.M.  Fayadh
     Mobile Pilot Plant Study of the  Wet  Limestone  Process
        for SO2 Control 	    575

     Robert J. Phillips
     Sulfur Dioxide Emission Control  for  Industrial Power
        Plants 	    603

     Ivor E. Campbell and James E. Foard
     Sulfur Oxide Control at the Copper Smelter 	    639
     H.W. Elder
     SUMMARY: PROTOTYPE AND FULL SCALE TESTS -  PARTS  I,  II,
        AND III 	    653

               PROTOTYPE AND FULL SCALE TESTS - PART  J

     James Jonakin and James Martin
     Applications of the C-E Air Pollution Control  System 	    657
                                  vii

-------
                        CONTENTS  (continued)

Title
     E.G. McKinney and A.F. Little
     Removal of Sulfur Dioxide from Stack Gases by
        Scrubbing with Limestone Slurry: Design Considerations
        for Demonstration Full-Scale System at TVA  ............    673
               PROTOTYPE AND FULL SCALE TESTS - PART II

     M. Epstein, F. Princiotta, R.M. Sherwin, L.  Szeibert,
     and I. A. Raben
     Test Program for the EPA AiKali Scrubbing Test Facility
        at the Shawnee Power Plant  ............................    693

     R.M. Sherwin, I. A. Raben, and  P.P. Anas
     Economics of Limestone Wet Scrubbing  Systems .............    745

     J.D. McKenna and R.S. Atkins
     The RC/Bahco System for Removal of Sulfur Oxides and
        Fly Ash from Flue Gases ...............................    765

     Gerhard Hausberg
     The Bischoff-Process — Initial  Results from a Full-Size
        Experimental Plant ....................................    785

     J.J. O'Donnell and A.G. Sliger
     Availability of Limestones and Dolomites  .................    799
               PROTOTYPE AND FULL SCALE TESTS  -  PART  III

     Tsukumo Uno, Masumi Atsukawa, and Kenzo Muramatsu
     The Pilot Scale R&D and Prototype Plant of  MHI Lime-
        Gypsum Process  ........................................    833

     Lyman K. Mundth
     Wet Scrubber Installations  at Arizona  Public  Service
        Company Power Plants  ..................................    851

     J.H. McCarthy and  J.J. Roosen
     Detroit Edison Full-Scale Development  Program for
        Alkali Scrubbing Systems  (The material  in this
        paper was included  in Paper  No. 4c.)  ..................    863
                                   Vlll

-------
                         CONTENTS  (continued)

Title                                                             Page

     Robert R. Padron and Kenneth  C. O'Brien
     A Full-Scale Limestone Wet  Scrubbing  System for  the
        Utility Board of the City  of Key West,  Florida  	    865

     J.A. Noer and A.E. Swanson
     Air Pollution Control at the  Northern States Power
        Company Sherburne County Generating Plant 	    877

     J.W. James
     Ontario Hydro's Prototype Limestone Scrubber for SO2
        Removal from Clean Flue  Gas  	    899

     D.T. McPhee
     La Cygne Station Air Quality  System	    907

     J.F. McLaughlin, Jr.
     Sulfur Dioxide Scrubber Service Record Union Electric
        Company—Meramec Unit 2  	    915

     B.C. Gifford
     Will County Unit 1 Limestone  Wet Scrubber  	    917

     H.P. Willett and I.S. Shah
     A Summary Report—Chemico's Commercial Systems
        Installations at Electric  Power Generating Stations  ...    931
     A.V. Slack
     SUMMARY: PROBLEMS RELATED TO SCALING  IN LIME/LIMESTONE WET
        SCRUBBING 	    943

        PROBLEMS RELATED TO SCALING IN LIME/LIMESTONE WET  SCRUBBING

     A.V. Slack and J.D. Hatfield
     Removal of Sulfur Dioxide from Stack  Gases by  Scrubbing
        with Limestone Slurry: Operational Aspects  of the
        Scaling Problem	    947

     Bela M. Fabuss
     Calcium Sulfate Scaling  	    965

     Joan B. Berkowitz
     Review of Scaling Problems in Limestone Based  Wet
        Scrubbing Processes	    975
                                   IX

-------
                    CONTENTS (continued)

                                                            Page
     Martin
J.K. r^r^ni
Deposition Problems and Solutions in the Combustion
   Engineering Lime/Limestone Wet Scrubbing Systems  	


!»hilip S. Lowell
Use of Chemical Analysis  and Solution Equilibria  in
   Predicting  Calcium Sulfate/Sulfite Scaling Potential  ..   1001
J.A. Dorsey                                                   .
SUMMARY:  SAMPLING  AND ANALYTICAL METHODS 	   101.3


              SAMPLING AND ANALYTICAL METHODS


Klaus  Schwitzgebci
Development and Field Verification of Sampling and
    Analytical Methods for Shawnee 	


E.A.  Burns and A.  Grunt
On-Stream Characterization of the Limestone/Dolomite
                 _                                     ....   I UDo
    Wet Scrubber Process  	


Terry Smith and Ronald Draftz
^articulate Emissions from Two Limestone Wet Scrubbers ...   IU/J


Terry Smith and Hsing-Chi Chang                   ^
 Design Criteria for  a Size-Selective Sampler for Lime/

    Limestone Wet  Scrubbers  	


 H.M. Statniclc and J.A. Dorsey
 Instrumental Methods for Flue Gas Analysis	
 Gene W. Smith
 EPA Recommended  Source  Test  Methods  for new  Source
     Performance Standards Testing
                                                              1109

-------
   FUNDAMENTAL RESEARCH
Philip S. Lowell, Chairman
       Participants:

   C.Y. Wen and S. Uchida
   M. Epstein, C.C. Leivo, C.H. Rowland
   Delbert M. Ottmers, Jr.
   James L. Phillips
   D.C. Drehmel
   J.M. Potts, A.V. Slack, J.D. Hatfield
   L.H. Garcia
   J.D. Hatfield and J.M. Potts
   J.S. Morris
   Linda Z. Condry, Richard B. Muter,
   William P. Lawrence

-------
                             SUMMARY
                       FUNDAMENTAL RESEARCH
                          (PARTS  I AND  II)
                    Philip  S. .Lowell, Chairman
          Lime and limestone based processes for control of
SOS emissions from boiler flue gases are nearing commercializa-
tion.  The overall process involves removing SOS from the flue
gas in a scrubber.  In auxiliary portions of the system the
lime or limestone is dissolved.  Calcium sulfite and sulfate
are precipitated and removed from the system.

          The objective of the research part of process
development is to provide an understanding of the process.
This understanding of lime/limestone wet scrubbing processes
will provide the basis for several functions;

              design of laboratory and pilot plant tests,

              provision of a framework within which experi-
              mental data may be interpreted,

              •provision of the basis for extrapolating
              from bench and pilot work to full size,
              i.e., process design data and methods,

              solution of operating problems..

          The major subjects of fundamental research involve
the chemical reactions, mass transfer steps, equipment charac-
teristics, and the interactions of the above three.  In addition
to the main thrust of process development there are potential
ancillary problems that must be solved.  Two of these are solid
waste disposal and prevention of ground and stream water
pollution.

-------
SIMULATION OF S02 ABSORPTION IN A VENTURI SCRUBBER




               BY ALKALINE SOLUTIONS
             CONTRACT NO.  EHS-D-71-20
                   Prepared for:




      Lime/Limestone Wet Scrubbing Symposium




              New Orleans,  Louisiana




                8-12, November 1971
                        By



               C. Y.  Wen and S.  Uchida




        Department of Chemical Engineering




             West Virginia University




             Morgantown, West Virginia

-------
                       CONTENTS



1.     Introduction

2.     Mechanics of Venturi Scrubber Operation

       a.   Pressure drop in a venturi scrubber

       b.   Mean diameter of a liquid droplet

       c.   Equation of motion of a liquid drop

3.     Gas Absorption Rate in a Venturi Scrubber

       a.   Overall rate of SO^ absorption into a
            droplet of alkaline solution

       b.   Mass transfer and heat transfer coefficients
            of liquid droplet

       c.   Equilibrium data and absorption rate data

4.     Development of Performance Equations for a
       Venturi Scrubber

5.     Simulation of Performances of Various Venturi
       Scrubbers

       a.   O.A.P. Experiments

       b.   Experiments of Johnstone, et.  al.

       c.   Cottrell Evnironmental Systems' Experiment

6.     Conclusion

-------
1.  Introduction




       Although venturi scrubbers are widely used for simultaneous




removal of gaseous and particulate pollutants, studies of mass transfer




and chemical reactions taking place in venturi scrubbers have been few.




Operations of venturi scrubbers usually involve gas, liquid,  and solid




phases with considerable pressure drops in the system.




       The work presented here is sponsored by the Office of  Air Programs,




Environmental Protection Agency, and is an attempt to elucidate the




phenomena taking place in the venturi-type absorber both from the




mechanical and absorption kinetic points of view.  Thp design and scale-




up criteria of the venturi scrubber can thus be developed so  that more




efficient and economic commerical scrubbers can be built.  To meet this




goal, mathematical models are developed and the performances  of various




venturi scrubbers are simulated and then compared with the experimental




data available.




       It is hoped that the models developed can be used as an aid in the




design and scale-up of venturi scrubbers for S02 absorption under different




operating conditions.






2.  Mechanics of Venturi Scrubber Operations




       a.  Pressure Drop in a Venturi Scrubbfc-




              Pressure drops for gas flowing through a venturi scrubber




are estimated by knowing the effects of friction along the wall of the




equipment and the acceleration of liquid.  Frictional loss depends largely




upon the geometry of the scrubber and must be determined experimentally.

-------
The effect of acceleration of liquid is, however, predictable theoretical-




iy.



              Several correlations available in the literature are




summarized in Table 1.




              Matrozov's correlation [16] and Volgin's correlation [22]




were obtained mainly on small size venturi scrubbers.  Calvert [2] derived




a pressure drop equation by use of Newton's Law to obtain the force •




required to change the momentum of liquid at a given rate.  This equation




and Volgin's correlation do not contain the term responsible for the




frictional loss on the wall of the venturi scrubber, which becomes negligibly




small at a high liquid rate in comparison with the acceleration loss.




              The fourth correlation has been obtained experimentally for




Flooded Disk scrubbers using the sodium carbonate solution by Cottrell




Environmental Systems, Inc. [6]




       b.  Mean^ Diameter of a Liquid Droplet




              Various correlations are available in the literature to




estimate the mean liquid drop diameter from different types of atomizers




under different operating conditions.  These correlations are applicable




within.a certain range of operating conditions and properties of fluids;




such as volume ratio of gas to liquid, relative velocity of gas to




liquid, type of nozzle, surface tension of liquid, etc.  In using one of




these correlations to estimate the droplet diameter, it is important to




select a proper correlation which takes these factors into consideration.




              Applicabilities of various correlations in terms of the




range of mass ratio of gas to liquid, the relative velocity of gas to
                                 6

-------
              Table 1.  Correlation for Pressure Drop
Investigator   (Reference)
                              Equation
Matrozov
(16)
Volgin, et.al.      (23)
Calvert
( 2)
Gleason & Mckenna   ( 6)
      APd + ISw1-08*0'63



      2.22 X 10-W-V'143
AP* « 5 X 10~5(v )2R
  •^             o




AP4 = 4.86 X 10~6v 2(L + 73.8)
                  O

-------
liquid, and the viscosity of liquid are shown in Table II.  Among these

correlations, probably the best known and the most widely used is that of

Nukiyaraa and Tanasawa (18].  Analysis of this correlation shows that its

applicability is limited to the range of high relative velocities and large

mass ratio of gas to liquid.

              Kim and Marshall [12] recently developed a correlation which

covers a wide range of important variables.  Their correlation has a form

similar to the Nukiyama-Tanasawa equation, but predicts smaller droplet

sizes under similar operating conditions.  The Nukiyama-Tanasawa equation

was based on data obtained by physical sampling technique which probably

introduced some errors due to the evaporation and target effect to the

sample collecting device.  Kim and Marshall [12] also reported that a

modified form of their correlation can be used to correlate Wetzel's

drop-size data [24] for venturi atomization.

       c.  Equation of Motion of a Liquid Drop

              Assuming that there is no change in the drop diameter, and

that every droplet has the same diameter, velocities of gas and liquid can

be calculated as follows.

              The average gas velocity at any point in a venturi scrubber

is calculated by the following relation:

                             £2
                     Vg   *  eS

              The velocity of a droplet can be calculated by
                              (l-e)S


                                   8

-------
Table 31  Empirical Equations for Droplet Diameter and their Applicable Ranges
Investigator
[Reference]
Nukiyama , S .
and Y. Tanasawa
(1939) [18]
Mugele, R.A.
(1960) [17]
Gretzinger, J.
and W.R. Marshall
(1961) [7]
Wigg, L.D.
(1964) [25]
Kim, K.Y.
and W.R. Marshall
(1971) [12]


Equation
rS Ul 0.45 L 1.5
D - 585 /— + 597 ( / — ) • (1000 }
32 v /p. •op, G
ro i» Jj o
D32 DnPLvro B ^Lvro °
Dn = A ^L °

l)m - 2600[( )'( )J
g Mg
yL 1/2 ML !/2 _3
a0.41y0.32
— . . — . ^J
D_ " 249 n
(v o ^0.5750. 3600. 16
+ 1 O £. t\ f ^* \ • ^^-^J i 	 f 1 ,N
IzoU ( ; 0~54^M
m » -1 for Mg/Mj^ < 3
m " -0.5 for Mg/ML > 3
Appld
v [ft/sec]
ro
300 to
sonic
velocity
300 to sonic
velocity
300 to
sonic
velocity
270 to
1150
250 to
sonic
velocity


.cable Range
(Mg/ML) [-1
1.8 to
15
1.8 to
15
1 to
15
0.5 to
20.2
0.06 to
40


i
UL[c.p.]
1 to
46
—
1 to
30
3.2 to
45
1 to
50



-------
              Equation of motion of a liquid drop can be written from  the

momentum balance of a drop within the flowing gas as follows [10]:

          dvL           TtD3       irD;J      TrD^ pg
              If the motion of the droplet is in the vertical direction,

the second term can be neglected in comparison to other terms.   Since,

dZ      , the above equation can be written as
dt B VL


              f!t   _ *  + I    fi  .  '£* /vg"vL>lvvL|
              dZ    " VL   4  '  Dp '  pL '       VL                 (2)


              If the motion is in the horizontal direction,  the term due

to gravity can be neglected and simplified to


              f*t   „  1 . - . £* . 
-------
These are:

       — No coalescence or breaking up of the droplets occurs after
          atomization.  Droplets keep the same mean diameter as that
          created at the nozzle point.

       — Liquid droplets are all spherical.

       — The variation in droplet diameter due to evaporation may be
          neglected.

       — Heat of reaction is negligibly small.

       a.  Overall Rate of SO? Absorption into Droplets of Alkaline

           Solutions.

              The kinetic theory of simultaneous diffusion and chemical

reaction in the liquid phase has been developed by Hatta [8], Davis, et.

al [3], and others based on the two-film theory for physical absorption

originally proposed by Lewis and Whitman [14].

The experimental results obtained by Linn, et. al. [15] and Hikita, et.

al.[9] indicate that the absorption of S02 by water may be considered as

physical absorption.  The agreement of data for SC^-alkaline systems

obtained by Hikita, et. al.  [9] and Onda, et. al.  [19] with that

calculated by the physical penetration theory suggest that the absorption

rates are predictable in the same way as that tor  the SC^-^O system.

The rate of S02 absorption may be given by

                            Ce ~ °L           Ce - CL
               NA   =	   «   —	              (4)
                               + (H/kg)        I/KL
                               11

-------
       b.  Mass Transfer and Heat Transfer Coefficients of Liquid


           Droplet


              Gas-phase mass transfer to or from a single sphere placed


in a moving fluid has been studied by many investigators [5, 13, 21].


One of the correlations having the same form as the Ranz-Marshall


equation obtained by Steinberger and Treybal [22] can be used to calculate


the gas-phase mass transfer coefficient for each drop in a venturi scrubber,


                 N
                  Sh

where NgjjO can be calculated by


                 NSho  ~   2 + 0.569(NGr NSc)°'25


                           for  < 1()8
                                              1/3 0.244
                 NSho  -   2 -f 0.0254(NGr NSc)1/JNSc

                           for (NGr NSc) > 108

              For all practical purpose, Ngko= 2, and the above equation


becomes

                 Nsh   =   2 + 0.347(NRe N^2)0'62                (6)


The ranges of applicability for this correlation are


                 1   <   NRe < 30,000


                 0.6 <   Ngc <  3,000


The mass transfer coefficient, kc, is related to kg by


                 kc   -    RTkg

              A similar type of correlation as equation (6) is used for


the mass transfer coefficient of water vapor -from the surface of liquid


drop to gas phase.

              In contrast  to the extensive studies made on mass transfer
                              12

-------
around a sphere, no correlation for the mass transfer coefficient



inside the sphere is available.  Although liquid circulation or turbulence



in the droplet may be possible, it is assumed that the drop is a rigid



sphere and that there is no movement of the liquid within the droplet.



For an unsteady state gas absorption of gas into a drop, the mass



transfer coefficient in the liquid phase can be obtained as follows:



                       aCT     D    3    3CL
                         jj         • _    O  *J
                I.C.:   t = 0; 0 < r  < r,,;  C, =  0
                                       O    Lt


                B.C.:   t > 0; r = rQ; CL = CL



              This  equation  can be solved analytically  to  obtain  the



     of absorption  at  time t as:


                             * -             22
                         2DCL   ro        DnTiTt
              The mass  transfer  coefficient k^ may be defined as:



                          2D   «      .
                 kL    =    —   Z   exp(-

                          r
                           o  m=l          o                      (9)



               The heat transfer coefficient in the gas-phase needed



 to calculate the temperature  profile is estimated from the following



 correlation [2 I ] .



                 "No  -   2 + 0.6N^2'N^3                        (10)





 As the heat transfer in the liquid-phase is much larger than that in



 the gas-phase, it is assumed that there is no temperature gradient



 in the liquid droplet.
                               13

-------
       c.  Equilibrium Data and Absorption Rate Data

              Equilibrium curves of the S02-H20 system,  the O.OlM-NaOH-

S02 system, and the 0.03M-NaOH-S02 system at 77°F are shown in

Figure I.  The curves for the NaOH-S02 system are obtained from the

computer program provided by the Radian Corporation [20].

              Experimental data obtained from absorption of S02 into

alkaline solutions by. liquid-jet type [9], and stop-cock type [19]  absorbers

have been correlated in the present study based on the penetration
theory as shown in Figure 2.  The fact that the rate of S02 absorption

in NaOH solution can be represented by a physical absorption mechanism,

and since the contact time for the liquid-jet absorber is almost the

same as that in a venturi scrubber, equation (9) developed for S02

absorption in a spherical droplet seems applicable for S02 absorption

in a venturi scrubber.

4.  Development of Performance Equations for a Venturi Scrubber

       From the material and heat balances across a small increment of

a venturi scrubber, the following equations can be set up.

                               dCL
                        c  V ^~  -  adV>NA                    

                  Mym  =  *gwa(V-yw>dz  =  -di^               (12)
                  hga(tg-tL)dZ  =  LmCpLdtL + $kgwa(y*-yw)dZ     (13)
And
                  dV  -  SdZ

                                14

-------
   2.0-
n
o

 x


"5.
 o

 2
 =J
 o
 a.
    0.5
      oxidation %/

            10 /
0.01M- NaOH soin
                                                            oxidation %

                                                                    50
                                                                 0.03M-NaOH soln
                               8
      9   K)*5""              Ho   *i3
        concentration  of   S02 absorbed in liquid,[g-moles/cc]
          Figure I.  Equilibrium  S02Partial Pressure over Various  Solutions(77°F)

-------
o
CD
V)


flO
w   8
\-4
O
 6

 4
  10
,-9

8

6
                                              T	1	1—n	1	r
                                            //

                             — O   H20
                             — V   0.938N  NaOH
                             	A   l.ON  NaOH (with surf ace active reagent)
                             — O   2.0N   NaOH

                              Lines: theoretical values calculated  by   	
            Liquid jet (9)
               (short contact time)
                                                     Stop cock type  (19)
                                                         (long contact  time)
               46 8|Q
                        "Z
"4   6 8|0H  £H>T
      time, [sec]
                                                       till
                                                                             i  i  i
                                                    4  6 8 |03
                                                                      46  8|Q4
       Figure  2.  Comparison of Experimental  Data  on Rate of
                    Absorption  with Theoretical  Values
                                                                      SO,
                                                                                       8

                                                                                       6
                                                                                    10
                                                                                    8
                                                                                    6
                                                                                      r«
                                                     10
                                                     8

                                                     6
                                                                                         ,-T

-------
              Figure 3 illustrates the differential volume used for the




material and heat balances and the concentration and temperature profile




across the liquid-gas interface.


                                                          *       *

              The water vapor pressure at the interface, pw  =




can be estimated from Antoine's correlation [1].




                    *      .     B
Where




       A  «=  7.7423  and  B  =  1554.16  for  tL  =  35 ^ 55 [°C]




       A  =  7.8097  and  B  =  1572.53  for  tL  =  55 ^ 85 [°C]




              From the heat and mass balance equations and the flow




equations developed, the following differential equations relating liquid




velocity, S(>2 concentration in liquid, 1^0 mole fraction in gas phase,




temperatures of gas and liquid, molar liquid rate, and contact time with




respect to Z are obtained as follows:





       ^L     8   '£.    C  .  P&


       *Z   =  VL^ 4  '  Dp '  pL




       dCL

                                                                  (is)
                  m
                                  17

-------
00
           tu
           tg
tg-dtg
       m
                   Z+dZ
               (a)
                                m
                                               gas
                                     liquid


— ._.-.._
w

S02
tg 	 J



x
/"

\
x%
^^^


Pw

^L
ftX



1







                                 (b)
                                                                 t,
       Figure  3   (a)  Schematic Diagram of a Differential Element in
                       a Venturi  Scrubber
                   (b)  Concentration  and Temperature Profiles  at
                       Interface

-------
          dt        h a
               =        
          dLm           a
          dz   ~  kgwa(yw~yw)                                    (21)


          
-------
PO
o
                                                        3"
                  Figure 4. Dimensions of  OAP Venturi Scrubber

-------
        corrected  Volgin's  eq.

     O  Calvert's eq.

                         G
                           10        12
         pressure drop obtained experimentally
                  [in.H203
Figure 13
Comparison  of   Calculated Pressure
Drop  with Experimental  Data
(QAR Venturi   Scrubber)
                   21

-------
o
0>
TO



-150

-------
                  Gas rate ACFM Liquid rate GPM  Cone of soln.
                    1037         10           0
                    IO70         10          0.03 M
                     900         10          0.01  M
                         Run MO
0
5        1C        15
 distance from  nozzle, [ in]
Figure 7  Concentration Profiles of  S02  Absorbed
          in  Liquid (OAP Exp.)
                   23

-------
  300
  200
2?



E

-------
              The correlation equation for pressure drops given by




Matrozov has two terns, one for frictional loss and the other for




acceleration loss.  Since the agreement of this correlation with




experimental data is rather poor and no trend with respect to the




operating condition can be found, his correlation is not included in




the figure.  The Calvert equation always gives higher values than




experimental data even though it does not contain friction loss term.




On the other hand, Volgin's correlation takes into account the frictional




loss along the throat, but his correlation gives lower values than the




experimental data.




              For the venturi scrubber used by O.A.P., the pressure




drop can be estimated by corrected Volgin equation shown below.




               AP2  =  3.5 X lO-Vq0-26!0'143




              The liquid concentration profiles presented in Figure 7




show that equilibrium is almost reached within a few inches from the




nozzle.  Also the temperature profiles in Figure 8 indicate that both




gas and liquid attain the same temperature rather quickly.  Therefore,




the operation can be considered isothermal except for a short section




near the nozzle.  As can be noticed from the concentration profiles of




the liquid shown in Figure 7, the mass transfer rate near the nozzle is




so fast that the temperature effect can not be neglected.




              The results of simulation of the O.A.P. venturi scrubber




experiment are shown in Figure 9 for three types of scrubber liquors,




indicating satisfactory agreement.
                               25

-------
  100
  80
o


I 60

£

d1
CO


"S 40
jo

o
o
  20
   0
          O  0.03M  NaOH Solution

          Q  0.01M  NaOH Solution

          A  H20
             20        40       60        80
              experimental  S02 removal  %
                                              100
Figure 9
                Comparison of  Calculated  S02
                Removal % with  OAP Experimental
                Data
                     26

-------
       b.  Experiments of Johnstone, et. al. [11]




              Johnstone, et. al., reported a venturi scrubber study in




which S02 was absorbed in 0.6N-NaOH solutions.  The dimensions of their




equipment are shown in Figure  10 .  The results of simulations of their




experiments are compared with the actual experimental data in Figures 11




and  12  .  The absorption rate profiles are plotted as a function of




distance from the nozzle, and also as a function of liquid flow rate




through  the venturi scrubber.




      c.  Cottrell Environmental ISysterns' Experiments




              The experimental data obtained by Cottrell Environmental




Systems, Inc. for the SC^-lime solution system by Flooded Disk Scrubber




are compared with the results obtained from simulation of SC>2 absorption.




The dimensions of the equipment and the flow diagram are shown in Figures




13, 14, and 15.




              The results of the simulation are given in Figures  16 ,




17, and  18.




              The pressure drop in the Flooded Disk Scrubber can be




predicted rather well by the Calvert equation as shown in Figure  16 .




This correlation seems  to apply for Flooded Disk-type or short throat




type scrubbers since the friction loss has been neglected in the equation.




The liquid concentration profile, not shown in this paper, has a similar




shape  as those obtained for the O.A.P. venturi scrubber.




              The results of simulations based on 13 runs from FDS




data are shown in Figure  18  .
                                 Z7

-------
PO
00
                          Fluid inlet pipe 3/32"ID
                                                                 Sampling position
Gas rate
Liquid rate
SO? inlet cone.
Nozzle
5OO ft/sec at throat
0.03-0-242 GPM
0.001 52 atm
4 holes of 0.046"dia.
               Figure 10   Dimensions of  Venturi  Scrubber  used  by Johnstone etal.

-------
r\>-
10
       o
       o

       X
         10
          8
       to
       Q>

       "o

       E
        i
       CT»
        0)


        "5
       I 4

        9-
        o
        CO

       •8
        =1

        O
0  Experimental data of Johnstons etal.


    Calculated  values
           Fiqure 11
              468
              distance from  nozzle,  tin]


        Profiles  of  Cumulative Absorption  Rate  in

        Venturi   Scrubber  (Experiment of  Johnstone et af.)

-------
        O  Experimental  data  of Johnstone et al.
           Calculated  values
                    0.1       0.15        0-2
                liquid rate,   t gal Ions/miniteD
Rgure 12
Comparison  of  S02 Removal  Rate  in the
Venturi Scrubber Calculated with Experimental
Data  Obtained by Johnstone  et  al
                       30

-------
                    I
                                 8
                                •Hoi
Figure 13.  Dimensions of  Cottrell's  Flooded

            Disk  Scrubber
                       31

-------
co
ro
        gas in
gas in
                                                   FDS/*
                                                       CA0
gas out
 t
             packed
             tower
                                                                           clorifier
                                                             H20'
                  slurry
                  tank
      Fiqure 14  Two Stage Lime Scrubber     Figure 15.
        9        Wet Slurry(FDS; TASK TT5)
     Dry Lime injection to  FDS
     (FDS :TASK TT6A)

-------
  25
 20
3
.5 15
 o


-------
                   velocity of liquid
                   velocity of gas
        5         10        15        20
            distance from flooded disk, [in3

Figure 17  Velocity Profiles  of  bas and Liquid
           in  FDS(Task-TT5-2)
                   34

-------
 100
  80
$60

8?
cS
V)

2 40
J2
o
o
  20
   0
            O  TASK TT5(
            A  TASK TT6A
            20       40       60       80
              experimental S02 removal  %
100
     Figure 18  Comparison of Calculated  Values of
               S02 Removal  % with  Those  Obtained
               Experimentally (Cottrell FDS)
                    35

-------
6.  Conclusion

       A mathematical model has been developed to elucidate the mechanism

of SO^absorption into alkaline solutions and to provide a guide to the

scale-up in design of venturi scrubbers.  In spite of the assumptions

made in the analysis, the results of the simulations show good agreement

with the experimental data.

       From the results obtained, the following conclusions can be drawn:

       — A significantly large amount of mass and heat transfers take
          place in a section near the nozzle.  This is primarily due to
          the small transfer resistances, large driving forces, and
          relatively long residence time of the liquid droplets.

       — The operations of the venturi scrubber can be considered as
          isothermal except the section within a few inches from the
          nozzle.  High efficiency of the heat transfer is observed
          in this type of equipment.

       — Near equilibrium concentrations of SC^ in the liquid were
          observed at the end of the scrubber in the case of SC^-I^O
          system.

       — Selection of a proper correlation for the mean diameter of a
          liquid droplet is important in the prediction of the venturi
          scrubber performance.

       Refinement along the following lines will improve the validity

of the model developed:

       -— The effect of the presence of the solid phase on absorptiort
          rate should be examined.

       — The effect of oxidation of sulfite in the liquid phase on
          absorption and equilibrium must be investigated in more detail.

       — The equilibrium data should be obtained for the systems of
          SOj-NaOH solution and SO^-lime solution to confirm the values
          estimated by the computer model of Radian cooperation.
                                    36

-------
The formation of the scale, particularly around the nozzle wall,
during operation which affects drop and the diameter of the
droplet is one of the causes for the deviation of the simulation
result from the experiments.   Studies to elucidate the cause
of scale formation and to develop means of reducing scale are
in need.
                         37

-------
                             Notation

a     contact area of gas and liquid per volume,  [I/ft]

C     drag coefficient for a particle,  [-]

Ce    concentration of S(>2 in liquid phase in equilibrium with partial
      pressure of S02 in bulk phase of  gas,  [Ib-mol/ft3]

CL    concentration of S(>2 in liquid phase,  [Ib-mol/ft^]

CL    surface concentration of S(>2 in equilibrium with partial pressure
      of S02 at interface, [lb-mol/ft3]

Cpg   specific heat of gas, [Btu/lb-mol. °F]

D     diffusivity of S02 in liquid phase,  [ft2/sec]

DJU    mass median diameter of a droplet, [micron]

Dn    diameter of a nozzle, [cm]

Dp    diameter of a liquid drop,  [ft]

dV    incremental volume of venturi, [ft^]

I>32   surface-volume mean diameter of a droplet,  [micron]

g     gravitational acceleration, [ft/sec2]

GJJJ    molar gas flow rate, [Ib-mol/sec]

G0    volumetric gas flow rate, [ftVsec]

H     Henry's law constant, [Ib-mol/ft3'atm]

hg    heat transfer coefficient in gas-phase, [Btu/ft2«sec*°F]

k     thermal conductivity, [Btu/ft-sec-°F]

kc    mass transfer coefficient in gas-phase, [ft/sec]

k_    mass transfer coefficient in gas-phase, [Ib-mol/ft2'atm-sec]
 O

kgw   mass transfer coefficient of water in  gas-phase,  [Ib-mol/ft2'atm'sec]

KL    overall mass transfer coefficient, [ft/sec]
                                  38

-------
 k,    mass transfer coefficient in liquid-phase, [ft/sec]


 L    liquid rate, [GPM]


 A    throat length, [mm]


 I^   molar liquid rate, [Ib-mol/sec]

                             o
 Lo   volume liquid rate, [ft /sec]

                          3         3
 m    specific wetting, [m -liquid/m -gas]


 M    mass flow rate of gas, [Ib/sec]
  o

 ML   mass flow rate of liquid, [Ib/sec]


 NA   rate of S(>2 absorption, [Ib-mol/ft2>sec]

                        •»     i   i 2
 Ngr  Grashof number, gD (p -p )/p v , [-]
                        P  g  g   8 8

 NJJU  Nusselt number, h D_/k» [-]


 Npr  Prandtl number, Cpgy /k, [-]


 NRe  Reynolds number, pg v -VL  'Dp/yg,  [-]



  Sc  Schmidt number, yg/pgD, [-]

 Ngh  Sherwood number, kcD /D, [-]


 P    total pressure in venturi, [atm] or[mmHg]


 PS02  partial pressure of S02 in bulk phase of gas, [atm]


PS02  Partial pressure of S02 at interface, [atm]


pw    partial pressure of water in bulk phase of gas, [atm]


Pw    partial pressure of water at interface, [atm]


q     liquid rate, [liters/in3]


R     liquid to gas ratio, [gallon/MACF]


r,j    radius of droplet, [ft]

                                          9
S     cross-sectional area of venturi,  [ft ]
                                  39

-------
T      absolute temperature, [°R] or[°K]



t      time, [sec]



t      temperature of gas,  [°F]
 o

t7     temperature of liquid,  [°F]
 Lt


Vg     velocity of gas, [ft/sec]



v»     velocity of liquid droplet,  [ft/sec]



vro    relative velocity of gas to  liquid at throat, [ft/sec]


v      relative velocity of gas to  liquid at throat, [m/sec]



w      gas velocity at throat,  [ft/sec]


x      mole fraction of solute in liquid, [-]



y      mole fraction of water  in gas-phase, [-]


y*     mole fraction of water  at interface, [-]



Z      distance along axial direction of venturi,  [ft]


                                Greek Letters



$      heat of vaporization, [Btu/lb-mol]



APd    friction loss,  [N/m2]



AP,    pressure drop in venturi, [N/m2]


AP-    pressure drop in venturi,



A?3    pressure drop in venturi, [in



e      void fraction in venturi, [-]



u      viscosity of gas,  [C.P.] or  [Ib/ft-sec]
 O

u,     viscosity of liquid,  [C.P.]

                                    *\
Vg     kinetic viscosity of gas, [ft^/sec]'

                             o
p      density of gas,  [lb/ftj]
. O
  .                                          
-------
PL
       density of liquid, lib/ft3] or [g/m3]




pM     molar density, [Ib-mol/ft3]




a      surface tension, [dyne/cm]
                               41

-------
                              References

 1.       Antoine,  T.  E., Vapor-Pressure of Organic Compounds, Interscience,
         New York (1954)

 2.       Calvert,  S., A. I. Ch. E. J. ;'16, 5, 392 (1971)

 3.       Davis, H. S., and G. S. Crandall, J. Am. Chem. Soc., 52, 3757,
         3769 (1930)

 4.       Feild, R. B., M. S. thesis in Chem. Engg., Univ. Illinois, (1950)

 5.       Froessling,  N., Gerland's Beit-zur Geophys., 52, 170 (1938)

 6.       Gleason,  R.  J., and J. D. Mckenna, A. 1. Ch. E. 69th Nat'l.
         Meeting,  Cincinnati, Ohio, May 16-19 (1971)

 7.       Gretzinger,  J.  and W. R. Marshall, A. I. Ch. E. J., ]_t 2,
         312 (1961)

 8.       Hatta, S., Techol. Repts. Tohonu  imp. Univ., £, 1  (1928-1929);
         ibid., 10, 119  (1932)

 9.       Hikita, H.,  S. Asai, and J. Tsuji, The 34th Annual Meeting of Soc.
         of Chem.  Engrs, Japan, Preprint A207 (1969)

10.       Inoue, I., Chem. Eng. Handbook (Japan), 869 (1968)

11.       Johnstone, H. F., T. B. Feild and M. C. Tassler, Ind. Eng. Chem.
         46_, 8, 1601  (1954)

12.       Kim, K. Y. and W. R. Marshall, A. I. Ch. E. J., 17_, 3, 575, (1971)

13.       Langmuir, I., Phys. Rev., ,12 (2), 368  (1918)

14.       Lewis, W. L., and W. G. Whitman, Ind. Eng. Chem., _16, 1215 (1924)

15.       Linn, S., J. R. Straatemeier and H. Kramers, Chem.  Eng. Sci.,
         4, 2, 49  (1955)

16.       Matrozov, V. I. Soobshcheniya o Nauchno-Tekhnicheskikh
         Rabotakh NIUIF  Nos. 6/7, 152 (1958)

17.      Mugele, R. A.,  A.  I. Ch. E. J., £,  3  (1960)
                                  42

-------
18.      Nukiyama, S. and Y. Tanasawa, Trans. J.  S. M.  E., ji,  14,  86
         (1937); ibid., 4., 15, 134  (1937);  ibid., J>,  18,  63,  (1938);
         ibid., _5, 18, 68  (1938); ibid., £,  22, 7  (1940);  ibid., £,
         26, 5  (1941)

19.      Onda, K. E. Sada, and Y. Maeda, Kagaku Kogaku  (Japan), 35,
         3, 345 (1971)                                          —

20.      Radian Corporation, private communication

21.      Ranz, W. E., and W. R. Marshall, Jr., Chem.  Eng. Progr.,  48,
         141, 173 (1952)                                           —

22.      Steinberger, R. L., and R. E. Treybal, A. I. Ch. E. J., 6,
         2, 227, (1960)                                          ~

23.      Volgin, B.  P., T. F. Efimova and M. S. Gofman, Int. Chem.
         Engg.,  £, 1, 113 (1968)

24.      Wetzel, H., Ph. D. thesis, Univ. Wisconsin,  Madison  (1951)

25.      Wigg, L.  D., J. Inst.  Fuel. 37, 286, 500 (1964)
                                   43

-------
Appendix:  Data of S02 Absorption by O.A.P.  Venturi Scrubber
Run
No.

1
1A
2
3
3A
4
5
6
7
I-1A
1-2
I-3A
1-4
1-5
1-6
1-7
1-8
1-9
1-10
Gas Temp., °F
Inlet Outlet

310 115
320 110
320 115
315 110
300 100
310 110
275 110
310 110
320 115
300 115
300 115
300 100
300 110
250 110
300 115
300 115
300 100
250 95
300 105
S02 Cone. , PPM
In Out
(or % removal)
1980 1750
2055 1790
2000 1800
2050 1780
j.993 1705
2060 1700
1980 1740
2000 1160
1980 710.
2000 (12.85)
2000 (12.60)
2000 (14.45)
2000 (17.50)
2000 (12.10)
2000 (42.0 )
2000 (64.2 )
2000 (37.9 )
2000 (33.0 )
2000 (24.7 )
Liquid
Rate
GPM
H20 10
* 10
11 10
.10
10
15
" 10
0.12%NaOH 10
15
H20 10
11 10
" 10
" 15
11 10
0.12%NaOH 10
15
" 10
" 10
0.045%NaOH 10
Inlet Gas
ACFM

1037
1087
1023
1229
1206
1006
1010
1070
1051
900
900
1100
900
900
900
900
1100
900
900
Pressure Drop
in H20
________———
8.75
8.6
7f\
.8
12.0
n/
.4
9.0
8f
.5
8*
.4
9.5
8f
.6
7O
.8
Ui
.4
9f\
.0
8g>
.5
8f
.4
9O
.2
MA /
12.4
i f\ n
10.2
8f
.6

-------
MATHEMATICAL MODELS FOR
PRESSURE DROP, PARTICULATE REMOVAL AND SO2 REMOVAL
IN VENTURI, TCA AND HYDRO-FILTER SCRUBBERS
By


M. Epstein
C.C. Leivo
C.H. Rowland
of Bechtel Corporation
San Francisco
To be Presented at the
Second International Lime/Limestone
Wet Scrubbing Symposium
New Orleans, Louisiana

November 8-12.1971
                              45

-------
                         INTRODUCTION

 The Office of Air Programs (OAP) is sponsoring a project to fully
 characterize wet limestone scrubbing for removal of sulfur dioxide
 and particulates from boiler flue gas in a large prototype  system.
 The test facility consists of three parallel scrubber systems, each
 capable of treating approximately 30, 000 acfm of flue gas, which are
 integrated into the flue gas ductwork of an existing coal-fired boiler
 at the TVA Shawnee Power Station.  The test facility and test pro-
 gram are  described in Ref. 55.

 Bechtel Corporation, as the major contractor, is  developing mathe-
 matical models to characterize the three scrubber systems, in order
 to allow for effective and economic scale-up to full-size scrubber
 systems.   The model predictions are to be compared with the test
 facility data  and best-fit values for uncertain constants and coefficients
 calculated.

 This paper presents, in abbreviated  form, the models which have
beed developed to date for pressure drop, particulate removal,  and
SO  removal in the venturi, TCA, and Hydro-Filter scrubbers.
                                46

-------
                      VENTURI SCRUBBER

A venturi scrubber offers the advantages of high particulate removal
efficiency (for relatively small particles), mechanical simplicity and
proven reliability for  slurry systems.

PRESSURE DROP

A number of investigators have developed models for predicting pres-
sure drop in venturi scrubbers (Refs. 1  through 5).  Recently, Boll
and Leeman (Ref. 1) presented a numeric (computer) solution to the
simulataneous differential equations of drop motion and momentum,

Calvert (Ref. 2) has proposed that the velocity head-loss in a venturi
is  a linear function of liquid-to-gas ratio:

                                                                  (I)
Clavert (Ref. 3) cites a simplified version of the above equation, which
is in good agreement with observations reported by McKenna (Ref. 6)
on a variable orifice Research Cottrell venturi (Ref. 7).
            A "13 =
                               47

-------
In Figure 1,  the Calvert equation is compared with the McKenna data
(Ref. 6) and with data from an OAP "in-house" venturi (Ref. 8).

Recently, Wen (Ref.  9) has compared results from the Matrozov
(Ref. 4) and  Volgin (Ref.  5) correlations for venturi pressure drops
with data taken in the OAP venturi.   The Matrozov and Volgin equations
can be represented by the form:*
The form of the expression for pressure drop which will be tested with
the data obtained at the test facility is given by Eq. 3, where the best
values for the constants fa , through/?^,  are to be determined.

PARTICULATE REMOVAL

Various investigators (Refs. 3, 11 and 12) have studies participate
collection in Venturis.  It seems generally agreed upon that the
"inertial impaction" mechanism (Ref. 12) is the important one.

Calvert (Ref.  3) has presented the following correlation for particulate
removal in the diverging section of a venturi,  based upon impact theory.
    In the Volgin equation  fa = 0 and fa is a function of venturi geo-
    metry.  In the Matrozov equation Q±  is a function of venturi
    geometry.
    The equation for Po will be given in the following section on
    absorption of SO_.
                                48

-------
               FIGURE 1

COMPARISON OF EXPERIMENTAL DATA AND

  PREDICTED VALUES OF PRESSURE DROP

        FOR VENTURI SCRUBBERS
 14
 12
 10
  8
1
a.
O

o
UJ
QC

to
to
UJ

Q_

O
5  4
UJ
QC
Q.
  0
       o
       o
COHRELL VENTURI

GAP VENTURI

G • 600-1200 cfm

L/G - 5-?0 gal/mcf
                  O
ox>
          2     4     6     8     10

            MEASURED PRESSURE DROP,  in.
                                   12
                               14
                  49

-------
 The collection efficiency, E£ ,  which is used in the model,  is an
 approximation of the experimental results of Walton and Woolcock
 (Ref. 13):
 The inertial impaction parameter,  HQ , is defined as:
 Calvert has found that Eq.  4 agrees with limited experimental data if
 the velocity ratio, (j(J^ -1%\/'U* »  is set equal to 0,4.  With this assump-
 tion, Eq. 4 becomes:
 The assumed fly ash size distribution curve for the test facility is
 shown in Figure 2 (Ref. 14).   The computer model for predicting
 particulate removal divides this curve into representative size ranges*,
 calculates removal efficiencies for each size, then predicts overall
 collection efficiency.
Overall particulate removal efficiencies and pressure drops, are pre
dicted by the Calvert equations (Eqs. 3, 7,  and 8), as shown in
Figure 3, as a function of liquid and gas flow rates.
* Size ranges of: 0-1,  1-2,  2-3,  3-6,  6-10,  and  > 10 microns.

-------
                  FIGURE  2
    TEST FACILITY FLYASH DISTRIBUTION
 99.9
   99 -
   95

   90
   70

   60
   50
   40
0  20

   10

    5
    1
   .1
I  I
                         FLYASH SIZE
                         DISTRIBUTION
                                     i  i
     1
345
10
20  30 4050
100
                PARTICULAR SIZE, micron
                   51

-------
                FIGURE  3
  PREDICTED VENTURI PARTICULATE REMOVAL
            AND PRESSURE DROP
         FROM CALVERT EQUATIONS
95.0
99.0
99.9 r
          8
10     12      14
PRESSURE DROP, in. H20
   52
16

-------
In Figure 4,  a comparison has been made between predicted (Eqs. 7
and 8) and measure overall particulate removal efficiencies in a
Chemico pilot venturi.   The data was obtained from tests with flue
gas at the Crane Station (Baltimore Gas and Electric) and Dickerson
Station (Potomac Electric Power Company), with a 1500 cfm venturi
absorber.

The form of the expression for particulate removal which will be tested
with the data obtained at the test facility is:
                                           (p0
where £:*£ is obtained from Eq. 5 and the coefficients  rt^ and R^  are
to be fitted to the data.  The parameter p^> is to be directly measured
from photographic data, if possible.

ABSORPTION OF SULFUR DIOXIDE
 The following are the major assumptions made in the formulations of
 the mathematical mode
 scrubbing in a venturi:
the mathematical models for SO absorption during limestone wet-
         The droplets are spherical
         The liquid droplets do not  either coalesce or
         re-shatter after atomization.
         The droplets can be presented as having an average
         diameter
         The changs in droplet diameter due to evaporation
         is negligible.
                                53

-------
                FIGURE  4
 COMPARISON OF EXPERIMENTAL DATA AND
PREDICTED VALUES OF OVERALL PARTICULATE
   REMOVAL FOR CHEMICO PILOT VENTURI
   100
    99
    97
   96
   95
     95
OPERATING RANGE
  G • 1500 cf m
  L/G • 10-17 gal/mcf
   O
                      O
   96     97      98      99
   MEASURED PARTICULATE REMOVAL, %
         54
100

-------
The original model for predicting SO  removal considered a distri-
                                   LJ

bution of sizes of droplets,  as predicted by a correlation of Nukiyama


and Tanasawa (Ref.  15).  Subsequently,  it was  shown that a model


based upon droplets of mean diameter Q> (diameter of a single drop


with same surface area to volume ratio as total spray) successfully


reproduced  the results from the multi-droplet model.





Droplet Size





Recently, Wen (Ref.  9) and Boll and Leeman (Ref.  1) have compared


predictions  from a number  of correlations for average droplet dia-


meter in venturi scrubbers.  The present development uses the


Nukiyama -Tanasawa equation to predict mean droplet size (Ref.  15):
Droplet Velocity





In accordance with the assumptions of constant drop mass and spherical


shape, the equation of motion for a droplet downstream of the venturi


throat can be written:
                      *  +  3-   — '&-     CD-IS') I
                     O     <+
 The gas velocity at any point is calculated from:




            U  =
                            55

-------
Void fraction,  £   , at any point is related to the droplet trajectory by:
The drag coefficient has experimentally been found to depend upon the
drop Reynolds number, f?g,  .   The functional relation has been
shown by Ingebo (Ref. 16) to be:
Rabin, et. al.  (Ref. 17) have shown that, for droplets accelerated from
rest, Eq. 14 is valid only for Re  <^  60.   For Re  >   60 they found:
Eqs.  11 through 16 are solved numerically (computer) for velocities of
droplets as a function of distance downstream from the point of atom-
ization.

Mass Transfer

The liquor- solid mixture entering the scrubber is assumed to be in
thermodynamic equilibrium.  As the mixture passes through the
scrubber the SO  (and CO ) which is adsorbed from the gas reacts with
               L*         C*
the ionic species present in the liquor and, depending upon the
                              56

-------
concentrations and equilibria, solid phase material can be either
formed or dissolved within the liquor.  For alkali wet-scrubbing of flue
gas, the assumption has been made that the liquor in the venturi is at
all times in equilibrium with an interfacial CO, vapor pressure of
                                              £
0. 1 atm of CO ,  i.e. the rate of CO, absorption (and desorption) from
              Lt                    £*
the flue gas is large.

The thermodynamic equilibria in the following models are obtained
from a table based upon results  from the Radian Equilibrium Computer
Program (Ref. 18).
         •
In accordance with the two-film theory,  the equations for removal of
SO in a differential element of the venturi may be written as:
   Lt
                            "  X                             (lf)
            ^0-       "*26-      -&L
    •'**'*•
            LA
A differential venturi element is shown in Figure 5.  Equations 17
through 20 are solved numerically (computer) for SO removal along
the venturi axis.  The height of the differential  elements, ^H  ,
are automatically set within the computer program to give a mathe-
matically stable solution.   Equations for predicting interfacial
                                 57

-------
               FIGURE  5
       DIFFERENTIAL VENTURI  ELEMENT
              GAS
DROPLET
CROSS
SECTIONAL
AREA
                    58

-------
temperature,  /  ,  and dissolution of solid species will be presented
in the following  sections.

Gas and Liquid-Side Mass Transfer Coefficients.  The gas -phase
mass transfer coefficient is obtained from a correlation developed
for mass transfer from gases flowing past single spheres (Ref.  19).
The gas-liquid interfacial area per unit volume is  given by:
          a, =

Combining Eqs.  21 and 22 gives:

       fl  -.  "'
Handlos and Barron (Ref. 20) have presented a correlation for liquid
phase mass transfer coefficient (in the absence of chemical reaction)
for a "circulating" liquid droplet.
where   ^   =2.88 for no continuous phase resistance to mass transfer.
Sherwood and Kolloway (Ref. 21) and Whitney and Vivian (Ref. 22)
                                59

-------
determined that temperature had a considerable effect upon liquid-
phase mass transfer coefficients.  Their results can be summarized
as:
                    _  & 0*013
Chemical reaction of the absorbed gas in the liquid phase enhances
mass transfer by decreasing the bulk liquor composition (and inter-
facial vapor pressure) of the absorbed gas and by increasing the
liquid-film mass transfer coefficient,  J£LL  .  Calvert (Ref. 23),
Perry (Ref. 24) and Sherwood and Pigford (Ref. 25) have presented
reviews of the theories and solutions presently available for deter-
mining the effect of chemical reaction on «#^  .  The analytic
results are generally complex and depend upon the rates of reaction,
the order of reaction,  the rate of diffusion of reactants through the
liquid film the concentration of reagents, etc.  There is presently,
however,  no analytic  solution available for  the case of obsorption
followed by a  series of reversible homogeneous liquid reactions of
varying order and,  further complicated, by heterogeneous liquid-
solid reactions.
As an example of an approximation to the actual conditions encountered
                                                      5J;
during alkali wet scrubbing of SO ,consider a very rapid  pseudo-
second order irreversible reaction between the dissolved gas A(SO )
    Calvert (Ref. 23) refers to this as the "instantaneous regime".
    The reaction is so fast that A, the absorbing component,  cannot
    coexist with B, the reacting component. Mass transfer is limited
    by the diffusion of A and B toward each other to a "reaction plane"
    where they meet and react (see Figure 4-3, Ref. 23).
                              60

-------
and "reagent" B.  If the diffusivities of A and B in the liquid are equal,
the following expression can be used to predict the effect of reaction
upon the liquid- film coefficient (see Refs.  23 and 25):
Combining Eqs. 22, 24, 25, and 26 gives:
For alkali wet- scrubbing of flue gas, X Q   has been defined as:
As will be shown in the following section, values for  A    an<^
of 20 and 1/2, respectively, were obtained by fitting mcdel predictions
to pilot plant data.

Heat Transfer
The following set of equations are solved numerically (computer) in
                                                         -r~*
each differential element for the interfacial temperature   /
between the liquid droplets and gas, as well as for the gas and droplet
temperatures and the evaporation rate of liquid.

       -±  -.   G-
                                   61

-------
A similar type of correlation to Eq. 21 is used to calculate the mass

transfer coefficient of water vapor  from the droplet surface to the

gas.  Nusselt-Sherwood analogs are used to calculate /1~ and  ^7.   ,

based on Eqs.  21 and 24, respectively.  The correlation of Antoine

(Ref. 26) was used to estimate the water vapor pressure at the gas-

liquid interface,  /  .


Dissolution of Solids


The following are the major assumptions made in the formulation of

the solids dissolution model.   The assumptions only have applicability

in that region of operation where the thermodynamic driving force is

in the direction of dissolution of solid species.
         The absorbed SO£ reacts with the dissolved species
         in solution, resulting in the formation of "reagent"
         H (hydronium ions) which diffuses through the liquid
         film surrounding the solid particle and reacts with
         the  solid species S according to:
                               a
                 S  +  xx H  —^  products

         The solid-liquid heterogeneous reactions (Eq.  33)
         are slow (rate controlling) with respect to liquid
         phase homogeneous reactions.
         Diffusion of reagent through the liquid film sur-
         rounding the solid particle is not rate  controlling,
         i.e. chemical reaction at the solid surface controls.*
    The data of Smiths on and Bakhski (Ref. 27) for dissolution
    of magnesia (MgO) indicated that chemical reaction was the
    controlling  mechanism.  This assumption can be tested by
    varying liquor temperature during the  screening experiments,
    i.e. if chemical reaction controls, and dissolution rate is
    measurable, then the rate of dissolution should be a strong
    (exponential) function of temperature.
                                62

-------
         The heterogeneous reaction occurs at the outer
         skin of a spherical solid particle of radius n, .
         As reaction proceeds the particle dissolves and
         shrinks in size. * The initial average radius
         of the particles is R .

         The heterogeneous reaction is first order with
         respect to  \-\  and zero-th order with respect
         to the solid species.  *
The following set of differential equations have been developed, based

upon the previous assumptions, for prediction of rate of solids dis-

solution.  The equations have been solved numerically (computer)

for each differential scrubber height element,
                               ev
                                             X* - X,A
    These assumptions are identical to those made by
    Levenspiei (Ref.  28) and Wen (Ref. 29).
                                63

-------
    XV; j'-h'c. /
         Jl=  R
          Xo =
 Preliminary venturi results with the dissolution model have indicated
 that, for values of  -&sXs X$o    of less than 200 sec", the fraction of
 solids dissolved is essentially zero.  For values of .fa_s <£ Xso
 greater than 1000, the fraction of solids dissolved is approximately
 equal to the fraction dissolved for an infinite rate of dissolution  (the
 equilibrium dissolution).  Therefore,  it can be concluded that, there
 will be essentially zero solids dissolution within  the venturi scrubber
 for initial solids concentrations of less than 200 Ib/ft   (-— 75% solids)
 and 20 Ib/ft  (/~ 25% solids), with  J^s t   values of approximately
 1 and 10 ft /lb sec, respectively.
The data of Smichson and Bakhski (Ref.  27) and the data of Battelle
(Ref. 30) indicate that the maximum values for   ~fa-s fils    f°r
dissolution of MgO and hydrated limes in an aqueous media is
approximately 1 ft /lb sec.  There  is presently no published  data for
the dissolution of solids  in sulfite-bisulfite liquors, from which values
r»f  J&flbf    for limes and limestones can be estimated to any degree
of accuracy.
                                  64

-------
Comparison of Predicted SO,, Removal with Data
Experimental data have been reported by Johnstone and Roberts


(Ref.  12), by Johnstone, Field and Tassler (Ref.  10), and by OAP


(Ref.  8) for SO -sodium and SO -water systems.   In addition,
              £               £

Cottrell Environmental systems (Ref. 31) has presented results of


experiments in which sodium carbonate,  lime (CaO), and limestone


were employed as the alkali scrubbing agent.  The sodium data cover


a range of SO removal between less than 10 percent (Ref.  10) and
             C*

80 percent (Ref. 12).
For the sodium scrubbing data,  it was assumed that the interfacial




and that the gas phase mass transfer was rate controlling,  i. e
mole fraction of SO0, U  , was zero every where inside the venturi
Johnstone and Roberts: Na CO  Scrubbing.   The experimental
                          C*   J

venturi of Johnstone and Roberts (Ref.  12) was 4 ft. in length and had


a rectangular cross section of 1.5x8 inches at the throat.  A


sodium-water solution was injected at the throat,  either as eight


vertical jets  at the top or  as two horizontal jets, one at each side.


Gas rates varied between  600  and 1100  cfrn corresponding to throat


velocities of  140-240 ft/sec.  The mole fraction of SO  in the inlet
                                                   L*

gas varied between 800 and 2000 ppm.  The range of liquid-to-gas-


ratio was 0.7 to 5.5 gal/mcf.
                                 65

-------
The comparison between predicted and measured SO  removals is
shown in Figure 6,

Johns tone,  Field, and Tassler: NaOH Scrubbing.     The experi-
mental venturi geometry of John stone, et al (Ref. 10) is shown in
Figure 7.  These experiments were carried out at very low liquor
rates (0.03 to 0. 3 gpm) and at relatively high gas throat velocities
(500 ft/sec), corresponding to a range of L/G from 0. 1 to 1 gal/mcf.

The comparison between predicted and measured SO0 removal is
shown in Figure 8.

OAP In-House Data;  Water and NaOH Scrubbing.  OAP conducted
in-house alkali wet scrubbing experiements in the Cincinnati labor-
atories of its Divi'sion of Process  Control Engineering  (Ref. 8).
Experimental conditions were 1000 cfm of flue gas at 300F contain-
ing 2000 ppm SO, and 10 to 15 gpm of either water or 0. 03 M NaOH
                l*
solution. The water runs (runs #1,  la,  2, and 4) are of value in
testing the correlation for liquid phase mass transfer coefficients in
the absence of chemical reaction (Eq. 24).  The usable NaOH runs
are those for which gas samples and concentrations were obtained
at the venturi exit  (runs #6 and 7).
The OAP venturi is shown in Figure 9.  The liquid is sprayed from
a 30  nozzle located 3. 5 in. in front of the throat.  Hence, the liquid
droplets and gas are actually in turbulent contact between the nozzle
    In all other runs, outlet gas samples were taken only at
    location several feet downstream of the venturi section,
    after a knock-cut-drum.
                               66

-------
                 FIGURE  6
  COMPARISON OF EXPERIMENTAL DATA AND
   PREDICTED VALUES OF S02 REMOVAL FOR
        JOHNSTONE-ROBERTS VENTURI
   80
<
o
   60
 CO
O
to
CO
<
UJ
   40
   20
     SOLUTIONS
G =600-1100 cfm
L/G • 0.7-5.5 gal/mcf
               40
                   60
80
               PREDICTEDS02 REMOVAL, %
                    67

-------
00
                                                            3"-CL-
                       FLUID INLET PIPE, 3/32" ID
                              FIGURE  7
         DIMENSIONS OF THE JOHNSTONE-FEILD-TASSLER VENTURI

-------
              FIGURE  8
COMPARISON OF EXPERIMENTAL DATA AND
 PREDICTED VALUES OF S02 REMOVAL FOR
   JOHNSTONE-FEILD-TASSLER VENTURI
     0.6NNaOH SOLUTIONS
     G - 350 cf m
     L/G - 0.1-1 gal/mcf
          PREDICTED S02 REMOVAL AT9 in.,
                 69

-------
LIQUID
         3-1/2" 2-1/2"
               30° CONE (ADJUSTABLE)
                                         •11-3/8"-
                 FIGURE  9
        DIMENSIONS OF OAP VENTURI

-------
and the throat, and mass transfer occurs in this region.  The predic-


tions of the venturi model -were accordingly based upon droplet forma-


tion  at the nozzle.





In Figure 10,  model predictions are compared with the water and NaOH


solution OAP data.  For the NaOH comparisons, the average droplet


diameter at the nozzle exit, as  predicted by the Nukiyama-Tanasawa


correlation (Eq. 9), was reduced by 30%, in order to give a  reason-


able  fit.





For  the water runs, a good fit to the data was obtained by using a value


of A = 20, for the Handles and Baron correlations for liquid side mass


transfer coefficient (Eq. 24).  This compares to a value of 2. 9 sug-


gested by the authors.  For these runs, the predicted SO concentra-
                                                       Ci

tions in the liquid  and gas phases were close  to equilibrium near the


exit  of the venturi. Hence, only a lower limit for the value of /A  was


obtained.





Profiles of the calculated overall and individual mass transfer coeffi-


cients for an OAP water run are presented  in Figure 11.  Although the


liquid phase resistance,/W-*2i-,  is greater than the gas phase resis-


tance,  i/Jh.&  ,  neither can be neglected.  These results are in agree-


ment with those of Whitney and Vivian (Ref. 22), who studied absorp-


tion  of SO  in water in one-inch ceramic ring packages at 70 F.
         £t




Cottrell (Tidd) Data:  Na^CO^ Scrubbing.  In a pilot plant operation at


the Tidd Power Station of Ohio Power,  Research - Cottrell (Ref. 31)


obtained experimental data for SO- removal in Na CO   solutions for
                                  71

-------
                 FIGURE  10
  COMPARISON  OF  EXPERIMENTAL DATA AND
     PREDICTED  VALUES OF S02 REMOVAL
              FOR OAP VENTURI

-------
                 FIGURE  11
PREDICTED MASS TRANSFER COEFFICIENTS FOR
   SO 2 REMOVAL IN WATER IN OAP VENTURI
10,000
                     G -1100 cf m
                     L/G -10 gal/mcf
            2468
              DISTANCE FROM NOZZLE, IN.
                   73

-------
 a venturi system operated over a wide range of conditions.  Flue gas
 rates were between 300 and 920 cfm and liquor rates varied from 4 to
 12  gpm.  The Cottrell venturi, which contains an adjustable "flooded
 disc" is shown  in Figure 12.

 As was the case for the OAP venturi, the predictions of the model for
 the Cottrell venturi were based upon droplet formation at the initial
 point of intimate turbulent contact between liquid and gas,  which was
 assumed to be the point 14 inches downstream from the liquid inlet
 (see  Figure 12). As little mass transfer was  expected beyond a bend
 in the line, located 2. 5 ft below the open position of the flooded disc,
 the model's calculations were terminated at that point.  The SO  re-
 movals predicted by the model are compared with the  experimental
 data  in Figure 13.  Although these predictions tend to be high for low
 SO, removal and low for high removal,  more than three-fourths of
   £•
 the data points are within 20 percent of predicted values.

 Cottrell (Tidd) Daca;   CaO Scrubbing.   For the Cottrell experiments
 in which the SO  was removed using lime  slurries, there was some
               Ct
 dissolution of solid CaO in the scrubber.  This was demonstrated by
 the relatively low removals  (13-15 percent) obtained in three experi-
 mental runs (Task VII-C) for which a clarified lime  solution was sent
 to the venturi.  Since  experimental data is lacking for  the lime dissolu-
 tion reaction rate constant and specific surface area, only the three
 clarified solution runs were  compared with predictions from the model.
These runs are  of interest because they provide an opportunity to exam-
ine  the predicted effect of chemical reaction in the correlation for
                                 74

-------
 LIQUID
  INLET
6-3/4"
     6-3/4
            8"
                          7/8"
>
6"
i
7
\
V

1

1
f
J
24-
^

- L
4"
i
3/4"
J
c

                               -CL-
           FIGURE  12
         DIMENSIONS OF
COTTRELL FLOODED DISC  SCRUBBER
              75

-------
                 FIGURE  13
   COMPARISON OF  EXPERIMENTAL DATA AND
   PREDICTED VALUES OF S02  REMOVAL FOR
      COTTRELL FLOODED DISC SCRUBBER
   80
LLJ

 CM
O
CO

UJ
   60
to
   40
   20
       Na2C03 SOLUTIONS
       G - 400-1300 cfm
       L/G • 6-14 gal/mcf
             O
               40
                          60
               P RED ICTEDS02 REMOVAL, %
80
                    76

-------
(see Eqs. 26 and 28).  As the outlet liquor pH readings for the clari-



fied lime runs -were fairly low (5. 2 in each case), the equilibrium inter-



facial mole  fraction of SO , U*~ ,  could not be assumed zero.  Values


     •¥•
for W   at each axial location were obtained from the table based upon



Radian Equilibrium Computer  Program.






For three clarified lime runs, the gas rate was maintained at 700 cfm



and the  liquid-to-gas ratio at 10 gal/mcf.  The inlet gas was at a tem-



perature of 330 F  and contained approximately 1600 ppm SO .  Throat



velocities were 105 ft/sec.







As Figure  14 indicates,  a reasonable correspondence is obtained be-



tween predicted and experimental values of SO  removal with values



of  X and I/ in the correlation for -£n_    (Eq. 27) of 20 and 1/2, re-



spectively.   Since the predicted SO  concentrations in the gas and
                                  £

liquid were near equilibrium at the exit of the  venturi, only (as with



the OAP water data) the lower limit for A (and "/ ) could be checked.







Profiles of the predicted overall and individual mass transfer coeffi-



cients for a clarified lime run are shown in Figure 15.  The predicted



overall mass transfer coefficient is gas phase controlled for only a



short distance after droplet formation. However, about two-thirds of



the calculated mass transfer occurs in this region, due to the rapid



approach to equilibrium in  the absence of appreciable liquid phase re-



sistance.  The sharp drop in liquid side coefficient is due to the sudden



depletion of "reagent" within  the liquid (see Eq. 28).
                                 77

-------
              FIGURE  14
COMPARISON OF EXPERIMENTAL DATA AND
 PREDICTED VALUES OF S02 REMOVAL FOR
    COTTRELL FLOODED DISC SCRUBBER
  20
  15
o
 CM,n
O 10
CO
O
UJ
CO
a
   o
LIME SOLUTION
G - 1000 cf m
L/G - 7 gal/mcf
                     10
                       15
20
             PREDICTED S02 REMOVAL, %
                   78

-------
                 FIGURE  15

   PREDICTED MASS TRANSFER COEFFICIENTS


     FOR S02 REMOVAL IN LIME  SOLUTION

     IN COTTRELL FLOODED DISC SCRUBBER


    1000
en
 o>

 o
 E
 i
    800 J
o
o
o
CO
CO

CO
    600 -
400 -
    200 -
     0
                         G = 1000 cf m


                         L/G -10 gal/mcf
       0      0.5      1       1.5      2      2.5


         DISTANCE FROM POINT OF DROPLET FORMATION, IN.
                     79

-------
Cottrell (Tidd) Data;  Limestone Scrubbing. Of the limestone wet
scrubbing experiments conducted by Cottrell at the Tidd pilot plant,
only those from Task C-6 are considered to be reliable (Ref. 33).
Except for slightly higher gas throat velocities (127 to 151 ft/sec),
experimental conditions for these tests were essentially the same as
those listed for the clarified lime solution runs in the previous section.

For the model comparisons,  the initial liquor  compositions were based
upon measurements by Radian Corporation during Task C-6 (Ref.  32),
with appropriate modification for inlet pH variation from run-to-run.
The previously determined values of ^  and I/ (of 20 and 1/2, respec-
tively) were  used in the liquid-side mass transfer coefficient correla-
tion (Eq.  27) and the assumption was made that there was no solids
dissolution.  In Figure 16 the experimental SO  removals for limestone
wet-scrubbing are  compared with the model predictions (the clarified
lime runs are also included for comparison).  Since,  for these lime-
stone runs, the SO  measurement error of about  .  160 pprn (see Ref. 31)
                  £*
is as large as the measured removal, the overall comparison is not
bad.  If an infinite  rate of solids dissolution had been assumed, the
venturi model would have predicted SO  removals of greater than 50%
                                     C*
for all of the runs.
                                 80

-------
              FIGURE  16
COMPARISON  OF EXPERIMENTAL DATA AND
 PREDICTED VALUES OF S02 REMOVAL FOR
    COTTRELL FLOODED DISC SCRUBBER
      A LIME SOLUTION
      O LIMESTONE SLURRY
        G -1000 cf m
        L/G - 7 gal/mcf
               O  O
   0
             PREDICTED S02 RBVIOVAL, %
                  81

-------
                   HYDRO-FILTER SCRUBBER

The Hydro-Filter (or flooded-bed) scrubber is shown schematically
in Figure 17 and is described in Refs.  34 and  35.  Normal operation
is as follows:  The main liquor stream (plus sn«««»«H»cl solids) is
sprayed through nozzles onto the bottom side  of a bed of glass
spheres (marbles); simultaneously, a  secondary flow of liquor is
sprayed onto the  top side of a "turbulent layer" formed above the
spheres.  Gas (with suspended solids) enters  the scrubber at the
bottom, concurrently contacts the main liquor stream,  bubbles
upward through the turbulent layer, and  leaves at the top.  The main
effluent liquor stream is collected in an  overflow-weir within the
turbulent laver region and  is directed  out of the scrubber.

PRESSURE DROP

The total gas-phase pressure drop is assumed to be the sum of the
pressure drop through the  glass-sphere  and turbulent-layer  regions
of the Hydro-Filter.

The pressure drop tht -  gh the glass-sphere region of the Hydro-
Filter may be estimated from the following correlation, which is
                               82

-------
             FIGURE  17
    SCHEMATIC OF HYDRO-FILTER
             SCRUBBER
INLET LIQUOR
    GAS IN
               GAS OUT
TURBULENT LAYER

  GLASS SPHERES
«—INLET LIQUOR
                        EFFLUENT LIQUOR
            EFFLUENT LIQUOR
                  83

-------
based upon the work of Leva (Ref. 36 )  and Ergun (Ref. 37),
where ft is a constant which depends upon the type of packing.  The

characterization factor of packing is given by:
The upper limit on pressure drop occurs when the bed becomes fully

fluidized, and it has been suggested by Fan (Ref. 38) and Leva (Ref.

39) that at this point the pressure drop is approximately equal to the
weight of the bed.  Therefore:
                            S-z
The pressure drop due to the turbulent layer above the spheres is

estimated from the following correlation given by Ludwig (Ref. 40):
    It has been assumed that the form of the Leva correlation
    (Ref.  36), which is based upon data taken below the flooding-
    line for dumped packings and with counter cur rent flow, is
    applicable for the glass-sphere region of the Hydro-Filter.
    In normal operation the glass-sphere section is within the
    flooded range of fixed beds and approaches fluidization.
    Also,  the flow through the spheres is primarily cocurrent.
                                 84

-------
It follows that the total pressure drop &  in the Hydro-Filter is
given by:
The constant/3 in Eq. 45 was estimated to be i. 5 x 10~8 from a fit
of overall pressure drop (Eq. 49) to Hydro-Filter vendor test data
(Ref.  41).  The comparison between predicted and measured overall
pressure drop is  shown in Figure 18.  For the fitted value of
Eq. 49 becomes:
                                '               „,0
In Figure 19 the total predicted pressure drop in the Hydro-Filter
(Eq. 50)  is shown as a function of the liquid and gas rates.


The form of the expression for pressure drop which will be tested
with the data obtained at the facility is:
where the coefficients /£.   through Q, are to be determined from
the data.
                                  85

-------
               FIGURE  18
COMPARISON OF EXPERIMENTAL DATA AND
  PREDICTED VALUES OF PRESSURE DROP
            FOR HYDRO-FILTER
      SPHERE DIAMETER » 0.625m.
      O SINGLE-STAGE HYDRO-FILTER
      ODOUBLE-STAGE HYDRO-FILTER
           246
          MEASURED PRESSURE DROP, in.
                   86

-------
                   FIGURE  19
     PREDICTED PRESSURE  DROP THROUGH
     GLASS-SPHERE REGION  AND TURBULENT
             LAYER OF HYDRO-FILTER
 10.0
  9.0
  8.0
  7.0

  6.0
  5.0
o
CM
  4.0
o.
o
o
£ 3.0
to
to
£ 2.5
QL
   1.5
   1.0
—i	1—\—I	r-
SPHERE DIAMETER = 0.625 in.
PACKING HEIGHT = 3.5 in.
OVERFLOW WEI R HEIGHT - 8.0 in.
     600
                       REGION OF
                        INTEREST
      I   I	L
1
                                   1
           1000               2000
            GAS FLOW RATE, Ib/hr ft2
                87
                3000

-------
PARTICULATE REMOVAL

Particulate removal in the Hydro-Filter scrubber is obtained within
the bed of glass spheres and also within the turbulent-layer region
directly above the glass spheres (see Figure  17). Calvert (Ref. 2 and
42) has found participate removal in packed beds to be consistent with
an inertial impact!on collection mechanism.  He has also presented
(Ref. 43) an equation based on this mechanism for particulate removal
in a frothing bed on a sieve tray.  Similarly,  Pozin et al. (Ref.  44)
have also found  removal in froth to follow an  impaction mechanism.

In accordance with the theories presented in Refs. 2 and 43, particu-
late removal within the glass-sphere and turbulent-layer regions of
the Hydro-Filter for particles of size i can be represented by:
The inertial impaction parameter is defined as:
    The term L/£- was not included in the original theory.
    The term-^  was not included in the original theory.
              art
                                 88

-------
The height of the turbulent layer is estimated from a correlation
presented by Lowry and Van Winkle (Ref. 45):
with the following constraint:
The total removal of participates of sizec. within each stage** (glass
sphere and turbulent layer region) of a Hydro-Filter is given by:
                                        J [, -
The numerical constants  Ql ,  and Py^ in Eqs. 52 and 53 were obtained
by fitting the overall particulate removal (Eq. 8) to vendor test data
(Ref.  41),  using the  assumed flyash distribution curve shown in
Figure 2.   The comparison between predicted and measured over-
all removals for values of /?j  , and Q^ of 0. 2 and 6. 3, respectively,
are shown in Figure 20.   In  Figure 21,  predicted overall particulate
removal efficiency is shown as a function of gas rate and liquid-to-
gas ratio, based on  the fitted coefficients and the assumed distribution
curve (Figure 2).
    The overflow weir height is measured from the base of the
    glass spheres.
##
    The Hydro-Filter is normally operated in either a single-
    stage (one ma,rble-bed) or double-stage (two marble-beds)
    configuration.  For a double-stage configuration, the
    particulate removal is given by:
                                  89

-------
               FIGURE  20
COMPARISON OF EXPERIMENTAL DATA AND
      PREDICTED VALUES OF OVERALL
       PARTICULATE REMOVAL FOR
               HYDRO-FILTER
           1       I       I
      SPHERE DIAMETER = 0.625 in.
      O SINGLE-STAGE HYDRO-FILTER
      O DOUBLE-STAGE HYDRO-FILTER
           96      97      98      99
           MEASURED PARTICULATE REMOVAL,
                   90

-------
                FIGURE  21
PREDICTED  OVERALL PARTICULATE  REMOVAL
EFFICIENCY  FOR SINGLE-STAGE HYDRO-FILTER
                   SPHERE DIAMETER-0.625 in.
                   PACKING HEIGHT -3.5 in.
                   OVERFLOW WEIR HEIGHT « 8.0 in.
                  20       25
                   L/G, gal/mcf
                     91

-------
 For the fitted values of the coefficients /?,  , and 7?^ ,  Eq. 57
 becomes:
 As an example, for a gas flow rate of 2000 Ib/hr f t , a liquor flow
 rate of 3000 Ib/hr ft2 (L/G = 12.7 gal/mcf), a sphere diameter  of
 0. 625 in. ,  a packing height of 3. 5 in. and an overflow weir height
 of 8 in. , the particulate removal efficiency from Eq. 58 is as fol-
 lows for various characteristic particle diameters:
               microns                      % Removal
               1                                15
               2                                48
               5                                98
               10                               100
 The predicted overall removal efficiency (see Figure 21) for the
 above example is 98. 2%.

 The form of the  expression for particulate removal which will be
 tested with the data obtained at the facility is:
                      -   J F * Ki H-r /dH 1
where the constants /j% ,  O^ , andy33  are to be fitted to the data.
The parameters f and /fi-will be directly measured from photo-
graphic data, if possible.
                                   92

-------
ABSORPTION OF SO.
                    b
The major assumptions made in the formulation of the mathematical

model for mass transfer of SO, are:


    (1)   Plug flow of gas phase

    (2)   Completely back-mixed bulk liquid phase throughout
         entire mass transfer region of Hydro-Filter, i.e.
         uniform composition (and temperature) within bulk
         liquid phase. Furthermore, the liquid "film" between
         vapor and bulk liquid is also assumed to be at a
         uniform composition (and temperature).
    (3)   Liquor is at all times in equilibrium with an interfa-
         cial CO, vapor pressure of 0.1 atm of CO,, i. e.
         rate of CO,  absorption (and desorption) is  very large.


The development of the equations for describing SO, mass transfer

in the glass-sphere and turbulent layer regions of the Hydro-Filter

will not be presented here,  due to the length of the venturi presenta-

tion.  The basic differential equations describing SO  mass transfer,
                                                  u
heat transfer and solid dissolution are similar to those presented for

the venturi.


As an example  of a result of the Hydro-Filter modeling,  for the inter-
facial SO mole fraction,  (A , equal to zero (gas phase controlling),

SO, removal becomes:
SO
         ,
                                             "7
                                             J
                                  93

-------
                                                                «•
The coefficients  tft , and 8* are to be fitted to the data obtained
                 r*       r
at the test facility for Na CO. aqueous scrubbing of mixtures of air
                       u   J
and S0_ (see Ref. 55)*.
  Figure 22 (which showed "Predicted Mass-Transfer-Limited
  SO. Removal") has been omitted in this 2nd edition of the
  paper  since discussions with the vendor have indicated that
  the figure was probably in error.  Figure 22 was to have ap-
  peared  on page 50.
                                 94

-------
        FIGURE Z2
Not included in this revision
                95

-------
                         TCA SCRUBBER

A two-stage TCA scrubber, shown schematically in Figure 23, uti-
lizes a packing of low density plastic spheres which are free to move
between retaining grids.  Normal operation is as follows  (Ref. 46):
Gas enters the scrubber at the bottom and passes upward through
the packing, while the liquor stream is sprayed downward from the
top.  The  countercurrent flow forces the spheres upward  in a random,
turbulent pattern and the degree of bed agitation can be changed by
controlling the liquid and gas rates.
PRESSURE DROP
Curves showing total pressure drop (include demister) for 1,  2 and
3-stage TCA pilot scrubbers have been furnished by UOP (Ref.  47).
Thefollowing equations were fit (multiple regression computer pro-
gram) to the pressure drop curves, under the assumption that the
form of the correlation for pressure drop in the packing section is
the same as that given by Leva (Ref. 36).
                                  96

-------
           FIGURE 23
   SCHEMATIC OF TWO-STAGE TCA
SCRUBBER WITHOUT TRAP-OUT TRAY
 INLET LIQUOR
    GAS IN
            GAS OUT
               I



k p-




1
)))W))))Yt.
////////////

fr & $
»\\ /t\ • :\
O O° QQO
o oo _ ooo
O^OQ 0 ^0

0~0 0 °~0
o n x*_L
°o%^°ol
o ^ o^o^8
».
.
, hCUICTCD
1 DcMldltK

J> RETAINING GRIDS
r
u/\nnr m^ifiii/^
MOBILE PACKING
SPHERES


                    * EFFLUENT LIQUOR
                 97

-------
A comparison between predicted and measured (Ref. 48) pressure
drops for a two-stage TCA is shown in Figure 24.

In Figure 25 the predicted total pressure drop is shown for a 2 -stage
TCA scrubber and demister, as a function of gas and liquid flow rates.
Typically, the pressure drop through the demister section is less
than 15% the total pressure drop.

PARTICULATE REMOVAL

Particulate removal in the TCA scrubber is obtained within the
fluid! zed- bed stages.  Calvert (Ref. 42) has found that inertial impac-
tion is the predominant mechanism for particle collection in packed-
beds.  With the assumption that the form of the expression for parti-
culate removal in a TCA stage is  the same as that for a packed-bed
stage, the following equation is proposed for estimating removal in
a TCA stage:
The inertial impaction parameter is defined in Eq. 6, and overall
removal is obtained from Eq. 8.  The function of  ys ( L./&J ^ s
was determined by using a multiple regression fit of Eq. 65 to
overall particulate removal data from Ref. 49, for a 2 -stage TCA
scrubber.  The  resultant form of Eq.  65 is:

                          - s*io"(
                                   98

-------
                  FIGURE  24
   COMPARISON OF EXPERIMENTAL DATA AND
     PREDICTED VALUES OF PRESSURE DROP
              FOR TWO-STAGE TCA
  14
  12
.E 10
Q_
O

o
LU
Q£
ID
CO
CO
LU

Q_
O
LU

O
O
   8
   0
   I     T      I      I
10" OF 1-1/2" SPHERES/STAGE
OPERATING RANGE:
   G = 450-1050 cf m
   L/G = 20- 90 gal/mcf
                                  O
                              O
              O
                                O
                                 o
                                          o
          2     4     6     8     10
             MEASURED PRESSURE DROP, in.
                                       12
                                     14
                     99

-------
                 FIGURE   25

        PREDICTED PRESSURE  DROP

    THROUGH TWO-STAGE TCA SCRUBBER
   12
   10
o  o
 CM O
Q_
O
CXL
OC

CO
to
LU
QC.
Q_
              I          I

       10" OF 11/2" SPHERES/STAGE

       128 °F GAS TEMPERATURE
   0
    1500
2000       2500

   GAS FLOW RATE, !b/hr ft
3000
2
3500
                     TOO

-------
A comparison between the predicted (Eqs. 66 and 8) and measured
(Ref. 49) overall removals is shown in Figure 26,  using the flyash
size distribution given in Ref. 48.  In Figure 27, the predicted over
all particulate removal efficiency for a 2- stage TCA is  shown as a
function of gas and liquid flow rates,  using the assumed Shawnee
flyash distribution (Figure 2).

The  form of the expression for particulate removal which will be
tested with the data obtained  at the  test facility is:
where the const  ants ~8t ,  and/S^are to be fitted to the data.

ABSORPTION OF SO-
                    £

The major assumptions made in the formulation of the mathematical
model for absorption of SO  are:
                          2

    (!)   Plug flow of gas phase.
    (2)  A dispersion model can be used to describe axial mixing
         within the liquid phase.  This dispersion model consi-
         ders all deviations from plug flow to be  centered in a
         term analogous to Fick's law of diffusion.
    (3)  Liquor is at all times in equilibrium with an interfacial
         CO_ vapor pressure of 0.1 atm of CO?, i. e. rate of
         CO2 absorption (and desorption) is large.
                                101

-------
                FIGURE  26
 COMPARISON OF EXPERIMENTAL DATA AND
      PREDICTED VALUES OF OVERALL
        PARTICULATE REMOVAL FOR
             TWO-STAGE TCA
  100
 .99
  98
o
cz.
UJ
§
o
< 07
Q_ VI
O
^
O

§96
Q_
  95
           I        I       I
       10" OF H/2" SPHERES/STAGE
       OPERATING RANGE:
         G = 550-1000 cf m
         L/G = 20-70 gal/mcf
                   I
                                 1
          96      97      98      99
          MEASURED PARTI CULATE REMOVAL, %
                                       100
                    102

-------
               FIGURE  27
PREDICTED OVERALL PARTICULATE REMOVAL
     EFFICIENCY FOR  TWO-STAGE TCA
                       12" OF 11/2" SPHERES/STAGE

                       128 °F GAS TEMPERATURE
                  L/G, gal/mcf
                    103

-------
As with the Hydro-Filter,  the development of the equations for
describing SO? absorption in the TCA will not be presented here.

Correlations for axial dispersion in the liquid phase,  liquid holdup,
mass transfer coefficient  and interfacial area have been presented
by Khanna (Ref.  50) and Chen (Refs.  51 and 52).  The  partial differ-
ential equations describing the mass transfer between liquid  and gas
have been discussed by Khanna (Ref. 50)  and Miyauchi (Ref.  53 and
54).
                              104

-------
O
                         NOMENCLATURE

              gas-liquid interfacial area per scrubber volume,
              I/ft
A            cross-sectional area,  ft

              drag coefficient
   ,           characteristic factor of packing, ft

  fo           specific heat,  BTU/lb°R
              diameter of particulate of size £ » micron

              diameter of glass spheres,  ft
              diameter of plastic spheres, ft
              mean diameter of droplets, micron
                                              2
              diffusivity of SO2 in nue gas, Ib /hr

 p .           collection efficiency for particulate of size L

 p"            specific gravity of turbulent layer
                                                      2
 O/           gravitational acceleration,  Ib   ft/lb  hr
 "                                   2
              gas flow rate, Ib/hr ft
                                                         2   o
              gas-side heat transfer coefficient, BTU/ft hr R
  ^           liquid- side heat transfer  coefficient,  BTU/ft hr  R

 U            height of glass -sphere region, ft
                                105

-------
np            static height of packing for a single-stage    TCA,  ft

//-r            height of turbulent layer, ft

/-/yy            height of overflow-weir, ft

   .            gas-side mass transfer coefficient,  Ib mol/hr ft

   i_            liquid-side mass transfer  coefficient, ft/hr

   ,            liquid-side mass transfer coefficient in absence of
               chemical reaction, ft/hr

   r            reaction rate constant based upon unit surface, ft/hr

 Lx'            inertial impaction parameter

               overall gas-phase mass transfer coefficient,
               Ib mol/hr ft^

               overall gas-side mass transfer coefficient for
               diffusion of water vapor, Ib mol/hr ft

 L/&         liquid-to-gas ratio, gal/mc f

              Henry's law constant,  ft / Ib mol

              number of stages in TCA

              total number of solid spheres per unit volume of
              liquid, I/ft3

              pressure drop, in-H O

/?) .          volumetric flow rate of gas,  cfs
*-**•<>

              volumetric flow rate of liquid, gpm

              raoius of solid particle, ft

              initial radius of solid particle, ft
                                 106

-------
              Reynolds number

 C-            Schmidt number

              Sherwood number

•£            tLne,  hr
_                            o
 l^            temperature,   F
   J&                                               o
 ~f~~           temperature of gas -liquid interface,  F

 (J            gas velocity, ft/sec

If           liquid droplet  velocity, ft/sec

 (/y •           weight fraction of particulate of size  c-

 X^           ccncentration  of dissolved A (SO ) at gas -liquid
              interface, Ib mol/ft3

 •^Q           initial concentration of dissolved reagent B,
              Ib mol/ft3

 r\cf\         concentration  of total calcium in solution,  Ib mol/ft

 X_           concentration  of total calcium in solution in
              equilibrium with liquid and solid phases,  Ib mol/ft

  H- $0-2      concentration  of H^SOj (or molecular SO2) in
                                                              3
          bulk liquid,  Ib mol/ft3

          concentration of all species in solution, Ib mol/ft
                                                            3
          concentration of magnesium in solution, Ib mol/ft

          concentration of solid species in slurry, Ib mol/ft

3         ccncentration of SO  species in  solution,  Ib mol/ft
                                                                3
X r.           concentration of SO  species in solution,  Ib mol/ft
                                                                3
                                 107

-------
               mole fraction of SO  in gas phase
                                  £t
   jf.
 Lf            mole fraction of SO  in gas phase in equilibrium with
 v             concentration of SO  (H SO ) in bulk phase of liquid
                                  £•   Ct   j

 Y            mole fraction of water vapor in gas phase

 y            mole fraction of water vapor at gas-liquid interface

 ^2.            distance, ft

 2?L            distance from point of droplet formation,  ft

^12           height of differential element, ft

Greek Letters

     04.       constants in Eq.

     >./•-?,      constants to be  determined from data
      r6
               specific  surface area,  ft /lb

               void fraction, i.e. fraction of scrubber volume
               occupied by gas

               particulate removal efficiency

 A            constant in Eq.  26

 A            heat of vaporization of HO,  BTU/lb mol
                                       Lf
 LA,            viscosity, Ib/ft sec

 /2/           stoichiometric coefficient relating the number of
               moles of B reacting with one mole of A
                            O
 @            density,  Ib/ft

 Cs            surface tension,  Ib/sec

/ft sf)f        equilibrium functions (Radian Computer Program)

 (U            function
                                 108

-------
Subscripts

 £D           refers to demister

 &           refers to gas
  «
  ^            refers to particulate of size  c,

  L           refers to liquid

 M           refers to glass-sphere region or glass spheres

 0            refers to point of droplet formation or point at ^ = 0

 ~f)            refers to solid particulates

 P            refers to plastic sphere packing in TCA

 S            refers to solids

              refers to venturi throat

 "T            refers to turbulent layer region
                                  109

-------
                         REFERENCES
 (1)    R. H. Boll and C. A.  Leeman, "Particle Collection and Pressure
       Drop in Venturi Scrubbers, " AIChE 69th National Meeting,
       Cincinnati, Ohio, May 16-19, 1971
 (2)    A. C. Stern, Air Pollution, Vol. 3, Chap. 46, "Source Control
       by Liquid Scrubbing, " Academic Press, New York, 1968
 (3)    S. Calvert, AIChE Journal, Vol. 16, No. 3, p.  392, May 1970
 (4)    V.I. Matrozov,  Soobahcheniya O Nauchno - Tekhicheskikh
       Rabotakh NIUIF, Nos. 6/7, p,  152,  1958
 (5)    B.P. Volgin et al, Int'l. Chem. Eng. , Vol. 8, No. 1,  p.  113,
       1968
 (6)    J. D.  McKenna,  "Evaluation of Two-Stage Participate Scrubber
       and Gas Absorber Applied to Power Plant Flue Gas, "  presented
       at Pensacola,  Florida,  March  16-20, 1970
 (7)    Research-Cottrell,  "Flooded-Disc Wet Scrubbers. " Bulletin
       RC-975
 (8)    B. N. Murthy and B. C.  Eusebio, Air Pollution Control Office
       (APCO), In-House Venturi Project for SOY Control by Wet
       Scrubbing, Annual Progress Report,  March 1971
 (9)    C. Y. Wen and S. Vchida, "Simulation of SO  Absorption in a
       Venturi Scrubber by Alkaline Solutions, "  presented at New
       Orleans, Louisiana,  November 8-12, 1971
(10)    H. F. Johnstone, R. B. Field,  and M. C.  Tassler, Ind. and Eng.
       Chem., Vol.  46, No. 8, p. 1601,  August 1954
(11)    R. B. Field, Ph.D.  Thesis, University of Illinois, Urbana,
       Illinois,  1952
(12)    H. F. Johnstone  and M. H. Roberts, Ind.  and Eng. Chem. ,
       Vol. 41, No.  11, p.  2417, November 1949
                              110

-------
(13)    W. H.  Walton and A. Woolcock, Intern.  J. Air Pollution,  Vol. 3,
       p.  129,  I960
(14)    Tennessee Valley Authority,  Report No. 64-70-6299-5 on
       Shawnee Steam Plant, February 3,  1970
(15)    S.  Nukiyama and Y.  Tanasawa, Trans.  Soc.  Mech. Engrs. ,
       (Japan), Vol. 4, No. 14, p. 86,  1938
(16)    R.  D. Ingebo, NACA Tech. Note 3762, September 1956
(17)    E.  Rabin and co-authors, AFSOR TR 60-15,  Rocketdyne, (A
       Division of North American Aviation, Inc.),  March I960
(18)    Radian Corporation, A Theoretical Description of the Lime-
       stone-Injection Wet Scrubbing Process, NAPCA Report, June 9
       June 9,  1970
(19)    R.E. Treybal,  Mass Transfer Operations, McGraw Hill, New
       York, 1968
(20)    A.  E. Handles and T. Baron, AIChE Journal, Vol. 3, No. 1,
       March 1957,  p.  127
(21)    T.  K. Sherwood and F. A. L. Holloway,  Trans. Amer, Inst.  Chem.
       Engrs., Vol. 36, p. 39, 1940
(22)    R.  P. Whitney and J. E.  Vivian, Chem.  Eng.  Prog., Vol. 45,
       No. 5,  p.  323,  May 1949
(23)    S.  Calvert,  Wet Scrubber Systems Study; First, j>econd, and
       Third Quarterly Progress Reports,  Ambient Purification
       Technology, Inc., 1971
(24)    J. H. Perry,  Chemical Engineers Handbook,   Fourth Edition,
       McGraw Hill, New York, 1963

(25)    T.  K. Sherwood and R. L. Pigford, Absorption and Extraction,
       McGraw Hill Book Co. ,  Inc. , New York,  1952

(26)    T.E. Antoine, Vapor Pressure of Organic Compounds,  Inter-
       science, New York, 1954

(27)    G.  L. Smithson and N. N.  Bakhshi, Can. J. Ch.  Eng. . Vol.  47,
       p.  508,  October 196.9
(28)    D.  Levenspiel, Chemical Reaction Engineering, John Wiley
       and Sons, Inc.,  New York, 1962
                                m

-------
(29)    C. Y.  Wen, Ind. and Eng. Chem. , Vol. 60, No. 9, September
       1969
(30)    Battelle Memorial Institute,  Final Report on Investigation of
       the Limestone - SOp Wet Scrubbing Process,  by R. W. Coutant,
       R.H.  Cherry, H. Rosenberg, J. Genco, and A. Levy, Novem-
       ber 30, 1969
(31)    Cottrell Environmental Systems, Inc. , Pilot Scale Investigation
       of a Venturi-Type Contactor for Removal of SO by the Limestone
       Wet-Scrubbing Process, (final report  draft)'  July 1971

(32)    Radian Corporation Tech.  Note 200-006-12,  Chemical and
       X-Ray Analysis of Samples  Taken During the  Runs C5 and C6
       at the Tidd Plant in Brilliant, Ohio,  February 26, 1971

(33)    R. J. Gleason,  private communication, September 1971

(34)    National Dust Collector Corporation, General Catalog,
       December 23, 1968
(35)    J. E. Hanway, E. B.  Henby, and G. R.  Smithson, TAPPI, Vol.  50,
       No. 10, October 1967

(36)    M.  Leva,  Chem. Eng. Prog.  Symp. Ser.,  Vol. 50,  No.  10,
       p. '51,  1954

(37)    S. Ergun, Chem. Eng.  Prog. , Vol.  48, No. 2, p. 89, 1952

(38)    L. T.  Fan, "Written Comments on Bechtel Monthly (October)
       Progress Report, " January 1971
(39)    M.  Leva,  Fluidization, McGraw Hill,  New York,  1959

(40)    E. E.  Ludwig, Applied Process Design for Chemical and Petro-
       chemical Plants, Vol. 2, Chap. 8, "Distillation,"?.  98,  Gulf
       Publishing Co., Houston, Texas,  I960
(41)    Combustion Engineering Inc. , private  communication

(42)    S. Jackson and S. Calvert, AIChE Journal, Vol. 12, p.  1075,
       1966
(43)    S. Calvert,  private communication,  Lime/Limestone Wet Scrub-
       bing Symposium, Pensacola,  Florida,  March 1970
                                 112

-------
(44)    M. E. Pozin, I. P. Mukhlenov,  and Ya.E. Tarat, J.  Appl.  Chem.
       U. S. S.R. , Vol. 30,  p.  293,  1957

(45)    R. P. Lowrey and M. Van Winkle, AIChE .Journal. Vol. 15,
       No.  5, p. 665, 1969

(46)    Universal Oil Products, Air  Correction Division, Bulletin No.
       608, "UOP Wet Scrubbers, "  1967

(47)    Universal Oil Products, Air  Correction Division, Erection and
       Operating Instructions for Two-Stage TCA, Wet Scrubber,
       June 30, 1970

(48)    Universal Oil Products, Air  Correction Division, The Turbulent
       Contact Absorber for the Absorption of Sulfur Dioxide,  October,
       1965

(49)    W. A. Pollack, J. P. Tomany, and G. Frieling,  "Removal  of
       Sulfur Dioxide and Fly Ash from Coal Burning Power Plant Flue
       Gasses," ASME Publication 66-WA/CD-4, 1966

(50)    R. T. Khanna. Ph.D. Thesis, McGill University,  Montreal,
       Canada,  197 T

(51)    B.H. Chen,   Ph.D. Thesis, McGill University,  Montreal,  Canada,
       1965
(52)    B.H. Chen and W. J. M. Douglas, Canadian J. Chem. Eng.,
       Vol. 47, p.  113,  April  1969
(53)    T. Miyauchi,  "Longitudinal Dispersion in Solvent-Extraction
       Columns: Mathematical Theory, "  VCRL-3911, August 15, 1957

(54)    T. Miyauchi  and T.  Verme ulen, "Longitudinal  Dispersion in
       Two-Phase Continuous-Flow Operations, " I fo EC Fundamentals,
       Vol. 2,  No.  2, p.  113,  May 1963

(55)    M.  Epstein,  et al.,  "Test Program for the EPA Alkali Scrubbing
       Test Facility at the Shawnee  Power Plant," presented at Second
       International Lime/Lime stone Wet Scrubbing Symposium,  New
       Orleans, Louisiana, November  8-12, 1971
                                 113

-------
      Radian    Radian  Corporation
                 8500 SHOAL CREEK BLVD. • P. O. BOX 9948 • AUSTIN, TEXAS 78757 • TELEPHONE 512/454-9535
             A MODEL FOR THE LIMESTONE INJECTION  -
           WET SCRUBBING PROCESS FOR SULFUR DIOXIDE
               REMOVAL FROM POWER PLANT FLUE GAS
                              By:
                    Delbert M. Ottmers,  Jr.
                         Presented at:

              SECOND INTERNATIONAL LIME/LIMESTONE
                    WET SCRUBBING SYMPOSIUM
                      8-12 November 1971
                    Sheraton-Charles Hotel
                    New Orleans, Louisiana
                         Sponsored by:
                Environmental Protection Agency
                     Office of Air Programs
                 Division of Control Systems
                             115
CUBICAL RESEARCH • SYSTEMS ANALYSIS • COMPUTER SCIENCE  •  CHEMICAL ENGINEERING

-------
Radian Corporation
8500 SHOAL CREEK BLVD. • P. O. BOX 994B • AUSTIN. TEXAS 78757 • TELEPHONE 512 - 454-9535
1.0       INTRODUCTION

          The growing problem  of  atmospheric  pollution by sulfur
oxides has promoted a large  amount  of  research and  development
on removal process for power plant  flue  gases.   Of  the several
processes that have been proposed,  use of  lime or limestone in
a wet scrubber is one of the more promising.

          The limestone injection - wet  scrubbing (LIWS)  process
involves injecting limestone into the  power plant boiler  and
catching it  in a wet scrubber  after the  air heater  (see Figure 1)
Proponents of the process claim that injection of limestone into
the boiler removes a portion of the sulfur dioxide  ahead  of the
scrubber, provides protection  from  corrosion  by sulfur trioxide
and alkali salts, and converts the  limestone  to quicklime,  a
more reactive form for use in  the scrubber (TE-001,  PL-002,
MA-014).  The limestone tail-end  addition  (I/TEA) process  involves
adding the limestone in the  slurry  handling system.   In Figure 1,
the recycle  hold tank would  be one  possible addition point.  The
LTEA process avoids the injection of limestone into the boiler.
The major components of the  LIWS  and LTEA  processes are the
scrubber and the liquid slurry handling  system.  The LIWS and
LTEA processes are termed "throwaway processes" in.  that the by-
product sludge containing primarily fly  ash,  calcium sulfite
and calcium  sulfate is currently  considered a waste stream.

          A  number of similar wet scrubbing processes are being
considered for the control of  SOZ emissions from stationary
sources (MA-038, GR-009, DE-036,  SL-004, TE-001, CL-005,  BA-019,
HA-027, BR-014, CH-029, CH-030, SH-031).   These include the use
of limestone, lime, and magnesia  addition  into the  control
process liquor.  A useful chemical  by-product is to be recovered
from several of these processing  schemes.
                           116

-------
 COAL
      LIMESTONE
        11
AIR
                   FURNACE
                             AIR
                           HEATER
COOLER
           STACK GAS
           REHEATER
SCRUBBER
            EFFLUENT
             HOLD
             TANK
                      WATER
                     MAKE-UP
RECYCLE
 HOLD
 TANK
                                                                       SOLIDS WASTE
FIGURE i  -  SCHEMATIC  OF LIMESTONE INJECTION-WET SCRUBBING PROCESS

-------
Radian Corporation
8500 SHOAL CREEK BLVD. • P.O. SOX 9948 •  AUSTIN. TEXAS 78757 • TELEPHONE 512 - 454-9535
           Although a number of pilot unit programs have been
 conducted  and several commercial installations have been
 operating  periodically,  processing problems such as scaling
 and  low  S0a  removal still remain unsolved.  This, of course,
 is evidence  that the process design of these wet scrubbing
 processes  is not as simple as was first anticipated.  Radian's
 approach to  process engineering lime/limestone wet scrubbing
 systems  has  been to develop an understanding of their process
 chemistry.   This technology then provides a sound basis for
 (1)  determining the technical feasibility of process alterna-
 tives,  (2) estimating process flow rates and compositions,
 (3)  correlating experimental data, and (4) defining major
 process  variables.

           The purpose of this paper is to describe Radian's
 development  of a lime/limestone wet scrubbing process model
 and  to report some  of the simulation results from a parameter
 study of the LIWS  prototype system.  A comparison of the LIWS
 and  LTEA processes  is made by comparing the results of several
 simulation cases of each process.   Details of the process model
will be  described in terms of the LIWS prototype system.  How-
 ever, the model  is  easily adapted to other wet scrubbing schemes
 employing  the same  chemical species.
2.0       PROCESS DESCRIPTION

          Numerous processing  schemes  are possible for LIWS
and LTEA processes.  One  of  the more promising wet scrubbing
schemes for the LIWS is shown  in  Figure  2.
                             118

-------
GAS SPECIES
   FG.SG
 I. S02
 2. C02
 3.NOX
 4. H20
 5. 02
 6. CO
 7. N2
FLUE GAS
  FG
STACK GAS
    SG
              WATER
              MAKEUP
                WM
 SCRUBBER
     S
     LIMESTONE
      FLY  ASH
      SOLIDS
        LA
      I. CoO
      2. MgO
      3. CaS04
      4. MgS04
      5. CoS03
      6. MgSOs
      7. CoC03
      8. MgC03
      9. FLY ASH
     10. SOLUBLE Na
     II. SOLUBLE Cl
SCRUBBER FEED
                                       SF
                                    1
 PROCESS
  WATER
HOLD TANK
    P
      SCRUBBER
      BOTTOMS
        SB
 SCRUBBER
 EFFLUENT
HOLD TANK
    E
                                             SLURRY RECYCLE  SR
                                                     CLARIFIER
                                                      LIQUID
                                                        CL
   CLARIFIER
      FEED  .
      CF
CLARIFIER
    C
                                  CLARIFIER
                                  BOTTOMS
                                    CB
                               FILTER
                                 F
                                 I
                            FILTER
                            LIQUID
                             FL
                     FILTER
                     BOTTOMS
                       FB
                                       PROCESS SOLID  SPECIES
                                        (CF.SR.CB, FB, SF)
                                  I. CaO
                                 2. Ca(OH)2
                                 3. Co CO 3
                                 4. CoS03 • xH20
                                 5. CoS04 • xH20
                           6. MgO
                           7. Mg(OH)2
                           8. MgC03 -xH20
                           9. MgS03 • xH20
                           10. FLY ASH
   PROCESS
    LIQUID
   SPECIES
 SB.CF.SR, CB,
 FB,CL,FL,SF

  I.-H*
  2. OH"
  3. HS03
  4. SOf
  5. SO}
  6. HC03
  7. COf
  8. HSO?
  9. H2S03
 10. H2C03
 II. Co++
 12. CaOH+
 13. CaS03
 14. CaC03
 15. CaHCO^
 16. CaS04
 17. CaNOj
 18. N03
 19. Mg++
20. MgOH +
21. MgS04
22. MgHC03
23. MgS03
24. MgC03
25. Na+
26. NaOH
27. NaC03
28. NoHC03
29. NoS04
30. NaN03
31.Cl"
         FIGURE 2  - WET  SCRUBBING  SCHEME
                                  119

-------
Radian Corporation
8500 SHOAL CREEK BLVD. • P. O. BOX 9948 •  AUSTIN. TEXAS 78757 • TELEPHONE 512 - 454-9535
           In  this  flow arrangement,  the slurry recirculation
system  involves  two  holding tanks  plus a clarifier and filter
for solids  removal.   In the LIWS process, flue gas containing
both fly ash  and boiler-reacted limestone enters the bottom of
the scrubber.

          The chemical constituents  entering the wet scrubbing
system  from the  gas-solid  stream are listed on Figure 2.  The
limestone injected into the boiler will be in the form of
(1) the oxide (calcium or  magnesium) if it has been calcined,
(2) the sulfate  or sulfite if  it has reacted with sulfates or
sulfites via  gas-solid reaction in the high temperature zone
of the  flue gas  train,  and (3)  the carbonate if it has not
reacted or has recombined  with  C0a in the lower temperature
zone of the flue gas.   The flue gas  stream contains S02, COS,
NOX , and 0P which may be absorbed.  Soluble sodium, potassium,
and chlorides can  enter the aqueous  phase from the fly ash.

          In  the scrubber,  the  flue  gas (FG) and limestone-
fly ash solids (LA)  are contacted  with a basic scrubber liquor.
The scrubber  feed  (SF)  stream can  be either a clear liquid or a
slurry.  Due  to the  chemical species entering the aqueous phase
and to the various chemical reactions occurring within the
aqueous phase, a number  of ionic and nonionic species are present
in the process liquid  stream.   The list of process liquid species
is also given in Figure  2.   A list of the possible process solid
species is also shown.

          If a clear  liquid SF  stream is  utilized (no slurry
recycle), the SF flow  rate  must  be large  enough that  Che liquid
phase contains sufficient  basic  species to neutralize the S0a
and C0a absorbed in  the  scrubber.  If a slurry SF stream is
utilized, additional basic  species in the solid-phase will be
                            120

-------
Radian Corporation
3SOO SHOAL CREEK BLVD. • P. O. BOX 9948 • AUSTIN. TEXAS 78757 • TELEPHONE 512 - 454-9535
available for reaction  in  the  scrubber and the SF flow rate
required to remove a given amount  of S02  will  be less than with
a clear liquid feed.

          The chemical  objective of the LIWS process is to
promote the overall reaction of gaseous S02 reacting with solid
calcium or magnesium oxides to give solid sulfites,  i.e.,

    S02(g) + MeO(s) + H20(aq)  - MeS05 • xH20 +  H20(aq).

Gaseous C02 will react  in  a similar fashion to form  solid car-
bonates.  Oxidation of  the sulfite will give rise to sulfates
in the system.

          The actual reaction  path involves a  number of serial
steps.  The gaseous species S0a, C02,  NOX, and Oa must  be
absorbed by the liquid  at  the  gas-liquid  interface (see Figure  3)
Water is simultaneously being  transferred in the opposite direc-
tion due to its vaporization occurring in the  scrubber.   The
driving force for the sorption of  the  acidic gas species  is
maintained by the presence of  sufficient  basic species  in the
liquid phase.  In the scrubber, this basicity  can be furnished
by the dissolution of solids such  as calcium hydroxide,  calcium
carbonate, magnesium hydroxide, and magnesium  carbonate.   In  the
LIWS process, the calcium  and  magnesium oxides most  probably
hydrate to form the solid  hydroxides,  which then dissolve to
provide calcium, magnesium, and hydroxyl  ions.

          The LIWS system  involves a three-phase reaction system
which could be controlled  by one or more  individual  steps.  The
                                                      •
resistances to mass transfer are represented by both sides  of
the vapor-liquid film,  liquid-phase reactions,  liquid-solid
films, diffusion through solid layers,  and reactions with
solids.  The oxidation  of  sulfites to  sulfates will  influence
both vapor-liquid and solid-liquid equilibria.

                             121

-------
                   BULK VAPOR

                   VAPOR FILM
                 LIQUID FILM
                  BULK LIQUID
x ^LIQUIDJ-ILM
   'SOLID 1
       FIGURE 3 - THREE PHASE SYSTEM
                    122

-------
Radian Corporation
8500 SHOAL CREEK BLVD. • P. O. BOX W8 • AUSTIN. TEXAS 78757 • TELEPHONE 512 - 454-9S35
          The hydration of the calcium and magnesium  oxides
may take place in any part of the wet scrubbing network.   The
dissolution of the hydroxides and carbonates depends  upon  the
chemical environment in a particular scrubbing vessel.   That
is to say, the driving force for dissolution will  be  largest
in the scrubber where it is most acidic.  The driving force
for dissolution in the holding tanks and  clarifier where it
is more basic will be small.  In fact, if the holding times
are sufficiently large, solid-liquid equilibrium with respect
to dissolution of these species will be closely approached.

          The liquor in the scrubber will become  supersaturated
with respect to calcium sulfite and calcium sulfate.   The
degree of supersaturation occurring in the scrubber will depend
upon the liquid residence time and the rates of  (1) sulfur
dioxide absorption, (2) sulfite oxidation,  (3) calcium hydroxide
and carbonate dissolution, and (4) calcium sulfite and sulfate
precipitation.  The rate of calcium sulfite and  sulfate precipi-
tation is in turn dependent upon the degree of supersaturation.
Precipitation of calcium sulfite and sulfate on  scrubber com-
ponents has been a major operating problem for the LIWS process.

          Thus, it is apparent that proper design  of  the LIWS
system should promote the vapor-liquid mass transfer  of SOS
absorption and the solid-liquid mass transfer of  calcium
hydroxide and carbonate dissolution in the  scrubber.   The
precipitation of calcium sulfite and sulfate should be dis-
couraged in the scrubber and encouraged  in  the ancillary slurry
handling system.  The slurry handling  system should be designed
to provide (1) a basic scrubber feed stream and  (2) method for
removing the waste solids without undue  loss of  reactive lime
or limestone,,
                               123

-------
Radian Corporation
9500 SHOAL CREEK BLVD. • P. O. BOX 9943 • AUSTIN, TEXAS 7B7S7 • TELEPHONE 512 - 454-9535
3.0       PROCESS MODEL

          Based upon the present knowledge  of  the process
chemistry and equipment operating characteristics, Radian  has
formulated a LIWS process model using appropriate material
balance considerations and processing assumptions.  Due  to
the absence of rate data for this process,  Radian formulated
the model based upon conversions for these  rate  steps  as
specified by model inputs or by equilibrium assumptions.

          The present Radian model  is a  series of computer
routines which consists of three major elements:  (1)  equip-
ment subroutines that model each process unit, (2) convergence
subroutines that force convergence  of the model's iterative
parameters, and (3) an executive system  that interconnects the
processing units in the appropriate fashion and  controls the
sequencing of computer operations,

3.1       Process Chemistry

          The heart of the process  model is its  ability  to
describe the process chemistry adequately.   Thus, Radian started
this program by developing a chemical equilibrium program  for
the vapor-liquid-solid system encountered in the LIWS  process.
Vapor-liquid and liquid-solid mass  transfer rates are  usually
calculated using a driving force term that  is  a measure  of the
distance from equilibrium.  For some pieces of equipment the
assumption of various streams being in equilibrium is  reasonable
Hence the ability to predict equilibrium for the vapor-liquid-
solid system is a prerequisite for  process  modeling, data
analysis, and process engineering calculations.
                              124

-------
Radian Corporation
8500 SHOAL CREEK BLVD. • P. O. BOX 7948 •  AUSTIN. TEXAS 78757 • TELEPHONE 512 • 454 9535
          The development of the equilibrium program began  with
an investigation of the literature to determine  the significant
chemical species, reactions, and equilibrium constants.

          The chemical species in the aqueous phase originate
from lime (CaO, MgO), flue gas (C02, S02,  S03, 03 , NOX),  and
fly ash (Na, K, Cl, plus others).  A list  of neutral and  ionic
species formed by these species upon interaction with  water
was screened to retain only the species  having non-negligible
influences.  Some species were treated in  a simplified manner
so as to increase computational efficiency.  The potassium
species have been treated as "effective" sodium  since  the
equilibrium behavior of the two is similar.  Also, the nitrogen
oxides have been treated as forming only nitrates  in solution.

          A chemical equilibrium program was formulated based
upon the specification of nine key species in the  limestone
wet scrubbing process (CaO, MgO, Na30, SOS, S03, C02,  N205, HC1,
and H20),  Sulfite oxidation is considered to be a rate con-
trolled step in the limestone wet scrubbing process so that
both the amounts of sulfite and sulfate  sulfur are specified
as key species.  The anhydride of nitric acid, N20S, is speci-
fied for the NOX contribution to the liquid phase.

          The chemical reactions considered in the equilibrium
formulation are listed in Table I.  There  are twenty-two
dissociation-type equations, seven solubility product  relations,
and two gas solubility equations.  These equilibrium relation-
ships were expressed as activities using molality  times an
activity coefficient.  These equations were combined with mass
balance equations.  The set of nonlinear equations was solved
in the log molality domain.
                               125

-------
             TABLE I - REACTIONS CONSIDERED
LIQUID PHASE REACTIONS:
1) HS0 *
2) HSOj ?
3) HSS03 £
4) HSOj  Kg44" + OH"
§ i ..
•• t "1 '!' i o /^~"
i Mg 4- S03
I I ^^
i Kg' ' 4- SOJ
* Mg44" + 00=
^ Mg 4- HC03
i* Na4" 4- OH"
fi Na4" 4- SOI
0 Na+ 4- COJ
0 Na4" 4--HC03
# Na4" 4- NO;

-------
                              TABLE  I  - REACTIONS CONSIDERED  (Cont'd.)
                             LIQUID-SOLID  EQUILIBRIA:
ro
23)
24)
25)
26)
27)
28)
29)
GAS -LIQUID
30)
3D
CaC03(S) ?
CaS04(S) ;
CaS03(S) ;
Ca(OH)a(S) ;
Mg(OH)3(S) ?
MgC03(S) ;
MgSOg(S) *
EQUILIBRIA -
S02(g) + H20 <
C02(g) + H20 ;
5 Ca^ + C03
I Ca"^" + SOr
t Ca4"1" + tiO^
i Ca4^ + 20H
i Mg"1"4" + 2 OH'
i Mg44" + CO^
* Kg4"1" + S03
* H2 S03
i H2C03

-------
Radian Corporation
8500 SHOAL CREEK BLVD. • P. O. BOX 9948 • AUSTIN. TEXAS 78757 • TELEPHONE 512 - 454-9535
          The program predicts  the  equilibrium vapor pressure of
S02, C02, and H20 for the CaO-MgO-Na20-SOa-S03-C02-NS05-HC1-H20
system.  It also predicts liquid-solid  equilibrium for CaC03,
CaS03, CaS04, Ca(OH)3, MgC03, MgS03,  and  Mg(OH)2.

          The accuracy of the equilibrium program  in predicting
the vapor pressures of S02  and  C02  is shown  in Figure 4.   These
data are in the temperature range of  25 to 55°C and represent
ionic strengths up to 1.5.

3.2       Equipment Operating Characteristics

          The open literature was surveyed to  obtain information
concerning the important operating  characteristics  for the
various pieces of process equipment.  The scrubber  is the
primary area of concern here.   As mentioned  earlier, the
scrubber serves as a gas-liquid contacting device  with some
dissolution of lime and/or  limestone  solids.

          Two scrubber types were considered:   (1)  counter-
current contactors, such as the turbulent contact  absorber and
the marble-bed contactors,  and  (2)  cocurrent contactors,  such
as venturi scrubbers.

          The turbulent contact absorber  (TCA),  designed  and
manufactured by Universal Oil Products  Company,  Air Correction
Division, is a moving bed of hollow plastic  spheres.  The
marble-bed contactor (Hydro-filter),  manufactured  by National
Dust Collector Corporation, uses a  bed  of glass  marbles  instead
of hollow spheres.  Combustion  Engineering,  Inc.,  has'installed
two commercial scrubber systems using marble-bed units.   In each
of these scrubber types, the liquid-solid bed  is characterized
by violently turbulent action.  Both  of these  scrubbers will  be
                              128

-------
o
o
o
z
o
(0
u
w
in
u
o:
i-
c
o
u
_l
u
   ID
      *
     '1
     "2
   ID
     '5
 A  S02-H20 (25°C)
 O  S02- S04-H20 (25°C)
 D  Co* - S02 - H20 (25 AND 50°C)
 A  Mg»- S02 - SO; - H20 (25 AND 50°C)
 •  Co*- C02 - H20 (40°C)
 +  Co*- S02 - S03 - C02 - H20(55°C)
 •  Mg*-S02-S03 -C02-H20(55°C)
-  EXACT CORRELATION
                                                               rC02
                                                             RANGE OF
                                                             INTEREST
                                                              I     i
                                                            p
                                                              I/A
                                         S02
       IO
        '6
                  IO
                    "5
                  IO
                    '4
IO
  '3
IO
  "2
10
                                                                  "'
                EXPERIMENTAL  PARTIAL  PRESSURE  OF  S02  AND  C02  (ATM)
    FIGURE 4  -
                  COMPARISON  OF CALCULATED  AND  EXPERIMENTAL  PARTIAL PRESSURES
                  OF  S02  AND C02 FOR THE  SYSTEMS STUDIED
                                   129

-------
Radian Corporation
8500 SHOAL CREEK BLVD. • P. O. SOX 9948 •  AUSTIN. TEXAS 78757 • TELEPHONE 512 - 454-9535
tested in demonstration runs planned  by OAP for  the  LIWS process
on TVA's Shawnee Power Plant near Paducah,  Kentucky.

          Experimental data reported  by B.  H.  Chen "and  W.  J.
           ^
M. Douglas (CH-003) show  that  the liquid-phase dispersion
observed in a 12-inch diameter TCA  unit corresponds  to  approxi-
mately 1.12 ideally backmixed  vessels in series  for  the liquid
and gas flow rates considered  important in  the LIWS  process.
No experimental data concerning  the residence  time distribution
of the liquid phase in marble-beds  has been reported.   The
present Radian model assumes that the liquid phase in both of
these contactors is ideally backmixed.  The implication of this
assumption is that the liquid  in the  scrubber  has the same
composition as the liquid leaving the scrubber.  Thus,  if
vapor-liquid equilibrium  is closely approached in the contactor,
the partial pressure of S02 in the  exit gas will be  equal to  the
vapor pressure of S0a above the  exit  liquid.

          H. R. Douglas,  et al., (DO-001) studied SQS absorption
by NaOH solutions in a pilot TCA contactor. If  the  liquid flow
is approximately backmixed as  indicated by  Chen  and  Douglas,
the SOa-NaOH scrubbing system  appears to be very close  to vapor-
liquid equilibrium.

          In cocurrent contactors such as the  venturi,  the gas
leaving the scrubber is last in  contact with the liquor leaving
the scrubber.  Thus, as above, if vapor-liquid equilibrium is
closely approached in the scrubber, the partial  pressure of S02
in the exit gas will be equal  to the  vapor  pressure  of  SOS of
the exit liquor.  Experimental results on a pilot Venturi
scrubber (LO-027) indicate that  vapor-liquid equilibrium is
not closely approached, however.
                              130

-------
Radian Corporation
8500 SHOAL CREEK BLVD. • P. O. BOX W8 •  AUSTIN. TEXAS 78757 • TELEPHONE 512 - 45-4-9535
          The scrubber -effluent hold  tank is  a well-agitated
mixing tank that allows time for CaO  to hydrate,  Ca(OH)s  to
dissolve, and sulfite and  sulfate  solids  to precipitate.   This
stirred tank should approximate idealized backmixed flow and
in practice should be designed to  provide enough  residence time
and agitation to hydrate and dissolve the major portion of the
CaO and MgO that reacts in the system.   In a  properly designed
system, a major portion of the precipitation  will occur in the
effluent hold tank.

          The clarifier and filter function primarily as solid-
liquid separators.  If the scrubber effluent  hold tank is
designed to accommodate most of the solid-liquid  mass transfer,
very little additional reaction should  occur  in the rather
stagnant clarifier.  The process water  hold tank  is essentially
another completely mixed tank where make-up water is added and
some additional hydration,  dissolution,  and precipitation can
occur.

3.3       Model Formulation

          The Radian process model described  in this paper is
based upon several simplifying assumptions.

          (1)  The partial pressures  of S02,  COS,
               and HaO in  the stack gas leaving the
               scrubber are in equilibrium with the
               scrubber liquid at  the scrubber tempera-
               ture (vapor-liquid  equilibrium is
               achieved and the gas leaving the
               scrubber- is last in contact with
               liquid having the composition  of
               the scrubber exit liquor).
                               131

-------
Radian Corporation
8500 SHOAL CREEK BLVD. • P. O. BOX 9949 • AUSTIN. TEXAS 78757 • TELEPHONE 512 - 454-9535
           (2)  After  withholding a portion of the
               input  lime (CaO and MgO) as being
               chemically unreacted in the system,
               solid-liquid equilibrium is achieved
               in  the scrubber-effluent and process-
               water  hold tanks.

           (3)  Hold tanks closely approach idealized
               backmixed  vessels.

           (4)  The clarifier and filter are solid-
               liquid separators only (no chemical
               change occurs).

           (5)  Ionic  reactions taking place in the
               liquid phase are rapid and thus are
               at  equilibrium.

           With the present absence of reliable rate data, the
extent of  reaction for the remaining "rate controlling" steps
is specified by model input.  That is to say, the fraction of
limestone  solids in the inlet flue gas entering the scrubber
and dissolving, sulfite oxidizing, NOX absorbing, and solids
in the scrubber feed  dissolving are set by model parameters.
Thus, this version of the Radian process model is appropriately
termed a "fractional  conversions" process model.

           The  compositions for the system inlet and exit
streams [limestone (LS),  fly ash (FA), flue gas (FG), stack
gas (SG),  and  water make-up (WM)] and operating conditions for
the process are also  specified by model parameters.
                               132

-------
Radian Corporation
8500 SHOAL CREEK BLVD. • P. O. BOX 9948 • AUSTIN. TEXAS 78757 • TELEPHONE 512 - 454-9535
          Process  simulations  are conducted by specifying the
amount of S02  to be  removed  from a flue gas of known composition
and adjusting  the  amount  of  slurry recycle (or the amount of
scrubber feed  in some  cases)  to obtain this S02 removal.

          All  of the liquid-phase species are allowed to react
in the scrubber.   However, as  mentioned above, the fractions of
limestone solids reacting in the scrubber and the fractions of
various solid  species  in  the scrubber feed that are available
for reaction are specified as  model inputs.  Thus, for a fixed
scrubber feed  (SF) rate,  the amount of slurry recycle (SR)
required to obtain the desired SOS removal is a measure of the
scrubbing efficiency of the  process liquid.  High slurry recycle
rates indicate a poor  set of scrubbing parameters.

4.0       SIMULATION RESULTS

          A number of  simulations have been conducted for the
.LIWS wet scrubbing scheme depicted in Figure 2.  Simulation
results showing the  effects  of (1) sulfite oxidation, (2) ionic
strength, (3)  limestone amount and composition, and (4) scrubber
feed rate will be  presented  here.  More recently, a limestone
tail-end wet scrubbing system with essentially the same flow
arrangement was simulated.   In this instance, limestone was
added to the process water hold tank instead of entering with
the flue gas.  Observations  concerning the basic difference
between lime and limestone systems will be presented later.

4.1       Sulfite  Oxidation

          The  effect of sulfite oxidation on the operation of
the limestone wet  scrubbing  process is shown in Table II.  Here
the percent oxidation,  was varied from 0 to 90% for a system
                               133

-------
                                                  TABLE II
CJ
.£»
EFFECT OF SULFITE OXIDATION
CASE NUMBER
QUANTITY
SULFITE OXIDATION, 7,
SCRUBBER L/G (GAL/1000
ACF)
SLURRY RECYCLE/SCRUBBER
FEED RATIO
mFB[CaS03- 1/2H20(S)3
m [CaSO,(S)]
CO *t
mSF[Total S02]
mSF[Total S03]
fnHl 	
PSN19
0
13.4
0
5.34
0.00
2.8xlO"4
0.00
4.37
PSN13
10
14.0
0.208
4.80
0.52
2.8xlO"4
l.OxlO"2
4.30
PSN18
30
14.0
0.577
3.73
1.59
2.8xlO~4
l.OxlO"2
4.20
PSN2B
50
14.0
0.710
2.65
2.64
2.8xlO"4
l.OxlO"2
4.06
PSN15
90
14.0
0.819
0.52
4.70
2.8xlO"4
l.OxlO"2
3.39

-------
Radian Corporation
8500 SHOAL CREEK BLVD. • P. O. BOX 9948 • AUSTIN, TEXAS 78757 • TELEPHONE 512 - 454-9535
utilizing a pure CaC03  limestone.   The scrubber liquid-to-gas
ratio was maintained  at  14  gallons  of scrubber feed per 1000
ACF of flue gas, except  for the  0%  oxidation case.  In this
case, only 13.4 gal/1000 ACF were needed for 90% SOS removal
from a 2000 ppm S02 flue gas.  As the percent oxidation increases
from 10 to 907o the fraction of slurry recycle required in the
scrubber feed progressively increases from 0.208 to 0.819.  This
loss of scrubbing efficiency with increased oxidation is evi-
dently caused by the  formation of the strongly acidic sulfuric
acid (the pH of the scrubber bottoms stream is lowered).

          Of course oxidation changes the type of solids leaving
the system.  As expected,  the amount of calcium sulfite solid
leaving the system in the filter bottoms (FB) stream progres-
sively 'decreased with oxidation  while the amount of calcium
sulfate increased.  The  concentrations of sulfites and sulfates
in solution do not follow this trend, however, since they are
controlled by solubility relationships.

4.2       Ionic Strength

          The effect  of  ionic strength for systems injecting a
pure CaC03 limestone  is  given in Table III.  The ionic strength
of the scrubbing solution was varied from 0.48 up to 1.00 by
the addition of sodium chloride  to  the model.  This would
correspond to different  degrees  of  fly ash solubility in the
actual process.  For  the cases presented here, the effect of
ionic strength on scrubbing efficiency was fairly minor.  As
the ionic strength increased from  0.48 to 1.00', the fraction
of slurry recycle required in the  scrubber feed only increased
from 0.710 to 0.733.   This increase can be explained in terms
of the total concentration of basic species dissolved in the
scrubber feed liquid.  Here basic  species are defined as all
                               135

-------
                                             TABLE  III
CO
CTl
EFFECT OF
QUANTITY
IONIC STRENGTH
SCRUBBER L/G (GAL/ 1000 ACF)
SLURRY RECYCLE/ SCRUBBER
FEED RATIO
LIQUID-PHASE MOLALITIES OF SF
OH"
CaOH4"
CaS03U)
NaOH U)
All Basic Species
Ca44"
ACTIVITY COEFFICIENT OF SF
SPECIES
Ca-H-
OH"
sor
IONIC STRENGTH
PSN2B
0.48
14.0
0.710
1.56xlO"2
1.17xlO"2
2.59xlO"4
1.95xlO"4
2.78xlO"2
_]_

0.238
0.721
0.188
CASE NUMBER
PSN23
0.75
14.0
0.717
1.50xlO"2
1.14xlO"2
2.48xlO"4
6.39xlO"4
2.73xlO"2
1.21X10"1

0.226
0.744
0.157
PSN20
1.00
14.0
0.733
1.38xlO"2
l.llxlO"2
2.38xlO"4
1.04xlO"3
2.62xlO"2
1.27X10"1

0.228
0.784
0.139

-------
Radian Corporation
9500 SHOAL CREEK BLVD. • P. O. BOX 9948 •  AUSTIN. TEXAS 78757 •  TELEPHONE 512 • 454-9535
species capable of liberating  a hydrogen ion due  to the first
ionization of sulfurous acid.  For  the cases presented in
Table III, the sum of  the basic-species molalities  in the
scrubber feed decreases from 2.78xlO~2 to 2.62xlO~3 g-moles/
kilogram HaO as the ionic strength  increases.

          Although the sum  of  the basic-species molalities  does
not change much in this instance, the  individual  molalities and
activity coefficients  do vary  significantly over  the ionic
strength range shown with OH", CaOH and CaS03(^) decreasing
while NaOH(^) increased.  The  deviation of activity coefficients
from the ideal value of unity  demonstrates their  need in these
aqueous equilibrium calculations and suggests  their considera-
tion in rate correlations.  The significant variation of indi-
vidual molalities and  activity coefficients with  ionic strength
suggests that other regions of system  operation may exist
whereby ionic strength will have a  significant effect on
scrubbing efficiency.

4.3       Limestone Amount  and Composition

          The effect of limestone amount and composition is
shown in Table IV.  In cases PSN2B  and PSN4B,  112.5% of the
limestone required for 100% S02 removal was considered avail-
able for reaction in the system.  Case PSN2B corresponds to a
process injecting limestone containing a negligible amount  of
MgO, whereas Case PSN4B corresponds to a process  injecting
limestone containing 20 mole percent of reactive  MgO.  To
remove 90% of the S02  in a  flue gas containing 2000 ppm S02 ,
71.0% of the scrubber  feed was slurry  recycle  in  Case PSN2B.
On the other hand, Case PSN4B  required the scrubber feed to
be 87.7% slurry recycle.
                              137

-------
                                              TABLE IV
CO
CO
EFFECT OF LIMESTONE AMOUNT AND COMPOSITION
CASE NUMBER
QUANTITY
LIMESTONE AVAILABLE,
7, THEORETICAL
LIMESTONE COMPOSITION,
7, MgO
SCRUBBER L/G, GAL/
1000 ACF
SLURRY RECYCLE/ SCRUBBER
FEED RATIO
FILTER SOLIDS STREAM
pH
IONIC STRENGTH
mFBL (Total CaO)
mFBL(Total MgO)
mFBS[Ca(OH)2]
mFBS[Mg(OH)2]
7o S03 AS LIQUID SULFATE
7o MgO AS Mg(OH)2
PSN2B
112.5
0
14
0.710
11.3
0.48
l.SxlO'1
0
9.6X10"1
0
0.36

PSN4B
112.5
20
14
0.877
8.4
0.70
2.2xlO~2
2.4X10'1
0
1.14
4.54
82.69
PSN21
135
0
14
0.510
11.3
0.48
l.SxlO'1
0
2.16
0
0.37

PSN22
135
20
14
0.510
11.3
0.48
l.SxlO'1
9.1xlO"7
6.1X10'1
1.59
0.36
99.99+

-------
Radian Corporation
9500 SHOAL CREEK BLVD. • P. O. BOX 994B •  AUSTIN. TEXAS 78757 •  TELEPHONE 512 - 454-9535
          An explanation  of  this  observation is  possible when
the nature of the  scrubbing  liquid  is  examined for Cases PSN2B
and PSN4B.  Enough calcium oxide  has been added  to the system
in PSN2B so that the  scrubbing  liquor  becomes a  saturated
solution with respect  to  calcium  hydroxide (pH = 11.3) and
Ca(OH)a leaves  the system as a  solid  in the waste solid.  How-
ever, all of the calcium  is  utilized  in PSN4B (no Ca(OH)2 in
the waste solids)  so  that the scrubbing liquor becomes saturated
with respect to magnesium hydroxide (pH = 8.4).   The preliminary
conclusion here is that a calcium-based scrubbing liquor is
capable of more scrubbing than  the  calcium plus  magnesium-
based liquor.

          If the amount of limestone  available for reaction is
increased (refer to Case  PSN21  for  a  100% CaO system and Case
PSN22 for an 80% CaO-20%  MgO system),  the amount of slurry
recycle required to remove 90%  of the  SQS in a 2000 ppm flue
gas is reduced.  In fact, for both  cases PSN21 and PSN22, only
51.0% slurry recycle  is required.  The reduction in amount of
slurry recycle  for both cases is  expected since with more
limestone available the solid contains more calcium and magne-
sium hydroxide.  (The simulation  cases are run with a specified
fraction of Ca(OH)2 and Mg(OH)2 dissolving in the scrubber.)
The fact that cases PSN21 and PSN22 require the same amount of
slurry .recycle  can be explained again by examining the compo-
sition of the scrubbing liquor.  It turns out that enough CaO
is added in PSN22  so  that the scrubbing liquor is saturated
with respect to Ca(OH)2.  Thus  the  liquid phase of Cases PSN2B,
PSN21, and PSN22 are  almost  the same.   The molality of total
magnesium species  in  the  liquid phase for Case PSN22 is only
9.1xlO~7 g-moles per  kilogram of  liquid water.  Almost all of
the magnesium (99.99+%)  leaves  the  system as Mg(OH)2 solid in
Case PSN22, whereas only  82.7%  leaves  as Mg(OH)2 solid in
Case PSN4B.

                              139

-------
Radian Corporation
8MO SHOAL CREEK BLVD. • P. O. BOX 9948 • AUSTIN. TEXAS 78757 • TELEPHONE 512 - 454-9535
          For  the  calcium plus  magnesium-type scrubber liquid
(PSN4B), the ionic strength of  process liquor is 0.70 g-moles
per kilogram H20 as compared to 0.48 for calcium-type liquids
(PSN2B, PSN21, PSN22).   Of course,  this would be expected
since magnesium species  are more soluble.  Along the same line,
4.54% of the sulfate leaves the system in the liquid phase for
Case PSN4B.  Less  than 170 of the sulfate leaves in the liquid
phase for the  calcium-type liquids.   The amount of sulfate
leaving the system as a  liquid  would continue to increase as
more magnesium and less  calcium is  made available to the
process.

4.4       Scrubber Feed  Rate

          The  effect of  scrubber feed rate is shown in Table V.
Here the cases shown in  the first two data colums are the
same as those  shown in Table IV, i.e., PSN2B and PSN4B.  Cases
PSN11 and PSN12 were made based upon high scrubber feed rates.
In the system  using a pure CaC03 limestone (PSN11), 20 gal/1000
ACF of scrubber feed containing no  slurry was sufficient to
remove 90% of  the  S02 from a 2000 ppm flue gas.  In a system
using an 80 mole%  CaO -  20 mole70 MgO limestone (PSN12) , 87.2%
slurry recycle was required at  a scrubber feed rate of 21 gal/
1000 ACF.  This can be explained in terms of the amount of
basic species  dissolved  in the  calcium-based and calcium plus
magnesium-based scrubbing solutions.

          Comparison of  Cases PSN2B and PSN11 indicates that
the liquid-phase of the  scrubber feed has made a significant
contribution toward absorbing S02 in the scrubber.  On the
other hand, comparison of PSN4B and PSN12 indicates that the
liquid-phase for Ca plus  Mg-based scrubbing solutions is not
as potent.  In the Ca plus Mg-based cases, dissolution of the
                              740

-------
                               TABLE  V
       QUANTITY
LIMESTONE COMPOSITION,
  °i \v»n
  /o ivie>u

SCRUBBER L/G, GAL/
  1000 ACF

SLURRY RECYCLE/ SCRUBBER
  FEED RATIO

Wt.% SOLIDS IN SF

Wt.% SOLIDS IN SR

mCT? (BASIC SPECIES  IN
       LIQUID)
"
 FBS
FRACTION  C02  ABSORBED
iF SCRUBBER

PSN93
0
14
0.710
2.1
2.9
2.8xlO"2
0.96
0
FEED RATE
CASE
PSN'fB
20
14
0.877
6.0
6.9
2.0xlO"3
0
1.14

NUMBER
FSN11
0
'20
0.00
0.01
0.0
2.7xlO"2
0.57
0


PSN12
20
21
0.872
3.8
4.3
2.6xlO~3
0
1.10
5.2x10
                                    -4
2.4x10
                 -4
1.4x10
                 -3
                                                               3.5x10
                 -4

-------
Radian Corporation
8500 SHOAL CREEK BLVD. • P. O. BOX «<8 •  AUSTIN, TEXAS 79757 • TELEPHONE 512 • 454-9535
SF solids in the scrubber account for  a major  portion  of  the
scrubbing.  These remarks are supported by (see  Table  V):
(1) the amount of solids in  the scrubber  feed  stream  (weight
per unit time) is about the  same for Cases PSN4B and PSN12  and
(2) the concentration of basic species in solution  is  approxi-
mately 10 to 15 times as much for the  Ca-based solutions  as for
the Ca plus Mg-based solutions (refer  to  the molalities of  basic
species in the SF liquid shown in Table V) .

4.5        LTEA Versus LIWS

           A comparison of the limestone  tail-end addition
(LTEA) process with the limestone injection wet  scrubbing
(LIWS) process is shown in Table VI„   Cases PSN2B and  PSN11
are simulations for the LIWS process using a pure CaO  limestone
feed.  Cases PSN32 and PSN34 are the LTEA simulations  correspond-
ing to Cases PSN2B and PSN11, respectively.  In  Cases  PSN2B and
PSN32, the scrubber L/G was  set to 14  gal/1000 ACF  and the
fraction of solids dissolving in the scrubber  was set  to  0.35.
In Cases PSN11 and PSN34, the slurry recycle was set  to zero.
It was found that the desired S02 removal was  obtained by feed-
ing 20 gal of SF/1000 ACF of FG for Case  PSN11 and  72.7 gal/1000
ACF for Case PSN34.  Case PSN33 is a LTEA simulation corresponding
to Case PSN32 except that only 0.10 of the basic solids were  allowed
to dissolve in the scrubber.  This case was run  to  take into  account
the fact that limestone is generally not  as reactive  as lime.

           Comparison of Cases PSN2B and  PSN32 indicates  that
approximately 67» more slurry recycle was  required to  converge
the LTEA case (SR/SF = 0.752) than was required  for the LIWS
case  (SR/SF = 0.710),  This  would indicate that  if  a  relatively
high  fraction of solids dissolve in the scrubber (0.35),  not
much more slurry recycle would be required for the  LTEA process
                              142

-------
                                              TABLE VI
CO
COMPARISON OF

QUANTITY
PROCESS TYPE
SCRUBBER L/G (GAL/1000
ACF)
FRACTION SOLIDS
DISSOLVING IN S
SLURRY RECYCLE/ SCRUBBER
FEED RATIO
SF STREAM PROPERTIES
WT. % SOLIDS
mSFL (OH~)
mSFL (ALL BASIC SPECIES)
FB STREAM PROPERTIES
pH
IONIC STRENGTH
mpBL (TOTAL CaO)
mFBL (TOTAL C02>
mFBL (TOTAL S02)
mFBL 
-------
Radian Corporation

-------
Radian Corporation
8500 SHOAL CREEK BLVD. • P. O. BOX 9948 • AUSTIN. TEXAS 78757 » TELEPHONE 512 - 454-9S3S
           Further comparison  of  the  LTEA cases  with the LIWS
cases shows that (1) the  liquor pH's  are  less  for the LTEA
process, (2) the ionic strengths  are  approximately the same,
(3) the molalities of total CaO,  total S0a ,  and  total S03 in
the effluent liquor (FB liquor) are approximately the same,
and (4) the molality of total  C0a  in  the  effluent liquor is
much larger for the LTEA  process.  As would  be expected, the
LTEA process desorbs C0a  in the scrubber  as  contrasted to the
absorption of C08 in the  LIWS  process.

5.0        SUMMARY

           Radian has developed a mathematical description of
lime/limestone wet scrubbing processes.   Vapor-liquid-solid
equilibrium for the CaO-MgO-Na20-S(X, -S03-C02 -N205 -HC1-ILO
system has been described on a thermodynamically sound basis.
Numerical results based upon equipment descriptions and process
assumptions were calculated.   The following  conclusions have
been drawn based upon these simulation results.

           (1)  Scrubbing solutions originating  from high
                calcium limestones are more  efficient than
                those originating from dolomites.

           (2)  Sulfite oxidation reduces the  S03 scrubbing
                efficiency of  the wet scrubbing  process„

           (3)  Ionic strength has an effect on  the concen-
                tration of basic  species  in  solution and thus
                the scrubbing  efficiency.

           (4)  LIWS process (lime) liquors  contain more
                basic species  in  solution than do LTEA
                             145

-------
                   8500 SHOAL CREEK BLVD. • P. O. BOX 9948 • AUSTIN. TEXAS 78757 • TELEPHONE 512-454.9535
                process (limestone) liquors.  Thus, in
                LTEA processes where the  less reactive
                raw limestone is the additive, either
                higher liquor rates or weight percent
                solid slurries must be utilized to
                obtain comparable S02 removals.

           The lime/limestone wet scrubbing model presented
here can easily be adapted to other scrubbing processes  contain-
ing the same chemical species.  Although  the model described  in
this paper is based upon certain equilibrium assumptions and  the
specification of fractional conversions for rate limiting steps,
this model provides an excellent basis for process design and
development studies, particularly in the  areas of (1) determining
the technical feasibility of process alternatives, (2) estimating
process flow rates and compositions, and  (3) correlating experi-
mental data.

           As rate data become available, the model can  be
revised to include the proper rate correlations.  Since  v-A-s
equilibrium has been described in terms of the activities of
individual ions or ion pairs, these rate  correlations can be
properly expressed in terms of the activities of the various
reacting species and the distance from thermodynamic equilibrium.
                      ACKNOWLEDGEMENT S
           The process model  development  and parameter  study
for the LIWS process were  sponsored  by the Office of Air
Programs, Environmental Protection Agency under Contract
CPA 70-45, Messrs. E. L. Plyler  and  J. W. Jones, Project Officers.
                                146

-------
Radian Corporation
(=500 SHOAL CREEK BLVD. • P. O. BOX 9948 •  AUSTIN. TEXAS 78757 •  TELEPHONE 512 - 454-9535
6.0       REFERENCES

BA-019    Bahco, "Bahco  S02  Scrubber for Collecting Sulfur
          Oxides and Dust  from Flue Gases",  Company Brochure;
          no date given.

BR-014    Brodsky, Yu  N. ,  V.  I. Lazarev, and V. A. Pinayev,
          "Removal of  Sulphur Oxides from Combustion Gases by
          Dry and Wet  Methods", Paper presented at the Seminar
          on the Desulphurization of Fuels and Combustion
          Gases, Geneva,  16-20 November 1970.

CH-003    Chen, B. H.  and  W.  J. M.  Douglas,  "Axial Mixing of
          Liquid in  a  Turbulent-Bed Contactor", Can. J.. Chem.
          Engr. 47,  113-118  (1969).

CH-029    Chemico, "Facts  about Air Pollution and Business".

CH-030    Chemico, "S0a  Removal and Recovery by the Central
          Process Concept".

CL-005    Clay, C. W., G,  G. Poe, and J. M.  Craig, "Wet Scrub-
          bing  of Sulphur Dioxide from Power Plant. Flue Gases",
          For presentation at the 63rd. Ann. Mtg. of OAP, St.
          Louis, Mo.,  June 14-18, 1970.

DE-036    Dennis, Carl S., "Potential Solutions to Utilities
          S0y Problems in the  '70's", Combustion Oct.  1970,
          12-21 (1970).

DO-001    Douglas, H.  P., et al., "The Turbulent Contact
          Absorber", Chem. Eng. Prog., 5£(12), 85-9  (1963).
                              147

-------
Radian Corporation
8500 SHOAL CREEK BLVD. • P. O. BOX 99-19 • AUSTIN, TEXAS 78757 • TELEPHONE 512 - 454-9535
GR-009    Gressingh, L. E.,  et  al.,  "The Development of New and/or
          Improved Aqueous Processes for the Removal of S02 from
          Flue Gases", Final Report, Vol.  I, Contract PH 86-68-77.
          Prepared by Aerojet-General,  Envirogenics Division and
          submitted to NAPCA, Oct.  1970.

HA-027    Hausberg, G., "The Removal of SOS  and Dust from Flue
          Gas, The Bischoff  Process", Paper  presented at the
          Lime/Lime stone Wet Scrubbing  Symposium,  Pensacola-.
          Florida, 16-20 March  1970.

LO-027    Lowell, P. S., et  al.,  "A Study of the Limestone
          Injection Wet Scrubbing Process",  Final  Report,
          Volume  I, GAP Contract  70-45, Radian Corporation,
          Austin, Texas, 1971.

MA-014    Martin, J. R., W.  C.  Taylor,  and A.  K. Plumley,
          "The CE Air Pollution Control System", Paper pre-
          sented at 1970 Industrial  Coal Conference, Univ.
          Kentucky, Lexington,  8-9 April 1970.

MA-038    Maurin, P. G. and  J.  Jonakin, "Removing  Sulfur
          Oxides from Stacks",  Chem. En&. ,. 77.(9) ,  173-80 (1970).

PL-002    Plumley, A. L., et al., "Removal of  S02  and Dust
          From Stack Gases",  A  progress report  on  the CE Air
          Pollution Control  System", Combustion, 40, 16-23
          (1968).

SH-031    Shah, I. S. and C.  P. Quigley, "Concrol  of Sulfur
          Oxides in Stack Gases.  Magnesium  Base SOS Recovery
          Process:  A Protective  Installation  at Boston Edison
          Company and Essex  Chemical Company",  Presented at 70th
          National AIChE Mtg.,  Atlantic City,  New  Jersey, 1971.

                               148

-------
                    3500 SHOAL CREEK BLVD. • P.O. BOX 9943 • AUSTIN. TEXAS 78757 • TELEPHONE 512-454-9535
SL-004    Slack, A. V,  and  R.  E.  Harrington,  "Removal  of
          Sulphur Dioxide from Power Plant Stack  Gas;  Status
          of Limestone  Processes", Presented  at 2nd.  Int'l.
          Clean Air Congress,  Washington, D.  C.,  Dec.  6-11,
          1970.

TE-001    Tennessee Valley Authority, "Sulfur Oxide Removal from
          Power Plant  Stack Gas:   Conceptual  Design and Cost
          Study, Use  of Limestone in Wet  Scrubbing Process",
          Prepared for  NAPCA by TVA under Contract No. TV-29233A,
          Knoxville,  Tenn., 1969.
                                149

-------
      Radian    Radian  Corporation
                 8500 SHOAL CREEK BLVD. • P. O. BOX 9948 • AUSTIN. TEXAS 78757 • TELEPHONE 512/454-9535
               PRECIPITATION  KINETICS OF CaS04- 2H20
                               By:
                         James L. Phillips
                           Presented at:

               SECOND INTERNATIONAL LIME/LIMESTONE
                     WET SCRUBBING SYMPOSIUM
                       8-12 November 1971
                     Sheraton-Charles Hotel
                     New Orleans, Louisiana
                           Sponsored by:
                  Environmental  Protection Agency
                       Office  of Air Programs
                   Division of Control  Systems
                                151
TOilCAL RESEARCH • SYSTEMS ANALYSIS  •  COMPUTER SCIENCE • CHEMICAL ENGINEERING

-------
       Corporation  ssoo SHOAL CREEK BLVD. • P. o. BOXWS • AU* UN, TEXAS ws? • TELEPHONE siz - 454-9535
1.0       INTRODUCTION

          Precipitation of solid products  is an  important
aspect of lime or limestone wet scrubbing  process  design.
Process performance and operability are directly related  co
several solid-liquid mass transfer rates.  Radian  Corporation,
under contract to the Office  of Air Programs (Contract  No.
68-02-0023) ; is conducting laboratory  studies  intended  Co
characterize precipitation kinetics of CaS04•  2H20 and
CaSOa• %HaO.  This paper presents results  of tests conducted
to date for the CaS04• 2H30 portion of the study.

2.0       PROCESS IMPLICATIONS OF PRECIPITATION  KINETICS

          The overall reaction for S03 removal in  limestone
wet scrubbing processes may be written

S0a(g) + CaC03(s) +f 02(g)   -  (l-x)CaS03(s)  +  xCaS04(s) + C0a(g)

Even though removal of S0a is accomplished by  absorption  in
alkaline liquid, ^he ultimate products must be solid  CaS03 or
CaS04 in order for waste disposal to be feasible.   Thus in
order to precipitate a solid  product,  supersaturated  conditions
must be intentionally maintained at some point in  the process.
A workable process design must solve the problem of promoting
precipitation in one process  vessel or waste disposal area
while preventing or controlling it in  others.

          In the development  of almost any industrial process,
data characterizing important fundamental  aspects  of  process
chemistry and kinetics may not be available.   A  decision  then
must be made regarding the best method of  developing  a  reliable
and economically feasible process design.  Process development
activities may be generally divided into bench scale, pilot
                              152

-------
 Radian Corporation
8500 SHOAL CREEK BLVD. • P. O. BOX W« • AUSTIN. TEXAS 78757 • TELEPHONE 512 - 454-7535
scale, and full scale tests.  A proper.blend of these activities
must be selected to result in a successful development over a
reasonable period of time.

          For example, full scale tests of a process are
ultimately necessary to demonstrate process operability.  On
the other hand, field conditions may make it virtually impossible
to gather meaningful design information.  Extrapolation of the
design to new operating conditions may then be difficult and
risky.  A pilot plant offers an improved opportunity to control
and measure process variables as compared to a full scale opera-
tion.  At the same time, care must be taken to insure that pilot
plant results are applicable to the full scale design.  A similar
situation exists with respect to laboratory or bench scale
studies.  Such work may be capable of producing accurate and
valuable data if means are available to apply this information
to a full scale design.

          In process development then, one must

          a.  select those aspects of  the process
              that are most suitably studied at the
              bench, pilot, and full scale levels;

          b.  develop a design scheme  capable of
              applying laboratory and  pilot scale
              data to a full scale installation.

          Precipitation kinetics are one area of  limestone
scrubbing process design  that is suitable for laboratory  study.
It is a vital aspect of the process.   Careful measurements
under controlled conditions are necessary to draw meaningful
conclusions.  Laboratory  determination of the form of  rate
constants and driving force terms can  be applied  to  full  scale
                              153

-------
 K3ut3n Corporation  ssoo SHOAL CREEK BLVD. • P.O. BOX ws •
AUSTIN. TEXAS 78757 • TELEPHONE 512 • «4-9535
process design.  Laboratory values of  the rate  constants may be
used for full scale design although values determined in a
pilot plant would probably be more meaningful.

3.0       THEORY

          Precipitation processes are  in general a  combination
of two rate.steps, nucleation and growth.  Nucleation is the
process by which a small solid particle first forms in a super-
saturated solution.  Once this nuclei  has formed, it will grow
at some rate.  In many supersaturated  systems growth will occur
in the absence of nucleation when crystal seeds are introduced
into the solution.  It is the intention of this study to charac-
terize the growth step rather than the nucleation step of
          »__
precipitation kinetics.

          The bulk volumetric crystal  growth rate can be
conveniently written as the product of a rate constant k, a
term M dependent in some way upon the  amount of seed present
(such as mass, area, or number) and £, a "driving force" term
which is some function of solution composition. Thus,

   Growth Rate  -  kM$     moles liter~1min~

If the growth rate were limited by a chemical reaction at the
crystal surface, k should be a relatively strong function of
temperature only.  If the rate of diffusion  of  ions to the
crystal surface is important, k may also be  a function of agi-
tation, particle size, and solution composition.

          In previous investigations of precipitation kinetics
the driving force function $ was commonly expressed in terms
of the difference between the actual concentration  of a  salt in
                               154

-------
 Radian Corporation
8500 SHOAL CREEK 8LVO. • P. O. BOX W« • AUSTIN. TEXAS 78757 • TELEPHONE 512 - «4-?535
a liquor and the equilibrium or  saturated  concentration.  This
often led to satisfactory correlation  of experimental  data.   In
the case of limestone scrubbing  systems, however,  a simple
driving force function such as this  cannot be  applied.

          First of all, the precipitating  ions are not present
in stoichiometric amounts.  Process  simulations have shown,
for example, that when a high  calcium  limestone is used,  the
ratio of Ca"*"*" to SO^ in solution may be more than  10:1.   Secondly,
limestone scrubbing liquors have a relatively  high ionic  strength
so that the activities of ions in solution differ  significantly
from their concentrations.  This behavior  makes calculation
of "supersaturation" a complicated function of solution composi-
tion.  Finally, several ion pairs including CaS04  and MgS04  are
present in* solution in significant amounts so  that perhaps  50%
of total CaO and S03 in a liquor is  not in the form of the  Ca++
and SO^ ions that are precipitating.

          These factors make a rather  sophisticated computation
of individual ion activities a necessity  for correlating  pre-
cipitation kinetics in limestone scrubbing liquors.  Under
previous contract to OAP, Radian Corporation has developed  a
computer routine for performing  the  necessary calculations.
This computational scheme in its most  general  form also enables
laboratory kinetics data to be applied directly to full scale
process design problems.

          The immediate objectives of  the  present study are:

          a.  determination of the product kM and its
              dependence on temperature,  agitation,
              and amount of seed;
                              155

-------
 Radian Corporation  HMO SHOAL CREEK BLVD. •
P.O. BOXW8 • AUSTIN, TEXAS 78757 • TELEPHONE 512 - 4S4-W35
          b.  determination of the driving  force
              function $ in terms of individual
              ion activities and equilibrium  solu-
              bility constants.

4.0       EXPERIMENTAL

          A. schematic of the apparatus used to study  CaS04'2HaO
precipitation rates is shown in Figure 4-1.   Supersaturated
solutions are produced by introducing two separate  feed  solu-
tions (e.g., CaCla and NaaS04) into a single  well-stirred  reactor.
Feed rates are carefully set and controlled automatically.   The
reactor is situated in a constant temperature bath  held  at the
desired experimental temperature.  A given  amount of  seed  crystals
          »,
is introduced into the reactor at the beginning of  each  run.  A
filter membrane is fixed across the reactor outlet  to contain
the seed within the reactor.  Since the volume of seed crystals
does not increase significantly during  the  course of  a run,  a
steady-state material balance may be written  for the  reactor for
either CaO or S03*.  Thus,


  (FS03X CS03)inlet ' (FS03X CS03)outlet  "  **te of  Precipitation

  Rate of  S03 enter-  Rate of S03 leav-
  ing Reactor         ing Reactor

The rate is calculated directly by measuring  the feed and
effluent flow rates and concentrations  (FSO  and cso  )*

          The steady state concentration  of reactant  may be
varied by  either varying mean residence time  or  feed
compositions.

 *The term "CaO" refers to total calcium in solution.  "S03M means
 total sulfur in the +6 oxidation state^  This will be distributed
 among various species such as HSO£, SO^, etc.  The symbol SO*
 means that particular ionic species.
                               156

-------
           EXPERIMENTAL SYSTEM FOR

          PRECIPITATION RATE  STUDY
                                r
CONSTANT PRESSURE DROP
             REGULATOR
	,  EFFLUENT
      TO ANALYSES
                                                             TWO-LITER

                                                             STIRRED

                                                             REACTOR
                                I   V^V.X^^	1                    I



                                    CONSTANT TEMPERATURE BATH
                   FIGURE 4-1

-------
 R3QI3n Corporation  &» SHOAL CREEK BLVD. • P.O. BOX ws • AUSTIN, TEXAS TBTS/ • TELEPHONE 512-^-9535
5.0       RESULTS AND DISCUSSION

          In Figure 5-1 precipitation rate  R in millimole  liter"1
min~l is plotted against the quotient of activity  product  aca++aSOSB
and solubility product K   for the case where aCa-H-
Figure 5?2 shows a similar plot for  the nonstoichiometric  case,
             20*
          The rate curves appear  to  incorporate  two  distinct
regions.  At relatively low  supersaturations ,  the  rate  is  linear
with calculated values for the activity product  of Ca^" and  SO^
ions.  At higher supersaturations the  growth rate  rises very
sharply above this linear correlation.  Comparative  photomicro-
graphs of seed and product crystals  show  that  this departure
from a linear driving force  is due to  nucleation.

          Figure 5-3 shows a photograph of  CaS04 •  2R,0  seed
crystal used for these experiments.  This seed was prepared
from a mixture of CaCla and  NaaS04 solutions and screened  to
+ 325 mesh.  Figures 5-4a and 5-4b show product  crystals from
two of the stoicHiometric precipitation experiments. Figure
5-4a corresponds to precipitation in the  low supersaturation
region and 5-4b the high region.   In both cases,  large  regular
crystals have grown from the seed.  Some  smaller nuclei are
apparent in 5-4b however.

          Figures 5-5a and 5-5b  show products  from two
nonstoichiometric runs at low and high supersaturation  res-
pectively.  In both cases, a number  of smaller crystals has
formed.  Again, however, the high supersaturation case  shows
a much greater amount of nucleation  than  the lower case.
Clearly, two factors influence the amount of nucleation
markedly.  As the supersaturation increases beyond a certain
                              158

-------
N-
(0»
•-•
23
 '3
I I
t z
s=
.'' •
                     FIGURE 5-1

            PRECIPITATION OF  CaS04•2HaO

                     At 45°C
            Stoichiometric Solution
                 0.5% Seed Slurry
                             Supersaturatlon -
                            ! mnn: n^mrTrrnTrrrrrrrrrrrrnTr
                                                                       1.6
                                           159

-------
     FIGURE 5-2

PRECIPITATION OF CaS04-2HaO
       at 45°C
Nonsnoichiometric Solution
     0.5% Seed Slurry

-------
             0-50
500 Microns
                      FIGURE 5-3



SEED CRYSTALS FOR CaS04 •  2HaO PRECIPITATION EXPERIMENTS
                         161

-------
FIGURE 5-4a - PRODUCT CRYSTALS FROM STOICHIOMETRIC RUN AT
              LOW SUPERSATURATION
 FIGURE 5-4b  - PRODUCT  CRYSTALS FROM STOICHIOMETRIC RUN AT
              HIGH  SUPERSATURATION
                            162

-------
FIGURE 5-5a - PRODUCT CRYSTALS FROM NONSTOICHIOMETRIC RUN
              AT LOW SUPERSATURATION
            . £v?V  &., ;-^>fc>JT>v. >>,w^V/i-*
            •*§$ ' V 
-------
 Radian Corporation
8500 SHOAL CREEK BLVD. • P. O. BOX W8 •  AUSTIN. TEXAS 78757 • TELEPHONE 512 - 4S4-V53S
level, nucleation increases rapidly.  Also, as the ratio  of  Ca4"4"
to SO^ ions in solution increases, nucleation increases.   For  a
given ratio of Ca^/SO^ the level of super saturation  at which
nucleation becomes important can be expressed quantitatively as
a ratio of calculated ion activity product to the solubility
product at a particular process temperature.

          This behavior could have significant impact on
limestone scrubbing process design.  As mentioned before,  super-
saturated conditions must exist in some portion of limestone
scrubbing processes in order for solid waste precipitation to
be accomplished.  Furthermore, in a closed-loop system (desir-
able from the standpoint of avoiding waste water treatment)  all
of the process streams except perhaps  the vessel where make-up
water is introduced, will be at or near saturation.   This means
that an operable closed loop process will require that the
flow sheet be designed so that all process streams are main-
tained in or below the region of supersaturation where nucleation
is negligible.  The ability to design  such a process  requires:
                                     •
          a.  chemical and thermodynamic  information
              characterizing equilibrium  conditions
              in limestone scrubbing processes;

          b.  quantitative data describing precipitation
              rates as a function of process stream
              composition and temperature;

          c.  a design oriented computational  scheme
              incorporating these precipitation  rate
              correlations.
                             164

-------
Radian Corporation
8500 SHOAL CREEK BLVD. • P. O. BOX 9748 •  *"""* TEXAS 78757 • TELEPHONE 512 • 454-7535
          To illustrate the importance  of  an  accurate formula-
tion of the process chemistry and  thermodynamics,  the
precipitation data shown in Figures  5-1 and 5-2  have been
replotted in Figure 5-6.  Here,  the  precipitation  rates have
been correlated usirig physically measurable quantities, total
CaO and total S03 in solution.   Even though the  stoichiometric
and nonstoichiometric cases are  self consistent  on this plot,
two completely different rate curves result when the ratio of
CaO to S03 changes.  Obviously these results  would be quite
misleading if used for process design calculations.

          Recall, on  the other hand, that  the data correlated
using calculated ion  activities  show similar  behavior for
precipitation rates even over this great range of solution
compositions.  Thus,  one could confidently extrapolate these
data to a variety of  process design conditions.
 6.0        CONCLUSIONS

           The following can be concluded from work completed
 to date.
           •   For the CaS04 •  2HaO system, a region
              of supersaturation exists where nuclea-
              tion is negligible.

              Within this region, the precipitation
              rates of CaS04 •  2HaO are proportional
                                           -H"       =
              to the activity product of Ca   and S04,

              Properly correlated precipitation  rate
              data can be extrapolated over a wide
              range of liquor compositions.
                               165

-------
1.0
                    FIGURE 5-6
             PRECIPITATION RATE VS.
             CONCENTRATION PRODUCT
             Stoichiometric
                 Case
      N on-Stoichiometric
             Case
                 r~-  Concentration Product -
      .4    .5    .6    .7    .8    .9    1.0   1
                                    166
1  1.2.  1.3  1.4  1.5   1.6

-------
 LIMESTONE TYPES FOR FLUE GAS SCRUBBING
           Dennis C. Drehmel
    Environmental Protection Agency
        Office of Air Programs
       Control Systems Division
       Special Projects Section
              Prepared for
Lime/Limestone Wet Scrubbing Symposium
        New Orleans, Louisiana
          November 8-12, 1971
                   167

-------
            LIMESTONE TYPES FOR FLUE GAS SCRUBBINC
Dennis C. Drehmel
Special Projects Section
Division of Control Systems

Abstract

Limestones vary dramatically in their reactivity as bases.  Opti-
mization of limestone scrubbing processes will involve selection
of a limestone or carbonate rock type.  This work compares carbonate
rock types as to their dissolution rates in acid media and as to
their sulfur oxide removal efficiencies in a batch scrubber.

It was found that dissolution rates of limestones could vary by a
factor of ten while dissolution rates of calcined limestone and
scrubbing efficiencies of calcined and uncalcined limestones could
vary by a factor of five.  Whether calcined or uncalcined, the best
of the carbonate rock types tested for batch scrubbing efficiency
were marl and chalk.  The worst were marble and magnesite.  The
dissolution rates of limestones were markedly greater at lower pH
but were an order of magnitude lower than that of calcined limestone.

For calcined limestone, the change in scrubbing efficiency with cal-
cination temperature was a function of the change in surface area
or pore volume.  At higher calcination temperatures where surface
area was lost, reactivity was lost.  The variance In scrubbing
efficiency with different calcined limestone types was a function
of both the surface area and pore volume.
                                768

-------
                       Limestone Types for



                       Flue Gas Scrubbing
Introduction
In response to the demand for control of air pollution, a great many




scrubbing processes have been proposed and investigated*  .  Approximately




$200 million were spent last year on scrubber design and construction.




However, a process most likely acceptable to the utility industry would




not involve operation of chemical recovery equipment.  Lime and limestone




scrubbing processes not only meet this requirement but also promise




high S02 and particulate removals efficiencies^,3).  Since limestones




vary dramatically in their reactivity   , optimization of these processes




will involve selection of a limestone or limestone type.








The purpose of the work presented in this report was to delineate the




possible effect of limestone type on scrubbing efficiency.  To this end,




dissolution rates and batch scrubber tests were conducted using both calcined




and uncalcined carbonate rocks.  Twelve basic carbonate rock types were




selected for this study.  The first nine types correspond identically




to the nine types of Dr. R. D. Harvey   .  Types 10, 11, and 12 correspond




to Georgia marble (#1336), Michigan marl (#2129), and Kansas chalk (#2081),




respectively.  One additional limestone was investigated for effects of




calcination temperature on scrubbing performance.  This fine grained oolitic




stone (#2061) was calcined at several temperatures from 1700°F to 3200°F.




The physical properties and reactivity of these calcines have been




discussed in a previous study  .  Complete petrographic and chemical




analysis of all of the carbonate rocks used in this study has been given
                                 169

-------
by Harvey and is summarized in  Table 1.








Procedure




1.  Dissolution Rate of Limestones in




Two approaches were taken to determine the dissolution rate of limestones




in sulfurous acid.  The first procedure referred herein as acid washing




allowed 1.8N sulfurous acid to trickle through a thin bed of 5.0 grams of 16/20




mesh sample placed between filter paper in a buchner funnel at room




temperature.  The reacted acid was collected in 50 ml portions of the




original 200 ml of acid.  The calcium concentration of each portion was



determined wet chemically and then divided by the time required for passage




of that portion of acid through the sample.  The resulting number giving




the weight of calcium per time is divided by the weight of sample dissolved




per time.  After comparison of this value for each portion, these weight




dissolution rates were averaged and divided by the appropriate equivalent



weight.








The second procedure was called dissolution rate at constant pH.  A 1 g.




sized mesh sample is slurried in 200 ml of distilled water and titrated




with 0.21 N sulfurous acid.  The titrat^on rate is continuously adjusted




to bring the slurry pH down to the desired pH and then to hold it there.




Quantities of acid cqnsumed at minute intervals were recorded and the



equivalents of acid consumed per timeiwere calculated.  Since the buret




used to deliver the acid and the beaker/which held the slurry were water




jacketed, desired timperatures of interest could be selected.
                                 170

-------
2.  Dissolution Rate of Calcined Limestone in




Dissolution rates for calcines could not be determined in exactly the




same fashion as they were for carbonates because dissolution for calcines




was too rapid.  Consequently, a modified procedure of the Eckhard coarse




grain acid titration test was developed.  In this method 1 gram of sized




sample is slurried in 200 ml of distilled water and titrated at minute intervals




with sulfurous acid tp pH 7.  From these data a  ci^rve of cumulative acid




consumed versus time is drawn.  In order to have a numerical estimate




of the dissolution rate during the initial portion of the curve, the




predicted consumption of acid at one minute from the smooth curve was




used to compute the equivalents of acid consumed per time and volume.








3.  Flue Gas Scrubbing wifch Calcined and Uncalined Limestone




Both limestone and lime were tested for their effectiveness as flue




gas scrubbing additives using the same apparatus (see Figure 1).  The




apparatus consisted of a 1-liter temperature regulated batch scrubber




and a small side vessel (approximately 25O ml); total volume of the




system was 1500 ml.  The liquor was continuously circulated between




the scrubber and side vessel at 3 liters/minute.  Flue gas with 3000 ppm




S02 was bubbled through the liquid in the batch scrubber at 10 SCFH.




A Beckman 315 IR was used to monitor the effluent SO2 concentration and




pH was measured with a probe in the side vessel.  After an initial period




of bubbling ,flue gas through the scrubber liquid,  the effluent S02




concentration would stabilize at a concentration slightly below the S02




concentration in the flue gas.
                                 171

-------
At this time approximately one half gram of limestone or one quarter




cram of lime (weight necessary for 0.01 alkaline equivalents) was added




to the side vessel and the change in effluent SO2 concentration and pH




are recorded.  The reduction in SO2 would rapidly reach a peak and




gradually fall t0 zero.  The peak reduction in SO2 concentration due




to the addition of 0.01 alkaline equivalents of sized sample was reported.






Results




1.  Dissolution Studies




Dissolution rates of carbonates as determined by acid washing and by




titration at constant pH are given in  Table 2 for the ISGS types of




carbonate rocks.  These data were obtained for 16/20 mesh material at




70°F.  The ranking of dissolution rates by acid washing is 11>2>1




> 3, 4, 9>10>5, 6 or marl > calcite > marble> dolomite.  The effect




of particle size on these data is illustrated in Figure 2.  The ranking




of dissolution rates at a constant pH 4 is ll>8>3, 4>1, 2 > 10




> 5, 9>7 or marl> aragonite> calo.ite> marble > dolomite > magnesite.




The effects of pH level, solution temperature, and particle size are




shown in Figures 3, 4, and 5.  Dissolution doubled if the pH were one




unit lower or the temperature 40°F higher.  Particle size had little effect




for marl.









Dissolution rates of calcined carbonate rocks were estimated from an acid titration




prpcedure as described above and are given in Table 3.  These data were




obtained at 70°F for 1 gram of 42/65 mesh stone calcined at 2000°F.




The ranking for estimated dissolution rates of oxides is 1, 8>5, 11,
                                    172

-------
12 > 2, 3, 4> 6, 9, 10> 7 or aragonite and spar> chalk and marl >




calcite>dolomite;and marble> magnesite.  The effects of calcination




temperature and particle size are illustrated in Figures 6 and 7.  These




rates include the rate of hydration and the effect of hydration on particle




size.  Hence, they may reflect hydration and decrepitation rates more than




"dissolution".









2.  Flue Gas Scrubbing




Results for the peak reduction in SC^ concentration due to the addition




of 0.01 alkaline  equivalents of carbonate rock in the batch scrubber




are shown in Table 4.  Since the amount of alkali added is small,




the peak reduction in ppm of SC^ is only an index of reactivity and




not a predictor for absolute scrubber efficiency.  The ranking of carbonate




rock types at 120°F solution temperature using 150/170 mesh




material is 11>12>2>1, 9>others>7  or marl> chalk>calcite




> others>magnesite. Raising solution temperature from 90°F to 150°F




only slightly improved the performance of the marl while decreasing




particle size increased  it sharply (see discussion of results).









Addition of 0.01  alkaline equivalents of oxides  (2000°F calcines) gave




a clear  ranking in reactivity using a 150/170 sample (see Table 4).  At




110°F solution temperature, this ranking is 11>12>9>1, 4> others




or marl> chalk>clacitic dolomite>calcite>others.  Using 42/65




mesh sample, the  performance of almost all the types was poor; marble and non




reef dolomite gave no response at, all.  The effect of particle size was




strongly dependent on carbonate rock type.  For  some the reactivity




tripled  or quadrupled by reducing the particle size while types 2, 3,
                                 173

-------
and 8 (calcite spar, coarse grained limestone, and aragonite) showed no




significant improvement.








Discussion of Results




1.  Dissolution Rates




Analysis of dissolution rates by titration at constant pH was pursued




extensively since these data were precise.  In particular,.information




on the effect of temperature and pH were used to estimate the activation




energy and order of the reaction.  In Figure 8 is shown the dissolution




rate data for stone #2203 (ISGS type 3) and Michigan marl #2129 (ISGS




type(11) plotted against reciprocal absolute temperature.  From this it




was calculated that the activation energies for the reaction OC03




(solid)-f H2S03 (aq.)-»»solution at pH 6 are 5.14 k cal/gmole for Type 3




and 6.54 k cal/gmole for Type 11.  By plotting dissolution rate data at



9QOF against hydrogen ion concentration on log-log graph paper the order of




the dissolution reaction was estimated.  For type 3 the order of the reaction




was determined to be 0.34; for type 11, 0.70.








2.  Flue Gas Scrubbing




For 150/170 mesh material the most effective carbonate rocks were type 11,




Michigan marl and type 12, Kansas chalk both, as the carbonate and the




oxide.  After marl and chalk, the rankings of calcined and uncalcined




carbonate rocks are pot directly comparable, however, in general calcites




performed better than dolomites.
                                     174

-------
Far both carbonates and oxides, SC>2 sorption rises sharply with




decreasing particle size.  It is similar to the effect of /'particle size




on dissolution rate by acid washing as shown in Figure 9.  Dissolution




rates determined by acid titcation to maintain a constant pH were practically




independent of particle size fpr type 11, Michigan marl.  Another dis-




similarity\ between dissolution rate at constant pH and peak S02 reduction




in the batfh scrubber is the effect of temperature.  The dissolution rate




doubled from 70 to 110°F while S02 reduction was weakly dependent on tem-




perature over this range.








3.  Correlation with Physical Properties




All samples tested for dissolution rate or flue gas scrubbing effectiveness




werei also analyzed by a contractor* for surface area by the B.E.T. method




and pore volume using a mercury porosimeter. , Correlation coefficients




between dissolution rate or scrubbing effectiveness and physical properties




were computed.  Data for uncalcined carbonate rocks yielded only insigni-




ficant correlation coefficients.  For calcined stone several correlations




were observed.  With calcination temperature as the parameter, the cor-




relation coefficient for peak S02 reduction versus surface area is 0.99




and versus total pore volume is 0.98 (refer to Figure 10).  This is in




agreement with the previous study' ' that found strong correlations




between surface area or pore volume and several measures of reactivity.
  International Minerals and Chemicals
                                   175

-------
With carbonate rock type as the parameter, peak S02 reduction shows a




dependence on both surface area and pore volume.  As can be  seen from




Figure 11, variance,  in surface area predicts the variance in reactivity of the




high calcium oxide stones except for marl and chalk.  On the other hand,




pore volume appears to account for the high reactivity of marl and chalk while the




data for calcites remains somewhat ^scattered  .In Figure/ 12.








Conclusions



1.  The most effective carbonate rock types for flue gas scrubbing using




a batch reactor are Michigan marl and Kansas chalk irrespective of whether




th«» comparison is among calcined or uncalcined stones.








2.  The dissolution rates are a strong function of pH and most rapid at




low pH.  Dissolution doubles from 7CK>F to  110°F.








3.  There are marked difference* in the dissolution rates and scrubbing




efficiencies among Different types of stone (at least 5 times).








4.  For calcined carbonate rocks, the change  in scrubbing efficiency with




calcination temperature is a function of  the  change in  surface area or pore




volume.  At higher calcination temperatures where surface area is  lost,




reactivity is  lost.








5.  The variance in scrubbing efficiency  with different calcined  limestone




types  is a function ofj both the surface area  and pore volume.
                                   176

-------
                              References
1.  Aeorjet-General Special Report No. 5-4850-01-1,  Phase I,  Contract
    No. PH 86-68-77, June 21, 1968.

2.  Plumley, A. L. et al, "Removal of S02 and Dust from Stack Gases",
    Combustion 40 (1):16-23, July 1968.

3.  Pollock, W. A. et al, "FluerGas Scrubber", Mech. Engr..  89 (18):
    21-25, Aug. 1967.

4.  Borgwardt, R. H., Drehmel,,D. C., Kittleman, T.  A.  MayfieId,  D.  R.
    and Bowen, J. S., "Alkaline Additives for S02 Control",  Annual
    Report DCS/RLB-71-1, March 29, 1971.

5.  Harvey, R. D., "Petrographic and Mineralogical Characteristics  of
    Carbonate ^cks Related to Sulfur Dioxide Sorp^ion in Flue Gases",
    Final Report CPA 22-69-65, July 15, 1971.

6.  Drehmel, D. C., "Test to Evaluate Reactivity of Boiler.Calcined
    Limestone Used in Air Pollution Control", Bulletin of American
    Ceramic .Society, 50 (8): 666-670, Aug. 1971.
                                  177

-------
                    Table 1:  Chemical Analysis in Weight Percentage and Petrographic Characteristics
oo
Sample
Type
Number
1
2
3
4
5.
6

7
8
9
10
11

12

Si02
Nil
Nil
Nil
1.53
0.03
11.8

0.47
0.19
5.88
0.85
3.63

2.48

A1203
Nil
0.01
Nil
0.01
0.02
1.77

0.08
0.27
0.69
0.20
0.95

1.09

Fe203
Nil
0.19
0.20
0.31
0.34
0.13

Nil
Nil
2.82
0.15
Nil

0.15

MgO
Nil
Nil
1.86
0.00
21.40
17.4

44.2
Nil
15.33
1.4
Nil

1.41


CaO
55.3
55.5
53
54
30
26

2
55
30
53
46

51
.4
.8
.30
.5

.93
.2
.82
.7
.6

.4


C02
43.95
43.35
43
43
47
40

50
42
40
43
37

39
.75
.35
.30
.27

.96
.10
.68
.4
.18

.44

*
Mineralogy
Calcite
Calcite
Calcite(dolomite)
Calcite( quartz)
Dolomite
Dolomite(quartz,
clay)
Magnesite
Aragonite
Dolomite(calcite,
quartz)
Calcite(tremolite)
Calcite (quartz,
clay)
Calcite

Petrographic

Type
Iceland spar rhombs
Spar

Coarse limestone
Fine limestone
Reef dolomite


Nonreef (impure) dolomite

Fine magnesite
Oolitic sand



Calcitic dolomite
Marble
Bog marl

Chalk




          Minor mineral constituents given in parenthesis

-------
                              Table 2
                   Dissolution Rates* of Carbonate Rocks

                          Using 16/20 Mesh Sample
Type
1
2
3
4
5
6
7
8
9
10
11
Description
Iceland Spar
Calcite Spar
Coarse grained
limestone
Fine grained
limestone
Reef Dolomite
Non reef dolomite
Magnesite
Aragonite
Calcitis dolomite
Marble
Marl
Acid Washing
1.54
3,29
1.08
1.11
0.45
0.52


1.29
0.73
4.61
Titttation
PH»4
0.12
0.10
0.27
0.27
0.04
0.14
0.00
0.84
0.04
0.08
3-0
at Constant pi
pH-6
0.00
0.00
0.15
0.07
0.00
0.00
0.00
0.26
0.00
0.04
0.28
*in milli-equivalents per min per gram into volume of 200 milliliters
                                 179

-------
                             Table  3
                      Estimated Dissolution   Rates
                     of  Calcined  Carbonate Rocks  *
 Type                        Description                Dissolution Rate
                                                           me^/min/g

 1                            Iceland  Spar                     20

 2                           Calcite  Spar                      9

 3                           Coarse grained limestone          10

 4                           Fine grained limestone           12

 5                            Reef Dolomite                    13

 6                            Non reef dolomite                 6

 7                            Magnesite                         4

 8                            Aragonite                        25

 9                            Calcitis Dolomite                 6

10                            Marble                            7

11                            Marl                             13

12                           Chalk                            15


 *Calcined at 2000°F,  42/65 mesh
                                   180

-------
                              Table 4
                Sulfur Dioxide Reduction* After Scrubbing,
                          Using 150/170 Mesh Sample

 Type           Description               Uncalcined            Calcined
                                                                at 2000°F

 1               Iceland spar                  450                  450

 2               Calcite/ spar                  540                  250

 3               Coarse^ grained limestone      210                  300

 4               Fine grained limestone        225                  400

 5               Reef dolomite                 250                  —

 6               Non reef dolomite             210                  300

 7               Magnesite                       0                  —

 8               Aragonite                     240                  350

 9               Calcitis dolomite             490                  750

10               Marble                        200                  250

11               Marl                          950                 1150

12               Chalk                         700                  950


 *sulfur dioxide concentration in parts per million in flue gas
                                  181

-------
                              Table of Figures
 1.  Batch Scrubbing Apparatus
 2.  Effect of Particle Size on Acid Washing Dissolution Rate
 3.  Dissolution Rate VS pH
 4.  Dissolution Rate VS Temperature
 5.  Dissolution Rate VS Particle Size
 6.  Effect of Calcination Temperature on Dissolution
 7.  Effect of Particle Size
 8.  Arehenius Plot of Dissolution Data
 9.  Comparison of the Effect of Particle Size on Dissolution and on
     S(>2 Reduction using Uncalcined Marl
10.  Correlation of S02 Reduction to Surface Area When Calcination Temper-
     ature is Varied
11.  Correlation of S02 Reduction to Surface Area for Different Calcined
     Limestone
12.  Correlation of SO- Reduction to Pore Volume for Different Calcined
     Limestones
                                    182

-------
                 	LIQUID
                 ROTOMETER
             CONDENSER
              OUTLET	
               VALVE
                  -0
               TO RECORDER
00
u>
         TEMPERATURE
         CONTROLLER
THERMOCOUPLE-
INLET GAS TUBE
LIQUID INLET TUBE-
LIQUID  OUTLET  TUBE-

REACTION FLASK-
                                                  -OUTLET  GAS
                                                   ROTOMETER
                                                                   r
                                                                 JNLET  GAS
                                                                 ROTOMETER
                                           MANO-
                                            STAT
                                                           INFRARED
                                                            ANALYZER
                                                            TO EXHAUST
                                                        (ANALYZED GAS) i
                                                                       \*r
                                                                TO RECORDER
                                      TO
                                   EXHAUST
                                                      	pH  PROBE
  REACTOR
—PRESSURE
  MANOMETER

(BLUE OIL TYPE)
  INLET  GAS
—PRESSURE
  MANOMETER
                                                                                                         (MERCURY TYPE)
                                                                                                          PRESSURE

                                                                                                          RECORDER
                                                                                                       RECORDER-
                                                                                                   IR   INPUT-

                                                                               T0                  QH  INPUT-
                                                                               RECORDER            P
                                                                                          TEMPERATURE  INPUT-
                 LIQUID
                 CIRCULATION
                 PUMP
                                                               ADDITrON
                                                               FUNNEL
                                                              SYSTEM DRAIN
                                                                                                         -C
                                                                                                         -c
                                                                                                         -c

-------
 0)
                                      150/170

                                        f
                                                   42/65
                                                                              16/20

                                                                                I
o
Tt
4J

I— I
o
CO
CO
oo
CO
cO


•O
C
o
o
0)
CN
C

E

I
x

6

LJ
h-
    
    5
T-«   ""^V
a   I
    Q
    O
4.0
         3.0-
                                                        MARL  (T-ll)
          2.0-
          1.0-
                                                     SPAR (T-l)
                      40
                       —T~
                       70
                                    100
200
500
1000
                          PARTICLE DIAMETER, microns

-------
     5.0
     3.O
      2.O
                                                  MICHIGAN  MARL   T-|J
UJ
o

I-
:D
_i
o
is*


o
       LO
0.7-
0.5-
COLUMblA


    LIMESTONE



         T-3
      0.3
      0.2
                      456



                                       PH
                                    185

-------
                      4.  Dissolution Rate VS Temperature
C

I
LU
H

o:
 
 O
2.2



2.0-



 1.8



 1.6-



 1.4-



 1.2^



 I.OJ



 OB-



 0.<



 O.4-



 0.2-
                               MICHIGAN MARL
                                  (T-IO
                                      COLUMBIA  LIMESTONE


                                            (T-3)
                          TEMPERATURE,   °F

                                    186

-------
-I
2
Hi
-I
O
2.2-



ZO-



 1.8-



 1.6-



 1.4-



 1.2
0.8



 0.6H



 0.4



 O.2
                             MICHIGAN  MARL  T-ll,  pH =
COLUMBIA LIMESTONE
         T- 3
                150/170              42/65                16/20


                      PARTICLE  SIZE, TYLER MESH
                                 187

-------
o
•H
4J
3
t-4
O
to
CO
 o

 0)
 p

 JJ
 Oj
 p
 0)
 o.

 4)
 H


 O
 1-1
 AJ

 c
 -H
 O
 f—I
 aJ
 4J
 o
 01
 «M
 MJ
 w
CALCINE  TEMPERATURE

 X   2000»F

 0   23CX)* F


     320 0°F
                                                          (170/270 MESH)
                                                               8
                                         TIME ,  MIN.

                                              188

-------
                                 CALCINATION TEMP
                                      2OOO °F
                                 PARTICLE  SIZE
                                     16/20   A
                                    42/65   0
                                     150/170  X
4-
                       8          12
                  TIME .  MINUTES
16
2O
                          189

-------
         4-0-
iJ
ol
Q

C
O
O
n
o
i-4
CM

oo
3


0)
JS
14
00
     (VI

     6
     £   2.O-
     o
     o
UJ
I-

<£


o



o


5
          I.O
          O.7
          O.5
          O.3
                    \
                                      x    MARL  T-ll
                                  T-3
          .O03OO
                      10       20       3O      4O     00350

                      RECIPROCAL  TEMPERATURE,  I/°K
                                 190

-------
                         Comparison o£ the ne+mc'T oC Paftlcle Size

                         SO2 Reduction using Uncalclned Marl
                                                             on Dissolution and on
    1000-
 5.0
Q.
Q.
o

o

o
UJ
a:



v>
UJ
CL
    8OO-
     600-
     400-
     200.
•4.0
-3,0
•2.0
•1.0
                                                                                                    E

                                                                                                    o»
o

x
u>

?

g
o
                                                                                                    00
                                                                                                    o
                                                                                                    H-
                                                                                                    O
                                                                                                    (O
                          -j	1	1	1        i     "
                           70   100        200       400     700   1000


                               PARTICLE  DIAMETER,  microns

-------
         1000-
o
•H
U
1-4


V)

o
C
o
•H
.u
U

TJ
 CSltrf
O id
CO >

U-i og
 O i-l

 C 
-------
c
0
JJ
g>
01

•H
J

•o
91

•H
u
H
id
o

u-
c
11
0

x
o
IH
li
3
01

0
u


C
0

-------
                 12.  Correlation of jSO« .Reduction to Pore Volume for Different
                     Calcined Limestones
        I40O-
 I
O
I-
u
        I2OO-
        1000-
        800-
         600'
                               0
 CM
O
UJ
Q.
         4OO
         2OO--
                               A  aragonite

                               C  calcite

                               D  dolomite

                               K  chalk

                               M  marl
0.6
           0.2
0.4
0.8
                                                       IX)
T
 1.2
-t
 1.4
                            PORE VOLUME,  cc/g

                                    194

-------
        REMOVAL OF SULFUR DIOXIDE FROM STAC'. GASES

           BY SCRUBBING WITH LIMESTONE SLURRY:

               SMALL-SCALE STUDIES AT TVA
                            By

       J. M.  Potts, A. V.  Slack, and J.  D.  Hatfield
             Division of Chemical Development
                Tennessee Valley Authority
                  Muscle Shoals, Alabama
               Prepared for Presentation at
Second International Lime/Limestone Wet Scrubbing Symposium
     Sponsored by the Environmental Protection Agency
                  New Orleans, Louisiana
                    November 8-12,  1971

                           195

-------
               REMOVAL OF SULFUR DIOXIDE FROM STACK GASES

                  BY SCRUBBING WITH LIMESTONE SLURRY;

                      SMALL-SCALE STUDIES AT TVA

                                   By

              J.  M.  Potts,  A.  V.  Slack,  and J.  D.  Hatfield
                    Division of Chemical Development
                       Tennessee Valley  Authority
                         Muscle Shoals,  Alabama
                                ABSTRACT


          Results from tests with a spray scrubber (two-stage,  counter-
current, k-in.  diameter, k-Q ft3/min gas flow) are presented.   The
following parameters were studied.

    'Surge tank volume was important; SOo removal increased with tank
     volume up to about 900 gal/(lb S02)(min).  Apparently the  effect
     was due to providing retention time for completing reactions be-
     gun in the scrubber.

    'increase in inlet gas temperature reduced absorption significantly,
     presumably because of increased S02 vapor pressure over the liquor
     leaving the scrubber.

    'High liquor;gas ratio was required for good absorption. The
     ratios used, on the order of 80 gal/Mcf of gas,  were probably
     higher—because of relative inefficiency of the spray scrubber—
     than needed in practice.   The beneficial effect of high liquor
     rate indicates the importance of high "effective stoichiometry"
     (moles CaC03 to which a mole of S02 is exposed during passage
     through the scrubber).

    *Use of excess limestone (beyond stoichiometric)  was effective;
     however, under the scrubber conditions used the stoichiometric
     amount was adequate (about 90$ removal).  Under less favorable
     conditions,  excess limestone might have a major beneficial effect
     by increasing the steady-state effective stoichiometry in  the
     scrubber recirculating loop.

    'increase in inlet S0g concentration gave higher exit concentration
     and slightly lower percentage removal.

    'Decrease in limestone particle size improved absorption but not
     to a major degree.
                                    196

-------
    *Limej>tone type was not an important variable.   Dolomitic
     limestone was somewhat less effective,  but performance of
     high-calcium limestones was generally similar when particle
     sizes were comparable.

    'increase in soj.ids content of the slurry improved absorption
     when other conditions were such as to give relatively low
     scrubbing efficiency.

    *Ionic strength of the scrubber solution was shown to be sig-
     nificant; increase in ionic strength by addition of magnesium
     sulfate improved absorption.

    *Use of an oxidation inhibitor was effective in reducing sulfite
     oxidation.

          Calcium sulfite, the preponderant solid species formed, crystallized
in very small platelets that settled slowly and tended to'blind limestone
surfaces when conditions were such as to limit limestone dissolution rate and
thus promote reduction in pH.  Oxidation to sulfate was an effective means
of avoiding these problems.

          In tests of alkali scrubbing followed by regeneration with lime or
limestone, alkali bisulfite was easily regenerated to sulfite with limestone.
Sulfate purging is still a problem, but if ammonia is used as the alkali the
sulfate can be separated and either sold or regenerated with lime.
                                   197

-------
               REMOVAL OF SULFUR DIOXIDE FROM STACK CASES

                  BY SCRUBBING WITH LIMESTONE SLURRY;

                      SMALL-SCALE STUDIES AT TVA

                                   By

               J. M. Potts, A. V. Slack, and J. D. Hatfield
                    Division of Chemical Development
                       Tennessee Valley Authority
                         Muscle Shoals, Alabama
          A test program on limestone scrubbing in a small (5-cfm) con-
tinuous scrubber has been under way at TVA for about 2 years.1  The
main purpose of the tests has been to support the pilot plant program
(2000-3000-cfm scrubber) being carried out at TVA1s Colbert Steam Plant,
which in turn is aimed at supplying information for design of a full-scale
550-mw installation at TVA1s Widows Creek Steam Plant.  Much of the work
with the small-scale test loop has been concerned with roughing out the
effect of variables as a guide for the pilot plant; for this the unit has
served the purpose well.  However, difficulty in controlling the continuous,
loop-type system closely, coupled with the small size of the unit, has made
it difficult to obtain precise data.   Nevertheless, the data obtained can
be used to throw some light on the chemical and kinetic factors involved.

          The scrubber loop assembly is shown in Figure 1.   Two important
changes in operating procedure have been made during the past year.  Easier
startup and better combustion control have been attained by burning straight
natural gas to provide stack gas instead of the previous admixture of gas
with oil containing carbon disulfide; sulfur dioxide is fed to the combustion
chamber from cylinders to provide about ^000 ppm in the stack gas.  Also,
commercial spray nozzles are now used instead of the previous impingement
type, making it possible to use lower and more conventional liquid:gas ratios.

          The stack gas is fed upward through the spray scrubber,  which con-
sists of two separate stages each with its own circulating slurry system.
Fresh limestone slurry fed to the surge tank for the upper slurry system
causes overflow to the lower surge tank.   Spent slurry overflows from this
tank and is filtered; the filtrate is used in making fresh limestone slurry
and the spent solids (wet) are discarded.


  Slack,  A.  V.   "Absorption of Sulfur Dioxide in Limestone Slurry:  Small-
  Scale Tests at TVA."  Paper presented at International Symposium on Lime/
  Limestone Scrubbing for S02 Control (sponsored by NAPCA), Pensacola, Florida,
  March 16-20,  1970.
                                 198

-------
STACK
 t
CONE NOZZLE
(2nd stage) ,
1(
LIQUOR ,
SCREEN
TC
CONE NOZZLE —k

>"
r


(1st stage) t
12"
LIQUOR 1
SCREEN
TC IR
1 1
6U raw \
OBBATOR' S
/



TRAP
^~i~~
\ * I
u
x^ 	
k j
v
V
f^

1 '
k J
v
TC
IH
GLASS TUBE
k" dia.



>

,1
SLURRY
MAKEUP
)
1 If




V^ >s

$ti


\
1'
SURGE —
, TANK
1
SURGE
TANK
^»

T
L JL
FIGURE 1
Small-Scale Scrubbing System Used in Study of Sulfur Dioxide Removal
from Stack Gas by Limestone Slurry Scrubbing
                        199

-------
          In operation, the unit is started up with the surge tanks full
of limestone slurry of the prescribed solids content.  The stoichiometry
is quite high in the beginning, of course, but the slurry is circulated
without limestone makeup until the S02 removal has declined to the ex-
pected level (from previous experience).  The limestone feed is then
started.  The S02 removal adjusts to the test conditions and usually levels
out quickly.  The scrubbing is then continued until continued uniform S02
removal indicates that steady-state operation has been attained.  By this
means, a datum point at steady state can be obtained in a day's operation.
Absorption Mechanism

          Recent work by various investigators has clarified the situation
somewhat in regard to the nature and extent of reactions occurring in the
scrubber, although there appears to be some disagreement still.  There is
some support for the following set of reactions:

                            S02(g) £ S02(aq)                           (l)

                   S02(aq) + H20 £ H2S03 £ H^ 4- HS03-                  (2)

                            HS03~ «» H+ + S03=                          (3)

                            CaC03(s) «»CaC03(aq)                       00

                            CaC03(aq) ji Ca^ 4- C03=                    (5)

                Ca4"4" + S03= + 0.5H20 «»CaS03-0.5H20(s)                 (6)

                            C03= + H+ ^ HC03~                          (7)

                H+ + HC03" £ H2C03(aq) «* C02(g) + H20                  (8)

          A diagram of the complicated set of relationships involved is
given in Figure 2.

          For such a set of reactions there are several resistances, both
diffusional and chemical, that can affect the rate of the overall reaction:
                           (s) + s°2(g) +

                      CaS03-0.5H20(s) + C02(g)                         (9)

These include (l) diffusion of S02 to and through the gas- liquid film, (2)
dissolution of S02 and CaC03, (3) hydrolysis of S02, (4) diffusion through
the liquid film and into the droplet interior, (5) reaction of H*~ and CaC03,
and (6) reaction of Ca++ and S03=.  Added to these are auxiliary rates such
as oxidation rate of dissolved sulfite species and crystallization rate of
CaS03-0.5H20 and CaS04-2H20 (which have a strong tendency to supersaturation. )


                                 200

-------
ro
o
                CaC03,   .    6
                     3(aq)  -,—
DISSOCIATION CONSTANTS AT 50°C
I. 0.78 x 10-2
2. O.42 x IO~I
3. 0.30 x ID'3
4. 0.84 x IO-~
5. 0.18 x IO-8
6. 0.47 x 10 -3
7. 0.46 x 10"'
8. 0.67 x 10-'°
9. 0.52 x I0~f
10. 0.36 x 10-'
II. 0.57 x I0-5
12.0.48 x IO'2
13. 0.35 x IQ'J
14. 0.22 x|0'4   	
         HENRYS LAW CONSTANT, 50°C
         15. 0.54
         16. 0.019
                                             Ca
                                             Ca
                                             Ca
                                           I	
                                                                           +H20
                                                                           ~H2O
                                                                                 CaS03-0.5H20(s)
+  co;
   OH
S04"
                                                                                    +H20
        1S^  CaOH*     ll "  Ca(OH)2(s)
           13
                                                       \
                       (aq)
    12  +H
•»-H20
"IT* CaS04- 2H20(S)

-H20
                HS04"
                                                       FIGURE 2
                                      Equilibria in the System CaO-SOP-SO.q-COg-H20

                                Added MgO gives  similar equilibria, with Mg"1^ replacing C
                                Added Na20 and N205 give the species  CaN03+,  NaOH, NaC03M
                                NaHC03, NaS04-,  and NaN03.  Added HC1 gives Cl"  ions only.

-------
          Work by Boll1 has indicated that both diffusion and chemical
reaction are rate controlling, and that rate of CaC03 dissolution and of
gas phase diffusion are likely to be the important parameters.   He also
concluded that the chemical reaction is first order with respect both to
the limestone surface area per unit of liquid volume and to the S02 partial
pressure in equilibrium with the liquid.


Distribution of Reactions in Scrubber Loop

          There is a further consideration in operation of a countercurrent
scrubber, in that pH can drop as the slurry flows down the scrubber and the
concentration of dissolved sulfite species can thereby be increased.  This
not only increases rate of CaC03 dissolution in the lower part of the
scrubber (because the equilibrium partial pressure of S02 over the liquid
is increased) but also may leave some "free S02" in solution that will react
with CaC03 outside the scrubber if given time before return of the slurry
to the scrubber.

          This is analogous to use of Mg(OH)2 for S02 absorption in the
paper pulp industry, where all the S02 absorbed during a scrubber pass can
be kept in solution (no crystallization in scrubber) because of the high
solubility of Mg(HS03)2 at low pH.  The main difference between this and
the calcium system is that in the latter the solubilities of the various
sulfite species are much lower, so that all the S02 absorbed in a scrubber
pass (under normal conditions of liquor rate and S02 inlet concentration)
cannot be held in solution (except perhaps by supersaturation).  Therefore
part of the absorption in a given pass must be accounted for by crystalli-
zation in the scrubber and the remainder by increase of sulfite concentration
in the solution.  The magnitude of the latter will depend on the pH level
at the scrubber exit, which in turn depends on inlet S02 partial pressure,
liquor flow rate, and type of scrubber (fully countercurrent vs partially
backmixed).

          Calculated data for the effect of pH on solubility of calcium,
sulfite, and sulfate in the system CaO-S02-S03-H20 are given in Figure ~$.
Sulfite solubility increases rapidly with decrease in pH.  Therefore the
net amount of dissolved sulfur species picked up in the scrubber depends
on the net change in pH, which depends on many factors.  A calculation was
made based on the following assumptions:

                 pH of inlet slurry               6.0
                 pH of slurry leaving scrubber    5-0
                 L/G, gal slurry per Mcf of gas   40
                 Inlet S02 concentration in gas   2500 ppm
                 Outlet S02 concentration in gas   250 ppm


1 Boll, R. H.  "A Mathematical Model of S02 Absorption by Limestone Slurry."
  Paper presented at International Symposium on Lime/Limestone Scrubbing for
  S02 Control (sponsored by NAPCA), Pensacola, Florida, March 16-20, 1970.

                                  202

-------
 30,000

 20,000



 10,000



  50OO


  3000


  2000
0.
o.
1000




500


300


200



 (00



 50


 30


 20
          T	T
                                0  Ca

                                D  SO2 (SULFITE)
                                A  S03 (SULFATE)

                              — SATURATED WITH BOTH CaS03+ CoS04
                              '	SATURATED WITH CoSOi
                                   (NO SULFATE PRESENT)
                                   SATURATED WITH CoS04
                                   (NO SULFITE PRESENT)
            2.5
                                       PH
       Solubility Relationships  in the System CaO-:SOg-S03-IfeO at 50°C (l22°F)


                                  203

-------
 Assuming saturation equilibrium  at both  scrubber  inlet  and outlet, about
 20$ of the  total S02 absorption  can be accounted  for by formation of
 dissolved species,  which  at  the  pH involved  is mainly bisulfite  (HS03~);
 the remaining 80$ can be  assumed either  to have crystallized in  the scrubber
 as  CaS03'0-5H20  or to be  held  in solution by supersaturation.

          The assumed decrease of one pH unit (from 6.0 to 5.0) may or
 may not be  close to the actual value.  The pH at  the scrubber outlet is
 a transient value,  because as  soon as the slurry  is no  longer exposed to
 the high driving force of the  S02 and C02 in the  inlet  gas, these gases
 can start escaping from the  liquid phase.  Moreover, dissolved sulfite
 species continue reacting with CaC03 and therefore raise the pH.  The value
 needed is the instantaneous  pH at the point  where incoming gas and effluent
 liquor part company,  which is  somewhat difficult  to measure.  In the ICI-
 Howden work in England,1  the pH  drop in  the  scrubber was reported to be
 0.4 pH unit (from 6-5 to  6.l)  for limestone  and 0.8 unit (6.9 to 6.l) for
 lime.   In the TVA pilot plant  tests with limestone slurry, the pH as measured
 runs about  6.0 in the feed to  the scrubber and about 5.7 in the receiving
 tank from the scrubber.   In  the  small-scale  TVA scrubber, tests were run
 in  which the pH  probes were  inserted into the scrubber  drain line to give
 a continuous reading,  thereby avoiding the possible increase in scrubber
 outlet pH caused by the delay between sampling and measuring when a grab
 sample is taken.   For an  L/G of  88 gal/Mcf,  97$ removal of 3200 ppm S02,
 1.0 limestone stoichiometry, and inlet gas temperature  of 150°F, the pH
 was 6.06  at the  scrubber  liquor  inlet and 5-22 at the outlet.  The drop of
 0.8 pH unit is close  to the  1.0  assumed  in the above calculation.

          The foregoing does not take into account the  possible effect of
 supersaturation  on distribution  of crystallization in the scrubber loop.
 To  the  extent that  calcium sulfite - sulfate  forms in the scrubber and does
 not crystallize  because of supersaturation,  the crystallization must take
 place  outside the  scrubber.   Calculated  data  on relationship between pH,
 degree  of calcium  sulfite supersaturation,  sulfate:sulfite ratio in solution,
 and S02 vapor pressure for the system CaO-S02-S03-H20 are given in Figure i*-.
 These  indicate that S02 vapor pressure does not increase much with super-
 saturation;   thus a  very high degree of supersaturation  would be required to
 interfere with absorption in the  pH range normally encountered.

          Figure 4  also indicates that under  the  conditions encountered in
 limestone scrubbing (pH of about 6 and S04=:total S ratio in solution of 0.60
 to  0.90), the solution may be highly supersaturated.   In the reports of the
 English work, it is indicated that at the circulation rates necessary to avoid
 scaling,  the  supersaturation developed (over  saturation concentration) was
 about 3*1-0 ppm each  for sulfite and sulfate.   At the pH  level involved (6.0 to
6.3),  this would be a high degree of sulfite  saturation.  In the TVA work,
 analyses of  the  scrubber effluent solution has indicated a considerable super-
 saturation of both calcium sulfite and sulfate, considerably higher than had
been expected.  The situation is being explored further.


1 Pearson, D. L., Nonhebel,  G., and Ulander, P. H. ».    J. Inst.  Fuel VTII
  (39), 119-156  (February 1935).
                                   204

-------
400
200 -
                                SATURATED
                               UNSATURATED
                       ALL SOLNS. SATURATED WITH CoS04-2H20

                           NUMBERS ON CURVES REPRESENT
                           PERCENT OF S PRESENT AS SULFATE
                       	NUMBERS ON CURVES REPRESENT
                           VAPOR PRESSURE OF S02
O.I
                          gIGURE
Supersaturation of CaS03° 0.
                                   in System
                          205

-------
          The effect of pH on liquid phase composition, assuming a given
degree of supersaturation for calcium sulfite and sulfate, is given in
the computer printouts shown in Table I.

          Surge Tank Retention Time;  In view of the foregoing, it appears
that the surge tank in the scrubber loop must serve the functions of (l)
completing reactions between limestone and dissolved acidic sulfite species
and (2) crystallizing calcium sulfite and calcium sulfate formed in the
scrubber but not yet precipitated.  The size of this tank and the resulting
retention time obviously are important design considerations.

          The effect of surge volume on S02 removal in the small-scale
scrubber is shown in Figure 5-  The data indicate a major benefit from
increase in surge tank volume up to about 500 gal/Mcfm (or about 1000 gal/
Ib S02*min); further increase had only a small effect.   At this surge volume:
gas ratio, the retention time in the surge tank at the 37 gal/Mcf rate was
about 7 ™in as compared with 3 min at the 100 gal rate.  Presumably the lower
reactive sulfite concentration at the higher liquor rate required less time
for sufficient limestone to dissolve and react.

          A retention time on the order of 2 to 3 min was specified in the
ICI-Howden work (about 1000 ppm S02 and 115 gal/Mcf) to control scaling.
This corresponds to a surge volume of about 1500 gal/lb S02«min.-

          Gas and Solution Temperature;  One of the more important variables
in the TVA work has been inlet gas temperature (Fig 6).  Presumably the
adverse effect of high temperature is due to stripping of dissolved sulfite
species not yet reacted with calcium, thus reducing the amount of reaction
that can be accomplished in the surge tank.

          The magnitude of the effect appears too large even for the maximum
amount of unreacted sulfite species that could be present in the solution
(on the basis of calculated values).  A possible explanation is that the high
temperature heats the liquid film on droplets and thus interferes with gas-
liquid transfer.   In other words,  the stripping may not involve dissolved
equilibrium species, which would require both heating the entire drop and
also a relatively low pH in the drop to develop any significant back pressure
of S02.  Instead, the dissolution of S02 in the liquid film and the hydrolysis
to HgS03 may be inhibited,  which would reduce not only the amount of dissolved
unreacted species but also the supply of sulfite for the Ca4"1"/SQ^ reaction
in that part of the scrubber affected.  For the equilibrium
                          S02(g) *S02-%0(aq)                          (10)
The effect of temperature on the Henry's law constant (h = aS02-H20/Pg0 )
is as follows:
                                206

-------
                             TABLE I

         Effect of pH on Liquid Phase Compound at Given Degrees

                of Sulflte and Sulfate Supersaturatlon
 i>U:
         b.O
SI*  4,0
                                S2»  1,56
                                                PC02» 0,0067
SPECIES     ACTIVITY     KOLAI.1TY   ACT.COCFf
   h20
      H
    OH
   H303
 CAS03
    CA
  CAOM
C,\r!C£>3
 CAC03
 C A304
  HS04
   SO 4

  r!CO.J
   COS
          0,1000f:-04
         o,
         0,liU9ti-01
         P,'.311C-04
    0,11906-04
    o!
                                      0.99917
          0.7741U-02
          0.1102E-08
          0.10S6E-05
          0.7100E-09
    0,110Utf-03
    Ci.ll22t-02
    0.1095001
    0.1409E-OB
0,793flE*PO
0.1012E+01
0.7926G-MIO
0,3947L: + 00
O.KH2E + 01
    0,7015^-09   0,1012H*01
         O.V193E-05  0,1158c-04
         0, J2P2C—03
         0,64536-05
    0.1252E-03
    0.8142E-0&
    0.1102E-09
0.3563E-0.0
rt,1000E*01
                                   0,39476*00
CALCjOfl
CAS.T3
C A
CACIH
C/HC03
CAC.M
CA30«
,*,8
63 , rt
0,9
0,0
0,0
32,4
Sl'LFITF. SULFAie
HZSOi
HSOX
SOX
CAS03


0.1
91,2
0,8
8,0


CAS04
HS04
S04



43.7
0 1 1
56,3



                                                              C02
                                                        CAIICfl.f    1,0
                                                         CAC03    fi.O
                                                         H2C03   P?,9
                                                          HC03    0,0
                                                           COS    0,0
f»- *tO Sl= 4,0 S2* 1,56 PC02»
SPECIES
H20
H
OH
HZS03
HS03
SOS
CAS03
CA
CAOH
CAKC03
CAC01
CAS04
KS04
S04
H2C03
HCOS
C03
ACTIVITY MOLALITY

0.10UOt-05. 0.118BE-05
0.547ttC-07 0.6eiflf--07
0.1570E-06 0.1554L-06
O.l??9fe-02 0.1531E-02
.0.5237E-04 n,1259t-03
0.1136F-02 O.J124E-02
0,6416F-i)2 O.l^OSE-Ol
n,9796C-OB 0.12I9F-0/
n,9015C"05 O.Jj?2K-04
0,*>8fl5E-n7 1,5824t-07
0.0753F-T2 0.9652I--OR

O.U CAS03 4P.4
n,u
37,5
ACT.COEFF
0,99939
0,8417E:*00
0 , 8034G>f»0
0,101()F»01
0,^031^*00
n , 4ii>9t* oo
o.ioion*ai
0 , 42"2£- + 09
0,flQ34!;«nO
0 , ft034B*00
0,1010F*Ol
n,ioioF*oi
o,«o34fc*no
0 , 3fllftr.»03
n , i oioii*oi
0,8031^*00
0,4159F:*00
STKJd'lT IOM, %
SULFAT6
CAS04 40,7
HS04 0,0
S04 59,3



> 0,0067


















C02
CAHCOJ 5.2
CACD3 0.0
H2C03 *>7.7
HC03 37,1
C03 0,3

*  Degree of eupersaturation:  for sulflte, h.O times saturation; for aulfate,
   1-56.
                           207

-------
          SURGE VOLUME: S02 RATIO, GAL /(LB S02)(MIN)
            374      748      1122      1496      1870
  80
  70
  60
  50
UJ
O
§
  40-
(A
  30
  20
   IO
             I
             1
                                   2244
 I
O LIQUID: GAS RATIO,  100 GAL/MCF
A LIQUID: GAS RATIO,  37 GAL/MGF

ALL TESTS WITH 2% SLURRY OF
COARSE (76% -200 MESH) SPRING
VALLEY LIMESTONE.  3000 PPM S02
IN GAS.  GAS TEMP:  I25-I35°F
TWO-STAGE, CLOSED LOOP OPERATION
 I
I
1
I
    >       200      400      600      800      1000
              SURGE VOLUME: GAS RATIO,  GAL /MCFM

                           FIGURE  5

     Limestone Slurry Scrubbing of S02  from Simulated Stack Gas-

               Effect of Surge Volume on SOp Removal

                             208
                                   J200

-------
   TWO-STAGE, CLOSED LOOP  SCRUBBING
   WITH 2% SLURRY OF SPRING  VALLEY
   LIMESTONE(76% -200 MESH ; 51.5% CaO).
   STOICHIOMETRY:I.O
         40
   50       60        70

LIQUOR: GAS RATIO, GAL/MCF
80
90
                         FIGURE 6


   Limestone Slurry Scrubbing of S0g from Simulated Stack Gas--

Effect of Liquor:Gas Ratio and Inlet Gas Temperature on S0£ S.emoval
                        209

-------
                       Temperature, "F        h

                             150            0.3^3
                             200            0.169
                             250            0.092
                             300            0.055

At the point of initial gas-liquid contact, the low pH in the liquid film
increases the S02 back pressure and the increase in film temperature de-
creases the Henry's law constant, thus reducing transfer through the film
and into the droplet.

          The effect of inlet gas temperature may not be as significant in
pilot or full-scale scrubbers as in the small-scale unit.  Presumably the
stripping effect of the hot gas would be limited to only a small portion of
the scrubber height, since mass transfer of water vapor and consequent
cooling of the gas is rapid.  Since the height of the small-scale scrubber
was quite small as compared with pilot and full-scale sizes,  the portion
affected by the hot gas would represent a much greater percentage of the
total height in the small-scale unit as compared with larger ones.

          Tests were run also in which the overall solution temperature was
varied and the inlet gas temperature held at 150-200°F (Fig 7).   The adverse
effect of higher solution temperature on removal was quite significant.

          It has been generally considered that large-scale scrubbers would
be operated at the wet bulb temperature of the gas, which is on the order
of 120-125°F in power plants.   However, there may be special situations in
which the wet bulb temperature would be higher or lower; the data in Figure 7
indicate the likely result.

          Sulfite Oxidation:  The degree of sulfite oxidation in and outside
the scrubber is important in several ways.

     1.  Degree of oxidation in the scrubber affects the degree of calcium
         sulfate supersaturation in the effluent liquid phase,  which is
         an important factor in scaling (see paper in this symposium:
         "Removal of Sulfur Dioxide from Stack Gases by Scrubbing with
         Limestone Slurry:  Operational Aspects of the Scaling  Problem"
         by A.  V. Slack and J.  D, Hatfield).  There should be no other
         effect on solution composition, however, as the presence of
         excess sulfite and sulfate crystals should keep the solution
         at least saturated at all times (as compared with oxidation of
         dissolved sulfite in a crystal-free slurry, in which case the
         sulfite content could drop below saturation and thus affect pH
         and species distribution).
                                  210

-------
   80
UJ
ec


g

x
o
u.
-I
3
(0
   60
   50
   40
                          L/G : 79 GAL /MCF

                          INLET GAS TEMP: 150-200°F.

                          STOICHIOMETRY: 1.0
                          PARTICLE  SIZE:

                          72% MINUS 200 MESH

                          SOLIDS IN SLURRY: 2%
     90       100        120       130       140

          AVERAGE TEMP IN SLURRY SURGE TANKS, °F
                                                     ISO
                          FIGURE 7



      Limestone Slurry Scrubbing of S0g from Simulated Stack Gas:



                  Effect of Slurry Temperature
                        211

-------
     2.  There  is  some evidence that under upset scrubber conditions
         which  promote low pH, the increased sulfite concentration
         in solution  resulting from the low pH causes sulfite cry-
         stallization on limestone surfaces and consequent blinding.
         Increase  in  oxidation rate would reduce the sulfite content,
         which  perhaps would decrease the blinding effect.

     J.  Degree of oxidation both in and out of the scrubber affects
         the sulf ate: sulfite ratio in the solid phase, which is
         important in solids settling because the sulfate crystals
         grow larger  and thus settle better (Fig 8).

     k.  Oxidation throughout the system, including the waste pond,
         reduces the  possibility of water pollution by sulfite.

     5-  If a recovery process were operated with CaO-CaC03 as the
         absorbent, oxidation would not be desirable because difficulty
         in decomposing sulfate (as compared with thermal decomposition
         of sulfite to give a rich stream of S02 and to regenerate CaO
         for recycling) would likely make it better to discard the
         sulfate and  replace it with makeup limestone.

          Data  on  degree of oxidation in the small-scale tests have been
somewhat erratic,  ranging from 25 to 100$ depending on type of spray nozzle
and other factors; with the present atomizing nozzles, the range is 70 to
100$.  The tests are  not applicable to the actual power plant situation
anyway, because of the absence of fly ash (which may or may not promote
oxidation), absence of phenolic impurities in the gas . (which are known to
inhibit oxidation), and possible differences in N02 content (which is known
to promote oxidation).

          Data on  the composition of solids from pilot plant runs are given
in Table II.

          The calculated species composition for run B is given in Table III.

          There was no opportunity in these runs to study the effect of
operating variables on oxidation.   Work reported in the literature indicates
the following general relationship for sulfite oxidation,

                          - d(SOg) = g[HSO~"]
                              dt
where the lefthand term is rate of sulfite oxidation, the term g includes
the effects of free radicals and other effects characteristic of the system
under study, and [H] is the activity of the hydrogen ion.  Thus a decrease
in pH may or may not reduce oxidation, depending on how rapidly the HS03~
concentration increases as the pH is reduced.  In the limestone system,
                                   212

-------
                                 FIGURE 8

    Electron Micrograph or Solids Taken from Limestone Scrubber Slurry

The large blocky crystal in the upper lefthand corner is Ca304-2H20>  the
small spheres art fly ash, and the flat platelets are CaS03-0.5H20,   Some
limestone is also present.  The rosette in the center is made up of calcium
sulfite crystals.  (Photograph from work by Dr. G. H. McClellan, TVA, )

                                 213

-------
a drop in pH from 6.0 to 5.0,  in solutions saturated with CaS03'0.5H20,
increases the value of the [HS03~]/[H]1 2 term from 2.10 to 2.35,  indicating
that pH reduction in this range promotes oxidation.   The effect of pH on
the term under various conditions of saturation is given in Table IV.   The
data indicate that pH increase inhibits oxidation except when the solution
is saturated with both calcium and sulfite.
                                TABLE II

   Composition of Solid Phase from Pilot Plant Limestone Slurry Tests3
                                   Weight percent in various runs
             Constituent                ABC
        Ca                             2k.Q     26.6     23.5
        Mg                              1.0      0.8      0.8
        Total S                        11.7     10.1      8.0
        S as sulfite                    9-8      8.2      6.8
        S as sulfateb                   1.9      1.9      1.2

        Degree of oxidation, %           16       19       15


   a Limestone composition,  %:  CaO, 52.3; MgO, 1.2.  S02 in inlet
     gas, 2800 ppm.  L/G = 60 gal/Mcf.   Solids content of slurry:
   b 15-20*.
     By difference.


                                TABLE III

                   Species Distribution in Solid Phase
                                                    a
                     from Limestone Slurry Scrubbing
                     Species            Weight percent

                   CaS03-0.5H20              33.0
                   CaMg(C03)2                 6.2
                   CaS04-2%0                10.2
                   CaC03                     31.5
                   Fly ash                   19-0
                     Stoichiometry:  1.5.  S02 removal,
                     about 65$.
                                   214

-------
                                TABLE IV
          Effect of pH on [HS03~ KH"!12 in System CaO-S02 H20
               At constant Ca
                moles/kg H20
£H_
6.0
5.0
4.0

6.0
5.0
4.o

6.0
5.0
4.0
Ca (. total) SOP (total)
0.00152p
0.001525
0.001525
At
0.001525
0.001314-2
0.001258

0.001525
0.0014.520
0.01664
0.002649
0.003005
0.003188
constant S02
0.002649
0.002649
0.002649
At constant
0.002649
0.008725
0.03342
Saturation [HS03~ "]/ [H4] V2
1.00
0.15
0.017
(total sulfite)
1.00
0.12
0.012
saturation
1.00
1.00
1.00
2.096
0.864
0.291

2.096
0.766
0.244

2.096
2.353
2.670
          The preceding discussion refers only to oxidation rate in solution
assuming that oxygen is already present.   The rate of oxygen absorption
should also be important in determining overall rate in a limestone slurry
scrubber.   Chertkov1 proposed a generalized empirical equation for oxygen
absorption based on data from several scrubbers and differing absorbents.


                          GO = 0.8Q°-7.a.(S/C)6                         (12)
                                     Y '  M-
     where GQ = g 02 absorbed per hour per m2 of liquid/gas contact
                surface (considered to be equivalent to rate of oxidation)

           Q  = liquid flow rate, m3/m2-hr

           a  = t )  where t is the average solution temperature in °C
                50
         S/C  = mole ratio of sulfite to absorbing species in solution

           Y  = solution density, kg/m3

           jj,  = solution viscosity, kg. sec/m2


1 Chertkov, B. A.  J.  Appl. Chem. USSR 3J4_ (4), 743-47 (l96l).
                                    215

-------
The data currently available from TVA work are not sufficient to evaluate
this equation.   However,  Q is the only variable that can be changed easily,
and the minimum level of Q is set by the S02 removal requirement.

          Another factor is the ratio of 02 and S02 partial pressures in the
gas phase, which affects the overall oxidation rate by changing the amount
of 02 absorbed per unit of S02.  Thus a relatively low S02 content in the gas
should give relatively high oxidation rate.  Comparison of current pilot plant
tests on gas from western coal (if-00-600 ppm)--plus results from the ICI-Howden
work (1000 ppm)--with the TVA tests (2500-3000 ppm) indicate that this is true.

          The situation in regard to the various promoters involved (fly ash,
N02, limestone constituents) and inhibitors (phenolics) is somewhat confused.
More work must be done before the relative importance of these factors can
be determined.

          The effect of added inhibitors was tested in the small-scale TVA
work.  Under conditions that gave 50$ oxidation of the solids, the oxidation
was reduced to 1$ by addition of 0.1$ hydroquinone.  At the lower pH level
resulting from use of benzoic acid  to promote S02 removal, use of the same
amount of hydroquinone reduced oxidation  from 95 to 20$.  Presumably the
larger amount of sulfite available  in solution for oxidation at the lower
pH increased the percentage oxidation of  the solids--both with and without
inhibitor.  This assumes that oxidation involves the steps of sulfite disso-
lution, oxidation of sulfite in solution,  and crystallization of  sulfate.
The higher  solubility at low pH is  taken  into account  in Table IV.

          The discussion thus  far has been concerned with oxidation in  the
scrubber, which may or may not be practicable to change in actual operation.
It  should be feasible to oxidize outside  the scrubber, however, if oxidation is
indicated as a means of coping with settling, blinding, and pollution problems.
This could  be done by passing  air or oxygen through the slurry, either  through
the  recycled  slurry or  through the  sidestream diverted for solids  separation.

          The problem is getting an adequate rate  of oxygen  absorption  in the
slurry, which in Japan  has  led to development of  a special "spinning  cup"
technique to  increase absorption.   This could be used,  or  the operation could
be  carried  out under pressure, or both.

          Exploratory tests  with  the spinning  cup  oxidizer have been made in
the TVA small-scale work.   Oxidation was  much  faster  than with  introduction of
the air through  fritted glass  into  a stirred  solution; about four times as  much
air throughput was  required with  the fritted  glass to  accomplish  complete oxi-
dation.   Activated charcoal and  titanium increased the rate  somewhat but not
 significantly.   Use of  ultrasonic  energy was  not  effective.

           In the pilot  plant,  one exploratory test was made in which air was
bled into the scrubber  (airrgas  volume ratio  about l:l) to increase oxidation
rate.   Oxidation was increased by a major degree,  as  indicated by increase  in
 calcium sulfate content of the slurry solids.

                                     216

-------
          An oxidizing tower has been installed in the pilot plant and
will be tested on the solids-separation sidestream when time is available.

          Tests have also been made to determine whether sulfite will
oxidize on standing in the waste pond.  Samples of moist product slurry
solids-were stored in 1-in.  vertical tubes, unstoppered but. taped to pre-
vent exposure to light.   No oxidation occvirrad in storage over a period
of 16 weeks.
Surface Area Holdup in Scrubber

          Since rate of limestone dissolution appears to be. a controlling
factor in S02 absorption,  it is important to have a large limestone surface
area per volume of slurry and to hold this area in contact with the gas in
the scrubber as long as possible.  Hence the requirement can be expressed
as surface area of limestone exposed to each unit of S02 passed through the
scrubber.   There are several variables that affect this.

          Solids Content of Slurry;  A simplified flowsheet for the scrubber
loop is shown in Figure 9.   A large body of crystals, mainly CaC03 and
CaS03*0.5H20 (plus smaller amounts of CaS04«2H20, CaMg(C03)2, and fly ash),
circulate around the loop,  to which is added a relatively small stream of
CaC03 and from which is withdrawn a relatively small stream made up of
calcium sulfite,  unreacted calcium carbonate, and the other solid con-
stituents.   Thus the gas is exposed to an amount of limestone in the recy-
cling solids that can be much larger (per unit of S02) than the amount of
limestone fed.   The concentration of limestone in the recycling slurry
therefore is an important variable, with a linear effect on the "effective
surface area" (limestone surface area in scrubber per unit of S02 per unit
of time)o

          Only limited data on effect of solids content in the slurry have
been obtained in the small-scale tests; examples are given in Table V.
The improvement from increasing solids concentration under conditions that
gave a low level of S02 removal (single-stage test) was much larger than
for those that gave a fairly high level (two-stage).   This has been generally
true for all variables; more benefit was realized, even on the basis of
percentage reduction of unremoved S02, when there was more room for improvement.

          The increase in limestone surface area per volume of slurry at
steady state brought, about, by doubling the solids content of the slurry
should have been about 35$ for the one- stage tests and rf^% for the two-stage.
          In the pilot plant,  a much higher solids content is used ( 12-
than is  feasible for the nozzles in the small-scale spray tower.  The high
level is used mainly to provide sulfite-sulfate crystals for scaling control,
but some improvement in S02 removal has been noted also.
                                     217

-------
  SO2
                     t
                       GAS  TO  STACK
                 CaSO3
                                                   < CaC03
CaS03i CaC03
                              FIGURE 9

           Simplified Flowsheet for Limestone Scrubber Loop
          Particle Size;  The limestone surface area is inversely proportional
to the diameter of the particles;  for example, breaking 200-mesh particles
      into 325-mesh particles (W^)  should increase surface area by about
         An example of the effect of particle size on S02 removal is given
in Table VI.  Some improvement resulted from finer grinding but the effect
was not a major one.

         Stoichiometry;  Feeding of excess limestone beyond the stoichiometric
amount should have a considerable effect on limestone content of the slurry
in the circulating loop at steady state.  For a Stoichiometry of 1.0 and S02
removal of  80$, 20$ of the calcium (molar basis) in the loop, at steady state,
should be in the form of unreacted CaC03.  For 1.2 Stoichiometry, assuming
the same removal, 33% should be CaC03.  Thus a 20$ increase in external
Stoichiometry gives a 33$ increase in "internal" Stoichiometry.
                               218

-------
                                TABLE V
          Effect of Limestone Slurry Concentration on S0g Removal
                                                         a
                 Slurry concentration, %     S0g removal,

                 One- stage, L/G = 37

                           1                       38
                           2                       59

                 Two- stage, L/G = 72

                           1                       75
                           2                       78
                           5                       79
            Feed stoichipmetry :  1.0.  Closed- loop operation.
            Gal/Mcf to each stage.
                                TABLE VI
                                                              ,a
              Effect of Limestone Type and Size on SOo Removal
                                    	Calcitic	  Dolomitic,
                                    Hard           Chalk        hard
       Particle size, mesh   76$ -200  92$ -325   96$ -325    92$ -325
       S02 removal, %                78°             83          78

       * L/G = 72.  Stoichiometry:  1.0.  Two-stage closed  loop.
         Three different quarries.
          Data on effect of Stoichiometry are given  in Figure  10.  For tests
 at less than stoichiometric, the removal increased almost  linearly with amount
rof limestone fed.  Above 1.0 the effect was  small; however,  test  conditions
 were such that excellent removal was obtained at  1.0 and there was not much
 room for further improvement.
                                    219

-------
  100
   80
O
5  60
a:
UJ
o
x
§40
U.
_J
20
                 TWO-STAGE,  CLOSED  LOOP
                 SOLIDS CONTENT,  2%
                 LIQUOR/GAS RATIO, 77 GAL/MCF
                 INLET GAS TEMP, I40-I75°F
1
                        i
                              1
     0        1/3       2/3        '
     LIMESTONE FED,  PROPORTION OF STOICHIOMETRIC
                         FIGURE 10
  Effect of Stoichiometry on S02 Removal by Limestone Slurry Scrubbing
                     220

-------
          In the pilot plant work (see paper in this symposium:  "Removal
of Sulfur Dioxide from Stack Gases by Scrubbing with Limestone Slurry:
TVA Pilot Plant Tests — Part I, Scrubber-Type Comparison" by T. M. Kelso,
P. C.  Williamson, and J.  J. Schultz), there has been some indication
that operation with excess limestone (perhaps as high as 1.5 stoichio-
metric) may be advisable  to ensure that conditions do not develop which
cause the pH to become unstable.  If this is borne out by further tests,
then the effect of stoichiometry on S02 removal becomes a secondary con-
sideration.

          Liquid:Gas Ratio:  The amount of slurry holdup in the scrubber,
which is the major factor in surface area holdup, depends on scrubber type
and on circulation rate.   Very few data are available regarding effect of
scrubber type on holdup and, in any event, scrubber type is likely to be
dictated by scaling considerations rather than S02 removal efficiency.

          Data on effect  of slurry circulation rate (l/G) are given in
Figure 6.  The effect is  major; in fact, slurry rate and inlet gas tempera-
ture have been the only variables with a major effect in the TVA small-scale
work.   This has also been the situation in the pilot plant tests.  Again,
high circulation may be essential for avoiding scaling, which makes its
effect on S02 removal secondary.


Other Rate Factors

          Limestone Type;  It might be expected that the physical and chemical
properties of the limestone, and their effect on rate of limestone dissolution,
would be important rate factors.  Such properties are important in the dry
sorption method.   Tests in the wet system with various limestones, however,
have not indicated any major effect of limestone type—although the test work
has not been extensive.

          Examples of data from such tests are given in Table VI.  The chalk
used was a soft limestone that had seemed in preliminary tests to give better
results than hard limestones.  However, because of its softness the chalk
tends to break down into  very small particle size during the scrubbing operation,
which may have contributed to its activity.   When hard calcitic limestone was
also ground to very small particle size, the S02 removal efficiency was the
same.

          Dolomite was somewhat less effective although not greatly so.  This
may be associated with the reportedly slow reaction of MgC03 minerals with
weak acids.

          The pilot plant work has supported these conclusions.  No significant
difference has been noted between the various hard limestones tested.
                                 221

-------
          Blinding:  Since pilot plant results had indicated the possi-
bility of limestone blinding during periods of unstable pH,  small-scale
tests were made to explore ways for avoiding the problem.   Blinding had
never occurred in the previous small-scale work, presumably because
conditions were such as to ensure steady pH (perhaps resulting from the
relatively high liquor rate necessary in the small-scale equipment).

          The pilot plant results indicated that if for some reason the
pH in the loop decreases to a low level, the slurry becomes unresponsive—
that is, further addition of limestone does not raise the pH to the normal
level of 6.0-6.1.   In tests with the small-scale scrubber,  it was found
that a similar unresponsive slurry could be developed by operating with  a
deficiency of limestone.   In further studies with batches in a gas bubbler,
a solution of CaS03-0.5H20 in H2S03 solution (pH, 2.6) was found to be also
unresponsive; additions of limestone would not raise the pH above $.6.  An
electron micrograph of solids from such an unresponsive slurry (Fig ll)
shows that blinding of the limestone surface by calcium sulfite apparently
had taken place.

          In further tests, the unresponsive slurry was oxidized by bubbling
air through it for ~$ hours, which precipitated about 97$ of the dissolved
calcium as CaS04-2H20.  The slurry was then responsive; addition of limestone
to the oxidized slurry (in which the pH had decreased to 1.9) brought the
pH up to 6.2.

          A possible explanation for the blinding is that if pH for some
reason decreases during operation, presumably because the rate of limestone
dissolution lags behind S02 absorption, the sulfite content of the solution
increases rapidly because of higher sulfite solubility at low pH and the
reservoir of sulfite present as sulfite crystals.  Then when a limestone
particle dissolves at its surface and forms a film around it of higher pH
solution, the pH increase causes rapid sulfite crystallization at the surface
and consequent blinding.

          One difficulty with this postulation is that adding S02 to a
saturated solution of CaS03'0.5H20 canjnake the solution unsaturated.  The
effect is shown in Table VII.  The S03~ concentration decreases rapidly  with
pH by shifting back to HS03-, with which it is in equilibrium (see Fig 2).
Thus if the Ca++ concentration remains constant,  or even if it increases
somewhat, the Ca++/S03~ solubility product is not exceeded because of the
very low S03~ concentration.

          Hence if limestone dissolution were inhibited and the pH were
lowered by S02 absorption, the solution might well become unsaturated and
would be less likely to precipitate CaS03-0.5H20  than under normal operating
conditions.
                                222

-------
                                 FIGURE  11

       Electron Micrograph of Solids Taken from Limestone Scrubber Slurry

            The particle at  the right is a crystal of calcium sulfite
            (CaS03-0.5Hao).  At the left is a fly ash particle and in
            the center is a  limestone particle apparently blinded by
            calcium sulfite  crystals growing edgewise on its surface.
            The scale shown  is 1 micron.  (photograph from work by
            Dr. G. H. McClellan, TVA).
          Another possibility is that the unsaturation developed at low pH may
actually cause dissolution of CaS03-0.5 H20 crystals and reduce the surface
area available for crystal growth.   Then when CaC03 is .added in the delay tank
and a surface film of relatively high pH developed, the sulfite supersaturation
induced by the higher pH may cause nucleation on the CaC03 surface rather than
on the sulfite crystals that would ordinarily have been present.  More data
must be gathered before any conclusions can be made.
                                223

-------
                                 TABLE VII

           Concentration of Sulfite Species at Constant Calcium

             Concentration (0.001525 m/l)  but with Varying pH


                     Concentration, moles/1
                 H2S03   HS03-    S03=    CaS03  Total S02 m/l
           pH    x 1Q6   x 10+3  x 10+ 5   x 105      x IP3
           6.0    0.27    2.2i»-   11.80    28.37
           5-0    3.49    2.94    1.57    fc-33       3.005
           ^•0   37.2     3.14    0.17    o.Vf       3-188
           The rapid undersaturation developed at lower pH may also explain
why  sulfite crystals are usually small.   Even the pH variation in normal
operation may cause wide swings from rapid dissolution to rapid precipitation
from supersaturated solutions.

           If an insufficient rate of limestone dissolution is the initial
cause of the problem,  then anything that increases the rate should be helpful.
In the pilot plant,  increase in limestone stoichiometry has been effective.
Length of operating period also seems to be a factor, presumably because the
ionic strength builds  up and makes calcium more soluble (see later discussion).

           It may be possible to ensure steady pH at lower limestone stoichi-
ometry by feeding a little lime or ammonia along with the limestone.   Tests
of this procedure are  planned.

           There may be some blinding even at normal operating pH; adequate
data for evaluating this have not yet been obtained.  If so, grinding of
the  recycling slurry to renew the limestone surface, as practiced by Bahco
/see paper by K.  A.  Gustavsson,  "Bahco S02 Scrubber," presented at Internatioal
Symposium on Lime/Limes tone Scrubbing for S02 Control (sponsored by NAPCA),
Pensacola,  Florida,  March 16-20,  1970/ may be indicated.

           Ionic Strength:  Increase in ionic strength by buildup in the
solution of constituents such as  Mg,  Cl,  Na, K, and N03 should make Ca more
soluble and therefore  increase the driving force for dissolution.  Calculated
solubilities  for various ionic  strengths are given in Table VIII.  The data
indicate  an optimum  level for calcium solubility.
                                224

-------
                                TABLE VIII

                  Solubility in the System CaO-SOg-HgO

                  as Affected by Ionic Strength and pH


                                Ionic strength
                    0.004-0. Oi<.a
                        ppm
              PH
0.004-0.04*
ppm
Ca
SOo
1.0°
ppm
Ca
S02
4. 0&
ppm
Ca
SOo
6.0
5-0
4.0
61
181
667
170
558
2139
135
353
1077
360
1093
3446
92
231
701
236
711
2334
              a
                In the system CaO-S02-H20.
                Ionic strength increased by adding noncom-
                plexing ions such as Cl~ and N(CH3)4+.


          In the small-scale work, the simulated nature of the stack gas—
and the consequent absence of ionizable impurities such as carried in actual
stack gas—keeps the ionic strength from building up.   Tests were run,
therefore, in which the ionic strength was varied by adding various amounts
of magnesium sulfate (Fig 12).  An optimum level of ionic strength is indi-
cated at about 1.0.

          The pilot plant tests have given the opportunity of determining
the degree to which ionic strength will build up in actual operation, con-
sidering the fact that part of the solution is purged with the wet solids
and replaced with water.   Extended round-the-clock runs have indicated that
steady state in regard to ionic strength is attained in a day or so; there
is little further change in the concentration of constituents such as Mg,
Fe, Na, Mn, K, and Cl.  There has also been little variation in Ca and total
S, although these are subject to change with minor variations in pH,  Average
composition over a 2-week run is given in Table IX.

          The ionic strength of this solution is about 0.120, considerably
below the indicated optimum.  The only apparent way to increase it is to bleed
less liquid phase with the solids, but the 60$ liquor content of the purged
solids obtained in the current tests may be difficult to reduce.

          Inlet S02 Concentration:  In the usual gas scrubbing situation, a
decrease in inlet concentration of the constituent to be absorbed would be
expected to decrease the percentage absorption because of the lower driving
force for gas-liquid transfer.  For limestone slurry scrubbing of S02, however,
the role of solid-liquid transfer complicates the situation.
                                 225

-------
         IONIC  STRENGTH
                                         1.77
  5         10        15        20
  MgS04 CONTENT OF SLURRY,  %
25
             FIGURE 12

 Effect of MgS04 Addition on S0g Removal

in Closed-Loop Limestone Slurry Scrubbing
         226

-------
                                TABLE IX

                                                                a
        Average Filtrate Composition Over 2-Week Pilot Plant Run


                    Constituent     Concentration, gpl
Ca
Mg
Total S
Fe
Na
Mn
K
Cl
Total N
1.06
0.46
0.91
0.001
0.038
0.002
0.059
1.32
o.ok
          Limestone stoichiometry:  1.5-  L/G:  About 40 gal/Mcf.
          Solids content of slurry:  12-15$*  Water content of
          purged solids:  About 60%.  Samples taken at scrubber
          outlet.
          Data from tests on effect of inlet S02 concentration are given
in Figure 13.  The percentage removal was actually higher for the lower
concentrations, probably because of the smaller amount of limestone
dissolution required per unit volume of slurry (the liquor rate was the
same for all of the tests).  The effective surge tank volume (gal/lb
SOg-min), which has been shown to be important, was much higher at the
lower S02 concentrations (range from 1240 to 4900 gal/lb S02-min).

          Thus it appears that low exit S02 concentrations, on the order
of 50 PP™ or less, can be attained with high liquor rate and surge tank
retention time.
Waste Disposal

          As noted earlier (Fig ll), the calcium sulfite crystals formed
in limestone scrubbing are very small, thin platelets.  When the slurry is
allowed to stand, the crystals settle rapidly for a short time—giving a
fairly clear supernatant layer—but the settling rate then drops off sharply.
It is postulated that the decreased rate begins when the crystals start
touching each other; a voluminous, gel-like structure develops in which
settling is relatively slow.
                                227

-------
PO
ro
oo
         100
                           I           I          I
                           %-200 MESH LIMESTONE
                                      O 95
                                      A 85%-200MESH LIMESTONE
                      500
        Effect of
               1000      1500      2000      2500      3000     3500


                     S02 IN  INLET GAS, PPM


                             FIGURE 13


Inlet SO, Concentration on SO. Removal with Two-Stage  Closed-Loop Limestone Slurry Scrubbing
                      The liquortgas ratio was 80 gal/Mcf, the solids content of the slurry 2*,

                      and the inlet gas  temperature 150-200°F.

-------
          In small-scale tests with the slurry, it was found that the
average rate of free settling was about 2 in./hr; this rate was fairly
uniform as long as the  free settling phase lasted.  A sample of the gel
formed (l6$ solids) after compressive settling began was tested with a
gelometer; gel strength increased with time, as follows.

                Time elapsed after
                  stirring, min        Yield point, g-cm

                          0                    0
                         30                    6
                         60                    9
                        1080                   25
                                                                     o
          Settling of the gel could be accelerated by periodic slow stirring
to destroy the gel structure.  A gel that had settled to 38$ solids and
showed no indication of further significant settling was settled to 48$
solids by three stirrings about a day apart.

          Oxidation was quite effective in promoting settling.  Further
work will be done on this approach.


Alkali - Lime-Limestone

          The many problems involved in lime-limestone scrubbing make it
quite desirable to find a better throwaway method.  Several organizations,
including TVA, are working on a process that involves scrubbing with an
alkali solution (NH4, Na, K.) and regenerating the alkali with lime or lime-
stone.  The waste product is calcium sulfite (plus calcium sulfate, hopefully)
as in straight lime-limestone scrubbing.  Keeping the lime-limestone out of
the scrubber should greatly reduce the problems of scaling, erosion, silting,
wet-dry deposition of solids, particulate entrainment, and high pumping cost.

          Of the various alkalies, Na and K have the advantage that they are
not volatile as is NH.a  -which simplifies scrubber design and operation.  How-
ever,  sulfate formed by oxidation in the scrubber is quite difficult to regen-
erate; there have been  efforts over the last several decades to convert byproduct
sodium sulfate to caustic soda without success.  Regeneration in an S02 scrubbing
loop may be feasible, however, because only a partial conversion per pass is
acceptable,  but it is likely that the resulting NaOH solution would be quite
dilute and that circulation rate might therefore have to be high.
                                       •
          If ammonia is used as the alkali, regeneration may be somewhat
easier since (NH4)2S04  can be crystallized from a sidestream and sold (whereas
Na2S04 would have a quite limited market).  Or the (NH4)2S04 can be reacted
with lime to regenerate the ammonia, as in the Kuhlmann Electricite de France
process.


                                 229

-------
          In the TVA small-scale work,  limestone rather than lime has
been used as the regenerant since it is cheaper.  It reacts well with
alkali bisulfite (Fig 14) but not with sulfite; however,  the sulfite can be
recirculated to the scrubber as is the common practice in alkali scrubbing
processes of the recovery type.

          The data in Figure 14 indicate that the regeneration temperature
should be somewhat elevated; fortunately,  in power plant stack gas scrubbing
heat from the stack gas gives an adequate  liquor temperature level,  which
may not be true for other types of stack gas.

          Since the calcium sulfite formed in the tests settled well, no
difficulty is expected in separating and washing the waste solids.  Further
work will be concerned with separation of  (NH4)2S04.   Data on the ammonia
scrubbing step is being obtained in the current EPA-TVA ammonia scrubbing
pilot plant project.   Testing of the integrated process is planned in a
TVA pilot plant now under construction.
                                  230

-------
                                                                                       B
     110
     •I fV*\    •^» ^__  ^^_ ^^M ••••§ *mmm M^ «••
ro
CO
          _  TESTS  AT  75  F
   1*

                                                  Theoretical maximum
                                               /^reaction
                                                             LEGEND (for both figures)
                                                               Q Sodium salts
                                                               A Potassium salts
                                                               Q AmtoDnium salts
Salts in solutions con-
taining 16$ "bisulfite and
2% each of sulfite and
sulfate.
                                                  110
                                                                                                             - 100
   1.25     0.5            0.75
Ca:2HS05 mole ratio
                                                       FIGURE Ik

                              Effect of Calcium Carbonate Proportion on Degree of Reaction

                            with Solutions of Alkali Bisulfite-Sulfite-Sulfate Salt Mixtures
                                                                                              1.0
                                                1.25

-------
Absorption Studies of Equimolar
Concentrations of NO and NCL In
Alkaline Solutions
            By:  L, H. Garcia
                 Division of Control Systems
                 Laboratory Research Branch
                 Special Projects Section
            233

-------
                       ABSTRACT
ABSORPTION STUDIES OF EQUIMOLAR CONCENTRATIONS OF NO & NO  IN
ALKALINE SOLUTIONS

Equtnolar concentrations of NO and NO- in flue gas were passed
countercurrently to various alkaline solutions in a bench-scale
packed column.  The concentrations of NO and NO. studies were from
250 ppm to 750 ppm each.  Liquid velocity, gas velocity, size of
packing, and liquid temperature were the parameters varied.

The results of this experimentation indicates that at a concentration
of 250 ppm each of NO and NO. in flue gas the percent removal for
NO and NO., expressed as total NO  was between 12 to 14 percent; at
500 ppm each of NO and NO. in flue gas, the percent removal was between
17 to 19 percent; at 750 ppm each of NO and NO. in flue gas, the percent
removal was between 27 to 29 percent.
                              234

-------
                    Absorption Studies of Low Concentrations

                      of NO and VK>2 in Alkaline Solutions
BACKGROUND
     The Federal Government has enacted legislation to regulate and control
the levels of nitrogen oxides in the ambient air and to limit their emission
from specific new stationary sources.  The control of nitrogen oxides is
important because these air pollutants are Involved in the complex photo-
chemical reactions in the atmosphere which result in smog formation and
because, by themselves, they have adverse physiological effects for all
forms of life.  Although, a number of oxides of nitrogen are recognized,
the most common forms are nitric oxide (NO) and nitrogen dioxide (NO-) on
account of their widespread production and their relatively high atmospheric
stability.  In this paper the term "nitrogen oxides (or NO )" refers to
either or a combination of these two gaseous air pollutants.

     Mobile sources, the largest single source category, contribute over 40
percent of all the man-made NO  emitted in the United States.  Current
knowledge on the methods of control are covered in detail in AP-66, "Control
Techniques for Carbon Monoxide, Nitrogen Oxide, and Hydrocarbon Emissions
from Mobile Sources."  The next largest source is electric power generation,
which is responsible for nearly 20 percent of all man-made NO .  About
40 percent of NO  emitted from stationary installations is attributed to
electric generating power plants.  Control of NO  from stationary sources
is discussed in detail in AP-67, "Control Techniques for Nitrogen Oxide Emissions
from Stationary Sources."

     About 1 percent of the total man-made NO  emitted to the ambient air
of the United States is formed by chemical sources, mainly related to the
manufacture and use of nitric acid.  Concentration from these sources are
usually much greater, however, than those from combustion sources, and as a
consequence, these sources often give rise to a highly visible, brown-red gas.


THEORETICAL CONSIDERATIONS
     Under proper conditions, nitrogen and  oxygen tend  to  combine in
accordance with  the  following equation:

                 N2  f °2"* 2N°
                                    235

-------
     The equilibrium concentration of NO varies with temperature; it ib
negligible below 1,000°F but quite significant above 3000°F.  In addition,
it is influenced by gas composition; at a given temperature, for example,
the equilibrium concentration of NO in air exceeds that of NO in flue gas
of 3 percent oxygen content by a factor of approximately 3.

     Nitric oxide tends to react with oxygen as follows:

              2NO + 0
This equation implies the coexistence of NO and NO-*  Calculated equilibria
indicate that the stability of NO. decreases with Increasing temperature.
Nevertheless, from an equilibrium standpoint, the absolute concentration
of NO. increases with temperature while the ratio of its concentration
to that of NO decreases with increasing temperature.

     Chemical equilibria depend on the initial and final conditions and
not at all on reaction mechanisms or Intermediate reaction steps.  Equilibrium
concentrations are obtained after the lapse of sufficient reaction time;
therefore, they are not necessarily observed experimentally.  Because of
the simplicity of the molecules involved In the previous equations, their
thermodynamic properties are accurately known and equilibrium calculations
are made easily.

     It has been suggested    that equimolar mixtures of NO and NO-
might be removed by scrubbing with alkaline solutions.  The use of the
alkaline materials (NaOH and Na.CO-) for NO  scrubbing was studied
originally by Sherwood and Pigford  ' ^  wh5 found, using a concentration
of one percent NO  in the gas phase, that a 46 percent NaOH solution absorbed
nitrogen oxides from a gas containing equimolar quantities of NO and NO-
at rates as much as ten times faster than if NO. were present alone.
When they compared the NO + N0~ absorption rates with the calculated
equilibrium concentration of N^O- (NO + N02 -»• N-0-), it was found that
the rate was approximately proportional to the First power of the N-O-
concentration, gindlcating that this was the reacting species.  Ganz and
co-workers       performed similar studies using sodium carbonate solutions
and reached similar conclusions.  Andrew and Hanson™ describe the dynamics
of nitrous gas absorption into water as the result of a number of separate
reaction mechanisms, and state that relative importance of these mechanisms is
primarily dependent on the gas composition.

     In order to determine the applicability of aqueous alkaline scrubbing
of equimolar NO/NO, mixtures at typical flue gas concentrations, the EPA
laboratories decided to study the absorption of NO and NO. at concentrations
in the range of 250 ppm to 750 ppm each in flue gas.

     Although control of most of the NO  emissions from combustion sources
is expected to be achieved by various combustion modification or control
methods, additional flue gas or process stream treatment methods may be
required for specific problem areas.  Consequently, a limited effort has been
                                  236

-------
directed to research, leading to the identification and development of effective
flue gas treatment techniques for NO  control.

     Rates of formation can be calculated by kinetic equations that rely
heavily on experimental measurements.  The reaction products depend to a
large measure on the relative speeds of the reactions that actually occur.
The rate of oxidation of nitrogen to NO is highly temperature-dependent;
it is very slow at 5QO°F, but fast at 4000°F.  The underlying reason is that
a high level of energy is needed to break the N-N (dissociation energy 225.3
kcal/mole) triple bond of molecular nitrogen so that oxygen can react.
Conversely, a smaller but still relatively large amount of energy is needed
to break the N-0 bond to permit decomposition of nitric oxide into its
elements.  This means that high temperatures are required to form NO; once
formed, it resists any breakdown into its elements.  Breakdown becomes more
and more unlikely as temperature decreases, because the energy available for
thermal breakdown diminishes rapidly with, decreasing temperature.  Thus, an
Initially high temperature followed by quick cooling, even to a relatively
high temperature level, produces large amounts of NO.
     At temperatures above 2000°F, both NO and NO, are formed, but the amount
of NO, Is usually less than 0.5 percent of the total NO .  The oxidation of
NO to NO2 ky oxygen, however, is peculiar in that the rate of reaction
decreases with increasing temperature.  This is one of the few known reactions
that exhibit an inverse temperature relationship.  The resultant slow
oxidation rate at high temperatures accounts in part for the negligible
amounts of NO, frequently found in hot combustion gases.  Another
characteristic of the oxidation of NO to NO, is the fact that the rate
varies with, the square of the NO concentration.  The rate of oxidation of NO
by oxygen in air falls off rapidly, therefore, with dilution of the NO.
A long period of time may be required to oxidize trace quantities of NO
by this mechanism.

EXPERIMENTAL

     A laboratory bench-scale packed scrubber was fabricated  (see Figure
1).  The scrubber consisted of a glass (Pyrex) column 2 1/4 inch I.D. packed
with eighteen inches of 3/8-inch Berl saddles.  The scrubber was equipped
with an inlet liquid port and an outlet gas port at both the  top and bottom
of the scrubber.  Because of this, the laboratory bench-scale scrubber was
capable of passing scrubbing liquor cocurrently as well as countercurrently
to the gas flow over a wide range of liquid to gas (L-to-G) flow rates.

     In. the bencft-scale packed scrubber, the following parameters can be
varied and measured:  gas and liquid flow rates, concentrations of the gaseous
pollutants, packing height, type of packing, temperature, and concentration
of alkali.
                                   237

-------
         FIGURE  I

 FLOW DIAGRAM  OF EQUIPMENT
USED FOR ABSORPTION OF NO 8 N 02
pQ












X-<











J FLOiV i
J McTER
1 -C

y
i
1
.




y
i ,




_»•
>
|







*
ri
*

RgCORDSR


'INFRARED
c"rt *"'
2


RECCROuR
— ..
f
IKrRAfJED
KO


t —..«/*
RECORDER \
v
V
— i
1
ULTRAVIOLET
ANALYZER
N02






r-r
1
* C
r~^-^

c*i i/->f i f /^r^
/v
\
< 	 CH/}
«•— •
•« ' AiH
'4 	 i
_^
^ "I
A A p.
ri L J
> FLUu GAS N'°2 _ NP ^s°2




YT »
R
^
it
A ^ i


' - • • H X-X
| PACstED FEED
••••^ Htf J ~-*"i"Cwwi3»ir» irsi ^- \

• .
           CATCH
                       238

-------
     The main objective of these studies was to determine the rate of
mass transfer for equimolar concentrations of NO and NO. in the range of
250 ppm to 750 ppm each.  The parameters affecting the rate of mass transfer
were determined by a study of the following: gas and liquid velocitiies,
types (i.e., CaCOHK, Mg(OH)2, and NaOH} and concentrations of scrubbing
solutions, temperature, and mole ratios of HO and N02, size of packing,
and height of packing.

TESTING

     These studies were performed using flue gas from the combustion
of natural gas (see Figure 1.).  Air, natural gas, and NO were blended
and burned, (N0» was added prior to .entry to the preheater to minimize
thermal decomposition of NO. to NO), to give a standard flue gas having
the following approximate composition:  C02, 6.4%; 0,, 3.0%; HjO, 12.7%;
N2, 74.9%; NO, 0.025 to 0.075%; and NO,, 07025 to 0.075%.  The two gas
flow rates used were 20 SCFH and 30 SCFH.  The two liquid flow rates were
170 ml/min and 350 ml/min.  These combinations of gas and liquid flow rates
correspond to a range of L-to-G ratios from 9.3 to 29.2.  A flue gas
preheater was used to raise the temperature of the gas entering the scrubber
to 350°F.  After passing the gas countercurrently to the liquid flow through
the scrubber, the gas was pulled through two condensers to remove water
since water interferes with the monitoring of NO and N02.  Several monitors
were used to follow the concentration changes of the gaseous components of
primary interest.  A non-dispersive infrared spectrophotometer, that
provided continuous SO. measurement, was installed in case of SO. fluctuations
in the natural gas supply.  During the course of these measurements, no
appreciable content of SO. was observed in the flue gas.  Another infrared
instrument provided continuous NO measurement, and an ultraviolet
spectrophotometer provided continuous NO. measurement.  The specific
instruments used in these tests were BecKman Model 315A Infrared Analyzers
for SO. and NO and a DuPont Model 400 Ultraviolet Analyzer for the NO.
measurement.  By proper selection of flow paths using a three-way valve
(see Figure 1), the gas could be monitored by the instruments prior to
entry to the scrubber or after passing through the scrubber and condensers.
One feed tank and one catch tank were used in these experiments.  Scrubbing
liquor, i.e., (_Ca(OH)2, Mg(OH)2, or NaOH solutions, were fed on a once-
through basis from the feed tank through the scrubber and into the catch
tank.  The temperature of the solutions in the feed tank was maintained by
means of a 1000-watt electric heater which was thermostatically controlled
to ±5°F.  The range of temperatures over which this heater operated was
from 100°F to 200°F.

     A run was performed in the following manner.  The feed tank was charged
with fifteen gallons of distilled water and heated to 100°F, 125°F, or 150°F,
while stirring.  After the water reached the desired temperature, the solid
alkaline material was added to the feed tank.  The amount of Ca(OH>2 or Mg(OH)2
                                   239

-------
added was equal to that amount necessary to make fifteen gallons of solution
saturated with the alkaline material at the particular operating temperature.
The concentrations of NaOH solutions were from 10""* N solutions to 5%
solutions.  After the pH of the solution reached a maximum. value, i.e., about
fifteen minutes, the alkaline solution (of cither Ca(OH)», Mg(OH)2, or NaOH)
was then pumped into the top of the scrubber at either 1/0 ml/min or 350 ml/min/
As soon as the temperature in the packed bed of the scrubber reached within
± 2 F of the liquid temperature of the feed tank, the three-way gas feed valve
was switched and the NO^ and NO^ containing flue gas was fed to the bottom of
the scrubber.  The gas leaving the top of the scrubber was monitored for NO,
NO 2 and S02 as mentioned previously.  The time for the S0~ and NO to go
from the furnace to the monitoring instruments was twelve seconds at 20 SCFH,
and six seconds at 30 SCFH.  The time for the NO. to go from the preheater
to the instruments was approximately ten seconds at 20 SCFH, and five
seconds at 30 SCFH.

ANALYTICAL PROBLEMS OF SCRUBBING STUDY

A.  Gas Analysis

    The results of engineering studies can be no better than the analytical
tools used to obtain these results.  In the study of NO  scrubbing with a
packed bed reactor, a proper interpretation of NO and NO- concentrations
Cas obtained from the non-dispersive infra red analyzer and the ultraviolet
spectrophotometer, respectively) is dependent upon the specificity of the
instrument for the species being studied.  NO  chemistry is very complex
and a large number of equilibria occur simultaneously, i.e.,
                  NO + N02^N20


                     2NO + 02^T

                  3N02 + HjO .? 2NOJ + 2H+ + NO

                     N2°3 + H2° ^ 2HN02

                  3HN023T & + N0~ + 2NO + ty

                        and others.

     During the course of packed bed scrubbing studies, scrubbing efficiency
is determined by comparing the data for the scrubbed gases with those obtained
from by-passing the scrubber and passing the NO  containing flue gas into
the analyzers directly.  It was found that the ?otal NO  determined by the
analyzers was greater than the sum of the entering NO and NO. concentrations.
It was subsequently deduced that the NDIR was sensing N^O, as well as the
NO which it was designed to determine.  This interference was eliminated by raising
the temperature of the gas going to the NDIR, thus, decomposing any NO


                                 240

-------
to N0~.  It was believed advisable also to raise the temperature of the
gas sample seen by the UV NO. analyzer.  This was done to avoid any possibility
of interference by N20> or N-05 with the determination of NO-.

     It was decided that, because of possible interference problems with the
process monitors, a basic standard method was needed to establish the accuracy
of the monitors when used under various experimental conditions.  The first
approach to a solution of this problem involved gas chromatography.  A
method was developed in which the gas sample was separated on a Porapak
Q column (i ft x 1/8 in.) at 30°C.  With this procedure it is possible to
determine NO in NO-^ mixtures if the NO concentration exceeds 8000 ppm.
This lack of sensitivity causes the gas chromatographic approach to be
inappropriate  for levels of NO normally found in flue gas.

     The second approach at a standard method to check out the accuracy
of the NO  monitors involved the use of chemiluminesence.  An instrument
built by the Division of Chemistry and Physics was borrowed and comparisons
made to the process monitors.  After the modifications of the process
monitors previously mentioned were completed, the concentrations determined
by the monitors were identical to results determined by chemilumenesence.
This calibration yielded satisfactory results for the process monitors
in that material balances in the gas phase could then be achieved.

B.  Liquid Analysis

                                                                    8
    A liquid analysis for N0~ and NO., described by Walters and Uglum
was performed on the scrubber solution when water or NaOH solutions were
used.  This liquid analysis did not yield acceptable results with Ca(OH).
and MgCOH^ solutions.  The chief problems which were encountered were
caused by precipitation from the saturated solutions being used for scrubbing
in the cuvette causing erroneous results.  No liquid analyses were reported
in this paper due to the variation in instrument readings caused by this
precipitation.

DISCUSSION AND RESULTS

     Figures 2 through 18 are plots of all the data.  The ordinate of these
figures is the rate of absorption of NO and N0_ per hour per scrubber cross-
sectional area.  The abscissa is the mole percent of NO and NO- in the flue
gas stream.

     Many of the figures indicate a slurry flow or liquid rate of 170 and
350 ml/min.  Two figures could have been drawn showing that the data points
would have been identical to each other at the two slurry or liquid flow
rates.  It was found during this work that slurry flow rate had no effect
on the adsorption rate of NO .  Consequently, many of the figures indicate
flow rates of both 170 and 350 ml/min where the other parameters and
resultant adsorptions are replicated.  Instead of making two different
                                  241

-------
figures, it was decided to make one figure.

     Figures 2 through 8 show the rate of absorption of NO and NO. in flue
gas with use of saturated solutions of -magnesium hydroxide under various
conditions.  The parameter changed in Figures 2 and 3 was the gas flow rate
i.e., from 20 SCFH to 30 SCFH.  As can be seen from these figures, there is
no apparent change in the rate of absorption of NO and NO. with gas flow
rate.  Also, it should be pointed out that changing the slurry or liquid
flow rate from 170 ml/min to 350 ml/min produced no change in absorption.
In Figures A and 5 the parameters changed were liquid or slurry temperature
and gas flow rate.  The data on Figures A and 5 indicate no apparent change
in absorption rate due to variation in slurry or liquid temperature and
gas flow rate.  Figures 6 through 8 shown one gross change relative to Figures
1 through 5, and that is a change in the size of packing.  Four millimeter
Berl saddles were used instead of the 3/8-inch Berl saddles.  This change
was made to see if a change in interfacial area would affect the rate of
absorption.

     Another change made within the data in Figures 6 through 8 was the
slurry or liquid temperature.  No statistical difference could be found
due to a temperature effect.  Comparing Figure 6 through 8 to Figures 2
through 5 where the gross change was the packing size, no statistical
difference was found between the slopes of the best fit lines.  Intuitively,
one might conclude that increasing the interfacial area would yield an
increase in the rate of absorption but there is no statistical evidence for
this.

     Figures 9 through 12 shown the rate of absorption of NO and NO. from
flue gas using saturated solution of calcium hydroxide under various conditions.

     The main parameter changed in Figures 9 and 10 was the gas flow rate,
i.e., from 20 SCFH to 30 SCFH.  No change in the absorption of NO and NO.
was noted.  Similar to the results with the magnesium hydroxide solutions,
no change in absorption was noted due to the change in slurry or liquid flow
rates, i.e., from 170 ml/min to 350 ml/min.

     Again one main parameter was changed between Figures 11 and 12 and that
was the gas flow rate i.e., 20 SCFH to 30 SCFH.  No significant change
in absorption was noted.

     Comparing Figures 9 and 10 to Figures 11 and 12, where the main parameter
changed was temperature (i.e., 125°F to 150°F}» no statistical change was
noted in the rate of absorption.

     Figure 13 shows the results of the use of calcium hydroxide as a
scrubbing agent while the column was packed with four-millimeter Berl saddles.
It was not long before the column plugged and started to flood, yielding
limited data.  The plugging was caused by the reaction of calcium
hydroxide with the carbon dioxide in the flue gas.
                                  242

-------
                   FIGURE  2
 RELATIONSHIP BETWEEN  RATE OF AQUEOUS ALKALINE
 SCRUBBING OF NO AND M>2 AND INITIAL CONCENTRATIONS
 OF  EQUIMOLAR   AMOUNTS IN FLUE  GAS














o
E
—
1
z
o
t
or
o
CO
GO
**
u.
0
UJ
<[
or







-3
10X10
9
8
7
6


5

4
evj

.1 3
\


2







-4
10X10
9
8
7

e

5

4
3
-5
1 f\ VO R
aCRUBBINO LIQUOR - MgfOHL
BERL SADDLE 3.IZE - 3/3"
LIQUID TEMPERATURE - •
» I faO r
UIOUID FLCW RATE -
1708 350 ml/mln
GAS FLOW RATE - 2Q SCFH O
D

D
_
0
o
-
o


D
0
O a



a

O
-
-
a
_

-
O NO
••
a NO,
2
_
i i i . . » i i I
IX10"
5  6  7  8 9  10X10
                                                -2
INITIAL CONCENTRATION of NOand of N02 , MOLE PERCENT
                                                    243

-------
                               FIGURE 3
           RELATIONSHIP  BETWEEN RATE OF AQUEOUS ALKALINE
            SCRUBBING OF NO AND N0g AND INITIAL  CONCENTRATIONS
           OF EQUIMOLAR  AMOUNTS IN FLUE GAS
     -3
    10X10
        9
        8
        7
        6
SCRUBBING  LIQUOR
BERL SADDLE S.IZE
LIQUID TEMPERATURE
LIQUID FLOW RATE
6AS FLOW  RATE
 I
z
O
E
or
O
CO
•**
u.
O
UJ
H
or

10X10
9
8
7

6
    -5
         •
~ Mg(OH)2
-  3/8"
   I25*F
""  I7O ft 350ml/min
   3O SCFH
 O
O
         ixlO
             -2
                O
                a
                                                   O  NO
                                                   a  NO«
                                                     i	i
                             5   6  7  B 9 10X10
                                                            .2
         INITIAL CONCENTRATION of NOandof NO2 , MOLE PERCENT
                                                                 244

-------
                              FIGURE 4

          RELATIONSHIP BETWEEN RATE OF AQUEOUS ALKALINE

          SCRUBBING OF NO AND N02 AND INITIAL  CONCENTRATIONS

          OF EQUIMOLAR  AMOUNTS  IN FLUE GAS
    -3
   10X10

       9

       8

       7

       6


       5
2
O

E
tc
O
O
li.
O

Ul
     -4
    10X10

        9

        8

        7

        6
     -5
    10x2.5
           SCRUBBIN8 LIQUOR

           BERL SADDLE 3.1 ZE

           LIQUID TEMPERATURE

           LIQUID  FLOW  RATE


           6AS  FLOW RATE
Mg(OH)2
    n
3/8

I50*F

I7OS 350m|/mln

20 SCFH
                                             a

                                             O
                                          a

                                          O
                                      a
                                      0
                                a
                                O
                            O
                            D
                                                    O  NO
                                                       NO,
                                                             .2
          I xlO
          INITI
                                          5   e  7  8 9  10X10
              AL  C ONCE NT RATION of NOand of NO2 , MOLE PERCENT
                                                                   245

-------
                              FIGURE  5
           RELATIONSHIP  BETWEEN RATE OF AQUEOUS ALKALINE
           SCRUBBING OF NO AND N02 AND INITIAL CONCENTRATIONS
           OF EQUIMOLAR  AMOUNTS IN FLUE  GAS
    -3
   10X10
        9
        8
        7
        6
        5
 I
z
o
fc
or
o
CD
QC
 -4
10X10
    9
    8
    7
    6
       SCRUBBING  LIQUOR
       BERL SADDLE S.IZE
       LIQUID TEMPERATURE
       LIQUID  FLOW  RATE
       GAS  FLOW  RATE
MgtOH)2
3/8*
I50*F
I7O8 350 ml/min
30 SCFH
                                                 a
                                                 O
                                                    O  NO
                                                    a  NO,
                                          5   6  7  Q 9  10X10
                                                             -2
          INITIAL CONCENTRATION of NOandof NOg, MOLE  PERCENT   246

-------
                     FIGURE 6

  RELATIONSHIP BETWEEN  RATE OF AQUEOUS ALKALINE
  SCRUBBING OF NO AND N0g AND INITIAL  CONCENTRATIONS
  OF EQUIMOLAR  AMOUNTS IN FLUE GAS









V
i
I
5

1
2
O
E
ft
o
tft
o
*
u.
0
iii
h

e





-3
10X10
9
8
7
6


c
4
N
£
L:


2






_4
10X10
9
8

7

6
5
4
3
-5
in*9 R
SCRUBBIN0 LIQUOR - Mg(OH>2
SADDLE 3JZE - 4 mnii _
LIQUID TEMPERATURE - Iftn9e °
•. IW r
. L10U1° FLOW RATE - .7oml/mln e
6AS FLOW RATE - 20SCFH

a
-

8
9




o












O NO
a wo2

• ,.,.••
ixlO
                    3.
5  6  7  8  9  IQXIO
INITIAL CONCENT RATION of NOandof N02 ,  MOLE PERCENT
                                                     247

-------
                           FIGURE  7
       RELATIONSHIP BETWEEN RATE OF AQUEOUS ALKALINE
       SCRUBBING OF NO AND N0g AND INITIAL  CONCENTRATIONS
       OF EOUIMOLAR   AMOUNTS IN FLUE  GAS
 -3
10X10 T SCRUBBING LIQUOR -
        BERL SADDLE  3.IZE -
    8p
        LIQUID TEMPERATURE -
     7
     6
-
LIQUID FLOW RATE -
GAS  FLOW  RATE
Mg(OH)2
4mm.
I25*F
I70ml/min
20SCFH
or
0
m
**
u.
0
UJ
^
or


-4
IOXIO
9
8
7

6
  -5
 10X2.5
                                          o
                                          O
                                       D
                                       O
                               o
                               D
                        o
                        D
                                                 O  NO
                                                 D  NO,
                                 56789 10X10
       IXIO
          -2
                                                               248
       INITIAL CONCENTRATION of NOandof NO2 , MOLE PERCENT

-------
                           FIGURE 8

       RELATIONSHIP  BETWEEN RATE OF AQUEOUS  ALKALINE

       SCRUBBING OF NO AND NOg AND INITIAL CONCENTRATIONS

       OF EQUIMOLAR  AMOUNTS  IN FLUE  GAS
 -3
10X10

    9

    8

    7

    6
SCRUBBING  LIQUOR

BERL SADDLE 3,1 ZE

LIQUID TEMPERATURE

LIQUID  FLOW  RATE


6AS  FLOW  RATE
Mg(OH)2

4mm.

150° F

170 ml/min

2OSCFH
                                               o
                                               O
z
0
E
a
o
in
CD
<
u.
O
Ul
fc
a:







-4
10X10
9
8
7

6
   -5
 10X2.5
                                                 O  NO
                                                  a  NO,
       IxlO
           -2
                                   6789 10X10
                                                   .2
        INITIAL CONCENTRATION of NOandof NO2, MOLE  PERCENT
                                                               249

-------
                               FIGURE  9

            RELATIONSHIP BETWEEN  RATE OF AQUEOUS ALKALINE

            SCRUBBING OF NO AND NOg AND INITIAL CONCENTRATIONS
            OF EQUIMOLAR  AMOUNTS  IN FLUE  GAS
     -3
    10X10

        9

        8

        7

        6
 I
z
o
t
or
O
CO
o
UJ
 10X10

     9

     8

     7

     6


     5


     4
    3
  -5
10X2.5
         SCRUiiINC  LIQUOR

         BERL SADDLE S.IZE

         LIQUID TEMPERATURE

         LIQUID FLOW RATE

         6 AS FLOW  RATE
-  CO (OH)
      M ^
-  3/8

~  I25*F

"~  1708-350

~  20 SCFH
                                                 Oo
                                               q,
                                 o o
                                    D
                                                    O  NO
                                                    a  NO,
         ixlO
             "2
                                      5   6  7  8  9 IOXIO
                                                             .2
          INITIAL CONCENTRATION of NOandof NO ,  MOLE  PERCENT
                                                                  25C

-------
                    FIGURE 10
  RELATIONSHIP BETWEEN RATE OF AQUEOUS ALKALINE
  SCRUBBING OF NO AND N0g AND INITIAL CONCENTRATIONS
  OF EQUIMOLAR  AMOUNTS  IN FLUE  GAS










1
V
i,
J)
75
E
^
1
z
o
E
te.
O

-------
                    FIGURE II
  RELATIONSHIP BETWEEN RATE OF AQUEOUS ALKALINE
  SCRUBBING OF NO AND N02 AND INITIAL CONCENTRATIONS
  OF EQUIMOLAR  AMOUNTS IN FLUE  GAS









V

Si
0
E
.0*
•M>

1
O
t
or
o
00
**
u.
o
tu
^
or








-3
10X10
9
8
7
6

5
4
CM
£
L:
^




2




_4
10X10
9
8
7

e



4
3
-5
10X2.5
1
SCRUBBINC LIQUOR - CflfoH)
BERL SADDLE SJZE - 3/9"
LIQUID TEMPERATURE - ,•„
- ISO F
LIQUID FLOW RATE -
170 ft 350 ml/mln
GAS FLOW RATE - 2Q SCFH
-
-

a
D



D
O


D 0
O
D
O
D
O
-
0


O NO
•M
D NO,



i i i i . » i i I
xlO"2 2 3 4567891
                                                 .2
INITIAL CONCENTRATION of NOandof NO2,MOLE PERCENT
                                                   '  252

-------
                           FIGURE  12
       RELATIONSHIP  BETWEEN RATE OF AQUEOUS ALKALINE
       SCRUBBING OF NO AND N02 AND INITIAL CONCENTRATIONS
       OF EQUIMOLAR  AMOUNTS  IN FLUE  GAS
 -3
10X10
    9
    8
    7
    6
SCRUBBIN8 LIQUOR -  Cfl(OH)
BERL SADDLE S.IZE •
LIQUID TEMPERATURE
LIQUID  FLOW  RATE '
GAS  FLOW RATE
z
o
E
a:
o
00
^
u.
o
Ul
^
tr






^
1 Ox 10
9
8
7

e
       •
3/8
   •
150 F
170 &  350 ml/min
30  SCFH
   -5
 10x2.5
                                                o
                                                O
                                             D

                                          D  O
                                      D
                                      O
                                                 O  NO
                                                  a  NO,
       IxlO
                               5   678  9 10*10
                                                   .2
        INITIAL CONCENTRATION of NOandof NO2 , MOLE  PERCENT
                                                                253

-------
                              FIGURE  13
           RELATIONSHIP BETWEEN  RATE OF AQUEOUS ALKALINE
           SCRUBBING OF NO AND NOg AND INITIAL CONCENTRATIONS
           OF EQUIMOLAR  AMOUNTS  IN FLUE  GAS
    -3
   10X10
        9
        8
        7
        6
 I
z
O
fc
or
o
oo
u.
O
UJ
     .4
    10X10
        9
        8
        7
     -5
    10X2.5
            8CRUSBIN8  LIQUOR -  Ca(OH)
            BERL  SADDLE  SJZE
            LIQUID TEMPERATURE
            LIQUID FLOW RATE '

            GAS  FLOW  RATE
4mm.
125* F
170 ml/min
20 SCFH
          IxlO
              -2
                                              FLOO DIN 6
                                              REGION
                            O
                            a
                                                    O  NO
                                                     a  NO,
                                           56789 10X10
          INITIAL CONCENT RATION of NO and of NO2,MOLE  PERCENT
                                                                  254

-------
     The data of Figures 14, 15, and 17 are for runs made with sodium
hydroxide.  Two concentrations of sodium hydroxide used were 10"^ molar
and a 52 solution.  Comparing these figures, no statistical difference
in the rate of absorption of NO and N02 relative to changes of temperature,
solution or slurry concentration, and packing size were observed.

     Figure 16 shows the rate of absorption of NO and NO- when a 10~  molar
solution of sodium hydroxide is spiked with sodium chloride to 2/3 saturation
at the temperature of 1.25°F.  The reason for this run was that the literature
quoted the possibility of HNO- mist formation during absorption of NO and NO .
The 2/3 saturated NaCL solution's H20 vapor pressure is well below that of
the inlet gas stream, thus providing a nucleating site for HNO- formation.
This approach resulted in a small decrease in the rate of absorption of NO
and N02.  The rate of absorption was found to be statistically different
(lower; from the rates of absorption using magnesium and calcium hydroxide.

     Figure 18 shows the rate of absorption of NO and NO- when distilled
vater is used as the scrubbing agent.  One can see that the absorption
data are very similar to the other data with the hydroxide solutions.

     A statistical analysis of the data  Ci-e. data of Figures 2 through 18)
indicates the following:

          (a)  The concentration of NO and NO- in the flue gas stream  signifi-
          cantly affects the rate of absorption; NO and N02 at 250 ppm
          in flue gas is absorbed at a lower rate that 750 ppm NO and N02.

          (hi  The type of hydroxide significantly affects the rate of NO and
          NO, absorption, i.e., NaOH is more effective than Mg(OH>2, Mg(OH)2
          is better than CaCOHK, Ca(OH)2 is better than water.  On this
          particular point we should like to mention that we believe this
          notable difference was probably associated with the presence or
          absence of precipitate formation mentioned previously.  Precipitate
          formation during a run changes the flow patterns and causes channeling
          which produces changes in the  rates of absorption.

          Ccl  Over the ranges employed, changes in slurry or liquid temperature,
          changes in slurry or liquid flow rates, and changes in packing produced
          no statistical differences in  the absorption of NO and N02«

     The percentage removals of NO and N0« expressed as total NO^ removed
 (shown  in Figures 2 through 18) are as follows for the feed concentration
studied:
                                 255

-------
                              FIGURE K
           RELATIONSHIP BETWEEN  RATE OF AQUEOUS ALKALINE
           SCRUBBING OF NO AND  NOg AND INITIAL CONCENTRATIONS
           OF EQUIMOLAR  AMOUNTS  IN FLUE  GAS
    -3
   10X10
        9
        8
        7
        6
I
O
fc
O
tn
oo
u.
o
UJ
10X10
    9
    8
    7
 -5
10X2.5
        SCRUBBIN0  LIQUOR
        BERL SADDLE 3.IZE
        LIQUID TEMPERATURE
        LIQUID FLOW RATE

        6AS  FLOW  RATE
 -4
10 M  NaOH
3/8-
IOO°F
170 ml/mln
20  SCFH
                                                O  O
                                                D
                                             9
                                        O
                                        D
                                                     O  NO
                                                        NO,
          IXIO
                                           56789 10X10
                                                               ,2
                                                                   256
           INITIAL CONCENTRATION of NOandof N02 , MOLE PERCENT

-------
                           r iouK£ I b

        RELATIONSHIP  BETWEEN RATE OF AQUEOUS ALKALINE
        SCRUBBING OF NO AND NOg AND INITIAL CONCENTRATIONS
        OF EQUIMOLAR  AMOUNTS  IN FLUE  GAS
 -3
10X10
    9
    8
    7

    6
SCRUBBING  LIQUOR

BERL SADDLE SJZE

LIQUID TEMPERATURE

LIQUID  FLOW  RATE

6AS  FLOW  RATE
     5 -
       .
z
O
E
tt
0
(A
CD
*
U.
0
III
^
Q:







^
10X10
9
8
7

e
   -5
 10X2.5
-4
10  M  NaOH
   it
3/8

125 F
I70ml/min

20 SCFH
                                                a
                                               O
                                            a
                                            O
                                         a
                                         O
                                     D
                                     O
                               a
                               O
                         e
                                                 O  NO
                                                  D   NO,
       I xlO
                               56789  10X10
                                                   .2
        INITIAL CONCENTRATION of NOandof NO2, MOLE  PERCENT
                                                               257

-------
                   FIGURE 16
 RELATIONSHIP BETWEEN  RATE OF AQUEOUS ALKALINE
 SCRUBBING OF NO AND NOg AND INITIAL  CONCENTRATIONS
 OF EQUIMOLAR  AMOUNTS  IN FLUE GAS










\
^
JW
o
E

.d
****

1
z
o
fc
or
o
m
^
•a
II.
0
UJ
K

or








-3
10X10
9
8
7

6

5
4
(VI

J= 3
\





2






^
10X10

9
8



e

5

4
3
-5
-4
aCRUSSINfl LIQUOR- |o M NoO H spiked to2/3 saturoted
BERL SADDLE S.IZE - 3/8"
LIQUID TEMPERATURE - ,ocrec
125 r
_ LIQOI° FLOW RATE - .70 ml/mln

GAS FLOW RATE ~ SCFH


a
O
^



O
_. a
o
-
O
a

a


O
.
O

a

•

0
O NO
-
a wo2
"
_

10X2.5 . _o „ , >i R e 7 B 9 1C
                                             »    -2
I X4O
INITIAL  CONCENTRATION of NO and of NOg , MOLE PERCENT
                                                     258

-------
                           FIGURE  17
        RELATIONSHIP BETWEEN  RATE OF AQUEOUS ALKALINE
        SCRUBBING OF NO AND NOg AND INITIAL  CONCENTRATIONS
        OF EQUIMOLAR  AMOUNTS  IN FLUE GAS
 -3
I0x|0
    9
    8
    7
    6
       SCRUBBINO  LIQUOR -  5% NoOH
       BERL  SADDLE  S.IZE
       LIQUID TEMPERATURE
       LIQUID FLOW RATE
       GAS  FLOW  RATE
4mm
100 F
l7Oml/mln
20 SCFH
0
E
a:
o
OQ
^
U.
O
Ui
Jj
(C





^
10X10
9
8
7

e.
 -5
10X2.5
       I xlO
                                             n
                                             O
                                                  O  NO
                                                  a  NO,
                                        56789  10X10
                                                            -2
        INITIAL CONCENTRATION of NOandof NO2 , MOLE PERCENT
                                                                259

-------
                    FIGURE 18
 RELATIONSHIP BETWEEN RATE OF AQUEOUS ALKALINE
 SCRUBBING OF NO AND NOg AND INITIAL CONCENTRATIONS
 OF EQUIMOLAR  AMOUNTS  IN FLUE  GAS

















0
e

£

|
I
z
0
t
or
o
OQ
**
U.
O
UJ
F-

or







-3
10X10
9



7

6

5
4
M
H-


\




2







•A
10X10
9
8

7

e

5

4
3
-5
I/N vO R 1
SCRUBBINO LIQUOR - DISTILLED H.,0
- u ^
BERL SADDLE 3.IZE - 3/8
" o
LIQUID TEMPERATURE - |25 F

LIQUID FLOW RATE - ,70 a 350 ml/min

GAS FLOW RATE ~ '£> SCFH
rf
9

O
O




O
a
-
a
0





0
-
a
-

»

-
O NO
~
a NOO
2

i i i t . t i t I
I xlO"
5  6  7  8  9  10X10
                                                 -2
INITIAL  CONCENTRATION of NOandof N02, MOLE PERCENT
                                                     260

-------
              NO and NO, Concentration in Flue Gas             % Removal

                             250 ppm                           12 to 14
                             500 ppm                           17 to 19
                             750 ppm                           27 to 29
CONCLUSIONS
     The main conclusion drawn at present from our test is that equimolar
concentrations of NO and NO- in flue gas between 250 ppm and 750 ppm are
scrubbed by aqueous solutions of CaCOH),,, Mg(OH)_, and of NaOH with poor
overall scrubbing efficiency.  Since no change in absorption was observed
due to changes in liquid flow rate, and the large effect of NO  partial
pressure on scrubbing efficiency was found, this may be indicative of gas
phase reaction limitations.

     A study of the kinetics of the reactions taking place in the gas
phase might yield the rate limiting mechanism.  Should such a mechanism
or mechanisms be found, then steps to catalyze these reactions possibly
might improve the alkaline scrubbing efficiency sufficiently to make alkaline
scrubbing an acceptable abatement process for low concentrations of equimolar
quantities of NO and NO,.
                                 261

-------
                    BIBLIOGRAPHY
1.  Bartok, W., et al, Final Report, "Systems Study of Nitrogen
    Oxide Control Methods For Stationary Sources" - Volume II,
    Contract NO. PH-22-68-55, November, 1969.

2.  Sherwood, T. K. and Pigford, R. L., "Absorption and Extraction",
    Chem. Eng. Series, McGraw Hill, 1952.

3.  Ganz, S. N. and Mamon, L. I., Zhur Priklad Khim, 30, p.  369,  1957.

4.  Ganz, S. N. and Kuznetzov, I. E., Tr Dnepropetr Khim - Takhol.,
    Inst, 16, p. 17, 1963.

5.  Ganz, S. N. and Lokshin, M. A., Zhur Priklad Khim, 30, p.  1525,
    1957.

6.  Ganz S. N. and Dravchinskaya, S. B., Zhur Priklad Khim,  28,
    p. 145, 1955.

7.  Andrew, S. P. S., and Hanson, D,, Chem. Engr. Sci., 14,  pp. 105-113,
    1961.

8.  Vfettera and Uglum, J. Analytical Chemistry. 42, (3), pp. 335-340,
    March 1970.
                                 262

-------
         REMOVAL OF  SULFUR DIOXIDE FROM STACK GASES
            BY  SCRUBBING WITH LIMESTONE SLURRY:
                  USE OF ORGANIC ACIDS
                             By
               J.  D. Hatfield and J. M. Potts
              Division of Chemical  Development
                 Tennessee Valley Authority
                   Muscle Shoals, Alabama
               Prepared for Presentation at
Second International Lime/Limestone Wet Scrubbing Symposium
     Sponsored by the Environmental Protection Agency
                  New Orleans,  Louisiana
                    November 8-12,  1971
                         263

-------
               REMOVAL OF SULFUR DIOXIDE FROM STACK GASES

                  BY SCRUBBING WITH LIMESTONE SLURRY;

                         USE OF ORGANIC ACIDS

                                   By

                     J. D. Hatfield and J. M. Potts
                    Division of Chemical Development
                       Tennessee Valley Authority
                         Muscle Shoals, Alabama
                                ABSTRACT


          A survey of weak organic acids showed that many of them are
capable of dissolving CaC03 and MgC03 and therefore are of potential
value in improving the  scrubbing of S02 by limestone slurries.

          The physicochemical properties of four selected acids—benzoic,
phthalic, glycolic, and adipic—and their calcium and magnesium salts were
determined with respect to stability, solubility, effect on the oxidation
rate of sulfite to sulfate, and the nature of aqueous species.

          Small-scale tests were made to demonstrate the effectiveness of
some of the organic acids in improving absorption of S02.  The proportion
of benzoic acid required to effect an improvement was determined in closed-
loop spray tower operation under conditions designed to remove about 50$
of the S02 without additive.  The addition of 10$ of the-stoichiometric
amount of benzoic acid  to solubilize the feed limestone increased the S02
removal by about 10 percentage points; 50$-°f tne stoichiometric acid gave
a 22-percentage point increase, but further improvement with added acid
was 'small.  The use of  benzoic acid resulted in a higher degree of oxidation
of the product solids in the scrubbing operation, although laboratory tests
with oxygen indicated that oxidation of dissolved sulfite is inhibited by
the acid.  The high oxidation in the scrubber tests presumably resulted
from the lower pH brought about by the benzoic acid addition and the con"
sequent larger amount of dissolved sulfite available for oxidation.

        .  Organic acids not only improve S02 removal but also offer the
opportunity of getting  complete utilization of limestone by using enough
acid to dissolve all the feed limestone, in which case the only solid
phases are calcium sulfite and sulfate.  The increased oxidation in the
scrubber should also improve slurry settling rate and avoid the danger of
limestone surface blinding.  Work on these possibilities is continuing.
                                 264

-------
               REMOVAL OF SULFUR DIOXIDE FROM STACK GASES

                  BY SCRUBBING WITH LIMESTONE SLURRY:

                         USE OF ORGANIC ACIDS

                                   By

                     J. D. Hatfield and J. M. Potts
                    Division of Chemical Development
                       Tennessee Valley Authority
                         Muscle Shoals, Alabama


          The dissolution of limestone is one of the limiting reactions in
the removal of sulfur dioxide from stack gases by scrubbing with a lime-
stone slurry.  The equilibrium solubility of limestone in water (Table l)
does not give a great driving force to effect solution; consequently,  it
has been found advantageous to have a surge vessel to permit dissolution
of limestone and precipitation of calcium sulfite - sulfate before sending
the slurry to the scrubber.  The low concentration of basic components in
the solution also necessitates a higher ratio of liquid to gas (L/G) in
the scrubber than is required for other scrubbing agents such as alkali
metal salts or ammonia.

                                 TABLE I

              Solubility of CaC03 in Water at 50°C (l22°F)
Ca
ppm
9-5
16.8
30.7
57.7
111.1
221.7
1020.3
PC02
atm
0.000136
0.000803
o.oo466
0.0271
0.159
0.950
36.3

PJL
8.5
8.0
T..5
7.0
6.5
6.0
5.0
              a Calculated using Radian1s equilibrium pro-
                gram (PB 193029) for the system CaO-C02-H20.

          The solubility of limestone increases rapidly with decrease in
pH as shown in Table I.  A similar increase can be achieved by adding another
acid, provided that:

          (l) the calcium salt of the added acid is soluble, and

          (2) the ionization constant of the added acid is greater
              than that of carbonic acid, H2C03 (K! = lj-,4 x 10~7)
                                     265

-------
By  stipulating also that  the added acid be weaker than sulfurous acid, H2S03
(KX =  1.3 x 10"2), the limestone slurry scrubbing of stack gases might well
be  improved significantly by the addition of a small amount of an extraneous
acid (an acid stronger than H2S03 would volatilize S02 from solution and
prevent its precipitation as the calcium salt).

          This paper describes a survey of possible acid additives, labora-
tory tests of properties  of selected acids and their salts, and small-scale
tests  of selected additives to determine their effectiveness in S02 removal
and other operational characteristics.
Survey of Acids and Exploratory Tests

          Of  the  132 acids- (intermediate in strength to H2S03 and H2C03)
that  are listed in chemical handbooks, many are organic acids whose physi-
cochemical  properties in aqueous systems have not been studied.  Many can
be  eliminated from consideration as additives to limestone slurries because
of  cost or  difficulty to produce; others are of too high molecular weight
to  be soluble in water; and others, such as oxalic acid, have extremely
insoluble calcium salts.   The ideal additive would be

          (l)  soluble in water and limestone scrubber slurry;

          (2)  nonvolatile and stable under scrubbing conditions;

          (3)  not precipitated by Ca, Mg, or other metals likely
               to be found  in the solution; and

               commercially available and of reasonable cost.

          Solubility of Carbonates at 25°C:  Measurements were made of the
ability of  several organic acids to dissolve calcium carbonate and magnesium
carbonate at  room temperature (~78°F).  The acids were those available in
our laboratory at the time and consisted of both aliphatic and aromatic
acids (monobasic and polybasic) and covered a wide range in molecular weight.
The procedure  consisted in adding 1 gram of the organic acid (or 1 ml if the
acid  was a  liquid) to 100  ml of water and then adding 5 grams of reagent
CaC03 or MgC03 to the solution or mixture.  The mixtures were allowed to
stand for 1 week with occasional agitation by hand to promote dissolution.
Aliquots of 10 ml of the clear solution were taken after 3 days and after
1 week from each mixture and were analyzed for the dissolved metal.  The
pH was measured before addition of the carbonate and after the 7-day sample.
The results are presented  in Table II.
                                  266

-------
                                                  TABLE II


                   Use of Organic Acids in the Dissolution of Calcium and Magnesium Carbonates
pH Of

Acid

acid
Formula solution
Sol., Ca, vt. %,
After 7 days Sol., MR, wt. 1>,
after mole ratio
3 days
7 daye
acld:Ca
After 7 days
after mole rat^o
pH 3 days
Aliphatic monocarboxyllc acids
Formic
Acetic
Chloroacetlc
(Rycolic
ftenylacetic
Proplonlc
Lactic
Butyric
Caproic
Qluconic

Acrylic
Oleic

HCOOR
C%COOH
CICHaCOOH
HOCHaCOOH
C.aHsCHaCOOH
CaH5COCH
CHgCHOHCOOH
CaHTCOOH
CsHnCOOH
CHeOHCCHOH)*
COOB
CHs=CHCOOH
CH3(CHa)7C=C
(CHZ)TCOOH
2.30
2.30
2.15
2.1*0
2.70
2.90
2.50
2.90
3.00
2.86*

2.6O

5.80*
O.U8
0.32
0.21
0.22
0.13
0.28
0.21
0.22
0.17
0.06

0.31

0.01
0.5U
0.37
0.21
0.25
0.15
0.28
0.23
0.21
0.17
0.06

0.31

0.01
1.7
1.9
2.0
2.1
2.0
1-9
1.9
2.2
2.0
3.U

1.8

16.7
7.00
7.20
7.15
7.>»0
7.^0
7.20
6.90
6.65
6.15
7.00

6.60

6.60
0.1*0
0.31
0.20
0.2U
0.11*
0.27
0.21
0.23
0.15
0.06

0.25

0.03
7 days

O.Ul
0.31
0.20
0.27
0.13
0.21
0.21
0.21
0.12
0.08

0.21

0.03
acld:Mg"

l.l*
1.1*
1.3
1.2
1.1*
1.6
1.3
1.3
1.7
1.6

1.6

3.fc
Price
pH dollars/lb.

7.60
7.65
7.85
7.50
8.15
7.75
7.70
7.50
7.90
8.00

7.70

8.1*5

0.11*7
0.09
0.21
0.10
0.68
0.1U7
0.275
0.33

0.1U5

0.31

0.23
Aliphatic polycarboxyllc acids
Oxalic
Succinlc
lartaric

Malic
Pumarlc
Malelc
Adipic
Citric

COOHCOOH
COOHCHaCHaCCOH
COOHCHOHCHOH
COOT
COOHCHOHCHsCOOH
COOHCH=CHCOOH
HOaCCH=CHCOaH
COQH(CH2)4COOH
(COOHCH2)eOOH
CCOH
1.75
2.30
2.25

2.1*0*
2.35a
1.95
2.80
2.25

0.01
0.27
0.01

0.09
0.25
0.23
0.28
0.02

0.01
0.32
0.01

0.09
0.28
0.31
0.28
0.02

1*1*. I*
1.1
26.7

3.3
1.2
1.1
1.0
10.

7.70
7.35
7.70

7.60
7.30
7.25
7.20
7.80

0.13
0.29
0.21*

0.22
0.28
0.31
0.26
0.2U

0.08
0.30
0.26

0.2U
0.27
0.30
0.21
0.21*

3-1*
0.69
0.62

0.76
0.78
0.70
0.79
0.53

8.35
7.80
7.85

8.1*0
8.00
7.70
7.80
8.50

0.22
0.62
O.U15

0.315
0.225
0.1*8
0.18
0.33

Aromatic acids
Benzole
Salicylic
P-Amino
benzole
Dinitro
benzole
GaUic

ftthalic
crKaphthoic
CeHgCOOH
HOCaRtCOOH
NHaCeBeCOOH

(HOaJaCeHg
COOH
(HO)3CeH7
COOH
CeRe(COOH)2
CioH7COOH
2.75*
2.60*
5.50s

2.80a

2.85

2.l*5B
3.75a
O.Ik
0.11
0.15

0.09

0.11

0.12
0.02
0.17
O.ll*
0.15

0.10

0.11

0.06
0.01
1.9
2.1
1.9

1.9

2.1

1*.0
23
7.35
7.35
7.M)

7.1*5

6.20

7.60
6.70
O.OU
0.21
o.it*

0.09

0:12

0.22
0.08
0.13
0.13
O.lU

0.09

0.11

0.21
0.06
1.5
l.U
1.3

1.3

1.3

0.70
2.3
8.20
8.15
8,00

8.10

8.15

8.05
8.30
0.215
O.H25
1.72



2.65

0.12

  1 gram of acid did not completely dissolve  in 100 ml.

b
  Based on solubilities  after 1 week,  assuming  all acid dissolved.
                                                  267

-------
           The pH of  the acid  solutions varied  from 1.75 to 3-80 (many
of  the  higher molecular weight  acids  did not completely dissolve before
the carbonate addition).   The pH of the final  solutions or slurries
ranged  from 6.2  to 7.8 for CaC03 and  from 7.5  to 8.5 for MgC03.

           Most of  the mixtures  dissolved very  little after the third day.
The Ca-Mg content  of a few samples actually appeared to decrease between
3 and 7 days.  Some  of thase  variations may be due to sampling and analytical
errors, while others may  be real.  Despite the apparent lack of equilibrium
in  many of the mixtures,  the  final mole ratios (acidrmetal) shown in
Table 2 are indicative of the potential of the acid to dissolve the car-
bonate.  For the monobasic acids of appreciable solr.bilizing power, this
ratio was approximately 2 for Ca and  1.5 for Mg; for the dibasic acids, the
corresponding ratios were 1 and 0.75«  The results indicate that those acids
having  a solubilization ratio of 2-COOH groups per atra of Ca should be
considered as potential limestone slurry additives as this is the stoichio-
metric  value.

           The data in Table IE also show that the low molecular weight acids
were more effective  in  solubilizing calcium or magnesium carbonate per gram
of  acid than were  the high molecular  weight acids.  This is due to (l)
more carboxy groups  per gram  and (2)  increased solubility of the low molecular
weight  acids in  water.  The substitution of chlorine, hydroxy, or phenyl
groups  for hydrogen  decreased the solubilizing power per gram as expected,
while acids whose  calcium salt  is insoluble were particularly ineffective
(e.g.,  oxalic,  tartaric,  and  citric acids).

           Batch  Exploratory Tests;  Laboratory tests in batch apparatus
were conducted  to  compare the effectiveness of various organic acids in
improving scrubbing  efficiency  and limestone utilization.  The tests were
made by adding  the acid to 1$ limestone slurry and using the mixture to scrub
a synthetic flue gas, with stirring,  in a gas  scrubber bottle.  The S02
content of the gas,  before and  after  scrubbing, was monitored by an infrared
analyzer.  The S0£ removals and CaO utilizations were calculated by two
methods,  one based on chemical  analyses of the spent scrubber liquor and
the other on evaluation of the  infrared recorder chart; the latter method
appeared to be more  reliable.  The acids tested were salicylic, citric,
formic, benzoic, and acetic.  The results are  presented in Table III.

           Although the data are somewhat scattered, there is an indication
that salicylic and citric acids caused a decrease in the utilization of
limestone; citric  acid, at very low concentration (acid: CaO mole ratio =
0.06),  appeared  to have no deleterious effect. Formic, benzoic, and acetic
acids increased  the  utilization of limestone,  the overall S02 recovery, and
the  time  to 50$  removal.   Reasons for the poor performance of salicylic and
citric  acids may be  the high  value of the ionization constant, the low solu-
bility  of salicylic  acid,  and the ineffectiveness of citric acid in dissolving
CaC03 ( Table II \  Formic  and  acetic acids are  quite volatile and could result
in substantial losses in  scrubbing a  stack gas.
                                    268

-------
                                                                   TABLE III

                                          Laboratory-Scale Batch Tests of the Use of Weak Organic Acids
ro
»
10
Test No.
Acid added

w« ft
lonizatlon constant •
Amount added
Grams
Males x ID3
Limestone fed
Grams
Moles CaO x 103
W«±fvr fed a
IICbwGJ. * CLLj g£ •
Scrubber conditions
Gas rate, l./min.
Solution temperature, °F.
Total S02 fed
Grams
Moles x ID3
Test results d
Maximum S02 removal, #
Time to 50? removal, ndn.
Scrubber solution pH x
Overall S02 removal, #
By chemical analysis f
By infrared analysis
CaO utilization based
ont S02 .removal, #g
By chemical analysis
By infrared . analysis
CaO utilization attributable
to acid addition, %
By chemical analysis
By infrared analysis
32o
None


—
-

1.0
9.2
100
4.375
121-138
0.675
10.5
84
20
3-1
82
65

94
74

-
3o4
Salicylic
1.06 x 10-3

1.0
7.2

1.0
9.2
99
3.826
120-125
0.620
9-7
81
21
4.2.
64

67

-7
330
}41
37)
Citric


2.1
10.9

1.0
9.2
100
4.273
110-119
0.509
7.9
88
15
3.8
56
75

48
64

-46
-ID
8.4 x. 10

1.0
5.2

1.0
9.2
99
4.253
122-124
0.443
6.9
83
13
3-9
83
74

62
55

-32
-19
-4

0.1
0.5

1.0
9.2
99-9
4.743
120-125
0.686
10.7
87
20
4.2
88
75

102
85

8
11
.w
Formic
1.76 x 10 "4

1.0
21-7

1.0
9-2
"•99
yyy
31**
y*j
Benzole








4.273 4.743
121-125 118-120
0.775
12.1
88
23
4.0
86
75

113
99

19
25
0.775
12.1
92
21
3.7
104
77

137
101

43
27
6.3 x

2.2
18.0

1.0
9-2
97-8
4.253
123-126
0.731
11.4
88
22
4.5
69
65

85
80

-9
6
ID'S

O.2
1.6

1.0
9-2
99.8
4.253
120-125
0.731
11.4
86
22
4.2
67
73

83
90

-11
16
.7fj
Acetic
1.75 x ID'

1.0
16.7

1.0
9-2
99
JJC-









4.253 4.743
120-125 116-125
0.797
12.4
88
24
5.9
58
76

78
102

-16
28
0.775
12.1
93
21
3-7
105
76

138
100

44
26
a First hydrogen ion at 25° C.; value for HsSOs is 1.7 x 10
,. PhyslcjB, 21st Edition. % Chemical Rubber Publishing Co.)
                                                                 w 9
                                                                    at 25° C., for
          ysj,             . %                              .
        CruSSeg (72% -200 mesh) Spring Valley limestone; 51-5$ CaO, 1.3# MgO,
      c Simulated flue gas; 0.3# S02, 3«5# Oa, 80.256 N, l6.0# C02  (dry basis).
      d Maximum proportion removed as indicated by infrared analysis.
      ® By material balance.
      1 By intagration of recorder chart.
      8 100  (moles S02 removed)/ (moles CaO input).
      n (CaO utilization with  acid) - (CaO utilization without acid).
     is 3-5 x  ID"7 .at 18° C. (Handbook of Chemistry and

COa, 4# S102, 0.05$ S.

-------
           Continuous Exploratory Tests;  Exploratory tests  with citric,
 formic,  benzoic and acetic acids were carried out in a small  pilot plant
 consisting of two unpacked spray towers in series with the  recirculating
 slurry impinging on a deflector plate to give a flat spray.   Each- scrubber
 was k inches in diameter by 2 feet long.  The slurry of 1$  limestone was
 recycled at 9 1/min to scrub about 5 cfm (L/G =—lj.00) of simulated flue
 gas.  The procedure in these tests was to start with fresh  slurry in the
 surge tanks and adjust system flows and conditions to give  maximum removal
 of S02.   Operation was carried out without slurry addition  until removal
 started falling off, and then a flow of fresh slurry was started in an
 open-loop manner; the slurry flow rate was varied as necessary  to maintain
 S02 removal at approximately the initial level.  Each test  represents
 about 5 hours of steady-state operation.  The amount of acid  added was
 10$ of the stoichiometric amount for reaction with the limestone feed.
 The results and conditions are presented in Table IV.

           Citric acid gave better removal at 10$ stoichiometric addition
 than did the control test, but utilization of limestone was poor and a
 high feed rate of makeup slurry was required to maintain S02  removal;  in-
 creasing the citric acid to 50$ of the stoichiometric limestone feed gave
 even worse results.  Formic, benzoic, and acetic acids were more effective;
 the S02 removal and limestone utilization were higher at lower  makeup  feed
 rate.  When the supernatant liquid from a previous acetate  run  (No.  6k)
 was used to slurry the limestone, the results were poorer,  indicating  a
 possible loss of acetic acid due to volatilization.

           The product solids contained a much higher proportion of sulfate
 in those tests in which organic additives gave increased utilization of  the
 limestone and improvement in S02 removal.  Since sulfate crystals are  larger
 than sulfite, this phenomenon may improve the settling rate of  the slurry  in
 waste ponds; it may also reduce the possibility of sulfite  pollution if  there
 is seepage into ground water or streams from the pond.


 Physicochemical Properties of Four Selected Acids

           Pour acids were chosen from the exploratory tests to  determine
 their properties (and those of their Ca and Mg salts)  that  are  important to
 use in limestone slurry scrubbing.

                    	Aliphatic                 Aromaf-.ir.
      Monobasic    Glycolic,  CH2OHCOOH      Benzoic,
      Dibasic       Adipic,  HDOC(CH2)4COOH   Phthalic,
These were selected on  the basis of potential stability and nonvolatility,
ability  to solubilize CaC03 and MgC03 (TableII)  and their commercial availa-
bility at a reasonable  price.
                                    270

-------
ro
                                                               TABLE IV

                     Effect  of Weak Acid Type in Increasing Solubility of Limestone Used in Scrubbing Stack Gasa
A
Acid Additive0
Test No.f
Makeup slurry feed rate, 8 l./hr.
Stack gasn
Inlet temperature, °F.
SOa content, p. p.m.
To first-stage scrubber
From second-stage scrubber
Indicated SOa removal, $
Slurry from first-stage scrubber
pH at end of period
Composition, J % by wt.
CaO
Total S
Sulfite S
Indicated oxidation, $
Indicated CaO utilization, %
Chalk
None
66
16.5

300

3395
4io
88

5.6

33
16.1
11.2
30
85
Limestone13
None
63
20.6

305

3200
615
8l

4.3

32
0.2
_
- 1
(38)
Citric
67
19.8

305

3290
290
91

4.6

4o
2.6
1.3
50
n
Citrlcu
69
21.7

302

3255
730
78

4.7

26
0.1
, 0.04
_
•
Formic
61
14.0

300

3300
380
89

4.1

26
14.2
0.1
99
97
Benzole
70
8.0

304

3235
305
91

3.8

39
21.5

78
96
Acetic
6k
12.6

310

3450
395
89

3.9

32
17.2
0.5
97
95

Acetic6'
65
18.0

295

3^50
520
85

4.2

34
7-0

36
36
          a Two  spray scrubbers  k-  inches  in diameter by 2 feet long  in  series, recycling 1$ limestone slurry at
            9 l./min.  to  each; about  5  c.f.m.  stack gas produced by  burning a mixture of natural gas and fuel oil to which
            carbon disulfide was added.
          13 Spring Valley limestone from  Ralph Rogers and Company, Inc., Tuscumbia, Alabama; 76$ -200 mesh.
          0 Amount used vas 10$  of the  stoichiometric amount for reaction with the feed limestone unless noted otherwise.
          ^ 50$  of stoichiometric  to  react  with  feed limestone.
          e Supernatant liquor separated  from  test  6^4- was used to slurry the limestone.
          f Tests were continued for  about  5 br. at steady state.
          S 1$ limestone  content.
          "• Values averaged over steady-state  period.
            Composite  over steady-state period.
          £ Solids separated by  filtration.
          ? Overall oxidation (both scrubbers).
            Estimated by  similarity to  other tests.

-------
          Stability of Acids;  Two series of tests were made to determine
 the stability of the four acids under conditions considered to be more
 drastic than those likely to be encountered in scrubbing operations.
 These accelerated tests were made with both solutions and melts1 of the
 acids as follows:

              	Solutions	Melts	
 Conditions   Benzoic  Phthalic  Adipic  Glycolic  Benzoic  Phthalic  Adipic.

Acid concen-                                                      a
 tration, %      0.3      0.1       2       70       100     112.2     100
 Temperature,
 °C             75       75        75       55       136     212       162
 Time, hr         5!       53"       59       28        18       0.5      10
 Composition
 of gas, $
  S02            50       67        50       67        67      67        67
  °2            25       33        50       33        33      33        33
  N2            25        0         00         00         0
 Gas flow rate,
 ml/min        200      120        40       45        45      45        45


  Phthalic anhydride was used rather than phthalic acid.

          The ultraviolet absorption of the solutions and melts after treatment
 was determined as a means of detecting any decomposition of the acids. The
 solutions after  treatment had a strong odor of S02.  Since the S02 interfered
 with the absorbance measurements, it was eliminated by boiling the solutions
 for about 10 minutes. . Results of the measurements, shown in Figure 1 for
 benzoic acid,2 indicate that there was no chemical attack on the acids.  There
 were no new absorption peaks observed in the treated samples as there should
 have been if decomposition products were present.  There was a possibility
 that some glycolic acid was volatilized from the 70$ solution, but this could
 have been due to our inability to attain the same concentration of the acid,
before and after S02 treatment, for the ultraviolet absorbance comparison.
 It was concluded that all four acids were relatively stable.


  Glycolic acid was available only as a 70$ aqueous solution; therefore it
  was not included in the melt tests.
o
  Only the curves for the melt and for benzoic acid volatilized from the melt
  are shown.  Similar results, however, were obtained for the solutions.
                                   272

-------
                 A -  Pure  bensoic  acid

               B.C -  Melt  and  condensed benzole acid
1.0
                                                   , 0.1
           200
                                                     0.0
       250
Wavelength, ran
                    FIGURE 1

      Ultraviolet Absorption of Benzole Acid
                        273

-------
           Solubility of Salts and Solubilizing Power of Acids:  The
 calcium saltsi of  the  four acids were solubilized in water at 25° and
 50°C with  the following results:

   Solubility and  pH of Calcium Salts of Organic Acids at 25° and 50°C

                Solubility, % salt by wt.     pH of saturated solution
     Acid          25°C           5(TC                         *
Benzoic
Glycolic
Phthalic
Adipic
2.80 (4000)
1.52 (3200)
0.33 (650)
3.U* (6850)
3-83 (5^50)
3- 2k (6850)
0.25 (500)
1.98 (4300)
7.38
T.P8
6.36
7.57
6.83
6.60
7.U
7.55
    o
     Values  in parentheses are ppm Ca in saturated solution.

           The  solubility of  the calcium salts of the monobasic acids (ben-
 zoic and glycolic)  increased and the pH decreased with temperature.   The
 salts of the dibasic acids (phthalic and adipic) decreased in solubility
 with temperature.   Benzoic,  glycolic, and adipic acids are especially
 promising at 50°C,  but phthalic acid gives only about 1/10 of the calcium
 concentration  as compared with the other three acids.

           The  solubilizing power of the four acids toward CaC03 and MgC03
 was determined at three concentrations of acid (0.05, 0.1, and 0.2$) at
 25°C and at  one concentration (0.1$) at 50°C.  The results are given in
 Table V.   In each case approximately: two carboxy groups were required to
 dissolve an  atom of calcium  from CaC03, regardless of temperature or concen-
 tration of acid; for MgC03,  somewhat less than two carboxy groups were re-
 quired.   The pH' s of solutions saturated with CaC03 were about 8 at 25°C and
 about 7 at 50°C; for the. solutions saturated with MgC03 the corresponding
 pH values were 9 and 8.3 at  25° and 50°C, respectively.

           Effect of Organic  Acids on the Rate of Oxidation of Sulfite to
 Sulfate:   The  rate  of oxidation of sulfite in solution to sulfate by pure
 oxygen  was determined in the presence of the various organic acids.   The
 tests were made at  pH = k by adding sufficient powdered CaC03 to maintain
 the pH  constant as  oxidation proceeded.  The comparison is made in Figure 2
 for initial  sulfite concentration of 0.0123 molal at an oxygen flow rate of
20 ml/min.


  The magnesium salts were not tested because they could not be prepared in
  sufficient purity for testing.
                                   274

-------
                                                    Acid,  0.1$
                                                 O  Benzole
                                                 A  phtnalic
                                                 D  Adipic
                                                     Glycolic
                                                 A  None
10
                                     Time, hours
                                       FIGURE 2

               Effect of Organic Acids on the Oxidation of  Sulfite  to
               I I I II • I I I   M .•< Mil • •• III I I  I I • •  I •• •' •  • •*• ^ I I 1 I • • «-™ -™- I ,« Ml  -III I I - - I —1- 1 -   - -  I _ _ "I

                     Sulfate in Calcium System at 50°C and  pH i<-.0
                                 275

-------
                                 TABLE V

            Solubllization of CaC03 and MgC03 by Organic Acids

                 Effect of Concentration of Acid at 25°C
                         ppm Ca when acid     ppm Mg when acid
                           concn., "jo, is        concn., $, is
              Acid       0.05   0.1   0.2     0.05   0.1   0.2
            Benzoic        97    180   340      140   210   330
            Glycolic       IJO    2TO   550      170   280   460
            Phthalic       l^O    240   480      160   250   420
            Adipic         150    290   570      170   280   470
           Effect of  Temperature When Acid Concentration is 0.1%

                       pH       ppm Ca        pH        ppm Mg
           Acid   25°C  50°C   25° C  50°C  25°C  50°C  25"C  50°C
Benzoic
Glycolic
Phthalic
Adipic
••••••—^•M
T-95
T.98
8.00
7-92
^•••^^•i^
7.18
6.69
6.7T
6.80
^•MM^HB
180
270
240
290
(^•••W^
180
350
250
320
««•«••••
9-1
9.1
9.2
9.1
8.5
8.2
8.4
8.3
210
280
250
280
200
300
260
290
           Benzoic,  glycolic,  and adipic  acids, at 0.1% in solution, re-
 tarded the oxidation of sulfite; less  than 40$ of the original sulfite was
 oxidized in 5 hours when these acids were present, whereas 42$ of the sulfite
 was  oxidized in 1 hour when no additive  was  present.  Adipic acid has some
 retarding effect on oxidation, but not as great as the other three acids.
 The  curves for adipic acid and no additive do not plot into 100$ at zero
 time,  indicating an initial quiescent  period before the oxidation proceeds.
 The  first order rate constants for oxidation as determined by the curves in
 Figure 2 are:

                           Acid     K  x  10s, sec"1

                         None             43.0
                         Benzoic          2.4
                         Glycolic         2,9
                         Phathalic        4-3
                         Adipic          32.0

          A similar comparison of benzoic acid with no additive was made at
an initial  pH of 4.0 but without holding the pH constant during the oxidation.
                                  276

-------
Three replicas of the test without additive were made to determine the
reproducibility of the oxidation rates and also because of the diffi-
culty in obtaining identical starting solutions.  The results are shown
in Figure 3.

          When benzoic acid was present at 0.16$ the pH changed much less
rapidly as oxidation proceeded than when it was absent.  Again, the amount
oxidized in 1 hour with no additive was greater than that oxidized in
5 hours with benzoic acid present—indicating a retardation effect of the
rate of oxidation by benzoic acid.

          These results do not appear compatible with -those obtained in
exploratory tests on S02 absorption (Table 3V), in which the sulfate:sulfite
ratio in the product solids was higher when organic acids were used.  This
may have been due to the fact that excess sulfite crystals were present in
the absorption tests.  At the lower pH resulting from the acid addition
(on the order of 1.5)* this would produce a much higher dissolved sulfite
concentration in the solutions containing organic acid, and therefore might
promote the transition CaS03°0.5H20(c) -» CaS03'0.5H20(]_) -> CaS04-2H20(1) -*
CaS04»2H20(s) even if the oxidation rate in the solution phase were reduced
by presence of the acid.

          Aqueous Species:  The thermodynamic dissociation constants of the
acids1'2 at 25°C are as follows:

                         Acid     log Kg   log Kg.
                       Benzoic    -k
                       Glycolic   -3.831
                       Phthalic   -3.14    -5.40
                       Adipic     -4.42    -5.41

These constants denote the pH at which the activities of the undissociated
acid and the organic anion are equal; for example, the activity of benzoic
acid, HBz, equals the activity of the benzoate ion, 'Bz~, at pH 4.2, etc.

          Analysis of the solubility and pH data of Table V indicate that
the benzoate ion forms a weak complex, CaBz"1", for which the following
equilibrium constant is obtained:

              CaBz+ ?! Ca   + Bz"   K* = 1CT2-13 ± °-15 at 25°C

                                   K* = 1CT1'0  ±0'5    at 50°C

This compares with the only reported literature value of ~10-°'2 by Bunting
and Thong3 at 3°°c and an ionic strength of 0.4; this latter value becomes
10~°'8 at I = 0, using the Davies activity coefficients-

I Stability Constants, Special Publication No. IT, The Chemical Society (1964).
  Smolyakov and Primanchuk, Russ. J. Phys. Chem. 40 (3), 331-3 (1966).
  Can.  J.  Chem. 48, 1654-6 (1969).
                                    277

-------
100
 90 -
                                      Initial Solutions
                                       pH  S as
                               I  O
                              II  D
                             III  A
                              IV  --   k.05
                             Soln. IV contained 0.16$
                             benzole  acid
                                Time,  hours
                          FIGURE 3

         Oxidation of Sulfite to Sulfate in the

       System CaO-SOa-S03-iy) at 50"C With Og, Gas
                           278

-------
          The calculation of stability constants from solubility data re-
quires very accurate data both with respect to the component concentrations
and the pH of the solution.  Benzoic acid has very sharp ultraviolet ab-
sorption peaks, and its chemical determination is much more precise than
that of glycolic, phthalic, or adipic acids.  Even so, the dissociation
constant calculated from the data in Table V was not confirmed by the
solubility of the salt, CaBz2—again due to the sensitivity of the data.
No consistent values of dissociation constants for complexes of Ca or Mg
with glycolic, phthalic, or adipic acids were obtained from either the
dissolution of carbonates (Table V) or the solubility of the salts; in
many cases the data gave negative values for some of the assumed species.

          The significance of the value of the dissociation constant of the
complex CaBz , K*, is shown in the following tabulation of the amount of
calcium required to complex various amounts of the benzoic acid for two
values of K* when the benzoic acid concentration is approximately 0.1$:

            Degree of complexing      ppm Ca required if
                as CaBx+, %        K* = IQ-**-"3   K* = 1Q-J-*"

                     10                  30   ,        300
                     50                 300          3,000
                     90               3,000         30,000

         A further attempt to determine the presence of an aqueous complex
was made using both ultraviolet absorption and calcium ion specific electrode
with Job's method of continuous variation of the ratio Bz:Ca in solution.
The ultraviolet absorption data showed a strictly linear increase in absorbance
with total benzoate concentration with no breaks or irregularities due  to  a
significant amount of CaBz+  species.  The  specific ion electrode measurements
were not definitive but did  indicate a small amount of calcium that was  other
than ionic Ca"1"1";  this could  have been CaOH+, CaHC03+,  or other known  species
in addition to CaBz .  The conclusion is that if a complex between calcium
and benzoic acid  is formed,  it is a weak complex (highly ionized) that  defies
accurate measurements by  the methods attempted.


Small-Scale Tests with Benzoic Acid

         Continuous scrubbing tests with benzoic acid were made in the  5-cfm
two-stage scrubber described earlier except that commercial spray nozzles
were used which permitted operation at a liquid rate of TOO ml/min (37  gal/Mcf)
in each tower.  Limestone slurry  (with filtrate recycled) was added  to  the
second tower  to  scrub  the gas, and water was added to the first tower  to
humidify and  cool the gas before  scrubbing.^ The flue gas was obtained by
burning natural  gas, with S02 added from a cylinder to obtain about  3000 ppm
in the gas.   This procedure  simulated closed-loop operation under  conditions
of about 50$  S02  removal  so  that  the effects of an organic acid added  to the
limestone slurry  could be tested.  The results are shown in Table  VI.
                                    279

-------
                               TABLE VI

           Effect of Benzole Acid, on Sulfur Dioxide Absorption

                           by Limestone Slurry3


                                                 Test No.
           Recycle  slurry
             Benzole acid added, %°            0     50     75
             pH at  end of period              ^.5   3.7    4.3

           Stack  gas
             Inlet  temperature, °F            205   210    210
             S02  content, ppm
                To first-stage scrubber       3150  3150   3125
                From second-stage scrubber    1565   915    7^0
             Indicated S02 removal, $          50    7!     76

           n
             First  scrubber  of water recycling at about 700 ml/
             min; second of  limestone slurry recycling at
           , 700  ml/min.
             Percent of that required to solubilize lime for
             reaction with all the S02 entering the system.


         Adding benzoic acid to the slurry in an amount equivalent to that
required to solubilize 50$ of the lime required for reacting with the S02
feed  increased  the  S02 removal from 50 to 71$; 75$ of stoichiometric gave
a  further  increase  to 76$ S02 removal.  Other runs (not shown) with lesser
fractions  of benzoic acid were made with intermediate results. The results
of the  improvement  in S02 removal with the percent of stoichiometric benzoic
acid  for lime solubilization are shown in Figure Ij..  The results indicate
a  rather steep  increase in S02 removal for the first 25$ of stoichiometric
acid  and a leveling of the response beyond this amount.

         An extended closed-loop test was made using 20$ of the stoichiometric
amount  of  benzoic acid.  No  apparent loss in the efficiency of the slurry to
remove  S02 was  noticed in 18 hours, indicating that the losses of benzoic
acid were  probably  low.  A similar test using 150$ of the stoichiometric
benzoic acid indicated, by ultraviolet absorption analysis of the liquid,
about 1% loss per hour.
                                  280

-------
             30
ro
oo
             20
             10
First scrubber:  Water recycled at TOO ml./min.
Second scrubber:  2$ slurry of local limestone
 (76# -200 mesh, 51-5# CaO) recycled at
 760 ml./min.; makeup at 5 l./hr.; pH 3-7-5-0
Inlet gas:  3150 p.p.a. S02
                                                                            I
                              I
                           10          20          30          40          50          60          70
                          Percent of  stoichlometric benzole add to  solubillze CaO for all  sulfur  in
                                                    80
                                                           FIGURE k

                                            Effect of Proportion of Benzole Acid

                                     Added to Limestone Slurry on Sulfur Dioxide Removal

-------
         The vapor pressure of benzoic acid at 0.16$ was determined be-
tween kO° and 65°C (l04° and 149°F) and was expressed by least squares
by

                       log Vmz = 9-20 - lj-787/T

where T is the absolute temperature, °K.  This implies an energy of
evaporation of benzoic acid of 22 kcal/mole from solution, compared with
21-9 kcal/mole for sublimation of the pure acid.  The dependence of
Henry's law constant, h, for benzoic acid absorption becomes

                       log h = -10.27 + ^500/T

and the vapor pressure of 0.01 molar benzoic acid at various temperatures
and pH's are as follows:

                 Vapor pressure, ppb when temperature is
             B§

              2

              k
              5
              6

These values indicate that losses from evaporation will be relatively low
in the range of scrubber conditions.  However, a vapor pressure of 100 ppb
of benzoic acid would result in the loss of about 2.5 Ib of benzoic acid
per hour if used to  scrub the gas from a 500-mw boiler.  It is likely £hat
other losses, through adsorption on solids and faulty mist eliminators,
will be more important than evaporation per se.


Summary

         The use of  certain weak organic acids, particularly benzoic, appears
promising as a means of improving the effectiveness of limestone slurry in
removing S02 from stack gas.  It has been determined that the stability,
volatility, and calcium salt solubility for benzoic, adipic, phthalic, and
glycolic acids are within ranges that appear acceptable for practical
operation.  Only small amounts, on the order of 20 to kO% of that needed
to dissolve the stoichiometric quantity of limestone (based on amount of
S02 to be removed), are required to effect a major increase in S02 absorption
efficiency.
45°C U13°F) 5
136
129
85
19
2
OiC U22"FJ
217
206
137
31
k
55°c U31°F;
355
33T
225
52
6
                                   282

-------
         The method has the further promise of increasing limestone
utilization.  If all the feed limestone were dissolved in the organic
acid, the limestone should remain in the scrubber loop until completely
reacted; without the organic acid,  part of the limestone is lost with
the waste solids because complete conversion of CaC03 to CaSOQ'O.^T^O
is never obtained in one pass through the scrubber.

         The potential advantages of organic acid addition are such
that further work is planned, both on a small scale and in pilot plant
tests.
                                 283

-------
POTENTIAL WATER QUALITY PROBLEMS ASSOCIATED WITH LIMESTONE

       WET SCRUBBING FOB SO? REMOVAL FROM STACK GAS
                       J. S. Jfcrris
    Division of Environmental Research and Development
                Tennessee Valley Authority
                  Chattanooga, Tennessee
               Prepared for Presentation at
 Second International Line/Limestone Wet Scrubbing  Symposium
      Sponsored by the Environmental Protection Agency
                  New Orleans, Louisiana
                    November 8-12, 1971
                      285

-------
         TOTEHTIAL WATER QUALITY PROBLEMS ASSOCIATED WITH LIMESTONE

                WET SCRUBBING FOR SOg REMOVAL FBCM STACK GAS
                                J. S. Morris
             Division of Environmental Research and Development
                         Tennessee Valley Authority
                           Chattanooga, Tennessee
          The present emphasis on rapid attainment of a higher quality total
environment has focused added attention on the interrelated effects on air,
water, and land resources that result from new technology for controlling
pollution.  An inescapable byproduct of any separation process is a concentrated
waste.  Without waste recovery it is rarely possible to develop practical con-
trol methods for  specific wastes in any one of these three natural resource
areas without affecting at least one of the other two areas.  In the limestone
scrubbing process for control of sulfur dioxide, land use is affected by the
necessity for disposing of large quantities of solid waste material, and dis-
charge of once-through scrubber water or bleed-off from recirculating systems
to receiving streams presents a potential problem of water quality degradation.

          A scrubber system based on a lime process has a somewhat different
effect on water quality than one based on limestone.  In evaluating the poten-
tial for each process, TVA has concluded that the use of line appears infeasible
because the system  requires  (l)  dilution of the scrubber liquor, which necessi-
tates an "open loop" operation with an overflow to a watercourse, or (2)  use
of a very expensive, high rate of recirculation.  Since most of TVA's efforts
have thus been confined to limestone scrubbing, the following discussion is of
this basic process.


Itotenf.;i a,1 "Pollution Problems

          Table 1 lists the soluble cation constituents of the effluent from
a typical limestone scrubber.  This semiquantitative spectrographic analysis
indicates the wide  range of cationic species found in such a system.  These
constituents, together with the associated anions—mostly 803, SOl^, and NOj—
and the pH and suspended solids, represent the potential water pollutants.  It
is assumed that operation of a scrubber system will provide adequate settling
for solids; thus, suspended material will not adversely affect water quality.

          The minor and trace elements shown in table 1 require quantitative
analysis because  of the low concentrations permissible under existing standards.
The maximum stream  concentrations of the following elements are based on allowable
limits for potable water:  manganese, 0.05 rag/1; zinc, 5-0 mg/1; silver, 0.05 mg/lj
copper, 1.0 mg/1; and lead, 0.05 mg/1.  Stream standards generally apply to the
concentrations existing after the effluent is fully mixed with the receiving waters.
                                     286

-------
Adverse  effects  from various  concentrations of the remaining trace cations are
included in the  Water Quality Criteria publication of the Resources Agency of
California referred to in the appendix.  With the exception of zinc, manganese,
and copper, TVA  has not quantitatively analyzed the minor and trace elements
listed.   Chemical  analysis of scrubber water during one preliminary pilot-plant
run resulted  in  the following maximum concentrations:  manganese, 80 mg/1;
zinc,  9.0 mg/1;  and copper, 5-0 mg/1.   Each of these concentrations would likely
exceed limits for  an effluent standard, and manganese would likely exceed appli-
cable  stream  standards.
                                   TABLE 1

                           SPECTROGMPEEC ANALYSIS

                  Element	Relative Distribution

                   Calcium                    Major
                   Iron                       Major
                   Aluminum                   Major
                   Silicon                    Major
                   Magnesium                  Major (-)*
                   Sodium                      Major (-)*
                   Potassium                  Major (-)*
                   Titanium                   Major '(-)*
                   Manganese                  Minor
                   Zinc                       Minor
                   Tin                        Trace
                   Copper                      Trace
                   Silver                      Trace
                   Vanadium                   Trace
                   Molybdenum                 Trace
                   Galium                      Trace
                   Mercury                    Trace*-*
                   Lead                       Trace**
                   Boron                      Trace
                   Nickel                      Trace

                  •*Less than other major components, but more
                   than minor quantities.

                 **Quantity is at the lower limit of detection
                   for emission spectrograph.


          Quantitative analysis of the major soluble constituents from flyash
and limestone  together with some preliminary data from pilot-plant operation
indicate that  the characteristics of the scrubber water that are of most concern,
relative to potential for  adverse effects on water quality, are those shown in
table 2.   Table 2 also indicates an approximate allowable concentration of each
constituent based on various regulatory standards and on criteria recommended
ty TVA.  Although limits generally apply to the concentrations after mixing,
strict adherence to a nondegradation standard or adoption of an effluent standard
                                      287

-------
might require that these limits not be exceeded in the discharge itself.
Effluent  standards,  such as those recently adopted by the Ohio River Sani-
tation  Commission for the Ohio River, are becoming more prevalent with
increased demand for more stringent controls.

           A general  discussion of the detrimental effects of the more important
of these  characteristics is presented in the  appendix.


                                   TABLE 2
                                            ^Maximum Concentration
                                             Milligrams "Per Liter

                    Calcium   )                   .
                    M     .    ) Hardness (as CaCOq)    200
                    Magnesium )          ^       J
                    Sulfate                           150
                    Chloride                          150
                     Total Dissolved Solids           500
                    PH                             6.5 - 8.5

                   *Based on various regulatory standards and TVA
                    criteria for receiving streams.
 Conceni
           The constituents listed in tables 1 and 2 are likely to be present
 in any scrubber solution; however, if one disregards the threat to the aqueous
 environment from system leaks and pond seepage,  the real potential for degrada-
 tion of water quality is due to any scrubber solution discharged to the receiving
 waters.  Since operation of a scrubber system with complete recycle increases
 the problems of scaling and may cause operational problems due to the buildup
 of dissolved salts, many systems are being designed to discharge some of the
 scrubber water.  Figure 1 shows a simplified diagram of a typical limestone
 scrubbing system.  Item D, the discharge from the settling pond, may be any
 portion of the total contents of the pond from no discharge to 100 percent dis-
 charge.  "No discharge would represent a fully closed-loop system with no potential
 for pollution, and this system is discussed only in the conclusion.  The following
 discussion is confined to partial bleed-offs and to 100 percent discharge.

           Table 3 shows the theoretical concentrations in scrubber water from  a
 limestone process utilizing a 3 percent WgO limestone, a 150 percent stoichio-
 metric limestone feed for 3-5 percent sulfur coal, 100 percent NOg absorption,
 100 percent MgO reaction, 25 percent formation of calcium sulfate, and 75 percent
 formation of calcium sulfite.  This theoretical analysis is based on the simul-
 taneous solution of assumed solubility equations.  The analysis represents the
 buildup of concentrations for a system with 100 percent recycle; a system with
 a partial bleed-off, with the necessary makeup of fresh water, would be expected
 to contain somewhat lower concentrations.  The reduction in concentrations is
 dependent on the percentage of scrubber water discharged; however, for bleed-off s
                                      288

-------
                               FIGURE I
                 TYPICAL SCRUBBER FLOW  SYSTEM
00
 GAS FROI^
  BOILER
                TO STACK
           SCRUBBER
                       A
B
                                MAKE UP H20
    CaC03
RECIRCULATION
   TANK
                                     POND
                                                           OVERFLOW
                                                           TO WATER-
                                                           COURSE
                                              ->-WET SOLIDS

-------
of 20 to 30 percent, the reduction is not expected to be .significant from the
standpoint of water quality.  Table h compares the theoretical concentrations
with preliminary data from a pilot plant with a similar operating system.  The
pilot-plant feed was approximately 100 percent stoichiometric, and the MgO
content was about 1 percent.  The major differences between these calculated
and measured concentrations are in the concentrations of sulfites and nitrates.
The theoretical assumptions regarding N02 absorption are apparently erroneous,
since no pilot "plant data reviewed to date indicate more than 10 mg/1 NOo.  The
higher sulfite concentration in the results for the pilot plant may indicate
a formation of MgS03, although no satisfactory balance of the complexes can be
made from the data.  The most important information from the pilot-plant data
is not obtained from an analysis of the differences between them and the calcu-
lated values, but from the conclusion that the high concentrations of potential
pollutants theoretically predicted may actually occur in the system.
                                    TABU;
             THEORETICAL CONCENTRATIONS IN SCRUBBER HMD EFFLUENT

                  Calcium                      950 mg/1
                  Magnesium                  3*500 mg/1
                  Hardness (as  CaC03)        16,500 mg/1
                  Sulfite                      150 mg/1
                  Sulfate                    9,500 mg/1
                  Nitrate                    7,750 mg/1
                  Sodium                       850 mg/1
              Total Dissolved Solids        25,000 mg/1
                  pH                           6-.0
                                    TABLE U

                                 THEORETICAL CONCENTRATIONS
                                     KITH
                         PRELIMINARY PILOT

                                       Theoretical             ELlot ELant
                                      Concentration           Concentration
                                 Milligrams Tier Liter    Milligrams "Per Liter

           Calcium                          950                   1,600
          Magnesium                     3,500                   1,150
          Hardness  (as  CaCOo)            16,500                   7,100
           Sulfite                          150                   5,1*00
           Sulfate                       9,500                   6,500
          Nitrate                       7,750                       1
          Sodium                            850                      26
           Total  Dissolved Solids        25,000                  13,000
          PH                               6.0                     5.0
                                       290

-------
          The high quantities of dissolved salts in these scrubber systems.
represent a significant potential for creating a problem of water pollution.
Utilization of other limestones in different modes of operation may create
greater or lesser potential problems; however, for partially closed-loop
operations, table h should generally be representative.  In situations where
stringent effluent standards are implemented or where strict interpretation
of a nondegradation standard is enforced, it is doubtful if sen  er waters
can be .discharged without a very high degree of treatment.  The potential
adverse effects from such discharges—regardless of standards—ate dependent
on the quantity of bleed-off and the flow of the receiving stream.

          The effects on a receiving stream from the release of such concen-
trations were calculated for a hypothetical 1500-megawatt plant with an
assumed bleed-off of 30 percent, equivalent to about k,500 gpm, and an assumed
dilution ratio of the receiving stream to plant discharge at critical low
stream flow  (? day-10 year minimum) of 135 to 1.  The ratio is obtained by
assuming that the critical low flow is equivalent to the condenser cooling
water required for plant operation, 600,000 gpm or 13*4-0 cubic feet per second.
Although this may seem extreme, a random survey of some 20  existing plants
indicated that two plants are located on streams that have much less relative
stream flow.  It is not actually necessary to assume any plant  size, since the
ratio of discharge to available dilution would remain the  same in  all  cases;
however, the  assumptions indicate that the critical flow levels are represen-
tative of  fairly large rivers.  Figure 2 shows the resulting percentage  of
total  allowable concentration  assuming typical background  concentrations of
100 mg/1 hardness as  CaCX^,  20 mg/1  S01+, 150 mg/1 total dissolved solids, and
0 mg/1 manganese, complete mixing in the stream, and a scrubber discharge
containing the theoretical  concentrations shown in table k less the nitrate
concentrations.   The manganese  concentration was taken from the pilot-plant
data.  Although  only  hardness  and manganese exceed the assumed limits,  the
percentage increase in the  concentration is quite significant.   Such  increases
in total stream  concentration  are not  generally acceptable.   Refer to the
appendix for a discussion of actual  adverse effects from the  resulting concen-
trations.

           One potential water-quality problem, that of the dissolved sulfite
 concentration,  appears to be much less severe than  originally anticipated  and
may,  in fact, be quite insignificant.   High concentrations of sulfite will
utilize large quantities of dissolved oxygen  in the oxidation to sulfate:
 0.2 milligrams of oxygen per milligram of sulfite.  The presence of certain
metals in the flyash, such as iron,  '.was- thought  to be sufficient catalyst for
 the oxidation of sulfite to proceed rapidly enough to significantly lower  the
 dissolved oxygen in the settling pond and ultimately in the receiving stream.
 From pilot-plant results, it now appears that this reaction proceeds quite
 slowly and, indeed, may be difficult to induce.

           A complete theoretical computation of the chemical characteristics
 of the effluent from the settling pond for a limestone scrubbing system operating
 with a 100-percent overflow has not been made; however, some pilot-plant data
 have been analyzed.  Table 5 compares pilot-plant data obtained during open-loop
 operation with those obtained during closed-loop operation.  The open loop
 operated with 1 percent MgO limestone and a limestone feed of approximately
 150 percent stoichiometric for a unit burning coal with 3 percent sulfur.   Although
 the concentrations are less for the open loop, they are surprisingly high and of


                                       291

-------
                              FIGURE  2
   HARDNESS
              STREAM CONCENTRATIONS AT LOW FLOW
                                 50
13
   SO,
ro
vo
ro
   DISSOLVED

   SOLIDS
   MANGANESE
                                           110
                                     60
       30
55
                                      J L
                                                 1200
                   PERCENT OF MAXIMUM ALLOWABLE CONCENTRATION
         BACKGROUND
     :•:•:{ SCRUBBER CONTRIBUTION

-------
such magnitude as to be of concern.  If the concentrations for closed-loop
operation are assumed to be representative of a 30~percent bleed-off, then
the total quantities of waste constituents discharged from each of the
operational modes are similar.  The greater flow of the open-loop process
'(more than three times that of the closed-loop system) produces somewhat
greater amounts and, therefore, a greater potential for adverse effects on
water quality.
                                   TABLE 5

              CONCENTRATIONS IN OPEN-AMD CLOSED-SCRUBBER SYSTEMS

                                     Closed-Loop               Open-Loop
                                        Mode                     Mode
                               , Milligrams per Liter     Milligrams "Per Liter

           Calcium                       1,600                    1,100
          Magnesium                     1,150                      *
           Hardness  (as  CaCOo)           7,100                  **2,700
           Sulfite                       5,^00                    1,700
           Sulfate                       6,500                    2,600
           Nitrate                           1                        6
             Total Dissolved Solids     13,000   -               **7,000
          PH                            5.0                      5-3

         *Not determined, MgO content of limestone less than 1%.

        **Not determined, estimated minimum values.


Factors That Affect the Quality of Scrubber Water

           Almost all input factors in a limestone scrubbing system will affect
the quality  of scrubber water and its potential for water pollution.  In partic-
ular, the  coal, limestone, feed rate, type of boiler, type of scrubber, ratio
of liquid  to gas, and pH of the scrubber water will have major effects on the
characteristics of the  scrubber water.

           The oxidation of sulfites to sulfates in the system will have a
major bearing on the quantity of dissolved salts in the system and, according
to TVA pilot plant results, will greatly affect the settling characteristics
of the reaction products.  If the solids are mainly calcium sulfite rather
than calcium sulfate, as has been the case in much of TVA's work on the closed-
loop system, settling is very poor.  The thin sulfite platelets form a gel-like
structure  that settles  very slowly.  Although this settling problem has no
direct effect on water  quality, the settling pond for the gel~like substance
would  certainly require careful design to assure containment adequate to prevent
any threat to nearby watercourses.  While formation of sulfate may be more desir-
able from  the standpoint of settling, it may create a greater potential for
direct water pollution. The sulfate reaction products are much more soluble
than the sulfites, more than 50 times as soluble in the case of the calcium pro-
ducts; thus, higher concentrations of dissolved salts result from increased
oxidation.

                                   293

-------
           The chemical makeup of the limestone is quite important in defining
the potential effects on water quality.   Among the major controlling charac-
teristics of the limestone is its magnesium content, which has an important
bearing on waste disposal because of the high solubility of magnesium sulfat.e.
There is some indication that small amounts of magnesium may appear as magnesium
hydroxide rather than sulfate in the products; this is desirable because the
hydroxide is extremely insoluble.   In  many areas of the country, however, calcitic
limestones are not readily available,  and dolomite  (CaC03'MgC(^) or dolomitic
-limestone would be the raw material.   Figure. 3 shows the concentration of sulf ates
 and hardness of the scrubber water  as  functions of the MgO content of the limestone,
 Conclusion

           A given concentration of a pollutant has the  same detrimental effect
 on water quality no matter what its source, but a minor detrimental effect from
 the residual waste discharge from an efficient production process and treatment
 facility might be considered acceptable while the same  effect, if it results
 from an attempt to provide control of wastes fr.om other sources sufficient to
 meet an applicable standard, might not be considered acceptable.  TVA believes
 that a solution to a problem of air pollution that in turn creates a problem
 of water pollution—even though possibly one of lesser  extent—is not satis-
 factory.  Thus, TVA's full-size experimental unit being designed for installation
 on unit 8 of the Widows Greek Steam ELant will be a  closed-loop operation with
 no degrading of water quality.

           If an open loop is absolutely necessary for other systems, their
 continued operation, in the immediate future, will be dependent on the charac-
 teristics of the waste from the process, the quantity of discharge, and the
 size of the receiving stream.  It seems increasingly evident, however, that
 more stringent controls, the implementation of effluent standards, and strict
 enforcement of nondegradation policies will require  a closed cycle or a high
 degree of treatment.  Although no thorough investigation of technically feasible
 methods of treatment has been made, wastewater containing such high concentrations
 of dissolved solids as are present in the scrubber flow generally requires ad-
 vanced, sophisticated, and often expensive treatment.  Treatment by distillation
 is considered technically feasible, but since the estimated annual capital and
 operating cost for treating a 30-percent bleed-off from a 1500-megawatt plant
 is $1.5 to $2.0 million, this method is economically infeasible.  Although other
 treatment methods may be technically adequate and more  economical, most processes
 still require disposing of a concentrated waste product.  Based on the present
 state of the art of treatment of such wastes and environmental considerations,
 complete recycle of the scrubber water, wherever technically feasible, should be
 the operational mode for limestone systems.

           Although it is quite apparent that limestone  scrubbing systems can have
 a potential for affecting water quality, the magnitude  of this potential is not
 adequately defined.  TVA is continuing its efforts to gain additional information.
 The Radian Corporation recently began a complete chemical analysis of TVA's pilot-
plant  scrubbing process, and the resulting data will be used to refine the theo-
 retical analysis of the system and to aid in evaluating effects on water quality.
 The large-scale lime/limestone scrubbing tests scheduled to begin in March 1972
 at. the TVA Shawnee Sfceam Hant will provide an opportunity to further define the
effects of various scrubbers and operating parameters.


                                         294

-------
                     FIGURE 3
       EFFECT  OF  MgO CONTENT OF LIMESTONE
                        ON
   HARDNESS AND  S04  OF  SCRUBBER POND EFFLUENT
   30
to
O
er
UJ

fe 20
or
uu
CO
CO

a:
o
CO
   10
CO
o
cc
o
V)
    0
     o
             HARDNESS

             S04
               PERCENT MgO IN LIMESTONE
                         295

-------
                                  APPENDIX
           The  feasibility of releasing dissolved or  suspended calcium salts
into watercourses  depends on how much the content of these impurities in'
the watercourse  is increased, which in turn depends  on the amount of the
particular pollutant released in the effluent,  the content in the watercourse
before influx  of the effluent, and the water volume  flow of the watercourse.
'Various regulatory agencies and advisory groups have set or recommended
limits above which the concentration of an impurity  in a watercourse should
not be increased.   Such limits for the dissolved salts released from a
limestone-wet  scrubbing process are summarized  herein.

                               Suspended Solids

           Most regulatory agencies do not place quantitative limits on either
 suspended solids or turbidity for raw water used as  a source for domestic
water, but do  generally specify that concentrations  shall not be sufficiently
 high as to be  objectionable or interfere with normal treatment processes.
 Both turbidity and suspended solids can be reduced to acceptable limits for
most purposes  by conventional treatment methods; however, high concentrations
 increase the cost of processing the water by increasing chemical requirements
 and the volume of sludges to be disposed of.  In some cases head losses
through the filters are increased with a resulting increase in the frequency
 of filter backwashing.

           Settleable suspended solids often blanket  stream bottoms, killing
fish eggs and  young and destroying much of the  normal benthic biota which
 serve as food  organisms required for propagation and growth of fish.  Turbid"
 ity in streams also interferes with light penetration and limits or diminishes
 the photo synthetic effects necessary for the primary productivity of fish-
 food organisms.   The Committee on Water Quality Criteria of the Federal Water
 Pollution Control Administration! recommends a  limit of 50 Jackson Turbidity
units in warm  water streams.
                                   issolved Solids
           The following information is quoted from Water Quality Criteria.
 issued by the Resources Agency of Calif ornia.^

           (a)  Domestic Water Supplies.  The 1962 USPHS Drinking Water
      Standards specify that the total dissolved solids should not exceed
      500 mg/1 if more suitable supplies are, or can be made  available.
      This limit was set primarily on the basis of taste thresholds.  The
      1958 WHO International Standards set the '^permissible limit" at


 Federal Water Pollution Control Administration.   Water Quality Hriteria
 (Report of the National Technical Advisory Committee to the Secretary of
 the  Interior).   U.S.  Government Printing Office, Washington, D.C.  (April 1, 1968).
2
 McKee, J. E. ,  and  Wolf,  H.  W.  Water Quality Criteria.   The  Resources Agency
 of California,  State  Water Quality Control Board, Publication No. 3~A (1963).
                                        296

-------
     500 mg/1 and the "excessive limit "  at 1500 mg/1,  but  no  "maximum
     allowable limit" is given.  The 1961 WHO European  Standards  do not
     include limits for total dissolved solids.  It is  generally  agreed
     that the salt concentration of good, palatable water should  not
     exceed 500 mg/1; however, higher concentrations may be consumed
     without harmful physiological effects and may indeed even be more
     beneficial.  Each water with a total salt concentration over 1000
     mg/1 should be judged on the basis of the local situation, alter-
     native supplies, and the reaction of the local population.

          (e)  Fish and Aquatic Life.  It has been reported that  among
     inland waters in the United States supporting a good mixed fish
     fauna, about 5% have a dissolved-solids concentration  under  72 mg/1;
     about 50% under 169 mg/1; and about 95% under kOO  mg/1.

          Water quality objectives adopted by TVA for surface  water in the
     Tennessee River basin include a limit of 500 mg/1  of dissolved solids.


                                      EH

          TVA has proposed upper and lower pH limits for the waters of the
Tennessee Valley of 8.5 and 6.55 respectively.  The pH  values  required for
industrial water supplies are covered in some detail in the FWPCA Water
Quality Criteria report.  The proposed TVA criteria should  be  adequate for
a.11  of these but a few extreme cases.
                                   Calcium

          Calcium in water supplies has been suspected of producing undesir-
able physiological reactions but no definite relationship has yet been
established.   The major concern is related more to its deleterious effects
in uses such as washing, bathing, and laundering and to the problem of
incrustations on cooking utensils and water heaters.  Some industrial processes
require a very soft or low calcium water.   Costs of conditioning water  for
such industries is directly related to source concentration.

          Thus the major detrimental effect of calcium hardness is economic.
TVA has proposed the following limitation  on water hardness:

          There shall be no substance added to the waters that will increase
     the hardness to such an extent as to  appreciably impair usefulness as
     a source of water supply or interfere with other reasonable and necesso,ry
     uses of the water.

          Rivers iri the Tennessee Valley on which steam plants are located
generally range from 50 to 125 mg/1 hardness as calcium carbonate.  In  general,
a concentration of 100 mg/1 would not be objectionable for most uses if this
concentration were not significantly greater than the normal range for  the  area.
                                     297

-------
                               Calcium Sulfate

          No limits have been placed on calcium sulfate by regulatory agencies.
The following summary of beneficial uses and reported acceptable  concentrations
is quoted from McKee and Wolf.

          (a)  Domestic Water Supplies.  High calcium sulj.c,_~ concentrations
     in water are disadvantageous for most household uses, but for drinking
     purposes, 300 mg/1 or mere is not harmful.  The taste threshold of calcium
     sulfate has been reported to be 250 to 900 mg/1.  Calcium sulfate signifi-
     cantly increases plumbosolvency; raising the calcium sulfate concentration
     from 25 to 250 mg/1 increased by 10 percent the amount of lead dissolved
     from lead pipes.

          (b)  In general, calcium sulfate is beneficial in the brewing industry,
     for it helps to maintain the acidity of the wort and therefore causes more
     complete coagulation of albuminous matter.  It also reduces  the solubility
     of the bitter substances of the hop.

          (c)  Irrigation.  Gypsum in irrigation waters improves  or restores
     the permeability and tilth of soil having an unfavorable sodium ratio.

          (d)  Stock and Wildlife Watering.  A saturated solution of•calcium
     sulfate, used as the sole source of drinking water for rats, permitted
     satisfactory growth.  A concentration of 2^00 mg/1 of calcium sulfate
     permitted normal growth and reproduction among rats.

          (e)  Fish and Other Aquatic Life.  Trama reported that  a saturated
     solution of calcium sulfate at 20° C did not produce significant mortali-
     ties among bluegills; but a later report from the same organization
     indicated that this concentration was the 96-hour TI^ at 18-20° C in soft
     water for the bluegill sunfish.  It was also reported that 3200 mg/1
     caused a 50 percent reduction in the rate of growth of the diatom,
     Navicula seminulum.  Using highly turbid water at 21-25° C and the mosquito-
     fish (Garobusia affinis) as the test animal, Wallen et al. indicated that
     the 96-hour Tim was greater than 56,000 mg/1, a concentration much larger
     than the solubility.

          TVA has proposed no specific limits for calcium sulfate; however,
the concentration is limited by the allowable sulfate concentration and the
permissible added calcium hardness.


                                  Sulfates

          McKee and Wolf discuss sulfate content as follows.

          (a)  Domestic Water Supplies.  The 1962 Drinking Water  Standards
     of the USIHS recommend that sulf ates not exceed 250 ing/1, except where
     a more suitable supply is not available.  This limit does not appear to
     be based on taste or physiological effects other than a laxative action
     toward new users.  Public water supplies with sulfate contents above this
     limit are commonly and constantly used without adverse effects.  The 1958
     WHO International Standards established a "permissible limit" of 200 mg/1


                                    298

-------
     and an "excessive limit" of 1*00 mg/1, but set no maximum allowable limit.
     The 1961 WHO European Standards include a recommended limit of 250 mg/1
     for sulfates.

          Korotchenok reported on heavily mineralized drinking waters in
     western Turkmenia (USSR), noting that no outbreaks of disease have been
     ascribed to waters in which the content of sulfate did not exceed 1295
     mg/1.  In British Somaliland well waters used for human consumption
     contain very high sulfates, many having 2000-3000 mg/1.  One village
     is using water of 1^00 ing/1 sulfates.  A survey in North Dakota indi-
     cated that water containing less than 600 mg/1 of sulfates is usually
     safe.  In reviewing the literature, Moore quotes Sollman to the effect
     that concentration of 1000 mg/1 of sulfates in water is harmless.  A
     cathartic dose is 1.0 to 2.0 grams, or a liter of water containing 1000
     to 2000 mg/1 of sulfates.

          Sulfates appear to have no detrimental effect on the corrosion of
     brass fittings in domestic water systems, nor do concentrations less
     than 200 mg/1 increase plumbosolvency.

          Whipple is quoted by Moore to the effect that the taste thresholds
     of sulfate salts were as follows:

                 Sodium sulfate          200-500 mg/1
                 Calcium sulfate         250-900 mg/1
                 Magnesium sulfate       1|-00-600 mg/1

          (b)  Industrial Water Supplies.  Limiting or threshold concentrations,
     or optimum ranges, for sulfates are assembled below:

                                  Limits of recommended values. mg/1

          Industrial process      SO^      CaSOk       Mgj30U.     _Ea2JSQ'.i_

          Brewing                  -      100-500     100-200       100
          Garb beverages          250           -           -
          Concrete corrosion       25           -           -
          Ice making                -         300     130-300       300
          Milk industry            60           -           -
          Hioto processes           -         100
          Sugar malting             20           -           -
          Textiles                100           -           -         -
          TVA has proposed a limit of 150 milligrams per liter for sulfates to
     provide a reasonably satisfactory water for both domestic and industrial
     use.  Normal water treatment has little effect on sulfate concentration.
                                   Sulfite,

          The major concern in regard to high sulfite concentration is oxidation
to sulfate, thus lowering the dissolved, oxygen content of the receiving water.
                                    299

-------
                             Magnesium

     The effect of dissolved magnesium has been treated by McKee and Wolf.

     (a)  Domestic Water Supplies.  Magnesium is an essential mineral
element for hviman beings; the daily requirement of magnesium is about
0.7 gram.  Magnesium is considered relatively non-toxic to man and not
a public health hazard because, before toxic concentrations are reached
in water, the taste becomes quite unpleasant.  At high concentrations,
magnesium salts have a laxative effect, particularly upon new users,
although the human body can develop a tolerance to magnesium over a
period of time.

     The 19J46 USTHS Drinking Water Standards recommended a limit of
125 mg/1 but there is no limit in the 1962 standards.  The 1958 WHO
International Standards have a "permissible Idjnit" of 50 mg/1 and an
"excessive limit" of 150 mg/1, but no maximum allowable concentration.
The 1961 WHO European Standards have a recommended limit of 125 rag/l>
but if the sulfate exceeds 250 mg/1, the magnesium is limited to 30 mg/1.

     The taste threshold for magnesium (in MgSO^) has been reported as
100 mg/1 and for the average individual it is given as about 500 mg/1.

     The negative correlation between hardness in water and cardiovascular
disease does not appear to hold for magnesium as it does for calcium; yet
one investigator reports the favorable use of magnesium gulf ate to treat
such cases and claims that magnesium, rather than calcium, is the beneficial
element in reducing cardiovascular attacks.

     (e)  Fish and Other Aquatic Life.  The relative concentrations of
magnesium and calcium in water may be one factor controlling the distri-
bution of certain crustacean fishfood organisms, such as copepods, in
streams.  Hart et al. cite a report that among U.S. waters supporting a
good fish fauna, ordinarily 5 percent have less than 3-5 nig/1 of magnesium;
50 percent have less than 7 ag/1; and 95 percent have less than Ik mg/1.

     Magnesium chloride and nitrate can be toxic to fish in distilled
water or tap water at concentrations between 100 and UOO mg/1 as magnesium.
However, magnesium chloride, nitrate, and sulfate, at concentrations be-
tween 1000 and 3000 mg/1 as magnesium have been tolerated for 2-11 days.
Some fresh-water fish have been found in very saline lake water containing
over 1000 mg/1 of magnesium as well as additional sodium and calcium salts.

     TVA has proposed no specific limit on magnesium.
                                  300

-------
                  POTENTIAL UTILIZATION OF SOLID WASTE
                  FROM LIME/LIMESTONE WET-SCRUBBING OF
                              FLUE GASES*

                                  by

                Linda Z.  Condry,  Research Technologist
              Richard B.  Muter,  Supervising Research Chemist
                                 and
               William F. Lawrence, Project Supervisor

                      Coal Research Bureau
                     West Virginia University
                     Morgantovn,  West Virginia
Prepared for presentation before  the  Second  International Lime/Limestone
Wet-Scrubbing Symposiun New Orleans,  Louisiana, November. 1971

*Work supported by the Environmental  Protection Agency  under Contract No.
CPA 70-66.
                                 301

-------
                              INTRODUCTION



      The increasing waste  disposal  problem  associated with coal combustion



 may be compounded by  the greater amount of  waste  generated if potential



 alkaline earth sulfur dioxide abatement systems are implemented.  Although



 utilization schemes for normal  flyash are available, "modified flyash," the solid



 by-product resulting  from  lime/limes tone wet-scrubbing programs, represents a



 new type of solid waste material having different chemical and physical properties



 from regular flyash.   In addition,  simple disposal of this ash by land fill or



 lagooning may cause serious water pollution problems especially if dolomite



 is used as a'modifying agent.  Previous research^  has shown that very little



 potential exists for  mineral separation and beneficiation of this new ash and



 Indicates that the most economical  utilization scheme would involve total or



 whole utilization of  the ash.  For  these reasons  the Coal Research Bureau of



 West Virginia University has been performing  research under partial support from


                                      (2)
 EPA (Environmental Protection Agency)    to develop and evaluate total



 utilization processess. As a result  of this  research, several promising



 utilization areas have emerged.  These include production of autoclaved



 materials such as calcium-silicate  products,  production of materials such as



 mineral wool using high temperature processes and agricultural uses such as



 soil amendments.



 Sources of Modified Flyash



      Several modified flyashes  were investigated  during the course of this



 research;  however,  only two were of a wet-collected nature.  They were



 obtained.from  the Kansas Power  and  Light  Co., Lawrence, Kansas (designated



KPL), a limestone wet-scrubbing system, and from  the Union Electric Company,



St. Louis, Missouri (designated SLO),  a dolomite  wet-scrubbing system.  In




                                 302

-------
addition, a sample from the Shawnee Steam Plant, Paducah, Kentucky (designated




TV A) was also examined.  This ash originated from a limestone injection, dry-



collection system of the type which could possibly be employed before a wet-



scrubber .



Autoclaved Products



     Because modified flyash liberates sorbed sulfur dioxide at elevated



temperatures, it was necessary to find a technique for the total utilization



of this ash which would either lock the sulfur into the product or require no



external application of heat.  Preliminary research examining the natural



pozzolanic activity of limestone modified flyash indicated that some



cementitious setting occurred.  Such setting is sufficient to obstruct power



plant process lines if water flow should stop but does not appear to impart



enough strength for structural products.  For this reason and because



the technology has been well developed in Europe using normal flyash, auto-



claving was selected for further investigation.  In addition to forming a stronger



material, this technique would also have cne advantage of locking sulfur into




the final product.



     Various autoclaved products were investigated using modified ash as



the principal raw material.  Those products exhibiting the greatest potential



for utilization of modified ash were:



     a.  brick or block,



     b.  aerated or foamed concrete, and



     c.  concrete materials.



A.  Calcium-Silicate Brick



     Calcium-silicate  (CS) brick production was examined as a potential large



tonnage application of limestone modified flyash.  Bricks produced by



autoclaving modified flyash have the added advantage of binding the sulfur
                                   303

-------
 components within a calcium-silicate matrix, and because they are full-




 strength after  curing, they are  immediately available for marketing.  In



 addition, bricks made using this  raw material surpass the standards for



 conventional  sand-lime brick.



      Figure 1 illustrates a simplified process flowsheet for the production



 of CS brick.  As illustrated,  this process involves dewatering of the ash,



 addition of silica sand  and lime, mixing of the resultant material, pressing



 to the  desired  form, humidity  curing and autoclaving.  Brick made with, on a



 dry basis, 40 percent KPL wet  collected limestone modified flyash, 39 percent



 silica  sand  (30 x 100 mesh) and  11 percent lime (95 percent pure CaO) exceed



 ASTM (C73-67, C73-51) standards  for grade SW (severe weathering) calcium-



 silicate brick. The compress ive breaking strengths are in excess of 5000 psi



 as compared  to  the ASTM  specification of 4500 psi minimum.



      Process  parameters  which  were found to be of importance in the production



 of CS brick include mixing and humidity curing.  It was found that surface



 cracking due  to insufficient calcium-silicate bonding occurred if the various



 components were not intimately mixed before pressing.  This problem was solved



 through the use of a muller-type mixer.  Also, it was found .that the



 compress ive strength of  the final product could be increased significantly



 if a humidity curing step was  incorporated in the process prior to autoclaving.



 The humidity  curing was  effected  by storing the pressed green brick over water



 at room temperatures for periods  up to twelve hours.



      Among the  promising areas for utilization of calcium-Silicate brick are:



 general low cost construction  materials, interior walls, decorative walls,



patios,  and thermal or acoustic insulatine walla or barriers.



B.  Aerated Concrete



     Originally patented in the U. S., aerated concrete has been widely



developed in Europe.  Practically all new factory buildings in Sweden and



80 percent in West Germany are constructed, in part, using this material.'
                                  304

-------
This lightweight structural material consists of small non-communicating gas



cells entrained in a calcium-silicate matrix.



     A process flowsheet for the production of aerated or foamed cellular



concrete from modified flyash is included as Figure 2.  Additions of



cement and a gas producing agent to the ash are required; however, for



this process, no pressing or humidity curing such as was required for brick



production is necessary.  Also, if a dry-collected modified ash is used,



additional lime or sand is not required.  On a dry basis the composition



of materials used for aerated concrete production was 90 percent dry-collected



modified ash, 10 percent portland cement (type I) and 0.16 percent aluminum



powder.



     The resulting aerated concrete has a density which may be varied from 50


            3                        '                        •
to 56 Ibs/ft  and a compressive strength, which may be controlled, between



400 and 850 pal.  Commercial aerated concrete varies from 20-50 Ibs/ft3 in



density and has compressive strengths on the order of 400-800 psi.  This



material can be sawn, nailed, drilled, screwed or glued as with wood but has the



.thermal and fire resistant properties of concrete.  A few potential areas



for utilization are in non-loadbearing walls, sandwich construction with



brick or concrete for insulating purposes and interior surfacing for exterior



walls.  This material can be formed in desired shapes or cut from panels



and should be highly suitable for modular construction.



     Research in this area was performed using both wet and dry-collected



modified ashes.  Chemically very few elemental differences are apparent between



the two types of ash; however, the greater amount of reactive lime in the



dry-collected material made it more amenable than wet-collected material to



this process.  As with foreign manufacture which uses normal flyash, aeration



and setting times are crucial in that setting must occur after aeratior is



completed and before the entrained bubbles collapse.  It is felt that a major



advantage of modified flyash production may lie in the use of electrostatic

-------
 precipitators in conjunction witn. wet  scrubbers.  A dry ash suitable for


 aerated concrete production would be obtained in  large quantities while the


 overall efficiency of the sulfur dioxide abatement  process would probably

 not be affected.  Such a process would greatly reduce the need for ash


 dewatering and would .act to reduce  the potential  threat of increased water

 pollution.  In addition, a predominantly calcium  sulface material would

 be produced in the scrubber system.

 C.  Poured Concrete


      The natural pozzolanic properties of limestone modified flyash indicated

 that a potential use might be as a  cementing  agent  in concrete materials.

 Because research on CS brick production showed that autoc laving increased

 calcium-silicate bond formation, further research was performed to investigate

 the feasibility of using modified ash  as the  cementing agent for the production

 of concrete block.


      Included as Figure 3 is  the process flowsheet  employed in this phase

 of research.  It should be noted that  the composition is the same as for CS


.brick with the exception that less  dewatering is  required.  Also, It should be

 pointed out that the material is poured into  molds  instead of pressed.  As

 with CS.brick, this composition consisted, on a dry basis, of 50% wet-collected

 modified flyash, 39 percent silica  sand (30 x 100 mesh)  and 11 percent lime

 (95 percent CaO).  The resulting poured concrete which is similar to
                                              3   .'•.--              o
 concrete block had a bulk density of 90 Ib/ft as compared to 150 Ibs/ft  for

 conventional concrete block and a compressive strength of approximately 900

psi as compared to 1000 psi for conventional  block.  Although addition of

 the necessary aggregate could  raise the bulk  density to 100 Ib/ft3, this would

still  be only 2/3 the bulk density  of standard block.  Storage requirements

could  also be reduced as this  is  a  full  strength, non-shrinking product

out of the autoclave  and does not require 3 to 4 weeks curing time before

sale.


                                  306

-------
     An advantage of the use of modified ash is that no prior grinding or

crushing of the raw material is required.  In addition, the possibility  is

being examined that limestone modified flyash could be used for the production

of an autoclaved aggregate to be used as coarse material for this  concrete.

     The advantages of producing autoclaved calcium-silicate products

from limestone modified flyash would Include:

     1.  No sulfur dioxide capture or marketing problems would result as
         processing occurs below the temperature for sulfur dioxide
         regeneration;

     2,  Little or no storage or curing time is required before use
         as the structural materials produced are durable,  pre-shrunk
         and pre-strengthened;

     3.  Mo exotic equipment is required for production as  common
         European technology is employed;

     4.  Ho pollution problems should occur at the plant as rejects
         and process waters may be recycled; and

     5.  A potential large tonnage market exists in areas of low cost
         housing and construction.

Fired Products

     High temperature processing was also examined to develop other possible
                                                                (4)
uses for modified flyash, the most promising being mineral  wool.    The

alkaline earth constituents in limestone modified flyash reduce the

amount of heat necessary to make the ash sufficiently fluid for the production

of mineral wool.  Comparative tests using normal flyash and bottom ash with

equivalent amounts of lime added as a fluxing agent gave a  poor grade of

mineral wool and required more heat for processing.  Conversely, mineral wool

fibers produced from modified ashes were uniform, quite resilient  and were

superior to commercial fibers in protecting steel from rust corrosion.

     AS mentioned previ< usly, the modified ash releases sorbed sulfur dioxide

upon heating to temperatures over 1700°F.  Since this could constitute a new
                                   307

-------
air pollution problem, it was felt that a study of sulfur dioxide  evolution

from heated modified ash was in order.  Accordingly,  several modified flyashes,

both wet and dry-collected, were examined to determine their sulfur evolution

characteristics.  In all instances, sulfur dioxide was evolved over a narrow

temperature range in a relatively pure form and in concentrations  similar to

those considered suitable for sulfuric acid manufacture.   For  example, the

theoretical total yield of sulfur dioxide in sample KPL is 10.34 percent by

weight.  At 2140eF, (melting temperature 2150-2160°F), 96 percent  of its

total yield is  given off in twenty minutes.  Even with this long heating

period,  concentrations of 9-13 percent sulfur dioxide in  the off gas could be

maintained.  The only impurities observed were carbon dioxide  and  water vapor,

both of which can be evolved at lower temperatures if desired.

    • After evolution of the sulfur gases, the parent  ash  may be  further

heated to  a molten state for production of mineral wool or it  may  be cooled

and utilized for the production of fired ceramic materials.

Soil Amendment

     Dried sample KPL (wet-collected,, limestone modified  flyash) is being

examined as a soil amendment agent.  This work is currently being  performed under

sub-contract to Virginia Polytechnic Institute.*  The results  to date indicate

that modified flyash with its increased calcareous components  may  find use as

a soil amendment in the area of pH control.  In addition, it was noted that

such uuage effects an increase in the boron supplying power of some soils

to which the ash was added.  As with regular flyash,  modified  ash  also has

the ability to  improve the texture and drainage characteristics  o? some soils.

     A potential advantage for wet-collected modified flyash is  that the slurry

may be sprayed directly on the soil area without the need for  prior drying.
*A report of this research, performed by Dr. David Martens, is included as an
 appendix in the final report of Contract No. CPA 70-66.

                                  308

-------
Before such utilization is recommended,  however,  further research will be



necessary to determine whether a possible pollution problem due  to water




leaching occurs.



                              SUMMARY




     Wet-collected modified flyash has been found to be a suitable raw



material for the production of autoclaved structural products such as




bricks, aerated concrete and cement materials.  Other areas of application



are in heat treating for mineral wool production and sulfur dioxide  collection



as well as in soil amendment for agricultural purposes.  All of  the  processes



discussed utilize the unique characteristics of modified flyash  to eliminate



a potential solid waste disposal problem.  A more detailed discussion of  the



processes described in this paper may be found in the final report for  Contract




No. CPA 70-66 which will be available from the Environmental Protection Agency




in the near future.
                                  309

-------
                               REFERENCES

1.  Anderson, Ronald E., Cockrell, Charles F., e£ ajL.,  "Study of Potential
    for Recovering Unreacted Line From. Lines tone Modified Flyash by
    Agglomerate Flotation," Final Report, Contract PH  22-68-18,  National
    Air Pollution .Control Administration, May, 1970.

2.  Contract numbers CPA 70-66 and EHS D 71-11 between West Virginia
    University and the Environmental Protection Agency.

3.  	___, PFA Data Book, Central Electricity Generating Board,
    England (1969).

4.  Cockrell, C. F., Muter, R. B., and Leonard, J. W.,  "Study of the
    Potential for Profitable Utilization of Pulverized Coal Flyash
    Modified by the Addition of Limestone-Dolomite Sulfur Dioxide
    Removal Additives," Final Report, Contract PH 86-67-122, National
    Air Pollution Control Administration, April, 1969.
                               310

-------
                         FIGURE 1
 Silica-§and
30 x 100 Mesh
                  Wet-Collected Limestone
                 Modified Flyash (1% Slurry)
Dewater Ash
to 58% H20
                    50% Ash + 11% CaO
                      + 39% Sand
                      Paddle Mixer
                      (10 Minutes)
                        Pour into
                          Molds
                            T
                    24 Hours Thermal
                      Set (110°C)
                    Autoclave, 16 Hours
                     190 psig, 185°C
                       72 Hours Air
                          Drying
Lime (CaO)
 96% Pure
                 FLOW SHEET:  FORMED CONCRETE
                             311

-------
                         FIGURE 2
              Mech. Precip.  - Dry-Collected
            Limestone Modified Flyash  (927.)
8% Portland Cement
  Type I, Normal
       0.16% Aluminum
          Powder
                      Paddle Mixer
                      (10 Minutes)
                     Pour into Molds
                     8 Hour
Air Set
                   Autoclave,  16 Hours
                     190 psig,  185°C
                       72  Hour   Air
                          Drying
             FLOW SHEET:   AERATED CONCRETE
                             312

-------
                                    FIGURE 3
 Silica Sand
30 x 100 Mesh
                          Wet-Collected Limestone
                         Modified Flyash (1% Slurry)
Dewater Ash
to 34% H20
                            50% Flyash + 11% CaO +
                              39% Silica Sand
                             Paddle Mixer (10 Min.)
                                @ 20.4% H20
                               a.i.aKing Reactor
                                 (1 Hour)
                              Forming Pressure
                              3000 psi,  17.7%
                                    H20
                           '. » Hour Humidity Storage
                           95% Rel. Humidity @ STP
                             Autoclave,  8  Hours
                              190 psig,  185°C
                                72  Hours  Air
                                   Drying
Lime (CaO)
 96% Pure
                    FLOW SHEET:   CALCIUM-SILICATE BRICK
                                  313

-------
PILOT SCALE RESEARCH AND DEVELOPMENT
     F.T. Princiotta, Chairman
           Participants:

    B.N. Murthy, D.B. Harris, and J.L. Phillips
    I.S. Shah
    R.A. Person, C.R. Allenbach, I.S. Shah, S.J. Sawyer
    R.J. Gleason
    T.M. Kelso, P.C. Williamson, J.J. Schultz
    N.D. Moore
    A. Saleem, D. Harrison, N. Sekhar
    J.L. Shapiro and W.L. Kuo
    J.H. McCarthy and J.J. Roosen
    A.L. Plumley and M.R. Gogineni
    D.E. Reedy
    John M. Craig, Burke Bell, J.M. Fayadh
    Robert J. Phillips
    Ivor E. Campbell and James E. Foard
                315

-------
             Pilot  Scale  Research  and  Development




                    (Parts I, H, and HI)



  Second International Lime/Limestone  Wet  Scrubbing Symposium



                     New  Orleans,  Louisiana



                       November 8-12,  1971




                   F.T.  Princiotta, Chairman



                            SUMMARY





     In light of the recent trend toward preserving and improving



man's environment,  legislation and genuine public  concern has put



increasing pressure on electrical utilities to  control the discharge



of gaseous and particulate pollutants.  Since wet  limestone processes



are generally considered closest to the state of the art in controlling



sulfur oxide emissions, research and development has been particularly



intense over the last several years in this area.  The number of papers



in this session and the depth of investigation  supports the conclusion



that more intensive and comprehensive investigation of wet limestone



processes has occurred relatively recently.








     The Pilot Scale Research and Development Session of the Second



International Lime/Limestone Wet Scrubbing Symposium included papers
                                 316

-------
which presented pilot plant results and pilot plant status for a

variety of scrubber types.   Emphasis has been noted  in selection of

scrubbers which are non-tortuous  and not overly prone to  scaling or

plugging when handling saturated  or supersaturated limestone or lime

slurries.  For the most part emphasis  has  also been  on closed loop

systems, which although are the most prone to reliability problems,

are also the most acceptable since they minimize  the extent  of water

pollution problems.



     The participants in the Pilot Scale Session  were as  follows:

        Name                       Organization

        B. N. Murthy               Environmental Protection Agency
        I. S.  Shah                Chemico
        C. R. Allenbach            Union  Carbide
        R. J. Gleason               Cottrell Environmental Systems
        T. M. Kelso                Tennessee Valley Authority
        N. D. Moore                 Tennessee Valley Authority
        A. Saieem                   Ontario Hydro
        J. L. Shapiro               Bechtel
        J. H. McCarthy             Detroit Edison
        A. L. Plumley              Combustion  Engineering
        D. E. Reedy                Universal Oil Products
        B. Bell                     Zurn  Environmental Engineers
        R. J. Phillips              General Motors Technical Center
        I. E. Campbell              Smelter Control  Research Association

     Tables 1 and 2 describe the pilot facilities and summarize

results of selected pilot plant studies presented in  the Pilot Scale

Session.  Please note, that only typical  results  were  included  in the

table,  since  it was infeasible to include all results  from all  the units

described.  Also,  the EPA Prototype Facility at the  Shawnee Steam Plant

and the Bisehoff Prototype were included to show how  these larger plants

compare to the pilot scale unit in terms  of size and scrubber types.
                                  317

-------
Also included  were some selected pilot plant results from earlier



studies by Howden-ICI in the 1930's and 1940's in order to help put



the most recent results, as presented in the Pilot Scale Session, in



reasonable perspective.








     Scrubber types which have received the most attention have been



the: venturi, Turbulent Contact Absorber, Hydrofilter (flooded marble



bed), spray tower and packed tower.  Generally, pilot plant results



have indicated that under carefully selected operating conditions, all of



these scrubbers, with the probable exception of the packed tower due its



inherent plugging tendencies, can be operated with relatively high S0_



and particulate removal efficiencies, with acceptable reliability.  Pilot



plant results have indicated that the desirable combination of good



removal efficiencies without excessive down-time can be achieved, under the



following  conditions:  (1) high liquid to gas flow rate ratios, (2) high



solids content in the scrubber slurry, (3) long residence times in a delay



tank following the scrubber and  (4) proper selection of scrubber type.
                                     318

-------
                  Table 1.  Selected Wet Limestone Pilot Plants and Prototype Facilities
                                     Status and Design Characteristics
Plant
Howden - ICI
(Fullham)

Howden - ICI
(Bankside)






EPA/Zurn at
Key West and
Shawnee Steam
Plant (Mobile
Unit)
Bischoff
Scrubber
Prototype at
Steag Power
Plant, Lunen,
W. Germany
Ontario-Hydro

Cottrell unit at
Ohio Power's
Tidd Plant


Status
Tested
in mid-
1930's
Tested
1948-49






Tested
1970-71



Started
Feb
1971



Tested
1971

Tested
1970-71



Design Characteristics
Size
acfm
3,150


11,700







1,500




110,000





4,000

1,000




Alkali
Lime and
limestone

Whiting
chalk
(limestone)
and lime

;
1

Various
Scrubber
Grid (wood) -
packed

Grid (wood) -
packed






Inspirating type;
limestone ! Zurn Dustraxtor
and lime


Lime and
calcium
hydroxide




Venturi in series
with spray tower

i
i
i
i
\
Limestone

Limestone




Spray tower

Venturi in series
with a wetted
film packed bed


Pt of Alkali
Addition
Scrubber
circuit

Scrubber
circuit



Fuel
Coal
1.7-2.0%S

Oil
/w
2.7%S


I



Scrubber



Oil (1-2%S)
circuit at Key Wes
! coal (1. 5-3^


at Shawnee
:
i
Gas inlet Coal
duct i 2-4. 5%S
;



Scrubber
circuit

Scrubber
circuit (alsc



Coal
1020-1600
ppm SO2 i
Coal
i 1100-2000
some sim- 1 ppm SO2 ir
ulated boil- \
*er injection)!
<*>
to

-------
                 Table 1  (Cont.). Selected Wet Limestone Pilot Plants and Prototype Facilities

                                   Status  and Design Characteristics
CO
ro
o
Plant
TVA unit at the
Colbert Steam
Plant




C.E. Prototype
Facility at
Windsor, Conn.








Mohave/Navajo
Pilot Facility




EPA Prototype
Facility at
Shawnee
Steam Plant


Status
Tested
1971





Started
up 4/70;
tests run lor
various
customers and
in-house tests
since; current-
ly EPA boiler
calcined lime-
stone tests
start 12^6/71
Started
up 7/71;
no data,
available due
to pilot
.plant problems
Testing
sched.
start
3/72


Desicrn Characteristics
Size
acfm
2,600






12,000








!
two
4,000
circuits


Alkali
Primarily
limestone





Na2C03 ,
lime,
limestone,
dolomite
Scrubber
1) Venturi-rod/
packed bed;
2) venturi-rod/
spray tower;
3) 3-stage TCA
(mobile bed)
scrubber
Hydrofilter
(flooded marble
bed) 1 or 2
beds
(






Limestone,
lime, soda
ash (w/ and
w/o regen.
by lime)
!
j
three !
30,000
circuits
|
Limestone
and
hydrated
lime

]






1) TCA ;
2) venturi;
3) polygrid
packed;
4) Lurgi
impingement
1) Venturi-spray
or packed bed;
2) versatile TCA;
3) hydrofilter
(marble flooded
bed)
Pt of Alkali
Addition
Scrubber
circuit





a) Scrubber
circuit;
b) boiler
injection
(simulated)






Scrubber
circuit




a) Scrubber
circuit:
y
b) boiler
injection

i
Fuel
Coal,
3000 ppm
SOo inlet,
4g/scf part



Oil
(S02, fly-
ash added)






,

Coal
0.3-0. 8%S




Coal
1. 5-3%S





-------
                   Table 2.  Selected Wet Limestone Pilot Plants and Prototype Facilities
                              Typical Performance and Operating Parameters
Plant
Howden - ICI
(Fullham)





Howden - ICI
(Banks ide)





EPA/Zurn at
Key West and
Shawnee Steam
Plant (Mobile
Unit)
Bischoff
Scrubber
Prototype at
Steag Power
Plant, Lunen,
W. Germany ]


Typica
SO2
Removal, %
97-99






90-99






70




1 Performance
Particulate
Removal, %
97-99






approx 98






Not
measured



89 I 99.7
I











j
i
Stoich. %
Operating Parameters
L/G
p-pm/1000 cfm
100 lime; ? 97
130 lime-
stone




110 138






100 lime- Approx 600
stone (induced by
scrubber)
\
I
110 i


—


>







AP
in. HaP
1.2






5.7






12




10
(in venturi)






Comments
Scaling was a major
problem, but design
changes enabled fair
operating reliability;
L/G and scrubber size
excessive by modern
s tandar ds .
Reliable operation a
continuous problem;
Settling difficulties
encountered;
L/G and scrubber size
excessive by modern
standards.
Open-loop system;
Results shown for fresh
water—salt water slurry
SO2 removal approx 80%
Operating reliability good
Large prototype system;
Early results indicate
scaling and plugging
problems and mechanic;
difficulties;
System is down now; to b<
restarted in early 1972.

CO
ro

-------
             Table 2 (Cont.).  Selected Wet Limestone Pilot Plants and Prototype Facilities

                               Typical Performance and Operating Parameters
GO
INS
ro


Plant
Ontario Hydro



Cottrell Unit at
Ohio Power's
Tidd Plant




TVA Unit at
;he Colbert
Steam Plant3^








C.E. Prototype
Facility at
Windsor, Conn



Typical Performance
S02
Removal, %
72



55-98






1) Varied due
to scaling;
2) 77 (avg);
3) 92 (avg).







95+ (Na2CO3
1 bed);
80 (reliably
with CaCO3
and Ca(OH)«
2 beds)
!
Particulate
Removal, %
No part, in
flue gas


Not
measured





1) 99. 3;
2) 98.9-
99.3;
3) 98.3.







99






Stoich. %
130



100-120






1) 150;
2) 150;
3) 150.








100-110
(Na2CO3,
Ca(OH)2);
120-150
(CaC03)


Operating Parameters
L/G
gpm/1000 cfm
72



40






1)40;
2) 60 (to both);
3) 38.








15
(Na2C03);
20 per bed
(CaC03,
Ca(OH)2)

AP
in. H20
0.5-1



7-9 (total)






1) Approx 10
(total);
2) 16 (total);
3) 9.







Approx 6
(per bed)



1


Comments
Reliability of operation
has been good;
No serious scaling or
plugging problems.
Performance data based on
an 80-hour test; some
SO2 variation due to SO2
inlet variation;
Considerable plugging of
packed bed: 23 Ibs after
80 hours.
1) Major scaling in packed
bed- -deemed unacceptable;
2) A 354-hour test was run
with only 20 hours down
time — reliability good--
only minor scaling;
3) A 172 -hour test indicated
no scaling problems --
ball wear, grid erosion,
and particulate entrainment
were problems.
Unit designed to solve
field unit problems and
to predict performance
of new units.




-------
         Table 2 (Cont.).  Selected Wet Limestone Pilot Plants and Prototype Facilities
Plant
Mohave/Navaj o
Pilot Facility









SPA prototype
Facility at
Shawnee
Steam Plant






Typical Performance
so2
Removal, %
wm —










__









Particulate
Removal, %
— ^










«.—









Stoich. %
^ ^










--









Operating Parameters
L/G
;pm/1000 cfm
— —










--









AP
in. H20
— —










--









Comments
Plant aimed at evaluating
s crubbing of low SC>2
level flue gas ;
Plant versatile with
extensive instrumentation;
Plant to evaluate use of
soda ash scrubber
solutions with lime
regeneration;
Data should be available
within s e ver al months .
Most versatile wet
limestone facility;
Extensive instrumentation;
Tests are planned to
fully characterize wet
limestone scrubbing —
break- in, screening, and
long-term tests are
scheduled;
Unit to start up '3/72.
aNumbers listed opposite TVA Unit at the Colbert Steam Plant represent:
   1) venturi-rod/packed bed;
   2) venturi-rod/spray tower; and
   3) three-stage TCA (mobile bed) scrubber.

-------
      SULFUR DIOXIDE ABSORPTION STUDIES WITH EPA




          IN-HOUSE PILOT SCALE VENTURI SCRUBBER









               B. N. MURTHY and D. B. HARRIS




    EPA, RESEARCH TRIANGLE PARK, NORTH CAROLINA 27711




    J.L. PHILLIPS, RADIAN'CORPORATION, AUSTIN, TEXAS
                 FOR PRESENTATION AT THE




SECOND INTERNATIONAL LIME/LIMESTONE WET SCRUBBING SYMPOSIUM




                 NEW ORLEANS, LOUISIANA




                  NOVEMBER 8-12,  1971
                          325

-------
 SULFUR PIOXIM ABSOHPTIOK 8TOH8 WITH EPA ft Pt-HOUSE PILOT-SCALE
 VEHTURI  SCRUBBER
                        B. K.  Murthy 4 D.  1.  Harris,  EPA
                                     and
                        J. L.  Phillips,  Kadlan Corporation

 Tha  in-house venturi pilot  plant systems  which was originally designed and
 in use for pertlculate  removal studios, was  modified to Investigate  the
 lines tone vet-scrubbing process for sulfur dioxide removal.  Changes included
 replacement of  ductwork with  fiberglass-plastic  piping, installation of
 slurry mixing and hold  tank*  and clarifian  and  instrumentation for measuring
 pi! values, liquid and gas flow rates, slurry sampling, and for gas analysis.
 The  venturi section was replaced by a stainless  steel unit of similar dimen-
 sions  and modified flow nossle.
 Vapor-liquid and solid-liquid mass transfer  in the venturi scrubber ware
 studied  by treating flue gas  containing sulfur dioxide with water, sodium
 hydroxide solution and  limestone slurry in separate  ser&esof experiments.
 The  results indicated that  the renturl scrubber  alone is not an efficient
 §02  absorption  device.   Liquid flew rate  and liquid  composition had an
 appreciable effect°nthe overall mass transfer coefficient, indicating that
 the  liquid phase reaction was significant under  the  experimental conditions.
 Experiments with limestone  slurry indicated  some solid-liquid mass transfer
 in the venturi  scrubber•  Calcined limestone-flyash  mixture removed more
 sulfur dioxide  in the wet-scrubber than lines tone.
 The  in-house venturl pilot  plant studied  also enabled the development of
 suitable  methods for slurry sampling and  evaluation  of laboratory analytical
methods  for practical application in scrubber development.
                                326

-------
            SULFUR DIOXIDE ABSORPTION STUDIES WITH EPA IN-HOUSE
                        PILOT SCALE VENTURI SCRUBBER


The  .limestone  wet  scrubbing  process  for:sulfur dioxide absorption will
be studied on a large scale by EPA early in 1972 using three pilot plants on
a coal fired power generating station.  The venturiscrubber is one of the three
systems planned for investigations.

To study the  imnw+ant  process  variables   in preparation for the large
scale test plans, an existing EPA in-house venturi pilot plant which was in use
for particulate removal studies, was modified.  The main objectives of this
study were:

     (1)  To determine the extent of vapor-liquid and solid-liquid mass trans-
          fer in the venturi scrubber.

     (2)  To investigate the factors influencing the important mass transfer
          steps in the scrubber.

     (3)  To gather preliminary information on the absorption of NO .
                                                                   3\

     (4)  To develop practical sampling and analytical methods for the proto-
          type studies through trials under the pilot plant conditions.

EQUIPMENT

A flow  diagram  of  the modified pilot  plant system is shown in Figure 1.
The major changes in the equipment consisted of the following:

     (1)  The venturi scrubber, de-entrainer and the cyclone were replaced with
          those of 304 stainless steel.

     (2)  The flow nozzle in the venturi was fitted with a 45° solid cone, to
          handle slurry feed.  The dimensions of the venturi scrubber are shown
          in Figure 2.
                                                               i
     (3)  The ductwork downstream from the venturi up to the induced draft fan
          was constructed of 8 in. diameter fiberglas reinforced plastic pipe.

     (4)  The I.D. fan was coated inside with a corrosion resistant paint.

     (5)  Two 55 gallon capacity sealed hold tanks fitted with stirrers and in-
          let-outlet connections were installed in the slurry recycle line.

     (6)  A rectangular clarifier of 450 gallon- capacity was added to the
          system to treat the slurry.
                                     327

-------
OPERATION AND ANALYSIS

Flue  gas,  produced by burning  natural   gas in an incinerator, was diluted
with excess air to make up for the required volumetric flow rate through the
scrubber.  The temperature and humidity of the gas could be adjusted by pre-
spraying with water in the incinerator.  S02» NO and N02 were metered into
the flue gas stream before the mixer, which was a dry venturi.  The limestone
or other dry solids were added to the flue gas at this point.  The gas stream
then entered the venturi scrubber where it was mixed with the liquid reagent.
Downstream from the scrubber, the entrained liquid and particulates were
removed in a knock-out drum followed by a cyclone from where the gas was dis-
charged to the atmosphere through the I.D. fan.

Gas Sampling;

A Semi-automatic sampling system was constructed to monitor the flue gas com-
position  at several points in the pilot plant.  The sampling train included
a glass cyclone to remove entrained particulates and droplets.

Flue Gas Flow Measurement:

An ASME standard venturi gas flow meter was used to measure the gas flow rate
through the scrubber with an estimated accuracy within ±5%.  The flow venturi
was located in a straight run of duct following the I.D. fan.

Liquid and Slurry Flow Controls:

Standard venturi flowmeters were connected to LSN differential pressure cells
and continuous recorders to measure liquid and slurry flow rates.  The recorder-
chart readings were used to manually control the flow rates.

Liquid and Slurry Sampling;

Sampling points at several points in the slurry flow system were connected to
1/1 and 1/2 inch stainless steel tubings for sample draw out.  Care was taken
during sampling to prevent oxidation or reaction with atmospheric carbon
dioxide.

Solid-Liquid Separation During Slurry Sampling;

Laboratory tests with a continuous centrifuge had, demonstrated that calcium
carbonate particles Gpuld-be_rapidly separated from a slurry of similar com-
position as expected in a scrubber, before analysis to determine the rate of
precipitation and dissolution in the system.  However, under the pilot scale
scrubbing conditions in which the slurry contained fly ash and calcined lime-
stone particles, the performance of the centrifuge was not satisfactory.
The outlet liquid stream would become cloudy with suspended particles and the
flow lines in the centrifuge plugged with solids.  Further work in developing
a satisfactory alternate method  of solid-liquid separation has progressed else-
where.
                                     328

-------
 Gas Analysis:

 S02 and NOx were analyzed by Whittaker Analyzers which were calibrated with
 Span gases supplied in cylinders.   The instrument output was connected to a
 continuous recorder.

 02 and C02 were analyzed by a laboratory gas chromatograph.  Sampling was
 done by the use of plastic bags.

 Liquid £ Slurry Analysis:

 Sulfite in the liquid sample w^  analyzed immediately by a modified iodine
 method in which the pH of the solution was kept at 6.0-6.2 to prevent-inter-
 ference from nitrites.  The iodine solution was generated as needed for each
 determination  using standard iodate and excess iodide ion at low pH.  The
 excess iodine  was back-titrated with arsenite solution using a dead-stop
 technique for  end point detection.

.Samples were shipped under stable conditions and analyzed at Radian Labor-
 atories fos» the following species:

             carbonate
             total stilfinr-as sulfate-
             total nitrogen
             Ca
             Mg
             Na
             K
             Cl

 The methods used for analysis are described in Reference 2.

 RESULTS AND DISCUSSION;

 Experiments we*»e conducted in Ifwo series.   In the first series, the
 vapor liquid mass transfer efficiency of the scrubber for sulfur dioxide
 absorption was studied by using only water and sodium hydroxide solution
 as reagents to circumvent problems of solids  dissolving or precipitating
 in the scrubber.  The concentration of sodium hydroxide was in the range
 of the dissolved  reagent concentration in the wet-limestone scrubbing pro-
 cess.  The experimental conditions are shown in Table 1.

      The effect of operating variables on the mass-transfer efficiency was
 determined by  calculating relative values of T%a'f the product of S02 mass
 transfer coefficient and the interfacial area, from  the familiar expression
 for dilute gas concentration:
                   Y in p            	
     z =^- •                dy        G     „
                   Y outV y - y*  =  Kga     *
                                  329

-------
                         TABLE    1
           COHDTIONS    FOR   SERIES    1.
FLITS  GAS;






FLOW  RATS




TEMFETIATUI32
             NO



            ' NOg




             COg




              02



             H20
                               (ACFM)
                              (MCZE
1000   TO   1300




 250   TO    325








0.20



 0     &    O.cA-5



 0     &    0.005



0.3    TO   1.3




20



0.6    TO   1*.2



    BALANCE
KaOH  IN  WATER
                            (G MUUES/L)



                               (GFM)
 0     TO   0.0^




10     TO   15
          FKESStJKS  PROP;   (BJ.
                                                   7.8    TO   12.6
                                330.

-------
                 Kg = overall mass transfer coefficient for S0«
                 a  = gas-liquid interfacial area per unit volume,
                 y  = gas phase S02 concentration
                 y* = gas concentration in equilibrium with tb  liquid
                 G  = Average Molar gas flow rate through the scrubber
                 z  = length of the scrubber
                 N  = Number of transfer units
Relative value of Kga is then calculated as:
     (K a) rel  =  Kga     =  Nt           &
                 (KgtTref              *
Equilibrium gas concentrations were calculated from liquid concentrations,
using a computer program.  Experimental results (Table 2) indicated that
sulfur dioxide absorption did not reach equilibrium with the venturi scrubber.

Figure 3 shows the effect of liquid composition on Kga.  The strong dependence
the rate of sulfur dioxide absorption on the liquid composition indicates that
liquid film resistance in the venturi scrubber is significant in the concentra-
tion range studied.

The effect of gas and liquid flow rates on Kga is shown in Figure 4.  The
increase in mass transfer with liquid flow rate is expected to be due to
increase in surface area from the additional droplets.  An increase in gas
flow rate did not appear to have an appreciable effect on mass transfer.

The pressure drop across the scrubber, however, was affected more by gas flow-
rate than the liquid flow rate (Figure 5).  These results suggest that vapor-
liquid mass transfer in the venturi scrubber can be increased without appreciabJ
power input by increasing the liquid flow rate for the same gas flow rate.

Figure 6 shows the effect of flue gas temperature on the absorption rates for
water and alkali solution.  Comparison of the results indicate that while a
decrease in temperature increased absorption with water, an increase in temper-
ature increased the absorption rate when the alkali solution was used.  This
indicates that alkali absorption in the scrubber is controlled by the chemical
reaction mechanism in the liquid phase, since chemical reaction rate is a
strong function of temperature.  However, under process conditions when the
equilibrium partial pressure of S02 is not zero, a lower temperature may favor
absorption.  Further studies are needed before firm conclusions on the effect
of scrubber temperature on absorption rate can be made .
                                     331

-------
                                                       TABLE 2



                                          EXPERIMENTAL RESULTS - SERIES 1
CO
CO
ro
RUN NO.
1
1 A
2
3
3 A
3 B
4
5
6
6 A
7
8
8 A '
9
9 A
io
10 A
INLET FLUE GAS , 	
FLOW RATE
(ACFM)
1040
1090
1020
1230.
1210
1210
1010
1010
1070
1120
1050
1270
1255
1060
1115
1090
1100
TEMP.
(oF)
300
320-
310
315
300
300
310
' 260
310
300
310'
325 .
3.00
260
250
• 315.
300
COMPOSITION, MOLE %.
S02 '
.198
.200
.211
.205
.199
.290
-.206
.197
.200
.196
.198
.246
-.202
.239
.200
.239
.207
NOX
0
0
.054
.056
.051
.05
.044
.039
.044
.05:
.051
,049
.05
.048
.05
.03
.05
C02
.85
.6
1.2
.9
.4
--
1..3
.6
1.1
—
.7
.5

.3
—
.10

o2
19.6
20.6
17.7
19.6
18.8
—
18.8
20.1
19.8
—
19.2
20.1
—
19.8
—
19.2


H20
2.4
2.3
4.0
2.5
2.3
--.
i.s
4.2
2.5
—
2.7
1.6

1.2
—
.6

.AQUEOUS
TtoOR
Moles/1
0
0
0
0.
0
0
0
0
y
.0282
.0416
.0365
.0370
.0349
.0393
.0280
' .0190
.0116
FEED
Feed
Rate
(SPM)
ip
10
10
10
10
10
15-
10
10
9.9
15
10
9.9
10
10
10
9.7
ous
s/1








282
416
365
370
349
393
280
190
1116
FEED
eed
Rate
(SPM)
ip
10
10
10
10
10
15-
10
10
9.9
15
10
9.9
10
10
10
9.7
SCRUBBER
(nH2o>
— 	 	
8.75
8.6
7.8
12.0
11.4
12.6
9.0
8.5
8.4
9.3
9.2
12.4
12.4
10.6
9.5
8.6
8.7
SD« OUT (MOLE %) -
EXPERIMENTAL
.1645
,1630
.1780
.1500
.1560
.2460
.1570
.1590
.0850
.0850
.0460
.1475
.0980
.1475
.1100
.158.0.
.1420
EQUILIBRIUM
.1210
..1260
.097
.111
.095
.228
.090
.086
0
0
0
0
0
0
0
0
.0775.

-------
                                TABLE 3
CONDITIONS FOR SERIES 2

FLUE GAS:

    Flow Rate
    Temperature
    Composition (Mole %):
                NO
                    NO,
                                   1000 ACFM
                                   300° F
H20
          '2     WVJ      "U2    w$   U2    "2W     "2

        0.20   0.045  0.005   0.8   20    1.0   BALANCE
SOLIDS:
    Feed Rate:   150% Stoichiometrie
    Type:         Piqua       Piqua
                Limestone   Limestone
                  (1)
                           (2)
   TVA
 Calcined
Limestone
 Flyash

   (3)
    Composition: 86% CaCOg  86% CaCOs    35.2% as CaO
Particle Size:
 (Microns)
            Range:

            Median:
                            2.0-50   2.0-32   2.0-40.3

                            15.5      12.7      6.5
LIQUID:

    Flow Rate Trough Scrubber:     10 GPM

    Composition:      Saturated Liquid from Clarifier

    Pressure Drop ("H?0);    7.0-8.5
                                333

-------
NOX Absorption:  No absorption of NO  was noticed from the gas phase analysis
since the change in NOX concentration was within the range of instrumental
error.  Liquid . pjhaae analyses also did not indicate any significant absorption
of NOX.

Oxidation of absorbed S02 to sulfate was uniformly 10 percent in all the runs.
Since the absorption rate of S02 was considerably more (about U times) with
the alkali solution than with water, the rate of oxidation was proportional
to the concentration of the sulfite in the liquid phase, indicating a chemical
reaction or a liquid film mass transfer controlled mechanism for oxygen transfer
in the scrubber.  However, the concentration of oxygen in a power plant flue gas
will be considerably less than that initfce!se_tests, and the same oxidation
conditions may not apply to the prototype scrubbing conditions.

The extent of solid-liquid mass transfer in the venturi scrubber was qualitatively
studied in the second series of experiments.  Commercial grade limestone and
calcined:., limestone-flyash mixture from dry limestone injection studies at
Shawnee (TVA) were used to compare the effect of limestone type and composition
on flue gas sulfur dioxide removal.  The experimental conditions are shown in
Table 3.

Three modes of operation were used to study the mass-transfer effects.  In the
first mode, clear saturated liquor from the clarifier was fed to the scrubber
with solids injection, under steady state conditions.  In the second mode,
solids injection was suddenly stopped and gas monitoring continued.  The differ-
ence in sulfur dioxide removal between the first and the second mode indicated
whether any of the solids were hydrated in the scrubber.  In the third mode
of operation, with the system under steady state conditions as in the first
mode, solids injection was suddenly stopped and the slurry from the hold tanks
was directly fed to the scrubber, bypassing the clarifier.  The difference in
sulfur dioxide absorption between the first and the third mode of operation
indicated whether any hydrated solids were dissolved in the scrubber.

The results, shown in Figure 7, indicate that very little hydration and dissolution
of limestone or flyash-calcined limestone takes place in the venturi scrubber.
The calcined limestone-flyash mixture removed appreciably more sulfur dioxide
from the flue gas than limestone alone.  The results also indicate that the
limestone part.icle size had a minor effect on the sulfur dioxide absorption
in the size range studied.

CONCLUSIONS;

The results of this investigation lead us to the following conclusions:

     (1)  The venturi scrubber aldne is not an effective device for sulfur-
          dioxide removal from power plant flue gases.

     (2)  Liquid recirculation rate and composition have an appreciable effect
          on the rate of sulfur dioxide absorption in the venturi scrubber.
                                  334

-------
(3)  Further work is needed to establish the effects of flue  gas  temper-
     ature and composition on the effectiveness of the scrubber.

(4)  Negligible hydration and dissolution of the solids take  place in
     the venturi scrubber.

(5)  Limestone type, particle size and composition have considerable
     effect on sulfur dioxide absorption in the venturi.  Detailed
     studies are needed to optimize the conditions.
                               33*

-------
REFERENCES:
     .1.  Technical Note 200-006-19, "Evaluation of Sampling Procedures for
         Limestone-Wet Scrubbing Systems - Colbert Test Series", by Radian
         Corporation, Austin, Texas

     2.  Technical Notes 200-004-03 to 09, Radian Corporation

     3.  Final Report for EPA Contract No. CPA 22-69-138, by Radian Corporation
                                      336

-------
                 SOLIDS
      S02,NO,N02
GO
CO
  FURNACE
           FLUE GAS
                                                                 CYCLONE
VENTURI
SCRUBBER
                                           MAKE UP
                                            WATER

                                                                       FAN
                      DE-ETRAINER
 CONCENTRATED
  No OH   t
 SOLUTION '
         1 water
                                   t>
        -i
                      
-------
                  10
00
    MATERIAL
     14 GA.
     304 S.S.
  ADJUSTABLE CONE
30°  /          3/4" LIQUID FEED PIPE
         FIGURE 2   VENTUR! DIMENSIONS

-------
   INLET  GAS  TEMP:  3OO °F
   GAS  FLOW- HOO (ACFM)
    LIQUID  FLOW-' |Q (GPM)
   "" SCRUBBER AP- 9" HoO
                                     -2
       I     2     3    4     5    X 10
   NQOH CONC-. (G-MOLES/L)
          FIG-3   EFFECT  OF
LIQUID COMPOSITION  ON   SCRUBBING
                 339

-------
LU
_J
LU
ft
    r
   1000
                     -O
                     INLET GAS TEMP:  300 F
     NQOH CONC-r
                                 04 G-MOLES/L-
                                      1500
                GAS FLOW, ACFM
    10
LIQUID FLOW,  GPM
 FIG-4-EFFECT  OF  GAS  AND  LIQUID  FLOW
             RATES   ON   KGQ
                       340

-------
  15
  10
X

 •
z


cu

-------
LJ
    GAS FLOW. IO50  ACFM
      LIQUID  FLOW;  io  GPM
    _ AP:  8- 9-5" H~o
                        lMaOH(-03GMOLES/L.)
                        •© WATER
  240         280          320

         TEMPERATURE, *F
360
  FIG. 6.  EFFECT OF INLET GAS TEMPERATURE
                      342

-------
  80
  70
  60
UJ
O
cr
UJ 50
a.
> 40
O
5
LU
cr
  20
           UJ
           l-
           V)
           bJ
FLY ASH PLUS

CALCINED L.S.

  6.5 MICRONS
                                            (QUA L.S.
                                              r MICRONS

                                              ; MICRONS
                MODE OF OPERATION
    FIG.7  EFFECT OF LIMESTONE TYPE  AND

           OPERATING CONDITIONS  ON  SULFUR

           DIOXIDE    ABSORPTION
                          343

-------
Chemical Construction Corporation

    Pollution Control Division
              SO;, REMOVAL USING CALCIUM BASED ALKALIES

                         PILOT PLANT EXPERIENCE
                                  I.  S. SHAH
                  Chief,  Process Engineering and Development
                           Pollution Control Division
                      Chemical Construction Corporation
                    320, Park Ave., New York,  N.  Y.  10022
                                  Presented at
                        2nd International lime -  limestone
                           Wet Scrubbing Symposium
                             New Orleans,  Louisiana
                             November 8 - 12,  1971
                                      345

-------
 Chemical Construction Corporation
     Pollution Control Division

            SQ2 REMOVAL USING CALCIUM BASED ALKALIES
                       PILOT PLANT EXPERIENCE
        I. S. SHAH, CHIEF, PROCESS ENGINEERING AND DEVELOPMENT

 Introduction
 Since January 1970, Chemical Construction Corporation in cooperation with various
 utility companies has undertaken extensive pilot plant work to obtain basic data for
 the design of commercial plants to remove sulfur dioxide from boiler flue gases
 using calcium based alkali namely limestone (CaCOs), lime (CaO), Dolomitic  lime-
 stone (CaCO3« MgCOs),  Dolomitic lime (CaO« MgO),  and carbide sludge.  In addition
 to obtaining basic design information,  extensive useful information regarding scaling-
 build up problems,  effects of delay tank, particle size, and quality of limestone on
 SO2 removal efficiency, were obtained.

 In this presentation, we would like to present our experiences and important findings
 from the pilot plant work done to remove sulfur dioxide from boiler flue gases, using
 calcium based alkali.


 Pilot Plant Work
 Chemical Construction Corporation,  has conducted several pilot plant  test programs
 for sulfur dioxide removal using calcium based alkali. Pilot Plants were installed at
 eight different power generating stations,  all of whom use  coal as the fuel.  Of the
total of eight, seven were balanced draft boilers, and one was a cyclone type boiler.
                                     346

-------
Chemical Construction Corporation



    Pollution Control Division




Pilot Plant Work (Cont'd)




Table I is the summary of boiler type, type of fuel and its sulfur and ash content,




SO2 and dust loadings in the flue gas, for the eight generating stations.  A total of




fourteen limestones,  five different types of lime,  one  dolomitic limestone, one




dolomitic lime, and one type of carbide sludge were tested.  Table III summarizes




the composition of limestones tested, Table IV summarizes the composition of




various limes, and Table V summarizes the composition of dolomitic lime and lime-




stone, and  carbide  sludge.  Both 32-5 mesh and 200 mesh limestones were tested.




Lime was always slaked in accordance with standard slaking procedure.









Pilot Plant Facility




A generalized pilot plant will include a two stage scrubber-absorber system, an




I.D. Fan,  recycle pumps, agitated delay tank, slaking system,  lime or limestone




slurry tank, a thickener,  and a filter.  Table II  provides a summary of pilot plant




equipment installed at various  power plants, the type of alkali used, and the mesh




size to which it was ground.









The flue gas from the generating unit was taken  downstream of the Air heater.  The




duct size and location were determined to obtain an isokinetic sample while the boiler




is  operating at or near capacity, and the gas volume measured at the outlet of the




Scrubber system is 1500 cu.ft. /min (saturated).  The  scrubber-absorber system




consisted of two W-A venturi scrubbers arranged  in series.
                                   347

-------
Chemical Construction Corporation



    Pollution Control Division




Pilot Plant Operation




Flue gas, isokinetically withdrawn downstream of the air heater, enters the first




stage scrubber, where SO9 and fly ash are simultaneously removed by intimate




contact with recycled slurry of CaSOg, CaSO., Ca(OH)2 or CaCOo and fly ash.




The recycled  slurry is introduced into the scrubber tangentially through  open pipes,




and  it swirls down the converging section of the scrubber,  thus completely wetting it.




This liquor forms a curtan of liquor across the throat, and the accelerating gas




shatters it into small droplets, which provide surface for dust collection and absorp-




tion of sulfur dioxide.









The cleaned flue gas and liquor enter the separator where liquor is separated by




gravity,  and any liquor carryover is separated by cyclonic action.









The flue gas now enters the second  stage-venturi absorber-where SO- is  absorbed




by contacting with recycled slurry.  The flue gas is then exhausted to atmosphere




after passing through an I.D. Fan.









The make up limestone or lime slurry is added to the suction side of the  second stage




recycle pump, and the bleed from the second stage enters the  suction side of the first




stage recycle pump.  The bleed from  first  stage is sent to  thickener, the overflow
                                      348

-------
 Chemical Construction Corporation
    Pollution Control Division
Pilot Plant Operation (Cont'd)
returning to the scrubber-absorber system.  The thickener underflow is then filtered,
the filterate returned to the system,  and the cake is piled.


In the case where thickener and/or filter is not included, the bleed from first stage
or thickener underflow is sent to sewer or ash pond for disposal.  In case a delay
tank is part of the pilot plant facility,  the bleed from second stage, and first stage
product liquor are sent to delay tank.   The delay tank liquor is then recycled to the
first  stage  scrubber.


Samples and Data
The flue gas samples were taken at the inlet of first and second stage, and outlet of
second stage, and SC»2,  NOx and dust loadings were determined.  Various liquor
samples were taken and analyzed-both liquid and solid phases.   Temperatures and
pressures of various streams were measured.

Fly ash samples were collected and particle size determinations and mineral analysis
were  made.  Lime and/or limestone samples were  analyzed and analysis compared
with suppliers'  data.  Coal samples  were collected  during pilot plant operation and
proximate and ultimate analysis were  performed.   Table VI  summarizes the analy-
tical data collected during each pilot plant operation.

Boiler operating data, namely Air heater temp., excess oxygen,  steam production, and
                                      349

-------
Chemical Construction Corporation
    Pollution Control Division
Samples and Data (Cont'd)
generating load, were also collected during the pilot plant operation.

SC>2 Removal Using Limestone
During the pilot plant work at various power plants,  fourteen different limestones
were tested.  The  composition of these limestones is presented in Table m.  The
CaCOs content  ranged from a low of 86% to a high of 99. 3%.  The effects of various
operating variables on SO2 removal efficiency were determined.  Table Vn sum-
marizes the various variables and their ranges studied.

Effect of Limestone Quality on SO2 Absorption
The quality of limestone varies from mine to mine, and from region to region.  Of
the limestones tested, limestones 2 through 10 were  from one geographical region.
12 through 14 were from another region, and 1 and 11  were from two other regions.
Table VIE summarizes the data  showing the effect of limestone quality on SO2 absorp-
tion.

For limestones 2 through 5, limestone 4 was found to be the least reactive, when
ground to 325 mesh, whereas limestone 4 was the most reactive when ground to 200
mesh.

For limestones  12,  13 and 14, the least reactive was  14 when ground to 325 mesh, whereas
it was the most  reactive when ground to 200 mesh.
                                       350

-------
 Chemical Construction Corporation
    Pollution Control Division
Effect of Limestone Quality on SO2 Absorption (Cont'd)
The SO2 removal efficiency ranged from 30 to 80 percent,  and of the 14 limestones
tested,  only .two were found to be reactive enough to provide 70%-fSC>2 removal
efficiency at a reasonable stoichiometry.   This is not adequate to meet present
pollution codes for sulfur dioxide.

There is a wide variation in SC>2 removal efficiency with the variations in limestone
quality.  From the data obtained so far,  it is still  not possible to accurately predict
SO2 removal efficiency, just from the knowledge of the composition of the limestones.

Effect of Particle Size on SC>2  Abosrption
Smaller the particle size, larger the surface area available for absorption of SO2-
The surface area for 325 mesh particles is approximately  70% more than that for
200 mesh particles.  This clearly shows that limestone ground to 325 mesh should
provide higher SC>2 removal efficiency than that ground to 200 mesh. Table IX is a
summary of data showing effect of particle size on SO2 absorption.  The data does
prove that 325 mesh limestones are more  reactive than 200 mesh limestones,  except
for limestone no. 14.  The extra operating and investment  cost to grind 325 mesh lime-
stDnes must be compared with the cost for a higher stoichiometric  requirement and/or
for delay tank, to attain the same SC>2 removal efficiency.

Effect of Limestone Stoichiometry on SO2  Absorption
The CaCOs reacts with SO2 according to the following chemical reaction,
                                       351

-------
 Chemical Construction Corporation
     Pollution Control Division
Effect of Limestone Stoichiometry on SO2 Absorption (Cont'd)
         CaCO3 + SO2 + 2H2O  	>  CaSO3- 2H2O + CO2
 One mole of CaCO3 reacts with one mole of SO2,  to form a mole of Calcium Sulfite
 (CaSOs* 2H2O).  Thus, a 100% stoichiometric limestone requirement means one mole
 of CaCOs for each mole of SO2.  The 150% stoichiometric requirement means 1. 5
 mole of CaCO3 for each mole of SO2.  Theoretically, higher the Stoichiometry,
 higher the SO2 absorption.  Table  X summarizes the data showing the effect of
 Stoichiometry on absorption of SO2.  The data clearly shows that SO2 absorption
 efficiency does increase with increase in Stoichiometry. The increase in SO2 absorp-
 tion efficiency is not significant enough, compared with the increase in Stoichiometry.

 Effect of Delay Tank on SC>2 Absorption Efficiency
 As a result  of SO2 absorption using limestone, calcium sulfite is form ed.  This cal-
 cium sulfite is presumed to form an inert coating on the limestone particle, thus mak-
 ing limestone  unavailable for further reaction.  The purpose of the delay tank is to
 remove this inert coating, and thus make more limestone available for absorption of
 SO2.  A delay tank has also been used in Howden - ICI process, to minimize and con-
 trol scaling  problems by holding the slurry leaving the scrubber for a period long
 enough for the supersaturation to dissipate on the surface of suspended crystals.

During the pilot plant work, a study was undertaken to determine the effectiveness of
the delay tank  in  improving the SO2 absorption efficiency. The delay tank was  included
either in the  first stage recycle liquor loop, or second stage recycle liquor loop,  or in
                                      352

-------
 Chemical Construction Corporation



    Pollution Control Division







Effect of Delay Tank on SO? Absorption Efficiency (Cont'd)




both recycle liquor loops.









In the case of delay tank in  the second recycle loop, the slurry leaving the scrub-




ber,  and make up limestone slurry were held in the delay tank.  The liquor from




the delay tank was sent to the scrubber for absorption of SO2.  The bleed from the




second stage delay tank liquor,  was sent to the suction side of first stage recycle




liquor pump.  In the case of a delay tank in the first stage recycle loop,  the bleed




from second stage,  and the  slurry leaving the scrubber,  were held in the delay




tank.









Table XI  summarizes the data showing the effect of delay time on SC>2 removal effi-




ciency increases with increase in delay time. In one case, the SC>2 removal efficiency




increased  by as much as 19  points, whereas in the other cases, the increase was about




12 points.









The recycle liquor pH increased with the increase in delay time.   In one case, the




pH increased  from 4. 9 to 6. 5,  and in the other cases it increased  from 5. 2 to 5. 9.









The use of a delay tank was found to be more effective in the first  stage recycle liquor




loop, as any percentage point increase in SC>2 removal efficiency is a full point increase
                                       353

-------
Chemical Construction Corporation
    Pollution Control Division
Effect of Delay Tank on SO9 Absorption Efficiency (Cont'd)
in efficiency, whereas  in the case of a second stage loop, it is only a fraction of
a point.  The use of delay tank in both the recycle liquor loops di d not produce
any significant increase in SO2 removal efficiency over the use of delay tank in
the first stage recycle  liquor loop.

Overall Summary - SO2 Removal Using Limestone
The quality of limestone does have a marked influence on SC>2 removal efficiency,
with increase  in stoichiometry, use.of a delay tank,  use of 325 mesh ground lime-
stone, and with a high  quality limestone, SC>2 removal efficiency as high as 80-85%
can be attained.

SC>2 Removal Using Dolomitic Limestone and Dolomitic Lime
The dolomitic limestone contains carbonates of both calcium and magnesium.   The
composition of the dolomatic limestone is presented in Table V.  A maximum SO2
removal efficiency of 30% was attained in a single stage venturi  scrubber system.
The pH values for the feed  slurry were about 8. 0 and for the recycle liquor between
4. 8 and 5. 5.

The composition of the dolomitic  lime is given in Table V. With the use of dolomitic
lime in a single stage scrubber system, the maximum SO2 removal efficiency  attained
was 92% (1800  ppm inlet) with 130% stoichiometric lime requirement.   The presence
of magnesium  oxide make dolomitic lime much more reactive.
                                       354

-------
 Chemical Construction Corporation

    Pollution Control Division

SC>2 Removal Using Dolomitic Limestone and Dolomitic Lime (Cont'd)

It is our opinion, that dolomitic lime or limestone should not be used as an absorb-

ing reagent in the case of a throw away process, as the soluble magnesium-sulfur

compounds will result in water pollution problems.



SQr> Removal Using Lime

Five different types of limes have been evaluated to determine SO2 removal efficiency.

The compositions of these five limes are presented in Table IV. The summary of

the data showing the effect of stoichiometry,  the effects of single stage and two stage

operations and effects of modified single and two stage operations on SC>2 removal

efficiency, is presented in Table XII.  The data clearly shows:

   (a)  That increasing the stoichiometry, increases the SC>2 removal efficiency.

   (b)  That with the use  of a two stage scrubber system a  better utilization of

       lime can be achieved.  At 100% stoichiometry, the SC>2 removal efficiencies

       for a  single stage and two stage operations  are respectively 60% and 83%.

   (c)  That the modified single or two stage systems provide still better utilization

        of lime, as can be seen from the following:

                                                    %  Stoichiometry     % Efficiency

               Single Stage                               100               60.0
               Modified Single Stage                     100               72.0
               Two Stage                                100               83.0
               Modified Two Stage                        100               88.0
   (d)  That with the use  of lime as an absorbing agent,  adequate SC>2 removal effi-

        ciency can be attained to reduce outlet SC>2 concentrations less than 150 ppm.
                                        355

-------
Chemical Construction Corporation
    Pollution Control Division
SO2 Removal Using Carbide Sludge
Carbide sludge is produced during the manufacture of acetylene using calcium car-
bide.
                 CaC2 + 2H2O •	> C2H2 + Ca(OH)2
The composition of the carbide sludge tested in the pilot scale is presented in Table
V.  The data for the single and two stage operations, showing the effects of stoichio-
metry, effects of modified one and two stage operations and effects of delay tank are
summarized in Table Xin.  The data clearly shows:
    (a)  That increasing the stoichiometry,  increases the SO2 removal efficiency.
    (b)  That use of two stage operation provides higher  SO2 removal efficiency than
        single stage operation.
    (c)  That modified system provides improved performance.
    (d)  That use of delay tank further improves SO2 removal efficiency,  and
    (e)  That with the use of carbide sludge as an absorbing agent,  adequate SO2
        removal efficiency can be attained to reduce outlet SO2 concentrations below
        150 ppm.

Thickener and Filter Tests
During every pilot plant operation,  equipment vendors performed necessary tests to
determine thickener and/or filter performance.  This was essential so that the vendors
can supply suitable equipment with guaranteed performance for the commercial plant.

Thickener underflow concentrations in the range of 40 - 60% solids  can be attained
                                      356

-------
 Chemical Construction Corporation



     Pollution Control Division




 Pilot Plant Work (Cont'd)




 with clean overflow.  In some cases use of flocculants was found necessary.   With




 the use of a suitable vacuum filter, filter cake containing 60% - 80% solids can be




 obtained.  The cake  characteristics were studied and were found suitable in most




 cases for truck transportation.









 Open Loop vs. Closed Loop Operations




 In the open loop operation, make up lime or limestone slurry was added to the scrub




 ber system, and the bleed from the scrubber system was discharged to the disposal




 area.  In the case of partially closed loop operation, the scrubber bleed was sent to




 a thickener, the  thickener overflow was returned to scrubber and the underflow was




 sent to disposal.  In the case of completely closed loop operation, the thickener




 underflow was sent to a filter, the filtrate was returned to the scrubber, and filtered




 cake was sent to disposal.









 Essentially the same SO2  removal efficiency was obtained during open and closed loo




 type operations.   During closed loop operations,  except for some scaling and build u




 no noticeable difference in operation was observed.









Scaling and Build-Up Problems




The use  of lime or limestone slurry as the absorbing agent results in a scrubbing li-




quor containing calcium sulfate,  calcium sulfite and unreacted calcium carbonate or




calcium hydroxide.  The recycle liquor being a slurry, the clear solution is always
                                      357

-------
Chemical Construction Corporation
    Pollution Control Division
Scaling and Build-Up Problems (Cont'd)
a saturated solution, containing CaSO3, CaSC>4 and CaCOg.  The mechanism for
SO2 absorption and precipitation of sulfite or sulfate is believed to be as follows:
                      S02(g) + H20(D	/ H2S03(i)
                      CaC03som + H2SO3— '-CaSOgscHn + H2O + CO2
                      CaSO3 excess    -  '- CaSO3 saturated + CaSO3 ppt.
                      CaCOs solid     -  '  CaCO3soln
Sulfur dioxide from the flue gas dissolves in water to form sulfurous acid, which reacts
with calcium carbonate in solution to form calcium sulfite. This causes a super sat-
uration in terms of calcium sulfite,  resulting in precipitation of calcium sulfite.  Some
of the calcium carbonate from the solid phase dissolves in the solution thus  causing
it to reach saturation.  Any oxidation of sulfite to sulfate causes sulfate to precipitate
out.  The inverse solubilities of both calcium sulfite and calcium sulfate, with in-
creasing temperature, does enhence the precipitation of both compounds.  This pre-
cipitation occurs within the scrubber on various surfaces, causing the scaling and build-
up problems.

The basic scaling characteristics  of solutions containing calcium sulfite and calcium
sulfate,  are inherent in nature.  Scaling and build-up will occur, regardless of the
type of scrubber-absorber selected, but how fast and how much will depend  on type
of scrubber and method of operation. The proper point of introduction of lime or
limestone in the scrubber recycle liquor loop, the presence of delay tank, and selection
of scrubbing equipment with minimum internals,  should minimize scaling and buildup
                                      358

-------
 Chemical Construction Corporation
     Pollution Control Division
Scaling and Build-Up Problems (Cont'd)
problems.  Based on pilot plant work done,  it is our opinion that scaling and buildup
problems will be more pronounced during a closed loop operation than during par-
tially open or open loop operations.  We are of the opinion that if and whenever poss-
ible,  the slurry  of calcium - sulfur compounds should be kept outside the path of gas
flow, to attain highest possible reliability for boiler operation, to generate  electricity.

Summary
The  limestone quality seems to vary from one region to another region,  and has  sig-
nificant effect on SC>2 absorption.   Smaller particle size and higher stoichimetry help
increase SC>2 absorption efficiency.  The presence of a delay tank increases effi-
ciency or help reduce limestone requirements for the same efficiency.  With a suit-
able  combination of limestone quality, particle size,  reasonable stoichiometry and dela
time, SC>2 removal efficiency as high as 80-85% can be attained.  With the use of lime
or carbide sludge as the absorbing agent, one  can attain 90+% efficiency  at 130% or less
stoichiometric lime requirement.   Selection of alkali should be made after proper econ
omic evaluation which should include stoichiometric requirement, alkali cost, grinding
and preparation costs,  disposal cost and installed equipment cost.

Scaling and build-up problems are inherent when using lime or limestone slurry as ab-
sorbing liquor.   Proper methods of operation,  and selection of scrubber-absorption
equipment, may  help reduce the frequency and extent of these problems.
                                      359

-------
                                      Table I.  Summary of Boiler Operating Data





                            (1)         (2)         (3)        (4)           (5)          (6)          (7)         (8)


    Boiler Type          Balanced   Balanced   Balanced   Cyclone     Balanced    Balanced   Balanced   Balanced

                           Draft      Draft      Draft                   Draft       Draft       Draft       Draft


    Fuel                 Coal        Coal       Coal       Coal        Coal        Coal       Coal       Coal




    % Ash               22-23     20-25     20-25     20-25      16.8-28.9    16.3-17.8  9-11       10.3-12.4



S
°   % Sulfur             0. 68-0. 71   5. 5  - 7.  3   5. 5  - 7. 3   5. 5 - 7. 3   3. 2 - 5. 5    2.1 -  2. 3  2. 5  - 2. 8   3. 3  - 4.1




    Inlet SO2, ppm       450-575   4000-5200  4000-5200  4000-5200   1500-3800    1100-1600  1500-2200   1700-2250




    Inlet Dust, Gr/SCFD 3. 9 -11. 7   4. 3  - 8.  3   4. 3  - 8. 3   0. 85 - 2. 5   4. 0 - 10. 0   3. 0 - 6. 0  3. 0  - 4. 5   3. 0  - 5. 0

-------
                      Table II Summary of Pilot Plant Facility at Various Power Plants

                           (1)           (2)          (3)           (4)          (5)         (6)        (7)        (8)
 Pilot Plant Facility

   I. D. Fan                V          V          /          \            V          \S
   Single Stage             \X                                                 ,        I/-        ^
   Two Stage                             V           * x'          « x          ^"                   ^^
   Recycle Pumps          ^-          ,
   Delay Tank                           . ,-
   Thickener                            ^
   Filter
   Closed Loop                                                                                     ^
   Open Loop               \X'          V-'"         W"          'vX'         ^
   Slaking System          v '          \^x
   Lime Slurry Tank        \x'          vX"                                   ^^
   Limestone Slurry Tank

•
'Alkali Used
   Limestone
   Mesh Size
     -325                  vX"         VX"         ux         v^--        v  --'       V
     -200                  v "'         Yx"         W-'"         \           t•-•''        V-
   Lime
   Dolomite
   Calcined Dolomite
   Carbide Sludge

-------
                               TABLE III - SUMMARY OF LIMESTONE COMPOSITIONS
(1) (2)
Si02 4.26
Fe2°3 °'63
A1203 0.36
CaC03 95-98.25 91.84
MgC03 2.96
R2°3
CO C
JET O
ro
(3)
5.74
0.80
0.53
88.98
4.45



(4)
7.51
0.90
0.60
85.78
5.12



(5)
6.13
0.52
0.61
91.67
1.08



(6)
6.07
0.55
0.48
91.72
1.05



(7)
3.75
0.58
0.24
94.22
1.78
0.82
0.068

(8)
4.84
0.92
0.38
88.56
5.18
1.30
0.07

(9)
7.07
0.92
0.44
86.71
3.78
1.36
0.06

(10)
1.60
0.09
0.28
96.81
0.80

0.06

(11) (12)
0.
0.
0.
95
3.
0.


82
16
18
.21 99.3
51
34


(13) (14)



86.8 93.4




P          0.03-0-1




Na-Kpxides 0.2-0-22




LOI                  41.50  40.56  40.20  40.83  40.84  41.98  41.73  40.56 42.93




Mesh-325             96     96     96     96     96     96     96     96    99+%          90+    90+   90+




    -200             98     98     98     98     98     98     98     98                   90+    90+   90+

-------
               TABLE IV - SUMMARY  OF LIME COMPOSITIONS
               (1)         (2)          (3)         (4)         (5)
Cao
Mgo
Fe20  ]
A120|!
*2o3

LOI
77.3
0.2
21.5
1.0



85.58
0.52
0.77
0.11
0.15
0.26
0.014
12.75
79.82
0.23
1.09
0.07)
0.12)
0.19
0.01
18.95
95.0
1.08
1.3
0.95)

0.03
1.75
95.0
1.08
1.3
0.95

0.03
1.75
                                   363

-------
TABLE V - COMPOSITIONS OF DOLOMATIC LIMESTONE & LIME & CARBIDE SLUDGE
Dolomatic Limestone       Dolomatic Lime       Carbide Sludge

Cao  60                    57                  Ca(OH)2     92.5
Mgo  40                    43                  CaC03        1.85
     97% thru              83% thru            Sio2         1.50
     325 mesh              200 mesh            MgO          0.70
                                                            1.60
                                364

-------
    Chemical Construction Corporation

       Pollution Control Division
                                 Table VI  Summary Of

                                     Analytical Data
Flue Gas                  Fly Ash                   Coal                 Liquor Analysis

Dust loading, gr/SCFD     Particle size Distribution  Ultimate Analysis    Slurry Analysis
SC>2 loading,  ppm         Mineral Analysis      (as received &dry basis)    % Solids
NOX, ppm                   P2°5                    Moisture                  pH
Temp., °F                  SiC>2                     Carbon
Pressure, "  W. G.          Fe2Os                   Hydrogen            Filtrate Analysis
   at                       Al2Os                   Nitrogen                  pH
1st stage inlet               TiO2                    Chlorine             dissolved Solids
2nd stage inlet              CaO                     Sulfur                    cl  ppm
2nd stage outlet             MgO                     Ash                      SO4 ppm
                           SOs                     Oxygen                  Ca  ppm
                           K2O                                              Mg  ppm
                                                    Proximate Analysis      SiO2 ppm
                                                /        •   j o  j   i.   • \     SO2 ppm
                                                (as received & dry basis) _,    . A1,   ,. ...
                                                    -,,_.,     J        Phenol Alkalinity
                                                    % Moisture                         J
                                                    % Ash                Total Alka inity
                                                    % Volatile            Conductivity mmhos
                                                    % Fixed Carbon
                                                      BTU               Dry Sollds
                                                    % Sulfur                % S°2
                                                                            % S04
                                            365

-------
Chemical Construction Corporation

    Pollution Control Division
                  Table VII variables and Their Ranges
         Variables

         Gas  Veloctiy, ft/sec

         Liquor to Gas Ratio,  gallons/1000 ACFM

         Limestone Stoichiometry

         Limestone Mesh Size


         Delay Time, Minutes

         Inlet SO2 » ppm
  Range

 50 - 250

 10 - 60

100 - 350

200 mesh and
325 mesh

 10 - 60

450 - 5200
                                   366

-------
                          TABLE VIII - EFFECT OF LIMESTONE QUALITY ON  SO,, ABSORPTION
CO
CM
% CaCo3
91.84
88.98
85.78
91.67
91.84
88.98
85.78
91.84
88.98
85.78
99.3
86.8
93.4
99.3
86.8
93.4
99.3
86.8
93.4
99.3
86.8
93.4
Limestone
2
3
4
5
2
3
4
2
3
4
12
13
14
12
13
14
12
13
14
12
13
14
Mesh Size % Stoichiometry Inlet SO2 ppm
96% thru 150 4150-4250
- 325 mesh


96% thru 350 4050-4200
- 325 mesh

98% thru 300 4400
- 200 mesh

90% thru 150 1500
- 325 mesh

90% thru 250 1500
- 325 mesh

90% thru 150 1500
- 200 mesh

90% thru 250 1500
- 200 mesh

Efficiency

    44.4
    47.0
    32.8
    41.2

    49.6
    55.4
    44.1

    39.4
    37.2
    44.7

    76.0
    57.0
    55.0

    80.0
    70.0
    60.0

    53.0
    49.0
    63.0

    65.0
    56.0
    68.0

-------
Chemical Construction Corporation

    Pollution Control Division
              Table DC.  Effect Of Particle Size On SO2 Absorption
 Limestone
    12
    13
    14
    11
%CaCO3

  99. 3


  86. 8


  93.4


  95.21


  91.84


  88.98


  85. 78


  91.67
  Mesh Size

90%+ thru 325
90%+ thru 200

90%+ thru 325
90%+ thru 200

90%+ thru 325
90%+ thru 200

90%+ thru 325
90%+ thru 200

96%  thru 325
98%  thru 200

96%  thru 325
98%  thru 200

96%  thru 325
98%  thru 200

96%  thru 325
98%  thru 200
Inlet SO2, ppm

   1500


   1500


   1500


   2900


   4000


   4200


   4200


   4200
%Efficiency

   75.0
   53.0

   57. 0
   49.0

   55.0
   62.0

   62.0
   54.0

   44.4
   26.1

   47. 0
   25.4

   32. 8
   30. 5

   41. 5
   34.0
                                       368

-------
Chemical Construction Corporation



    Pollution Control Division








       TABLE  X -  EFFECT  OF LIMESTONE STOICHIOMETRY  ON  SO,, ABSORPTION
% CaCo,
91.84

91.84

85.78

85.78

96.81


99.30


99.30


86.8


86.8


93.4


93.4


95.21


Limestone
2-325 mesh

2-200 mesh

4-325 mesh

4-200 mesh

10-325 mesh


12-200 mesh


12-325 mesh


13-200 mesh


13-325 mesh


14-200 mesh


14-325 mesh


11-325 mesh


% Stoichiometry
167
321
162
317
150
329
132
298
157
232
312
150
200
250
150
200
250
150
200
250
150
200
250
150
200
250
150
200
250
100
150
200
	 z 	
% SO? Efficiei
33.0
45.0
27.0
39.0
32.8
49.4
30.5
44.7
44.5
54.5
58.8
52.5
61.0
65.0
75.0
78.0
79.5
49.0
56.5
56.5
57.0
61.5
69.5
63.8
64.0
70.3
55.0
59.0
60.5
60.0
64.0
68.0
                                         369

-------
         TABLE XI - Effect of Delay Tank on SO? Removal


Delay Tank in l_st_Stage recycle liquor loop

Delay Time       % SO? Removal Efficiency    Recycle liquor pH    Limestone

Tx = 0 min                46.0                     4.9                3
T2                        58.5                     5.4
T3                        65.4                     6.5

T-L = 0 min                64.0                     5.2               11
T2                        74.0                     5.6
T3                        76.0                     5.9


Effect of Delay Tank location on SO^ Removal Efficiency

Delay Tank Location       Delay Time      % SO? Removal Efficiency

1st stage                    T2                   54.0
2nd stage                    T,                   54.0

1st stage                    T3                   65.0
2nd stage                    T~                   64.0
                                  370

-------
                     TABLE XII - SO? Removal Using Lime
Lime
% Stoichiometry    Inlet SO? ppm   % Efficiency
              100
              140
              220

              100
              100
              130
              145
              225

              100
              130
              100
              130
              150

              100
              140
                        1500
                        1500
                        1500

                        1500
                        1500
                        1500
                        1500
                        1500

                        1500
                        1500
                   4000-6400
                   4000-6400
                   4000-6400

                   4000-6400
                   4000-6400
83.0
94.0
95.0

88.0
60.0
64.0
71.0
77.0

72.0
88.0
66.0
75.0
80.0

80.0
89.0
Two Stage
scrubber
system

Modified Two
stage scrubber
system

Single stage
scrubber
system
Modified single
stage scrubber
system

Single stage
scrubber system
Two stage
scrubber system
                                371

-------
Chemical Construction Corporation

    Pollution Control Division
                   Table Xin SC>2 Removal Using Carbide Sludge
 % Stoichiometry

     100

     100

     100


     130

     130

     130


     100

     100

     100


     120

     120

     120
Single Stage

Modified Single Stage

Modified Single Stage
and Delay Tank

Single Stage

Modified Single Stage

Modified Single Stage
and Delay Tank

Two Stage

Modified Two Stage

Modified Two Stage
and Delay Tank

Two Stage

Modified Two Stage

Modified Two Stage
and Delay Tank
Inlet SO2 ppm
2500 -
2500 -
2500 -
2500 -
2500 -
2500 -
2500 -
2500 -
2500 -
2500 -
2500 -
2500 -
3000
3000
3000
3000
3000
3000
3000
3000
3000
3000
3000
3000
% Efficiency
62.0
69.0
72. 5
73
78
79. 5
74.0
78.0
92. 0
85. 0
92. 0
93. 0
                                         372

-------
              A  PILOT  PLANT  TEST PROGRAM FOR

                SULFUR DIOXIDE  REMOVAL FROM

           BOILER FLUE GASES  USING  LIMESTONE

                       AND  HYDRATED  LIME
                                  AT
                     UNION  CARBIDE CORPORATION

                           POWER STATION

                           MARIETTA,  OHIO
Union Carbide Corporation

Dr. R. A. Person
Manager-Environmental Control
  and Scope  Engineering
Engineering Department
Ferroalloys Division
Niagara Falls, New York
Dr. C.  R.  Allenbach
Staff Engineer-
  Environmental  Control
Engineering Department
Ferroalloys Division
Niagara Falls, New York
Chemical  Construction Corporation

I.  S. Shah
Chief, Process Engineering  and
  Development
Pollution  Control Division
New York, New York

S. J. Sawyer
Senior Process Engineer
Pollution- Control Division
New York, New York
                             Presented at
                 Second International Lime - Limestone
                       Wet Scrubbing Symposium
                        New  Orleans, Louisiana
                         November 8 - 12,. 1971
                                373

-------
                           ABSTRACT
      A pilot plant test  program to determine the  performance  and
optimum operating conditions of a Chemico Venturi Scrubber System
for the simultaneous removal of fly ash and sulfur dioxide was
performed at Union  Carbide  Corporation's coal-fired power station
near Marietta, Ohio.
      Hydrated lime and limestone were evaluated for sulfur dioxide
removal  during the test  program.  Fly ash removal efficiency  of the
venturi scrubber was determined  simultaneously with sulfur dioxide
removal.
      The test program was  performed in three phases:
      1.  Fly ash scrubbing  using water as  the  scrubbing  media.
      2.  Sulfur dioxide absorption using a hydrated lime  slurry.
      3.  Sulfur dioxide absorption using a limestone slurry.
      This paper discusses the  effect on sulfur dioxide absorption of
such  operating variables as liquid to  gas  ratio,  stoichiometry,  number
of stages, and  delay tank.
      Based upon the results obtained from the pilot plant tests,  it
was concluded that satisfactory technical  performance  using limestone
or hydrated lime in a two-stage  venturi system could be achieved on
a  full-scale  plant.   Estimated  capital and operating costs for a full-
scale  installation are presented.
                                 374

-------
                        INTRODUCTION

      Many coal-burning power generating plants  are being  required
to reduce particulate and  sulfur dioxide  emissions.  Some approaches
being taken to control  the particulate  emission problem are wet
scrubbers, more  efficient  electrostatic precipitators  and  conversion
to oil.   To decrease sulfur  dioxide emissions, either the use  of low
sulfur content fuels or various methods of desulfurizing the flue gases
can be  utilized.   Each  approach  poses many technical and  economic
problems,  and  considerable  effort is being  expended to develop the
most  acceptable  solutions.
      At U-nion Carbide's  Marietta, Ohio  power station, a captive  high
ash and sulfur content coal has  been  burned since the start of opera-
tions  in 1951.  The  station is equipped with mechanical  collectors and
two-field electrostatic  precipitators to control particulate emissions.
      The Parkersburg,  West Virginia/Marietta, Ohio Interstate
Abatement Conference issued  specific  recommendations in March and
April,  1970, calling  for a 70% reduction  of  sulfur oxide emissions and
a decrease in fly ash  emissions  to a  0.14 Ib./mm Btu heat input  level.
Compliance dates for these  recommendations  are  April,  1972 for sulfur
oxides  and March,  1973 for particulate.   Proposed State  of Ohio
regulations,  now pending, are more stringent.  .Particulate  emissions
of 0.10 Ib./mm Btu heat  input and sulfur dioxide emissions of 1.6 lb./
mm Btu heat input,  equivalent to about a 1%  sulfur coal  (12,500 Btu/lb.),
are proposed for this size plant.
      A survey of possible methods of reducing emissions pointed  out
the desirability of evaluating the use  of  wet scrubbing,  with a calcium
compound throw-away system, for control of both the  particulate and
sulfur dioxide emissions.  To evaluate such a system, a two-stage
venturi scrubber  pilot plant  supplied by Chemical Construction
Corporation was  installed at the power station.   Major aims  of the
                                  375

-------
program  were to determine particulate removal capabilities of the
venturi scrubber and the sulfur dioxide removal  efficiencies  of
two alkaline materials—limestone and hydrated lime.

                 DESCRIPTION OF POWER STATION

      The  Union Carbide power station is a high-load factor station,
rated at 200 megawatts, located on the Ohio River near Marietta, Ohio.
This  station supplies power to  all units  of Union Carbide's  manufacturing
complex in Marietta.  In addition  to power,  the  station also provides
process  steam  to the complex.   This additional  requirement  results in
a  coal consumption  equivalent to that of a 250  megawatt station.
      The  power station  has  four identical  boilers, each currently
equipped with a cyclone-type mechanical collector and an electrostatic
precipitator.  At full capacity,  the flue gas quantity  from  each boiler
is  approximately 280,000 ACFM at 320°  F.  entering the collectors.
      The  coal  burned at the  plant has been,  until recently, Meigs #9
Seam from Noble County, Ohio, with a typical ash and  sulfur content
of  about 20% and 5%, respectively.  Tests  were made with  this coal,
although currently  higher quality blend coal is being  used to comply
with  interim emission reductions recommended by the Interstate
Abatement Conference.

                    OBJECTIVE  OF TEST PROGRAM

      The objective of the pilot plant test  program was to confirm
engineering design parameters  and  develop  design data for a commercial
installation.  The key elements  of the program were:
                                  376

-------
          To  correlate particulate and sulfur dioxide removal
          efficiencies  with  various  scrubber operating conditions.
          To  establish a relationship  between outlet dust loadings
          and pressure drop across  the  venturi throat.
          To  determine sulfur dioxide absorption  efficiencies using
          limestone and  hydrated lime,  and to evaluate the effect
          of the following  variables:
          Limestone
          (a)   particle size - nominal 200  mesh by  Down and
               325 mesh  by Down
          *(b)   stoichiometric feed rate
          (c)   liquid to gas ratio
          (d)   delay tank
                                              s
          Hydrated  Lime
          (a)   single-stage  versus  two-stage scrubbing
          (b)   stoichiometric feed rate
          (c)   liquid to gas ratio
          (d)   delay tank
      4.  To  confirm the optimum design conditions  using limestone
          and  observe the  degree of scaling or depositing when
          operating for an  extended test period.

                            PILOT PLANT

      A slip stream for the  pilot plant was withdrawn after the air
preheater and  before the mechanical collectors.   Location and  size
of the gas withdrawal  duct  were determined by traverses to  assure
that  an isokinetic flue gas  sample was  taken for the  pilot plant.
The isokinetic condition  corresponded to capacity  operation of  the
boiler and 1500 cfm saturated gas  flow  from the pilot plant.
                                 377

-------
      A flow sheet of the 1500 cfm two-stage  pilot plant  is  shown
in Figure 1.  Figures 2 and 3 illustrate the  installation.
      The alkali feed, for a two-stage  configuration,  is fed  to the
second stage recycle loop with the second stage  bleed going to  the
first  stage  recycle loop.  This gives an alkali flow countercurrent
to the flue gas producing the  most efficient  absorption system.
      The delay tank is part of the first stage recycle loop.  Its
purpose  is  to allow  for additional  reaction time and increased  utilization
of the alkali by providing time for removal of  reaction products that coat
the unreacted alkali.  During  two-stage testing, the second stage bleed
is also fed to the delay tank.
      The main components  of the  pilot plant are  two  wet approach
venturi scrubber-separator units.   The flue gas is drawn into the scrubber
at the top and  the recycle  liquid is  fed through tangential nozzles.  The
liquid creates  a curtain across the venturi throat  and  the  accelerated
flue gas breaks up the  liquid into fine  droplets which  effectively interact
with  the fly ash and sulfur dioxide.
      The flue gases are  disengaged from the slurry, in both separator
stages, by  centrifugal and  gravitational forces.
      The recycle slurry  is  circulated by stainless steel pumps and  all
recycle piping is of stainless  steel construction.   Magnetic rotameters
are used to measure the flow.
      An induced draft fan is used to pull the  stack gas through  the
system.   It is of stainless  steel construction  and is capable of
pulling 1800 cfm with a  maximum pressure drop of 56  inches w.g.
                                 378

-------
      The other accessory equipment consists of:
      (a)  Two feed tanks of 5000 and 7000  gallon capacity
      (b)  A delay  tank  of 4500  gallon capacity
      (c)  Feed pumps with  20 gpm capacity
      (d)  Feed rotameters
      (e)  Stainless steel orifice plate and manometer, located
          upstream of the fan,  for flue gas flow measurement
      (f)  Delay tank pump identical to the recycle  pumps
      (g)  Stack gas  reheater
                           RESULTS

Fly Ash Scrubbing
      Phase I of the test program was to determine  particulate  removal
when scrubbing with water.   The throat  velocity and liquid to gas ratios
were  adjusted  to vary the pressure  drop  across the  throat over a wide
pressure range.  The particulate sampling procedures employed were
the methods  recommended by  the Industrial Gas Cleaning Institute.
The results were as follows:
      (a)  At inlet  dust  loadings ranging from 5.2 to 9.6 grains/SCFD
          and  averaging 7-8  grains/SCFD, the outlet dust loadings
          ranged from about 0.08 to 0.02  grains/SCFD over the range
          of 1.5 to 18.5 inches w.g.  pressure drop.
      (b)  Particulate sampling was  continued  during the sulfur  dioxide
          absorption program.  The results showed that the outlet dust
          loadings  were comparable to those obtained when scrubbing
          with water.  During the two-stage scrubbing test work,
          additional particulate removal  occurred in the second stage.
                                379

-------
Sulfur Dioxide Absorption with Hydrated Lime
      Phase II of the test program was to determif.j the sulfur dioxide
absorption capabilities of hydrated  lime.  The hydrated lime, a by-
product of the generation of acetylene  from calcium  carbide, was
obtained as a nominal 50% solids content sludge.  This sludge, if
used  on a commercial basis, would be 90-95% calcium hydroxide
on  a  dry basis.   However,  the  sludge fed to  the pilot plant was
only  80-85%  calcium hydroxide, with  5-10%  being calcium carbonate.
      The tests performed were to determine the optimum operating
conditions, to remove 70% and 90%  of the inlet sulfur dioxide, which
ranged during the program from 2000-3200 ppm sulfur dioxide on a
dry basis.  Both single-stage and  two-stage tests were performed.
The results of the single-stage  tests  were as follows:
       (a) At  stoichiometries ranging from  90% to 150%, based upon
          inlet sulfur dioxide concentration,  absorption efficiency,
          or  removal of sulfur dioxide,  ranged from  60-80%.
       (b) Use of a  delay tank improved the absorption efficiency
          at  the lower stoichiometries but had less  effect at the
          higher hydrated lime feed rates.
       (c) For the delay tank tests  no  improvements were obtained
          for residence times beyond  25 minutes.
       (d) Increasing  the liquid to gas ratio  from 20 to 60 gpm/1000
          ACFM, significantly improved the  sulfur dioxide removal
          efficiency.
After completion of the single-stage tests, two-stage  tests were performed,
In  general, the effects of the various operating parameters  were similar
to  that of single-stage tests with all values of sulfur dioxide removal
about  10-20% higher.  Values above 90%  were obtained.
                                 380

-------
      The use of a delay tank was found more beneficial  at  the
lower stoichiometries just as it  was for single-stage operation.
Higher reagent utilization was found to occur when the delay tank
was  used, and therefore, a commercial design would be significantly
benefited by its use.

Sulfur Dioxide Absorption with Limestone
      Phase III of the test  program was to  determine  the absorption
capabilities of an available limestone  analyzing  95%  CaCO  and
                                                          O
3.5% MgCO  with the  remainder being mineral inerts.  This  limestone,
            O
from Carntown,  Kentucky, is a material currently utilized in  other
operations at Marietta  and  is  normally received  by barge.   Only two-
stage tests were performed  since past experience indicated that  70%
sulfur dioxide removal  could not be obtained in  a  single-stage unit.
      The results of the limestone  tests  were as follows:
      (a)   Increasing the stoichiometric feed rate from 130%  to 230%
          increased  the removal efficiency  about 10%.
      (b)   The effect of liquid to gas ratio  on sulfur  dioxide removal
          efficiency was significant.  Increasing the liquid to gas
          ratio by 30% improved the sulfur  dioxide removal efficiency
          by better than 5%.
      (c)   The gain in  sulfur dioxide removal efficiency was  about
          10% higher when  using 325 mesh limestone than when  using
          200 mesh  limestone.
      (d)   The use  of a delay tank increased sulfur dioxide removal
          by as much as 10%.
      (e)  Values in excess of 75%  sulfur dioxide  removal were
          obtained  using a  delay tank and  325 mesh  limestone.
                                 381

-------
Extended Limestone Test
      The final portion of the  program  was to operate the pilot plant
continuously for three days using  325  mesh limestone.  The  system
was run in  an open loop configuration with no attempt made  to recover
the bleed stream liquid.   The results  of the tests  were as follows:
      (a)  A removal  efficiency of  better than 70%  for the optimum
          design conditions selected was confirmed.
      (b)  The unit operated smoothly  and  continuously during the
          entire test period.   No  changes  in the venturi or separator
          pressure drops were observed.
      (c)  Inspection of the piping, rotameters, pumps and tanks
          showed that scale  formation or depositing was insignificant.

                      SUMMARY OF RESULTS

      The results  of  the  test program for the  Meigs Seam coal and the
specific removal reagents are summarized below:
      1.  Fly ash levels down to  0.02 gr/SCFD were obtained.
          This value is well below the Interstate Abatement  Conference
          Recommendations and the proposed  State of Ohio Regulations.
      2.  A sulfur dioxide removal efficiency of  70%, the requirement of
          the Interstate Abatement Conference, was readily attainable
          with a single-stage  scrubber using hvdrated  lime or a  two-
          stage unit  using limestone.
      3.  A sulfur dioxide removal efficiency of  90%, which  would
          exceed  the proposed State of Ohio  Regulations, was attained
          using hydrated lime in a two-stage scrubber system.
      4.  A continuously operated three-day test using limestone
          developed  no  scaling or depositing problems.
                                 382

-------
       PROJECTED ECONOMICS  OF FULL-SCALE  PARTICULATE
            AND  SULFUR DIOXIDE SCRUBBING SYSTEM

      It was concluded from the pilot plant tests that satisfactory
technical  performance  using  limestone or hydrated lime in a two-stage
venturi  system could be achieved on a  full-scale  plant.   Studies  for
the projected  full-scale installation were based upon an individual
scrubbing train for each boiler  in order to  assure  maximum reliability
and on-time performance.   Predicted removal efficiencies for  the
full-scale system were 99%"for particulate, 90% for sulfur  dioxide
with hydrated  lime,  and 70% for Carntown  limestone  at a  system
pressure drop  of  20 inches w.g.
      Design and  engineering work by Chemical Construction  Corporation
under way since  January,  1971  and updated for information obtained
from the pilot plant program, yielded estimated costs for a complete
installation  utilizing a two-stage hydrated lime system as  follows:
Capital Costs
      Initial Scrubbing  System on One Boiler             $ 3,900,000
      Settling Basin for Scrubber Effluent                      900,000
      Hydrated Lime Materials  Handling System               800,000
                                       Subtotal           $ 5,600,000
      Three  Additional Scrubbing Systems                    9,000,000
                                       Subtotal           $14,600,000
Development Costs
      First Scrubbing System                             $ 1,500,000
      Remaining Scrubbing Systems                          1,500,000
                                       Subtotal           $ 3,000,000
Capital Plus Development Costs              Total        $17,600,000
                                 383

-------
Annual Operating Costs
      One Scrubbing System
          Energy                                        $    80,000
          Maintenance                                      150,000
          Absorbent (Hydrated Lime)  for
            90% Sulfur Dioxide Removal                      250,000
                                             Total      $   480,000
      Four Scrubbing Systems                 Total      $1,920,000

      The preceding cost estimates included the following specific
items for each  scrubbing system:
      1.  Chemico integral venturi scrubbers-mist eliminator units,
          with  each stage a  separate vessel.
      2.  -Delay tank.
      3.  Flue  gas  reheater using process  steam.
      4.  Steam turbine-driven fan to supply system pressure drop.
      5.  Recycle of pond water  to system.
The estimates include connection into an exhaust stack,  but  not  the
stack itself.

-------
Absorbent Cost  Comparison
     The following table  shows a material  cost  comparison of hydrated
lime vs.  limestone for a scrubbing system  on one  boiler burning 250,000
tons of 5%  sulfur content-coal per year.
     Basis:  Total Sulfur  -   12,500 tons/year
             Total  SO
-  25,000 tons/year
             70%  Removal of Total SO  - 17,500 tons/year
             90%  Removal of Total SO  - 22,500 tons/year
                                  70% Sulfur Dioxide  Removal
 Hydrated  Lime
  (50% solids)
 Limestone
Tons
Required
58,000
58,000
53,360
53,360

Cost Aon, $
3.40 (1)
3.90 (2)
4.10 (3)
4.30 (4)
90%
Total Cost/Pound of
Cost Cost/Ton of Sulfur Dioxide
$ Coal Burned, $ Absorbed, £
197,
226,
218,
229,
200 0
200 0
776 0
448 0
Sulfur Dioxide
.79
.91
.88
.92
Removal
0.56
0.65
0.63
0.66

 Hydrated  Lime  70,900
3.40
241,060
0.96
0.53
               Cost Assumptions:
               (1)  Hydrated Lime @  $1.00/ton
               (2)   Hydrated Lime @ $1.50/ton
               (3)   Limestone @  $1.70/ton and grinding @ $0.80/ton
               (4)   Limestone @  $1.70/ton and grinding @ $1.00/ton
 At the  most favorable material costs  for 70% sulfur dioxide removal,  hydrated
 lime has about a 10%  lower cost  advantage.
 A salient feature  of hydrated  lime favoring  its selection  over  limestone  is
 that about  55% as much absorbent solids and  65% as much absorbent -
 sulfur dioxide reaction  product solids  are handled in  the system.  Less
 waste product storage  capacity would be required.
                                385

-------
  ACTION TAKEN. BY FERROALLOYS  DIVISION OF UNION CARBIDE

      Concurrent with the activity discussed in this  presentation,
changing business  conditions indicated that a  high quality (10%  ash,
0.8%  sulfur) captive coal supply could be made available for the
Marietta power  station.  Utilization of this coal,  together with
precipitator upgrading, has been determined to be  the  minimum capital
solution for reducing Marietta power  station emissions to meet existing
recommendations and proposed regulations.
                                386

-------
Flue Gas

   f
                      Sample Point
              T7
                  Venturi
       I
         Meter
00
00
    XXX
       Bleed
                                          Sample! Point
                                 Separator
                           Pump
                           Tank

                       Delay
                       Tank
                          HXl
                                             4=iventuti
                                            T7
                                    I
                                       Meter
                                                        Sepa-
                                                        rator
                                            —txj-
                                                         Pump
                                                         Tank
                                                      3	MX
                                          Reagent —.
                                                          Water
                          IMeter
Feed Tank
                                                Cleaned
                                               Flue Gas
                                Sample Point
                               +-O
                                   Orifice Reheater
                                    Plate
                                                                                       Fan
                                                         Reagent
                                      Water
                                                                  I
Meter
                                                                Feed Tank
                                         -IX]
                                -CXJ
                                                 FIGURE 1
                                    TWO-STAGE VENTURI SCRUBBER SYSTEM

                                          SCHEMATIC FLOW SHEET

-------
                       Figure 2
ENTRIES TO VENTURI SECTIONS - MARIETTA PILOT INSTALLATION

-------
                         FIGURE  3
TWO-STAGE VENTURI ':'\rSTEM - MARIETTA PILOT INSTALLATION
                          389

-------
  LIMESTONE SCRUBBING EFFICIENCY OF
   SULFUR DIOXIDE IN A WETTED FILM
    PACKED TOWER IN SERIES WITH A
          VENTURI SCRUBBER
                 By

            R. J. Gleason
 COTTRELL ENVIRONMENTAL SYSTEMS, INC.
   A Division of Research-Cottrell
       Bound Brook, New Jersey
            Prepared For:

 Presentation at Second International
Lime/Limestone-Wet Scrubbing Symposium

        New Orleans, Louisiana
          November 8-12, 1971
                 391

-------
                     ACKNOWLEDGEMENT S
      The work upon which this publication is based was
performed pursuant to Contract EHS-D-71-24 with the
Environmental Protection Agency.

      Experiments were carried out at the Tidd Power Station
of Ohio Power, a subsidiary of American Electric Power.  The
cooperation of AEP and the Tidd Power Station personnel
played a key part in executing this study.

      Part of the work reported here was conceived and per-
formed by the Tennessee Valley Authority personnel.  The
progress resulting from the limestone tests is due, in part,
to TVA participation.
                               392

-------
          LIMESTONE SCRUBBING EFFICIENCY OF

           SULFUR DIOXIDE IN A WETTED FILM

            PACKED TOWER IN SERIES WITH A

                  VENTURI SCRUBBER
Sulfur dioxide removal from a power plant flue gas
was studied in a 1,000 CFM pilot plant containing a
packed tower and venturi scrubber combination.
Absorption efficiencies were determined for a range
of SO~ concentrations in the gas-phase and limestone
concentrations in the liquid slurry.  Results of the
absorption tests indicate removal efficiencies to
be strongly affected by the inlet SO- concentration
and the limestone stoichiometry.  Reaction products
deposition in the absorption tower were determined
for various operating modes and/or conditions.
Scaling of the tower absorber was controlled by
maintaining a high liquid-to-gas ratio and high
slurry concentration in the circulating liquor.
                       393

-------
I.  INTRODUCTION

          Control of power plant sulfur dioxide emission has
    been under study at Research-Cottrell since the early 1960's.
    Initially, R-C investigators were concerned with the absorp-
    tion characteristics of the Flooded Disc Scrubber (FDS) , a
    venturi-type wet scrubber.  Results of this experiment work
    demonstrated a low degree of SO2 removal (40 to 60%) when
    lime was employed as a scrubbing agent.  Scaling within the
    scrubber and auxiliary equipment was difficult to control
    during maximum absorption conditions.  To achieve greater
    sulfur dioxide removal and controlled scaling/ other conven-
    tional tower absorbers were screened.  Bench-scale studies of
    select packing types showed that a particular wetted film
    packing with exceptionally low pressure drop characteristics
    and high specific surface was well suited for SO- scrubbing
    systems.  In 1969, a 1,000 CFM pilot plant consisting of a
    FDS in series with a wetted-film packed tower was installed
    in Ohio Power's Tidd Plant.  The results of earlier bench
    work and the status of the subsequent pilot plant program
    were reported in March, 1970 at the First International
                                         1»2
    Symposium on Wet Limestone Scrubbing.

          Since the First Limestone-Wet Scrubbing Symposium,
    Cottrell Environmental Systems was contracted by EPA to
    execute a pilot plant study at the Tidd Power Station in-
    volving both lime and limestone.  One objective of this pro-
    gram was the evaluation of the wetted-film packed tower
    absorption capabilities with limestone wet scrubbing.

          The process conditions for the limestone tests were
    based upon bench scale studies performed by the Chemical
                               394

-------
Development Division of the Tennessee Valley Authority,
Muscle Shoals, Alabama.  The absorption system simulated,
to some respect, the Howden-ICI actual plant operation
of the Fulham Power Station, London.
      Test results reported in this paper were previously
presented at the 69th National Meeting of the American
Institute of Chemical Engineers, Cincinnati, Ohio, May
16-19, 1971.
                               395

-------
II.  EQUIPMENT
           A layout of the pilot plant system is given in Figure
     II-l.  The absorption section of the pilot plant contained a
     Flooded Disc Scrubber  (venturi scrubber) in series with a
     packed tower.  The flue gas, containing both particulates and
     sulfur dioxide, passed first through the venturi where the
     entering gas quickly cooled to its dew point  (#120°F).  Flue
     gas  was removed isokinetically from the suction of the power
     station I.D. Fan.*  Scrubbing solution or slurry entering the
     absorber tangentially  above the  throat flowed cocurrently
     through the venturi and into a cyclonic demister.  The fly-ash-
     stripped gas then passed  vertically through a conical hat
     gas/liquid  splitter before entering the packed tower.

            Slurry flow rates to the scrubber and other process units
     were measured  by venturi-type flow meters.  In-line pH probes
     were installed at the  discharge  of each scrubber.  Also,
     immersion-type pH elements were  placed in the clarifier and the
     hold tank.

            The venturi scrubber is a  cocurrent absorber with a
     variable throat orifice.  Pressure drop across the scrubber
     can  be varied  by adjusting the disc position within a venturi
     throat. Figure II-2.

            The packed tower is a countercurrent absorption device
     containing a packing with low pressure drop characteristics
     and  high specific surface (68 sq.ft./cu.ft.).  The packing,
     a rigid material fabricated from corrugated sheets of asbestos
      * Flue  sampling contained particulate concentration  of
       approximately 4.0  gr/scf from a pulverized coal  combustion,
                                   396

-------
V-VENTURI SCRUBBER
D-DEMISTER
A-ABSORBER
C-CLARIFIER
H-HOLDING TANK
M-MIXING TANK
B-BULK FEEDER
GIV-GAS INLET VENTURI
GOV-GAS OUTLET VENTURI
GOO-GAS OUTLET ORIFICE
VV -VENTURI  VENTURI
AV -ABSORBER VENTURI
CV -CLARIFIER VENTURI
VBV-VENTURI  BY-PASS VENTURI
ABV-ABSORBER BY-PASS VENTURI
VP -VENTURI  PUMP
AP -ABSORBER PUMP
   -CLARIFIER PUMP
   -HOLDING TANK PUMP
   -MIXING  TANK PUMP
   -HAND VALVE
                                                   CP
                                                   HP
                                                   MP
                            FIG.U-1  PILOT PLANT  LAYOUT

-------
                       GAS  FLOW
HEIGHT
  13
VARIABLE
                LIQUID
                 INLET
                  r CONDUIT
                      /LIQUID DISENGAGES
                    /    FROM WALL
FIG.E-2 DIMENSIONS FOR FLOODED DISC SCRUBBER
                          398

-------
coated with neoprene, was five feet in height and sixteen
inches in diameter.  Pressure drop across the packing at gas
velocities between 8 and 10 feet per second is approximately
one inch of water with well-irrigated conditions.

      Flooding characteristics of this packing were approxi-
mated by using the Eckerts flooding line correlation and a
                       i*
packing factor of 12.0.   The calculated flooding profile of
the wetted film packing are compared to other high capacity
packing in Figure II-3.  As illustrated graphically, the wetted-
film packing allows  significantly greater liquid rates than
either 3 inch plastic Pall rings or 3h inch Intalox saddles.
                      PROCESS EQUIPMENT
    Process Units
Size
 Material Of
Construction
FDS Scrubber
Packed Tower
Tower Slurry
Tank
Mixing Tank
Venturi Slurry
Tank
Cyclonic
Demister
Clarif ier
Blower
6" to 8" diameter
See Figure 2.
16" diameter x 5'
1,500 gallons agitated
1,500 gallons agitated
55 gallons
4 ' diameter x 4 '
1,200 gallons
1,000 cfm at Ap = 40
inches H2O
316 SS
316 SS
CS
CS
CS
SS
CS
CS

                               399

-------
                   CALCULATED FLOODING MASS VELOCITIES
           FOR THE WETTED-FIU1 AND OTHER HIGH CAPACITY PACKING
                                                   ji I
102
               103
GAS MASS VELOCITY, LB/(HR)(SQ,FT.)
10*
                                    400

-------
III.   RESULTS
            The limestone programs were carried out in two separate
      test series:   several types of limestone were first processed
      in an open-loop system and absorption efficiencies were com-
      pared for each alkali type.  In the second program, scale
      accumulation  within the tower absorber was investigated
      using a limestone selected from the first test result.   For
      the scaling runs, a closed-loop operating mode was used to
      minimize the  process makeup water.

            A.  EFFICIENCY TESTS - OPEN-LOOP

            Calcium carbonate materials (chalk, cement kiln dust,
      and two limestone grinds)  were processed under similar  opera-
      ting conditions.  Chalk (89%-200 mesh) and finely ground lime-
      stone (89%-325 mesh) allowed approximately the same absorption
      i.e. 96% removal.  Limestone ground to 75%-200 mesh gave an
      efficiency between 81 to 88% while the cement dust showed the
      poorest results with 73% absorption.  A chemical analysis for
      each of these materials is given in Table III-l.

            Process modes selected for the pilot tests utilized the
      packed tower  as the main contacting device.  The venturi
      scrubber removed fly ash and a small portion of the inlet SO2
      by operating  at a moderate pressure drop  (6 to 7 inches of
      water).  Fresh limestone slurry was fed first to the tower
      circulating tank where the bulk of the absorption took place.
      For maximum calcium carbonate utilization, the reaction
      products and  the residual limestone were bled from the tower
      slurry tank and passed through a venturi circulation tank.
      The input limestone and the bleed were balanced so that a
      constant calcium concentration remained in the tower hold tank
                                     401

-------
                 TABLE III-l
CALCIUM CARBONATE MATERIALS RECEIVED FROM TVA
Type
Tiftona
Limestone
Tiftona
Limestone
Cement
Kiln Dust
Selma
Chalk

Particle
Size
75%-200M
89%-325M
90%-200M
89%-200M

Amount
Received
Ibs.
4700
350
1000
1000

CaO
50.5
50.5
41.5
43.1

MgO
1.5
1.5
2.4
0.59

K00
0.4
0.4
3.1
0.49

Na.,0
0.3
0.3
0.19
0.18

Ign.
Loss
41.5
41.4
22.8
35.1

                     402

-------
      Block diagrams illustrating the operating modes are
shown in Figure III-l.  A limestone slurry containing 2% by
weight calcium carbonate was pumped to the Tower Slurry
Tank from the Mixing Tank at a flow equivalent to the SO2
stoichiometry desired.

      For six of the seven efficiency tests, the stoichio-
metric input was between 100 and 110%.*  One run was performed
at 160% of the equivalent S02 to determine crudely the lime-
stone stoichiometry effects.  The sulfur dioxide concentration
to the FDS was between 1,135 and 1,980 ppm.  Although no
particulate measurements were made, the normal fly-ash loading
was from 3 to 4 grains per SCF.

      The tower gas velocity for these tests was between 9.3
and 10.7 ft/sec.  Liquid flows for most of the tests were
maintained at 40 gallons per 1000 cu.ft..  A single test
operated at L/G = 20 gallons per 1000 cu.ft..  Pressure drop
on the tower was around 1.0 inches of water for most tests.
Tests with chalk and cement were operated at slightly higher
tower pressure differentials.

      Operating results for these experiments are given in
Table III-2.  Each efficiency measurement listed is an average
of four readings over a three-hour period.  Liquid-to-gas
ratio in the tower had considerable influence on SO2 reduction.
At L/G of 40 gallons per 1000 cf and a coarse grind limestone,
the efficiency across the tower was 81.6%; however, at L/G of
20 for the same operating mode, the SO2 removal dropped to
58.2%.
 * Based on total SO- feed to system.
                                403

-------
             GAS IN
         ^•SLURRY
           RECYCLE
                 VENTURI
                 SLURRY
                  TANK
"r
           GAS OUT
                                   PACKED
                                   TOWER
                                  DEMISTER
                                        1
                                               ^-SLURRY
                                                 RECYCLE
                                                  CAC03

                                                  WATER
TOWER
SLURRY
 TANK
        Z
MIXING
TANK
                                             SLURRY FEED
                   COUNTER-CURRENT SLURRY FLOW
    WATER
             GAS IN
                4-
           GAS OUT
                                   PACKED
                                   TOWER
                                  DEMISTER
t   ,r
TOWER
SLURRY
TANK
                                        1
                                               ^SLURRY
                                               'RECYCLE
                                                  WATER
                                                  MIXING
                                                   TANK
                        WATER FEED TO FDS
FIG, IBM OPERATING  MOOES USED FOR LIMESTONE EFFICIENCY TESTS
                             404

-------
                                                                          TABLE III-2
O
01
Material Used
Test No.

Tower Liquid Rate, GPM
FDS Liquid Rate, GPM
Gas Velocity, Tower, Ft/Sec.
Gas Velocity, FDS, Ft/Sec.
Tower Pressure Drop, inches H-0
FDS Pressure Drop, inches H«O
CaO/SO2 Ratio
SO- Concentrations, PPM
     FDS in
     FDS out
     Tower out
Fraction of SO2 Removed, %
     FDS
     Tower
     Overall
Gas Temperature,  °F
     FDS in
     FDS out
     Tower out
Liquid  Temperature, °F
     FDS in
     FDS out
     Tower  in
     Tower  out
 L/G ratio,  FDS, Gal.  per 1000  cf
 L/G ratio,  Tower, Gal.  per  1000 cf
1
iUMMARY DATA SHEET FOR THE OPEN-LOOP EFFICIENCY TESTS
Limestone
Selma
Chalk
A2
35
9.4
10.7
99
3.9
12.7
1.11
1550
1405
48
9.4
96.6
96.9
371
122
114
119
126
112
122
10.4
39
Cement
Dust
A3
36
9
10.7
105
6.5
6.4
1.06
1135
1025
268
9.7
73.5
76.4
377
122
116
121
124
114
122
10
40

Fine
A4
24
6
9.3
132
1.1
7.3
1.16
1650
1288
62
21.9
95.1
.96.1
361
118
110
118
125
110
117
10
40

Coarse
AS
24
6
9.3
132
0.9
7.3
1.00
1980
1795
209
9.4
88.4
89.3
361
116
109
116
124
110
116
10
40

Coarse
A6
3i
8
9.6
121
1.2
7.1
1.00
1467
1155
212
21.2 2
81.6
85.5
348
92
91
40
75
89
124
10
40

Coarse
A7
18
9
10.7
98
1.3
7.9
1.10
1637
1415
592
13. 62
58.2
63.8
391
102
99
40
97
95
106
10
20

Coarse
AS
32
8
9.6
87
1.3
9.4
1.62
1485
1402
199
5.6
85.8
86.6
369
101
98
41
95
96
101
10
40
                                  1. Operating conditions for each task shown are an average of four readingsmeasured over a
                                    three hour period.
                                  2.  Water  was  fed  to  the  FDS  for  this  test.

-------
      B.  SCALING EXPERIMENTS

      During  the limestone  efficiency  tests,  some scaling
 did  occur in  the packed  tower  after  the  chalk/ cement dust,
 and  limestone were  used  as  sorbate.  Limestone alone did not
 show appreciable scale accumulation  after  40  hours of contact.
 To determine  definitively the  scaling  characteristics of the
 limestone/SO- absorption, a second test  program was undertaken.

      The operating conditions for minimum scaling were  studied
 in four continuous  40-hour  tests.  Process parameters such as
 liquid-to-gas ratio, tower  slurry  tank residence time,  and
 tower slurry  tank temperature  were varied.  Scale deposition
 was  measured  by weighing the packing before and after each
 run. In most cases, new packing was installed for the
 subsequent  test.  Following these  four preliminary experiments,
 an 80-hour  continuous operation was  executed  and the scale
 accumulation  measured.   The closed-loop  mode  for the major
 portion of  this test series is shown in  Figure III-2.

      1. Scaling

      The conditions for each  run  are  summarized in Table A-l
 of Appendix A. Although some  operational  difficulties  were
 encountered,  and several process parameters were not consis-
 tent throughout any run, a  general trend in scale accumulation
 could be seen. With high tower liquid-to-gas ratio, lower
 solids  buildup was  measured than with  low  L/G.  Residence
 time for the  slurry in the  tower hold  tank showed no effect
 as the  hold time was varied from 30  to 10  minutes.

      Profile of solids  buildup on the packing going from the
top to  the  bottom section did  show a pattern  of scale.  Very
little  scale  deposited in the  top  section  while a consistent
                              406

-------
                   CLARIFIERJ
    VENTURI
    SLURRY
     TANK
TOWER
SLURRY
 TANK
FIG.TH-2  TWO STAGE CALCIUM CARBONATE SCRUBBER
                         407

-------
quantity precipitated  in  the  bottom four  elements.   This
profile of  solids  buildup suggested an  absorption  super-
saturation  taking  place within the tower  with  an induced
encrustation after one or two feet in the packing.   From  the
profile of  solids  deposited and the reduced  deposition  at high
L/G,  one could conclude that  incoming solution from  the hold
tank  was at a lower supersaturation, but  once  supersaturation
did develop within the tower, the rate  of encrustation  was
constant.

       Based on the aforementioned reasoning, the operating
conditions  selected for the long-term demonstration  tests
combined a  low level tower slurry tank  volume  with a high
tower liquid-to-gas ratio, i.e. a ten minute residence  time  in
the hold tank and  a L/G of 40 gallons per 1,000 cubic feet.
Stoichiometry at the start of the run was near 100%  for the
first 40 hours and 120% for the last 40 hours.  Two  limestone
grinds were used throughout the run; for  the first 40 hours,
a limestone having 75%-200 mesh was employed,  while  during
the second  40 hours the same  material having 61%-200 mesh was
fed.   Solids concentration in the slurry  was held  between 4.4
and 8.9%.   Chemical and particle analysis for  these  limestone
materials are given in Table  III-3.

       2. Absorption

       The absorption efficiency varied  from  a  low  of 55%  to  a
high  of 98%.   Near the end of the run,  the absorption was
higher.  A  profile plot of the Stoichiometry,  slurry concen-
tration, and  tower efficiency is shown  in Figure III-3.

      During  the continuous scaling run,  a blower motor failure
interrupted the test half way through the experiment.   The
                              408

-------
                          TABLE III-3

                Limestone Materials Supplied By

                TVA For The Scaling Test Series*
Designation
Particle Size
Amount Received, Ibs
Chemical Analysis
  Shipment

     A


61%-200 Mesh


    2000
                                                 Shipment

                                                    B


                                               75%-200 Mesh


                                                   4700
CaO %
MgO %
K2O %
Na20 %
Ign. Loss
50.8
—
—
—
41.5
50.5
1.5
0.4
0.3
41.5

       * Limestone supplied by TVA called "Tiftona
         Limestone".  Shipment "A" was a special
         supply of CaCO3 for the scaling test series;
         Material "B" was received from TVA for the
         first test series.
                                409

-------

                                             FIGURE 111-3
                     TASK C-6  EFFICIENCY PRQEHE, SLURRY CQWBURATIQN AMD STDlCHIQMEtfiy
                                     DURING THE RUN,    SEE TABLE M
-pi
o
   \i
   <*   04
0-
                                   l          I         '  •
                                            ' !  i
                „__.          L::J.-"n,_-jTJ     |  .'  r    '
"^V"
J.

\
• •'
::••..
^, ^x"
S ii ilppV r
•~vL.l|l\IM *^
: -i
- -gS^
.„.: 	 i 	
QNCi
- 4-

                                                                   D STOICHldMETRY

                                                                 !
                                            ;
                                                          !  -   .   . '


                                                      /
                                                      .s   '
                          !
                 :r
                     10
20
                             30        40       50

                                HOURS IN OPERATION
60
                                                                             70
                                                        -1,2

                                                        -1-.-0-


                                                       80

-------
packing was removed from the tower and each section was
examined and weighed.  Most of the encrustation deposited
during the 40 hours was on the packing periphery.  The en-
crustation had a mud-like consistency, not the hard scale
observed with hydrated lime systems.  Scale was not evident
on the well-irrigated surfaces.  For the second 40 hours, the
same packing was used; encrustation that did develop was again
predominantly at the periphery.  Measured weight gain after
40 and 80 hours is given in Table III-4.

     Within the first 40 hours, 10.5 pounds of solids had built
up and during the next 40 hours an additional 12.5 pounds were
deposited, weight gain determined on dried basis.  No pressure
increase could be measured throughout the 80-hour test.  The
amount of solids clinging to the packing was a small fraction
of the packing void volume.  Pressure measurements on the
tower were approximately 1 inch of water at the start and
finish of the tests, as given in Table A-l in Appendix A.
                                411

-------
                                            TABLE III-4

                            Measured Scaling Accumulation On The Packing
PO
Test Number
Tower L/G
gals/1000 cf 	
Residence Time
For Slurry 	
Run Time, hrs.
Packing Number
1 weight gain, Ibs.
2 weight gain, Ibs.
3 weight gain, Ibs.
4 weight gain, Ibs.
5 weight gain, Ibs.
TOTAL :

C-2
30
30
40

7
14
13
16
14
64

C-3
40-60
30
40

2
9
10
10
12
43

C-4 *
40
10
40

4
13
15
14
20
66

C-5 **
40
10
40

2
5
8
8
8
31 .

C-6
45
1C
40

.5
1.5
1.5
3.5
3.5
10.5

**
..i

)
80

3
3
7
6
4
23

                 * Temperature on the  tower  circulating  slurry  not  controlled
                  at  the  inlet gas dew point.   Heat  losses  on  the  slurry—10
                  to  15°F.

                ** Packing was weighed slightly  wet.
                                                                                     Packing
                                                                                     Position

-------
IV.   DISCUSSION

           A.  LIMESTONE EFFICIENCY TESTS - OPEN-LOOP

           The venturi  scrubber removed only a small fraction of
     the  inlet S02 with each type of  limestone.   Between 6  to 22%
     of the SO2 was sorbed  for an inlet stoichiometry from  1  to
     1.6,  a liquid-to-gas ratio of 10 gallons per 1,000 cu.ft.,
     and  a pressure drop across the throat of 6.5 to 12.7 in.  of
     H20.   Such low scrubbing efficiency would not be practical for
     commercial application.

           The packed tower performed outstandingly good with the
     limestone and chalk.   The finely ground limestone (89%-325 mesh)
     removed nearly 95% of  the tower  inlet SO2 (1288 ppm).  Chalk
     with  a through 200 mesh fraction of 89% absorbed 97% of  the
     S02  at 1405 ppm inlet.   Presumably, the high surface area
     characteristics of the chalk improved the absorption con-
     siderably.   Cement kiln dust absorption of  SO2 was the lowest
     for the three limestone types tested.   Although the cement dusts
     mesh  size was very close to the  chalk (90%-200 mesh) only 73%
     of the inlet SO2 was removed.  The cement dust may have  been
     "dead burned" by the calcination conditions in the cement kiln,
     hence an  appreciable portion could have been slow dissolving.

           Liquid-to-gas ratio in the tower had  a dramatic  affect
     on the SO2 efficiency.   For a L/G of 40 gallons per 1,000
     cubic feet and an  inlet SO2 at 1402 ppm,  the SO2 absorption
     was 85.8% while at L/G of 20 and an inlet S02 of 1415  ppm,
     the removal dropped to 58.2%.  Such sensitivity to liquid
     flow  implies a significant liquid-phase absorption resistance.

           The flow rates through the tower at the liquid-to-gas
     ratio of  40 gals/1000  cf WPT-» well below the calculated
                                    413

-------
flooding conditions indicated in Figure II-3.  For example,
at a L/G =40 and a gas velocity of 9.6 ft/sec  (Test No. A8
of Table III-2), the liquid and gas rates were ll,700(lbs)/
 (hr)(sq.ft.) and 2,160(Ibs)/(hr)(sq.ft.), respectively.  These
values fall well below the flood condition.  In fact, the
liquid rate could have been doubled to 23,400(Ibs)/(hr)(sq.ft.)
or an L/G = 80  and still remain below 70% of the flooding
liquid rate.

      B.  SCALING EXPERIMENTS

      1. Scaling

      The experimental technique used in measuring the  scale
accumulation on the packing was designed to qualitatively
indicate the operating conditions affecting the deposition.
Chemical composition and physical properties of the solid
deposits were  not analyzed.  As previously mentioned, solids
that did precipitate on the packing surface had a mud-like
consistency, not the hard encrustation normally observed with
lime.

      Because  of the packing design, a small diameter tower
such as the pilot unit has a significant fraction of surface
area that is not well-irrigated.  When the corrugated sheets
are cut in circular shapes, several peripheral layers are not
exposed to the  main irrigated pattern.  It is in this peripheral
region that much of the solids accumulated.

      The scaling experiments, Tests C2 to C5, were originally
designed to screen several tower process variables, i.e.
liquid-to-gas ratio and tower slurry tank residence time.
Tower slurry concentration and limestone stoichiometry  were
                               414

-------
planned at 5% by weight and 100 to 110%, respectively.  How-
ever, throughout the preliminary tests these variables were
poorly controlled.  Slurry concentration for tests C2 to C5
fluctuated considerably in some cases and stoichiometry was
sporadic because of poor limestone feed control.  No
quantitative conclusions could be made with the measured
scaling for the 40-hour tests, C2 to C5.  Yet, two general
trends could be seen:  1) the high liquid-to-gas ratio showed
lower scaling, and 2) residence time in the tower slurry tank
did not affect the solids accumulation.  The operation and
problems experienced in tests C2 and C5 were either eliminated
or closely scrutinized during the final continuous test.

      For the 80-hour run, solids buildup on the packing was
reduced considerably as shown in Table III-4.  The 40-hour
preliminary experiment showed 30 to 66 pounds of solids
accumulation in the five feet of packing while the final 80
hour run had 23 pounds.  More important, however, the deposi-
tion that did occur was not a hard encrustation but a soft
sludge precipitate and a considerable portion of this sludge
collected on the packing periphery.

      2. Absorption Efficiencies

      The absorption efficiency for the preliminary scaling
runs varied considerably because of the fluxuations in lime-
stone stoichiometry, slurry concentration, inlet gas composi-
tion, and liquid-to-gas ratio.  To avoid process variations in
the long-term run, a strict control of the operation was
maintained for the planned 80-hour tests.  However, even with
close scrutiny of these process variables, the absorption
efficiency fluxuated considerably.  To explain the efficiency
                           415

-------
variation, the operating conditions for the 80-hour run were
examined carefully.  Slurry chemical analysis performed by
TVA and Radian Corporation were combined with the absorption
efficiency, inlet SO2 concentration, and slurry concentrations
                    * 5,6
measured in the field.

      As a first step in the data analysis, the calcium car-
bonate concentrations were determined for the absorbate over
the entire 80-hour run by digital computer simulation of the
absorption process.  With a known CaCO3 analysis as a start-
ing point, the CaC03 slurry concentration was calculated for
each point in time that an efficiency measurement was made.
The predicted carbonate slurry concentration fit well with the
chemical analyses.  The computer simulation did not take into
consideration process changes such as spills, leaks, or
uncontrolled water addition.  A comparison of the computed
and analyzed carbonate concentration is given in Table IV-1.

                        TABLE IV-1
            LIMESTONE CONCENTRATION IN THE HOLD
                    TANK DURING TASK C6

                       Calcium Carbonate Concentration, %
Date
1/25
1/26
1/26
1/26
1/29
1/29
1/30
1/30
Time
2230
1300
1500
2100
1400
2100
1230
0103
Computer Predicted
1.89(Start)
1.19
1.19
1.46
0.36 (Start)
1.44
1.32
1.78
TVA
1.89
0.54
0.94
0.36
2.16
1.65
1.60
Radian
1.13
—
                           416

-------
      Using the computer-estimated value of limestone concen-
tration, the analyzed sulfur dioxide absorption efficiency in
the tower, and the measured tower hold tank slurry concentra-
tion, a linear correlation was developed which predicts the
absorption efficiency for each of the limestone materials
employed.

      For the first 40 hours, where limestone ground to 75%
-200 mesh was used, an outstandingly good correlation was
realized.  The absorption efficiency, predicted to within
+ 1.9%, showed sensitivity to inlet SO2 concentration and lime-
stone concentration as seen below:
    Y = 165.05 - 0.0463(ppm) + 30.48(% CaC03) - 9.126(SL)
where  Y       = SO- absorption efficiency, %,
       ppm     = tower inlet SO- concentration in ppm,
       % CaCO- = concentration of limestone slurry in
                 the hold tank, %,
       SL      = concentration of all solid in hold
                 tank, %.

Statistical parameters and the variable range for this correla-
tion are listed in Table IV-2.

      For the second half of the run, a similar linear correla-
tion having a precision of + 3% efficiency was developed for
limestone with 61%-200 mesh.  Here the last 26 hours of
operation were studied so that a mixture of the two limestone
types could be avoided.  The predicted efficiency showed less
                           417

-------
                        TABLE IV-2
                 STATISTICAL PARAMETERS FOR
                   EFFICIENCY CORRELATION
                    (First 40 Hours Run*)
Number of Data Points
Correlation Coefficient
Standard Error For Estimate
Significance of Regression (F)
% Efficiency Range, Y
Sulfur Dioxide Inlet Range
% CaCO, Range
% Total Solids Range, SL
30
0.987
1.9
349
53% to 97%
1160 to 1900 ppm
1.131% to 1.89%
5.4% to 9%
 * Limestone Used - Tiftona Limestone 50.5% CaO,  75% - 200
   Mesh.
                         418

-------
sensitivity to inlet SO2 concentration and qreater sensitivity
to the limestone concentration, i.e.

                                                           (IV-2)
    Y = 56.273 - 0.0178{ppm) + 50.313(% CaCO) - 4.15(SL)
See Table IV-3 for statistical limitations.

      To make use of these efficiency correlations, the lime-
stone slurry concentrations must be known.  Three factors in-
fluence the residual limestone concentration in the tower
slurry liquor:  1) the actual absorption efficiency for the
process, 2) the stoichiometric feed ratio of CaCO^/SO^/ and
3) the overall slurry concentration.  For a system with 6%
total slurry, the limestone concentration can be predicted by:

     % CaC03 = 1.23 - 0.033(% Y) +  2.236 R                (IV-3)

where  R = stoichiometric feed ratio, moles of CaCO^/moles of SO^

Table IV-4 presents the condition for equation  (IV-3) .

      Using this expression and equation  (IV-1) or  (IV-2) ,
the absorption efficiency can be predicted for either  lime-
stone material for a liquid-to-gas  ratio of 45 gallons per
1,000 cf., and a total slurry of 6% by weight.

      Clearly, if absorption efficiency is dependent  upon
the limestone slurry concentration  and the inlet S02  concen-
tration, then one or both of these  conditions must be  con-
trolled for a desired SO2 removal.  To illustrate this point,
equations  (IV-1) and  (IV-3) were combined and the resulting
relationship was plotted in Figure  IV-1.
                             419

-------
                        TABLE IV-3
                STATISTICAL PARAMETERS FOR
                  EFFICIENCY CORRELATION
                   (Last 26 Hours Run*)
Number of Data Points
Correlation Coefficient
Standard Error For Estimate
Significance of Regression (F)
% Efficiency Range, Y
Sulfur Dioxide Inlet Range
% CaCO-, Range
% Total Solids Range, SL
26
0.95
3.08
69
75.5% to 97.9%
960 to 1380 ppm.
1.317% to 1.847%
5.5% to 8%
 * Limestone Used - Tiftona Limestone 50.8% CaO, 61% - 200
   Mesh.
                        420

-------
                   TABLE IV-4
           STATISTICAL PARAMETERS FOR
     STEADY STATE WT.% LIMESTONE CORRELATION
Number of Data Points          =  16
Correlation Coefficient        =  0.995
Standard Error For Estimate    =  0.054
Significance of Regression (F) =  741.8
% CaCO3 Range                  =  0.1616% to 2.07%
% Efficiency Range, Y          =  53% to 92%
Range For Stoichiometric Ratio -  1.0 to 1.3
                      421

-------
ND. 341 -2D  DIETZGEN  GRAPH
      2O X 2C »EO ISCH
                                         EUGENE DIETZGEN CD.
                                OF CA€03/ltoL£ OF -S0;?

-------
As the graph illustrates, the efficiency is very sensitive
to the stoichiometry and the inlet SO- concentration.  For
a stoichiometric addition of limestone, SO- removal can vary
between 85 and 62% as the inlet SO, concentration changes
from 1000 to 2000 ppm.  This effect on inlet concentration
indicates a significant liquid-phase resistance.  It is
evident in equation  (IV-1) and (IV-2) that the total solid
concentration has an adverse effect on the absorption
efficiency.  This is logical in that the viscosity and flow
characteristics would change with increasing solid concentra-
tion.  A similar slurry influence was observed by Chertkov
with hydrated lime slurries in a board packing.   He observed
decreasing mass-transfer coefficients as a slurry concentration
increased.

      Simultaneous to this negative coefficient for slurry
concentration (SL), the residual calcium carbonate in slurry
shows a positive coefficient.  Hence, with increasing calcium
carbonate concentration, the absorption efficiency improves.
The net result of increasing or decreasing the slurry concen-
tration must be analyzed and the effects of total slurry and
calcium carbonate concentration should be taken into considera-
tion.

      The absorption efficiency for the two limestone grinds,
61% and 75%-200 mesh, showed a considerable difference in
absorption efficiency.  For a limestone feed of 120% of
stoichiometric and a SO2 inlet concentration of approximately
1300 parts per million, the absorption efficiency expected
with the 61%-200 mesh would be approximately 77%.  On the other
hand, the 75%-200 mesh material for the same conditions would
allow near 85% removal.  These differences in efficiencies
are demonstrated graphically in Figure IV-2.  The curves shown
                              423

-------
                        NO. 341-2C DIETZQCN GRAPH

                             20 X 2- = E9 ISCM
                        EUGENE DiETZGEN CO.

                          .«»C€ N j 5. *
ro
i-
                                         •fafiURE
                                          ABseimeNWiTt
                         -twesw
SLURRY IN A -HETTEfr-Fttff mm T0WER
                 •IV-2
              11
                                                      12      1}3
.4..     1 5
               i
                             TOICHIOWE'RY, MOLES
                OF CA^/ftJLE OF

-------
represent the expected absorption for a gas containing 1288 ppm
S02 inlet concentration and a total slurry concentration of 6%.

      To illustrate further the influence of the limestone
grind, a single point measurement for 89%-325 mesh was also
plotted in Figure IV-2.  The finer material allowed approxi-
mately 11 to 12% more SO2 removal for the same operating con-
ditions as the 75%-200 mesh.  One could conclude from these
results that the surface area available for limestone
dissolution within the slurry tank and tower and absorber played
an important role in the overall mass-transfer resistance.
                              425

-------
V.  SUMMARY AND CONCLUSIONS
          1.  A limestone  slurry, circulating through a
              high  specific-surface packed tower can
              absorb greater  than 90%  of  the flue gas SO-.
              Absorption efficiency is adversely affected
              by  increasing S0~  concentration and by high
              slurry concentration.  SO-  absorption can
              be  improved  by  increasing the calcium car-
              bonate slurry concentration in the absorbing
              liquor.

          2.  A finely ground limestone (90%-325 mesh)
              increases the SO-  absorption by 11 or 12% over
              a material with (75%-200 mesh).  Absorption
              increases with  higher liquor-to-gas ratio
              in  the tower.

          3.  Scale formation in a limestone/SO- scrubbing
              system can be controlled by maintaining a
              reaction product slurry  in  the absorbing
              liquor and by circulating a high liquid rate.

          4.  A significant liquid phase  resistance exists
              for the limestone/S02 system while absorbing
              the S02 with inlet concentration up to 1,000
              ppm.

          5.  Sulfur dioxide  absorption in a venturi type
              scrubber using  limestone absorbate is limited
              to  an efficiency between 10 and 20% for Ap = 7"
              of  water.
                                   426

-------
VI.  REFERENCES

     1. Walker, A. B., "Mass-Transfer Characteristics of Variable
        Annular Throat Venturi Scrubbers", Paper presented to the
        National Air Pollution Control Administration International
        Symposium of Wet Limestone Scrubbing, March 16-20, 1970.

     2. McKenna, J. D., "Evaluation of a Two-Stage Particulate
        Scrubber and Gas Absorber Applied to Power Plant Flue
        Gas", Paper presented to the National Air Pollution
        Control Administration International Symposium of Wet
        Limestone Scrubbing, March 16-20, 1970.

     3. Lessing, R., "The Development of a Process of Flue Gas
        Washing Without Effluent", Journal of the Society of
        Chemical Industry, November, 1938.

     4. Frisch, N. W., "Calculated Flooding Velocities For A
        Wetted Film Packing and Several Commercial Packings",
        Cottrell Environmental Systems, Inc., Technical Memorandum
        TM70-19, December 15, 1970.

     5. Letter from Potts, J. M. of TVA to Gleason, R. J. of
        CES, January 19, 1971.

     6. Barkley, J.,  (TVA), Schwitzgebel, K.,  (Radian), et. al.,
        "Chemical and X-ray Analysis of Samples Taken During The
        Runs:  C5(ll: p.m. 1/21/71) and C6(3:00 p.m. 1/6/71) at
        the Tidd Plant in Brilliant, Ohio", Technical Note
        200-006-12, February  26, 1971.

     7. Chertkov, B. A., "Coefficients of Mass-Transfer During
        Absorption  of SO, from Gases by Lime  Suspension", Khim.
        Prom. No. 7, pp  533-36  (1962).
                                   427

-------
APPENDIX
             428

-------
                                                                TABLE A-l
                                                              OCSCITZOTtS TOR TASK. C-2
1/13
2130
0
700
10
30
110
8.35
1.2
6.9
1
1350
1525
695

54. 5

329
118
111
110
115
115
115
4.8
6.3
5.3
4.7
1/14
0100

700
10
30
115
8.35
2.2
7.0
0.6
1580
1600
1020

36.. 3

342
122
110
122
122
122
120
4.6
6.2
5.0
4.6
1/14
0300
B
700
i.0
30
130
8.35
2.2
7.0
1
1500
1550
880

43.3

350
122
112
122
120
120
118
4.8
6.4
5.2
4.6
1/14
0600
B
700
10
30
137
8.35
3.6
7.0
1
1550
1230
850

30.9
~
350
120
112
122
120
115
120
4.6
6.5
5.2
4.6
1/14
1000
B
700
10
30
108
8.35
4.3
7.0
0.9
1930
1050

__
4r.6
349
122
110
122
119
122
4.8
6.4
5.9
4.8
1/14
1200
B
700
10
30
8.35
6.0
6.8
1500
1560
660
•»•»
57.7
~
—
—
__
—
--
--
1/14
1420
B
700
10
30
110
8.35
7.0
8.0
1
1325
1500
336
— —
77.6
—
345
122
115
105
121
118
120
5.2
5.3
5.5
4.7
1/14
2030
B
700
10
30
115
8.35
8
7.0
1.7**
1032
1225
520
<••*
57.6
— ~
340
120
116
116
120
118
118
5.6
6.6
5.2
4.7
1/14
2300
B
700
10
30
116
8.35
6.8
6.9
1.0
1063
1250
570
••
54.4
*"•
335
122
118
115
120
120
118
5.2
6.6
5.5
4.7
1/15
0100
B
700
10
30
98
8.35
6.0
7.0
1.1
1030
875
320
15.1
63.5
69.9
338
120
110
112
120
115
118
4.9
6.5
5.6
4.7
1/15
0330

700
10
30
88
8.35
13.2
7.0
1.1
1030
950
460
8.8
SI. 6
55.4
335
110
10.2
110
115
108
110
4.2
6.4
5.4
4.6
   Date
   Time
   Limestone Used*
   Gas Flow, cfm
   FDS L/C ratio/ gal/mcf
   Tower L/C ratio, gal/mcf
   Gas velocity FDS, ft/sec.***
   Gas velocity Tower, ft/sec.
   Tower pressure drop, inches B.O
   PDS pressure drop, inches n*20
   CaO/SO2 ratio   (inlet analysis)
   SO? Concentration, ppm
        FDS in **••
        FDS out
        Tower out
   Fraction of SO* removed, %
        FDS
        Tower
        Overall
   Gas temperatures, TT
        FDS in
ro      FOS out
10      Tower out
   Liquid  temperatures, °F
        FDS in
        FDS out
        Tower in
        Tower out
   PH measurements
        Tower  outlet
        FDS outlet
        Hold tank
        Clarifier tank

    •Limestone "received from TVA for  the  "around-the-clock* tests designated as A}Old Tiftona from TVA shown as Type B.
   ••Stoichiometry was 1.0 at 2015
  •••FDS gas velocity based on outlet gas volume.
  *••• Some difficulties in SO, analyses were experienced during this test
      High efficiencies and negative absorption on FDS should not be con-
      sidered valid.

-------
                                                             TABLE A-l cont'd
                                                        OPERATING OONPITIOK8  FOR TASK C-3
CO
o
Date
TilM
Limestone Used*
Gas Flow, cfra
FDS L/G ratio, gal/mcf
Tower L/G ratio, gal/mcf
Gas velocity FDS, ft/sec.***
Gas velocity tower, ft/sec.
Tower pressure drop, inches  H20
FDS pressure drop, inches H20
CaO/SGj ratio (inlet analysis)
SO2 concentration, ppa
     FDS in****
     FDS out
     Tower out
Fraction of SO* removed. %
     FDS      *
     Tower
     Overall
Gas temperatures. °F
     FDS in
     FDS out
     Tower out
Liquid temperatures. °F
     FDS in
     FDS out
     Tower in
     Tower out
pH nasurements
     Tower outlet
     FDS outlet
     Hold tank
     Clarifier tank
Hold tank volume
1/11
1600
A
700
10
50
76
8.35
1.0
_.
1
_—
__
— •
•»••
__
—
——
._
-.
— «.
._
_•
—
— —
._
--
4»^
600
1/12
0930
A
700
10.0
45. 5
76
8.35
1.0
5
1
1350
500
150
63.0
60.0
88.9
330
125
127
130
122
12
125
6.0
4.0
6.3
-_
600
1/12
1730
A
700
10.3
61.0
76
8.35
1.0
4.9
1
182S
17*5
625
2.2
65.0
65.8
34?
122
118
122
122
118
121
5.2
4.4
5.6
5.2
760
1/12
2230
A
700
11
40
76
8.35
1.2
4.4
1
1920
733
429
61.9
41.5
77.7
164
122
121
122
122
121
122
5.6
4.1
5.7
5.1
700
1/13
0100
A
700
—
..
76
8.35
1.7
4.4
1
2000
—
500
__
-—
75.0
242
122
118
122
122
120
122
5.3
3.8
5.4
5.0
700
     'Limestone received from TVA for the "around-the-clock" tests  designated as A Old Tiftona from TVA shown as Type B
   ***FDS gas velocity based on outlet gas volume.
  ••** Some difficulties in SO, analyses were experienced during this test.
       High efficiencies and negative absorption on FDS should not  be con-
       sidered valid.

-------
                                                        TABLE  A-l  cont'd
                                                 OPERATING CONOITIOHS FOR TASK C-4
Ci)
Date
Tine
Limestone Used
Gas Flow, cfra
FDS L/G ratio, gal/mcf
Tower r./G ratio, gal/mcf
Gas velocity FDS, ft/sec.***
Gas velocity Tower, ft/sec.
Tower pressure drop, inches H,0
FOS pressure drop, inches H.O
CaO/S02 ratio
SO, Concentration, ppm
  2  FDS in •*•*
     FDS out
     Tower out
Fraction of SO, removed, %
     FDS
     Tower
     Overall
Gas  temperatures,  *P
     FDS in
     FDS out
     Tower out
 Liquid  temperatures,  *F
     FDS in
     FDS out
     Tower  in
     Tower  out
 pU measurements
      Tower  outlet
      FDS outlet
      Hold tank
      Clarifier tank
1/18
1330
B
700
10
40
138
8.35

• «
—
__
•»•
—

_ —
—

^ —
—
— ^
_—
__
—
__
— —
__
—
1/18
1700
B
700
10
40
138
8.35
0.8
7.8
1.6
1150
1390
400

71.3

335
110
105
112
125
108
111
4.9
6.6
5.6
7.1
1/18
190C
B
700
10
40
138
8.35
0.5
6.4
1.0
1750
1900
150

92.2
—
335
90
80
100
120
83
100
5.7
6.9
6.3
7.1
1/18
ilJO
B
700
10
40
138
8.35
0.7
7.0
.9
1700
1900
1125

40.8
—
330
105
100
105
115
100
110
5.5
7.0
5.7
7.1
1/18
2330
B
700
10
40
138
8.35
0.7
7.0
1.0
1325
1500
320

78.7
—
330
105
100
105
110
100
-
5.2
5.4
5.7
7.1
1/19
0130
B
700
10
40
138
8.35
0.7
7.0
0.9
1230
1250
350

72.0
—
330
105
100
105
110
100
110
5.3
7.2
5.9
7.2
1/19
0530
B
700
10
40
138
8.35
0.7
7.0
1.2
950
1025
488

52.4
—
330
105
102
105
112
100
110
4.2
7.4
5.8
—
1/19
1030
B
700
10
40
138
8.35
0.8
7.8
0.9
1250
1500
—
	
—
-—
335
115
105
112
120
105
-
4.7
6.7
5.7
—
1/19
1440
B
700
10
40
138
8.35
1.4
7.9
0.9
--
—
700
— ^
~
— -
330
110
105
110
118
105
-
4.9
6.3
5.5
-
1/19
1600
B
700
10
40
138
8.35
1.2
7.2
1.4
1350
1350
680
**
49.7
~~
333
112
105
112
120
105
112
4.9
6.7
5.5
~
1/19
2230
B
700
10
40
138
8.35
0.9
7.0
• MM
1475
1400
800
5.1
42.9
45.8
338
112
105
108
120
105
~
4.8
6.5
5.5
™
      •••FDS gas velocity based on outlet gas volume.
     •••« some difficulties in SO. analyses were experienced during this test.
          High efficiencies and negative absorption on FDS should not be con-
           « *	j *. « 1 * «3
          sidered valid.

-------
                                                        TABLE A-l cont'd
                                                          COIPITTOIIS FOR TASK C-4
CO
ro
Date
Time
Limestone Used
Gas Flow, cfm
FDS L/G ratio, gal/racf
Tower L/C ratio, gal/mcf
Gas velocity FDS, ft/sec. **•
Gas velocity Tower, ft/see.
Tower pressure drop, inches  H-O
FOS pressure drop, inches H-O*
CaO/SO, ratio
SO. Concentration, ppa
  2  FDS in ••••
     FDS out
     Tower out
Fraction of SO. removed* %
     FDS
     Tower
     Overall
Gas temperatures,  *F
     FDS in
     FDS out
     Tower out
 Liquid temperatures,  *r
     FDS in
     FDS out
     Tower  in
     Tower out
 pR measurements
     Tower . outlet
      FDS outlet
      Hold tank
      Clarifier tank
1/19
2230
B
700
10
40
138
8.35
1.6
•7.0
—
16SO
1625
1040
1.6
36.0
37.0
325
110
103
105
120.
105
-
4.7
7.1
5.3
1/20
0330
B
700
10
40
138
8.35
1.6
7.0
1.0
850
702
400
17.4
42.4
52.9
332
105
104
102
112
102
102
4.6
6.4
5.4
1/20
0330
a
700
10
40
138
8.35
1.6
7.0
• •
500
425
100
15.0
76.5
80.0
325
100
98
102
112
102
—
-
—
—
           *.*FDS gas velocity based on outlet gas volume
               as
               sidered valid.
                                                                                 '

-------
                                                         TABLE A-l cont'd

                                                 OPEH?TTT«; CONDITIONS FOR TASK C-5
u>
D.-ito
Time
Limestone Used
Gas Flow, cfm
FDS L/G ratio, gal/mcf
Tower L/G ratio, gal/mcf
Gas velocity FDS, ft/sec.***
Gas velocity Tower, ft/sec.
Tower pressure drop, inches  H
FDS pressure drop, inches H,O
CaO/S02 ratio
SO, Concentration, ppa
  2  FDS in  •••*
     FDS out
     Tower out
Fraction of  SO- removed, %
     FDS
     Tower
     Overall
Gas temperatures,  *F
     FDS  in
     FDS  out
     Tower out
Liquid temperature*,  *P
     FDS in
     FDS out
     Tower  in
     Tower  out
 pH measurements
     Tower  outlet
     FDS outlet
      Hold tank
      Clarifier tank
1/20
1130
•*
700
10
40
141
8.35
OR
• o
6£
• w
1250
800
17
36.0
97.8
98.6
335
85
70
100
110
75
80
5.6
50
• V
6.3

1/20
1230
g
700
10
40
141
8.35
1.0
6.4

1075
600
200
44.2
66.7
81.4
330
108
102
105
115
102
105
5.0
6.0
7.1

1/20
1400
B
700
10
40
141
8.35
1.0
7.2
1.1
1050
425
235
59.5
44.7
77.6
330
110
105
105
115
105
110
5.2
4.9
6.2

1/20
1700
B
700
10
40
141
8.35
0.8
6.6
1.1
950
350
100
63.1
71.4
89.4
330
105
102
105
115
100
105
4.8
5.2
6.4

1/20
1900
B
700
10
40
141
8.35
1.1
7.2
2.1
850
300
185
64.7
38.3
78.3
333
115
112
108
118
112
115
5.1
4.9
6.6
6.4
1/20
2100
B
700
10
40
141
8.35
1.1
6.7
1.4
750
250
155
6».6
38.0
79.3
332
115
112
111
118
112
115
4.5
4.9
6.4
5.9
1/20
2300
B
700
10
40
141
8.35
1.1
7.0
1.1
850
300
285
64.7
S.OO
66.5
328
117
113
112
118
115
118
5.3
4.8
6.4
5.7
1/21
0100
B
700
10
40
141
8.35
1.1
7.0
2.0
775
378
330
SI. 3
12.7
57.4
330
118
112
110
118
115
118
4.6
4.8
6.2
5.6
1/21
0300
B
700
10
40
115
8.35
1.9
7.0
1.7
600
250
160
58.3
36.0
73.4
332
112
112
108
118
112
118
5.0
4.8
6.4
5.5
1/21
0500

700
10
40
115
8.35
1.9
7.0
1.45
500
234
125
53.2
53.4
75.0
332
110
105
102
126
106
™
4.8
4.7
6.0
5.5
1/21
0730

700
10
40
138
8.35
3.1
7.0
— —
600
300
ISO
50.0
50.0
75.0
--
MO*
	
—
^^
—
~
•—
~~
1/21
0915

700
10
40
108
8.35
5.4
5.8
1.45
750
275
120
63.3
54.6
84.0
335
115
110
85
110
""
4.8
4.3
5.7
4.3
      •**FDS gas velocity based on outlet gas volume.
     •••• some difficulties in SO, analyses were  experienced during this test.
          High efficiencies and negative absorption on FDS  should not be con-
          sidered valid.

-------
                                                   TABLE A-l  cont'd
Date
Tine
Limestone Used
Gas Flow, efm
FDS L/C ratio, gal/mcC
Tower L/G ratio, gal/mcf
Gas velocity FDS, ft/sec.***
Gas velocity Tower, ft/sec.
Tower pr ensure drop,' inches H,O
FDS pressure drop, inches H.O
CaO/SO, ratio
SO, Concentration, ppm
   2  FDS  in  ••••
     FDS  out
     Tower out
  Fraction of  SO, removed, %
     FDS        *
     Tower
     Overall
 Gas temperature*,  *F
     FDS in
     FDS out
     Tower out
 Liquid temperatures,  *F
      FDS in
      FDS out
      Tower in
      Tower out
  pB measurements
      Tower outlet
      FDS outlet
      Hold tank
      Clarifier tank

1/21
1100
B
700
10
40
111
8.35
5.0
5.8
1.27
825
225
210
72.8
6.7
74.6
335
120
115
110
*•
120
4.9
4.5
5.7
5.2
1/21 1/21
1300 1500
B
700
10
40
132
8.35
5.0
6.0
1.3
890
225
240
74.8
73.1
340
115
70
105
115
115
4.8
3.7
5.4
5.1
B
700
10
40
115
8.35
3.2
5.2
0.8
1125
1050
330
7.7
68.6
70.7
335
120
120
118
120
120
3.0
3.V
5.7
5.0
IJTTOHS FOR TASK
1/21
1700
•*
700
10
40
115
8.35
3.3
4.6
1.0
1300
1250
370
5.7
70.4
71.6
335
118
119
114
118
119
3.8
4.1
5.8
4.8
1/21
1900
g
700
10
40
115
8.35
3.6
4.2
0.9
1500
1375
355
8.4
74.2
76.4
335
119
119
115
118
119
5.9
4.6
6.2
4.2
C-5
1/21
2100
B
700
10
40
115
8.35
3.9
4.6
C.9
1600
1550
425
3.1
72.6
73.5
335
121
120
118
120
121
5.8
4.9
6.0
4.2

1/21
2300
B
700
10
40
115
8.35
4.3
4.0
0.9
1650
1525
360
7.6
76.4
78.2
333
120
120
118
119
121
5.3
5.1
5.9
4.2
   ••• FDS g*s velocity based on outlet gas volume.

       sidered valid.

-------
                                                     TABLE  A-l  cont'd
                                               OPERATING CONDITIONS FOR TASK C-6
en
    Date
    Time
    Limestone Used
    Gas Flow, cfm
    FDS t/G ratio, gal/mcf
    Tower L/G ratio,  gal/mcf
    Gas.velocity FDS, ft/sec.***
    Gas velocity Tower,  ft/sec.
    Tower pressure drop, inches l
    FDS pressure drop, inches H2
    CaO/SO- ratio
    SO. Concentration, ppm
         FDS in
         FDS out
         Tower out
    Fraction of S03 removed,  %
         L-ne      *
         Tower
         Overall
    Gas temperatures, *F
         FDS  in
         FDS  out
         Tower out
     Liquid  temperatures,
         FDS  in
         FDS  out
         Tower  in
         Tower  out
     pH measurements
         Tower  outlet
         FDS  outlet
          Hold tank
          Clarificr  tank
1/25
1830
B
700
10
45
127
8.35
1.0
6.4
0.9
1700
1580
540
7.1
65.9
68.3
335
105
108
40
85
108
108
5.4
4.6
5.3
3.2
1/25
2130
B
700
10
45
141
8.35
1.0
7.0
1.0
1700
1635
620
3.9
62.1
63.5
328
120
120
114
118
118
118
5.4
4.4
5.5
3.4
1/25
2400
B
700
10
45
141
8.35
1.1
6.5
0.9
1920
1900
880
1.0
53.7
54.2
335
120
120
112
120
121
121
4.9
5.2
5.3
3.4
1/26
0200
B
700
10
45
141
8.35
o.«»
6.8
1.0
1930
1760
560
8.8
68.2
-1.0
330
118
118
115
118
118
120
5.2
5.0
5.7
4.7
1/26
0400
B
700
10
45
141
8.35
1.0
7.3
1.0
1890
1735
460
8.2
73.5
75.7
330
120
120
118
,120
'120
122
5.4
4.8
5.4
4.6
1/26
1115
B
700
10
45
141
8.35
1 0
6.0
1.0
1680
1560
320
7.2
79.5
81.0
118
118
112
120
118
118
5.5
5.4
6.2
4.3
1/26
1230
*»
700
10
45
135
8.35
l.C
6.2
1.2
1360
1220
100
10.3
91.8
92.7
118
118
112
120
118
120
5.9
6.8
6.4
4.5
1/26
1500
B
700
10
45
135
8.35
1.0
5.8
1.2
1260
1160
34
8.0
97.1
97.4
330
118
118
112
120
118
120
6.6
6.8
6.1
4.9
1/26
1715
B
700
10
45
135
8.35
0.9
6.2
0.9
1380
—
335
118
115
110
118
115
118
4.3
6.6
6.0
4.9
1/26
2230
B
700
10
45
141
8.35
0.9
7.0
1.1
1400
1320
240
5.8
81.9
82.9
333
118
115
114
118
115
118
5.7
4.8
6.4
4.6
1/27
0300
B
700
10
45
141
8.35
0.9
7.0
0.9
1565
1440
280
8.0
80.6
82.2
335
115
116
114
118
114
118
4.8
4.8
6.3
4.5
      •••FDS gas velocity based on outlet gas volume.

-------
                                                TABLE  A-l  cont'd
Dat«
Tim*
Limestone Used
Gas Flow, cfm
FDS L/C ratio, gal/mcf
Tower L/C ratio, gal/mcf
Ga« valocity FDS, ft/sac.***
Gas velocity Tower, ft/sec.
Tower pressure drop, inches H.O
FDS pressurw drop, inches H.O
CaO/SO, ratio
SO, Concentration, pp»
   *  FDS in.
     FDS out
     Tower  out
Fraction of SO- removed, %
     FDS      *
     Tower
     Overall
Gas temperature,  *F
     FDS in
     FDS out
     Tower  out
 Liquid temperatures,  *F
      FDS in
      FDS out
      lower in
      Tower out
 pH Bcasurements
      Tower outlet
      FOS outlet
      Hold  tank
      Clarifiex tank
1/29
1405
A
700
10
45
127
8.35
0.8
6.4
1.2
1440
1320
360
8.4
72.8
75.0
342
122
122
119
126
122
125
5 8
** • W
5 4
W * ^
7.6
3.6
1/29
1620
A
700
10
45
127
8.35
0.8
7.0
1.2
1440
1340
220
8.0
83.5
84.8
340
112
110
100
118
112
115
3.5
4.2
5.4
5.0
1/29
2100
A
700
10
45
127
6.35
0.8
7.0
1.2
1360
1280
210
5.9
83.6
84.6
342
122
120
118
125
121
123
5.8
5.4
S.4
5.0
i/30
0200
A
700
10
45
141
S.35
0.7
7.0
1.2
1280
1140
280
11.0
75.5
78.2
335
120
118
118
122
118
120
5.6
5.4
6.2
5.2
1/30
0600
A
700
10
45
141
6.3i
0.6
7.0
1.1
1480
1380
280
6.8
79.8
81.1
330
120
118
116
122
118
120
5.4
4.7
6.2
4.9
1/30
1230
A
700
10
45
151
8.35
0.6
6.4
1.3
1450
1340
400
7.6
70.2
72.5
325
120
120
118
122
122
5.4
5.5
5.8
5 A
.2
1/30
1330
A
700
10
45
141
8.35

1<9
• «
1420
1300
320
8.5
75.4
77.5
—
*"*
— **

1/30
1415
A
700
10
45
141
£.35
~~.
1 2
J» • •»
1240
120
90.4
•••»
•—

zz

1/30
1505
A
700
10
45
141
8.35
0.6
6.0
1.2
1260
1120
140
11.2
87.5
88.9
330
120
118
115
122
118
122
5.5
6.0
6 2
Q • •
5.2

1/30
1830
A
700
10
45
141
8.35
1.1
7.2
1.2
1160
1100
84
5.2
92.4
92.8
330
116
112
115
122
113
118
5.1
4.4
6.4
5.1

1/30
2100
A
700
10
45
141
8.35
1.1
7.6
1.2
1000
960
20
4 .0
97.9
98.0
320
120
118
114
120
120
120
5.3
4.4
6.3
4.9

1/30
2330
A
700
10
45
141
8.35
0.9
6.4
1M
.2
1000
979
20
4n
. V
97.9
98.0
325
118
112
115
121
112
118
4.4
4.0
6.2
5.0

1/31
0100
A
700
10
45
141
8.35
0.9
•7 O
/ . u
i 3
x . *
1080
1000
72
2.9
92.8
93.3
330
118
112
116
120
114
120
4 A
.8
4.9
6.0
5.0

   •••FDS gas velocity based on outlet gas volume.
 co

-------
        REMOVAL OF SULFUR DIOXIDE FROM STACK GASES
           BY SCRUBBING WITH LIMESTONE SLURRY:
                  TVA PILOT PLANT TESTS
               I.   SCRUBBER-TYPE COMPARISON

                            By

      T.  M.  Kelso,  P.  C.  Williamson,  and J.  J.  Schultz
              Division of Chemical Development
                 Tennessee Valley Authority
                   Muscle Shoals, Alabama
          II.  EXPERIMENTAL DESIGN AND DATA ANALYSIS
               FOR SPRAY AND MOBILE-BED SCRUBBERS
                            By
                        N. D.  Moore
             Division of Power Resource Planning
                 Tennessee Valley Authority
                   Chattanooga, Tennessee
               Prepared for Presentation at
Second International Lime/Limestone Wet Scrubbing Symposium
     Sponsored by the Environmental Protection Agency
                  New Orleans, Louisiana
                    November 8-12, 19J1
                          437

-------
               REMOVAL OF SULFUR DIOXIDE FROM STACK GASES
                  BY SCRUBBING WITH LIMESTONE SLURRY:
                         TVA PILOT PLANT TESTS
                                ABSTRACT
                       I.  SCRUBBER-TYPE COMPARISON

                                   By

            T. M.  Kelso, P. C. Williamson, and J. J. Schultz
                    Division of Chemical Development
                       Tennessee Valley Authority
                         Muscle Shoals, Alabama


           TVA plans to install limestone slurry scrubbing on a 500-mw
 boiler (No. 8 unit at  the Widows Creek station) as a demonstration project.
 Since  adequate design  data for limestone scrubbing are not available, a
 pilot  plant (2000-cfm  capacity) has been constructed at the Colbert Steam
 Plant  and  operated for about 10 months.  The limestone, pulverized to about
 70$ minus  200 mesh, is a high-calcium type from the Widows Creek area.  Most
 of  the tests  have  been at a CaO:S02 mole ratio of 1.5, an L/G (gal slurry/
 Mcf of gas) of 50  to 60, and a slurry solids content of 10-15$; the stack gas
 contains about 5000 ppm of S02.  All long-term tests have been with a closed-
 liquor loop.

           Since others have shown that packed scrubbers are effective con-
 tactors but subject to scaling, three scrubber types were tested—ranging
 from fully packed  (with fixed packing) to an empty tower.  These systems
 were:

     1.  Ventri-Rod (a type of venturi) followed by a packed crossflow
         scrubber.

     2.  Three-stage mobile-bed scrubber (bouncing bed of hollow plastic
         balls).

     J.  Ventri-Rod followed by open spray tower.

          With the Ventri-Rod - packed crossflow scrubber configuration,
about 80$ of the S02 was removed from the gas stream.  This system could not
be operated for more than about 120 hours because of plugging of the packing
either by deposition of slurry solids or by scaling.
                                438

-------
          With the Ventri-Rod - open-spray tower configuration,  about
77$ of the S02 was absorbed.  This system was operated continuously
without difficulty for over 350 hours and routinely shut down.

          Tests are in progress with the mobile-bed scrubber.   In the
tests that have been completed, 90-92$ of the S02 in the flue gas has
been removed.

          Several problems other than scaling have been encountered,  in-
cluding (l) slurry solids deposition on surfaces, (2) poor settling of
solids, (3) unstable operation (irreversible decline in pH) under some con-
ditions, (4) corrosion and erosion of surfaces, and (5) mist carryover.
All of these have been fairly well resolved (although further improvement
would be desirable) except the last two.  Test programs aimed specifically
at erosion and mist carryover are under way.

          The following conclusions can be drawn from these studies:

     1.  Good absorption (75-90$) can be obtained without scaling
         if the proper scrubber type is used.  Longer test runs  are
         desirable, however, to confirm the nonscaling operation.

     2.  At the high liquor rates required for S02 removal in
         limestone slurry scrubbing, and for fly ash of the type
         involved in these tests, excellent collection of the dust
         in the incoming gas is achieved with no special provisions.
         Solids carryover in mist is a major problem, however,  and
         must be solved if limestone slurry scrubbing is to be a
         usable process.

     3.  Care must be taken to maintain an adequate rate of limestone
         dissolution in order to avoid unstable operation.   Other
         variables may also be effective in accomplishing this but
         have not been tested.
                                  439

-------
                II.  EXPERIMENTAL DESIGN AND DATA. ANALYSIS

                     FOR SPRAY AND MOBILE-BED SCRUBBERS

                                    By

                                N. D. Moore

                     Division of Power Resource Planning
                         Tennessee Valley Authority
                           Chattanooga, Tennessee
          The experimental designs and statistical analyses of the data
from several pilot plant runs on limestone scrubbing are discussed.
Scrubber types involved are a mobile bed (three beds of hollow plastic
balls) and three configurations of an assembly consisting of a Ventri-Rod
followed by sprays in an open tower.  The significant variables for each
type are identified and linear regression equations derived.

          A mathematical model for fly ash removal is developed from data
obtained in a mobile-bed scrubber operated at a power plant with water as
the scrubbing medium.  The fly ash model is compared with experimental
data obtained in scrubbing with limestone slurry.
                               440

-------
               REMOVAL OF SULFUR DIOXIDE FROM STACK GASES
                  BY SCRUBBING WITH LIMESTONE SLURRY:
                         TVA PILOT PLANT TESTS
                      I.  SCRUBBER-TYPE COMPARISON

                                   By

             T. M. Kelso, P. C. Williamson, and J.  J. Schultz
                     Division of Chemical Development
                        Tennessee Valley Authority
                          Muscle Shoals,  Alabama
          In addition to the small-scale work described earlier in this
symposium ("Removal of Sulfur Dioxide from Stack Gases by Scrubbing with
Limestone Slurry:  Small-Scale Studies at TVA" by J. M. Potts, A.  V. Slack,
and J. D. Hatfield), pilot plant studies on limestone slurry scrubbing have
been carried out.  The purpose of both studies is to supply information
needed for design of the full-scale demonstration scrubbing facility that
TVA plans to install on the No. 8 generating unit (550-mw) at the Widows
Creek Steam Plant (near Chattanooga, Tennessee).   The pilot plant is located
at the Colbert Steam Plant (on the No. 3 unit), which is near TVA1s chemical
research facilities at Muscle Shoals, Alabama.

          The main purpose of the pilot plant work has been to identify the
problems that affect operational reliability and to find ways of solving
them.  (Scaling, for example, is known to be a major problem of this type.)
Part I of this paper deals mainly with this aspect of the test program.  Some
effort was also made toward optimizing the operating parameters; Part II covers
the experimental design and data analysis associated with this phase of the
work.

          The pilot plant, which is still in operation, receives a sidestream
of gas of about 2000 cfm (measured at scrubber outlet conditions--saturated
at 120°F) from the outlet duct of the boiler (Babcock and Wilcox,  front-fired,
pulverized coal), in which coal containing about k^> S and 15$ ash is burned.
The stack gas entering the pilot plant contains about 3000 ppm S02 and 3 to
k grains per scf (dry basis) of fly ash.

          In the pilot plant the gas is scrubbed with a circulating limestone
slurry containing 12 to 15$ solids.  This level of solids was selected because
a higher content increases pumping difficulty and increases settling tendencies
in the scrubber; a lower one reduces the beneficial effect of high solids
content on scaling and S02 removal.
                                     441

-------
          Pulverized limestone is slurried and fed into the liquor circu-
lation loop continuously.  Typical analyses of the three limestones used
are given in Table I.  Preliminary tests in the pilot plant did not indicate
any significant difference in the reactivity of the three; consequently,
most of the tests were made with limestone from Tiftonia,  Tennessee, which
is near the Widows Creek Steam Plant.
                                TABLE I

                       Typical Limestone Analyses


                                                Screen analyses,
     Quarry      Chemical analyses, %       cumulative % retained on
     location     CaO   MgO  AlgQ3  SiQ2   100 mesh  200 mesh  325 mesh

   Birmingham,
    Alabama      53.8 1.2  0.07   1.^       14       28        38
   Scottsboro,
    Alabama      50.0 k.O  0-52   2.2        8       22        37
   Tiftonia,
     Tennessee    52.3 1.2  0.13   3.4       10       27        UO


           Three different types of scrubbers have been tested; fixed packing,
mobile  packing,  and spray tower.  The objective of the tests was to determine
which of the  three would give  the best combination of reliability and S02
removal efficiency.


Packed  Scrubber (Fixed Packing)

           The  packed scrubber  selected for testing was the crossflow type
(Fig l); the horizontal configuration would have the advantage of fitting
better  into the duct system from boiler to stack.  The scrubber, manufactured
by the  Ceilcote Company, consisted of a horizontal housing (glass-reinforced
plastic) filled with spiral-type polypropylene packing (Tellerettes) irrigated
across  the top with nozzles set at right angles to the gas flow.  A section
of the  packing at  the gas outlet was left unirrigated to serve as an integral
entrainment separator.

           The  crossflow was preceded by a venturi scrubber, the primary function
of which was to remove fly  ash.  The type used was the Ventri-Rod (manufactured
by Environeering,  Inc.), essentially a horizontal housing containing a "window"
section of irrigated  parallel  rods (3/^-in. diameter; stainless steel) spaced
to develop the  pressure drop necessary for particulate removal (Fig 2).  The
pressure  drop across  the rod section could be adjusted by two guillotine-type
movable  gates.   The rod section was followed by a chevron-type entrainment
separator.  The  entire scrubber, except for the rods, was made of mild steel
coated with bitumastic material.

                                    442

-------
                     LttUOR
                     OUTLET
         FIGURE  1
Packed Crossflow Scrubber

-------
INLET
                         VENTRI-ROO
                          ELEMENT,
                                                                                                                 OUTLET
                                                                                                                  GAS
                                                          ENTRAPMENT
                                                           SEPARATOR
                                  SLURRY1
                                   INLET
                                            DRAIN
                                                          FIGURE 2
                                                  Horizontal Ventri-Rod

-------
          The overall pilot plant (Fig 3) consisted of the scrubbers
connected in series, recirculation tanks equipped with agitators and
pumps, settling tanks, and a slurry preparation tank.   Separate slurry
circulation loops were provided for the two scrubbers.  Pulverized lime-
stone was slurried in the slurry preparation tank, with liquor from the
recirculation tank in the crossflow scrubber loop, and pumped continuously
at a controlled rate back into the tank.  From there it was pumped to the
scrubber and returned to the tank by gravity.   The recirculation tank
had a working capacity of about 500 gallons.  In most of the tests, the
L/G ratio (gal slurry/1000 scf of gas) was maintained at about 40; conse-
quently slurry residence time in the tank was about 6 minutes.  This was
about twice the retention time used in the ICI-Howden experiments.

          Slurry from the crossflow recirculation tank flowed by gravity
to the Ventri-Rod recirculation tank.  The working capacity of the latter
was 500 gallons; at the usual L/G of 15, residence time in the tank was
about 17 minutes.

          Slurry from the Ventri-Rod tank flowed by gravity to the settling
tanks.  Supernatant liquor from the settling tanks was returned to the
crossflow recirculation tank, at a withdrawal rate regulated to maintain
the solids content in the recirculated slurry at  12 to 15%.  The product
solids and fly ash were periodically pumped from  the  settling tanks and
discarded.

          Unstable pH:  In preliminary  tests with this scrubber configuration,
two major problems developed  that prevented sustained operation:   (l) decline
in the pH of the scrubbing slurry finally resulted in an unresponsive slurry
and (2) uncontrollable solids buildup in the circulating slurry.

          In the initial attempts to start  the pilot  plant on a continuous
multishift operation, the crossflow circulation tank  was charged with lime-
stone and water proportioned  to form a  slurry containing 5$ solids.  The
flow of gas was started through the crossflow scrubber and the slurry circu-
lated from the recirculation  tank through the scrubber and back to  the  tank.
No fresh limestone was added  to the  system until  the  initial batch of limestone
in the slurry was exhausted.  A forward  flow of slurry to  the Ventri-Rod
scrubber was then established, the scrubbing loop closed, and the  addition of
fresh limestone started (stoichiometry,  1.0).  The pH of the  slurry  immediately
began to drop, despite the continual addition of  fresh limestone,  to such an
extent that the slurries became completely  unresponsive.   The pH  declined
finally to 3.7 and  the S02 removal decreased to about 30$.  Addition of lime-
stone in large quantities to  samples of  the slurry did not appreciably  increase
the pH.  Electron microscopic analyses  indicated  that the  limestone  particles
in the slurry were  covered with a thin  crystalline layer of material which
appeared to be calcium sulfite.  The  layer  of sulfite apparently  prevented
the limestone substrate from  further  dissolution  and  reaction.
                                    445

-------
INLET
 FLUE-
 GAS
                  1
                    VENTRI-ROD
                    SCRUBBER
•4
	I
V
or
01
                *»*
           RECIRCULATION
               TANK
                 f
               SOLIDS
             DISCARDED
                                              CROSS-FLOW
                                               SCRUBBER
                    \
                   RECIRCULATION
                       TANK
     SETTLING
      TANKS
                                      FLUE GAS
                                         TO
                                        STACK
                                                PULVERIZED
                                                LIMESTONE
                                                    SLURRY
                                                  PREPARATION
                                                     TANK

                                                MAKE-UP
                                                 WATER
                                         FIGURE 3

                   Pilot Plant with Ventri-Rod and Crossflow Scrubber Configuration

-------
          The mechanism of the limestone blinding is not yet known (see
paper in this symposium, "Removal of Sulfur Dioxide from Stack Gases by
Scrubbing with Limestone Slurry:  Small-Scale Studies at TVA" by J. M.  Potts,
A. V. Slack, and J. D. Hatfield); however, it is obvious that low pH triggers
it.  The exact pH at which the slurry becomes irreversibly unresponsive has
not been clearly defined but there is some indication from the pilot plant
work that if a level no lower than 4-5 can be maintained, the slurry will
return to full responsiveness.

          If the pH falls below about 4.0, the pH cannot be restored by
any procedure practical in large plant operation, and the slurry must be
discarded.

          The procedure developed to maintain pH at a high level during
startup involved filling the circulation tanks with fresh limestone slurry
and then starting the forward feed flow, slurry circulation, and limestone
feed simultaneously with the flow of gas.  Consequently, during the initial
startup period before plant operation was stabilized the circulating slurries
contained a large excess of limestone.  Moreover, the stoichiometry was
increased to 1.5 in an effort to maintain a stable pH.

          Even with a large excess of limestone in the scrubbing circuits,
the pH of the slurries and the removal of S02 tended to decline during the
first 20 to 30 hours of operation.  After this initial period, however,
there was a rapid inc-rease and then stabilization.  This phenomenon may be
a function of the ionic strength of the circulating slurry.  A typical de-
cline and recovery is shown in Figure 4.

          Unstable pH also occurred in two other tests in which the plant
had to be shut down for a period of about 1 hour during each test for emergency
maintenance.  The flow of slurry to the crossflow scrubber was maintained
during the shutdown to prevent drying out of the papking.  During both periods,
a small amount of flue gas (estimated to be JOO cfm ) leaked through the
scrubber.  Even this small amount of gas depleted the alkalinity of the slurry
and made it necessary to abort one of the tests because the slurry had become
completely unresponsive.  In the second test, lime was added to the scrubber
slurry to raise the pH to about 5-0 in the crossflow scrubber loop and 4-8
in the Ventri-Rod loop.  The pH rapidly dropped back to about 4.0 and remained
in the range of 4.0 to 4.5 for several hours.  After approximately 7 hours of
operation,  the pH of the makeup limestone slurry (made with circulation tank
liquor) increased from 4-5 to 6.1 within an hour.  After 3 hours of additional
operation,  the pH of the slurry to the crossflow scrubber increased from 4.2
to 5.0 within an hour and immediately the removal of S02 increased from 49 to
    (Table II).
                                   447

-------
-p.
-pi
CO
                   100
                    90
                 5  80
                 o
                 2
                 UJ

                 *  70
                  M

                 O
                 CO
                    60
                             5     10     15    20    25   30    35

                                OPERATING HOURS  SINCE START-UP
40
                                               FIGURE
                           SOP Removal  and yK in Crossflow Recirculation Liquor

-------
                          TABLE II
Increase in
Sulfur Dioxide
Removal

pH of liquor, from

March Time
17 5 p.m.
o p.m.
7 p.m.
8 p.m.
9 p.m.
10 p.m.
11 p.m.
12 p.m.
18 1 a.m.
2 a.m.
3 a.m.
4 a.m.
5 a.m.
6 a.m.
7 a.m.
8 a.m.
9 a.m.
10 a.m.
11 a.m.
12 m.
1 p.m.
2 p.m.
3 p.m.
S02
removal, f
54
59
64
49
50
48
47
45
46
47
50
_42. 	
80
86
87
84
83
83
82
77
77
76
78
F-ll
a Mixing tank
4-7
4.6
6.6
6.1
5-5
6.8
4.7
4.6
4-.L
r~ 6.1 "I
! 6.0 i
J 6.1 [
6-5
6.8
6.8
6.5
6.8
6.7
6.7
6.9
6.7
6.6
6.8
F-12
Crossf low
4.1
4-5
4.4
4.3
4.2
4.2
4.3
4.0
4.0
4.1
4.2
4.2
5-0
5-1
5-2
5.8
5-6
5-4
5-5
5-4
5-4
5-6
5.4
F-13
Ventri-Rod
4.2
4.6
4.5
4-3
4-3
4-3
4.3
4.1
4.2
4.2
4.3
4.4
1 4.3
1 4.6
[_ 41£__
"~4-9
4.9
4.9
4.9
4.9
4.9
5-0
4.9
a Lime added:  40 Ib to circulation tank of crossflow scrubber
  and 20 Ib to circulation tank of Ventri-Rod.
                                449

-------
           The combination of high  initial  limestone content in the slurry
and a continuing makeup at 1.5 times  stoichiometric made it possible to
operate without trouble from unstable pH.  Lower stoichiometry may be
feasible but has not yet been adequately demonstrated.

           Solids Settling Rate;  The  second problem that prevented sustained
operation  in the preliminary tests was  the uncontrollable solids buildup
in the circulating slurry.   In the original pilot plant configuration, the
settling tanks had a total capacity of  about 800 gallons but due to the
slow settling characteristics of the  slurry, the resulting 3-1/2 hours'
retention  time was inadequate.   The solids content in the slurry would in-
crease linearly from about 5$ in the  starting slurry to approximately 25
to yyf> after about 30 hours of operation,  a level that caused plugging of
the packing in the cross flow scrubber.

           The settling tank was  replaced with three tanks having a usable
capacity of about  4,000 gallons, equivalent to about 20 hours'  retention
time.  In  subsequent tests with the additional settling capacity, the solids
content of the slurry could easily be controlled at the desired 12 to l$%
level.

           Although the additional  settling time made it possible to operate
the pilot  plant satisfactorily in  these and the later tests, settling remains
a problem.  Even after several weeks, the  settled solids layer contains only
about k-0%  solids and the volume occupied in a waste pond therefore would be
relatively large.   Tests on methods for getting a more compact settled layer
are planned.

           Scaling;  Because of the uncontrolled pH decline and the buildup
of solids, the initial pilot plant tests were usually of only 30 to kO hours'
duration.  After these problems  were  corrected, attempts were made to operate
the pilot  plant for longer periods.   The longest period of sustained operation
(about 125 hr) was terminated because of scaling of the scrubber packing.

           Before this test was made,  the packing was removed from the crossflow
scrubber and cleaned.   At the outset  of the test, the pressure drop across
the scrubber was 4 in.  t^O.   There was  a steady increase during the test until
after 125  nr of operation (Fig 5)  the pressure drop exceeded 17 in., which
caused the seals in the scrubber sump to blow.  Many of the Tellerettes were
cemented together  into a large mass by  a hard, tenacious scale.  Also, much
of the open area in the individual Tellerettes was blocked or obstructed by
the scale  formation.

           Several  attempts were made  to wash the packing in place with high-
velocity streams of water.   Practically none of the tightly adhering scale
could be dislodged.   The packing was  removed, new packing installed, and the
test resumed.  The results were  essentially the same as in the previous test.
                                      450

-------
a:
UJ
o ii  -
CO
UJ
        12  24 36  48  60 72 84  96 108 120

                OPERATING  HOURS
                     FIGURE 3



    Ventri-Rod Crossflow Scrubber Pressure Differential



                and Slurry Solids Content

                         451

-------
          Fetrographic examination of  the  scale formed on the packing in-
dicated a bulk phase  (70-95$) of CaS04-2H20, a minor phase (10-20$) of
CaS03-0.5H20, and very minor phases of CaC03 and CaMg(C03)2.

          These  tests indicated that a packed scrubber of the type tested
is very prone to plugging,  either by solids deposition or by chemical scaling.
Although the S02 removal  efficiency was probably acceptable (about 80$ under
the conditions tested), it  was concluded that such a scrubber should not be
selected for the Widows Creek project  because operation would likely be un-
reliable.
Spray  Tower

           In  the  spray  tower design used  (Fig 6), the venturi scrubber was
integrated into the  scrubber by placing a Ventri-Rod element near the bottom
of  the tower  with the spray pointed upward.  The objective was to get more
S02 removal in the venturi (which was quite low in the venturi - packed scrubber
arrangement),  by  inducing a fountain effect that would give more slurry holdup
in  the scrubber.   The auxiliary pilot plant equipment, i.e., tanks, pumps, and
blowers, were essentially unaltered.

           The spray  tower was 32 in. square and 20 ft tall, with a superficial
 gas velocity  of 5 ft/sec at a total gas flow of 2000 scfm.  The Ventri-Rod
element, mounted  in  the inlet end of the  32-in. housing, was irrigated from
the bottom side cocurrent with the upward gas flow.  Two banks of spray nozzles
were located  in the  open tower at 'y-f.t intervals above the Ventri-Rod element.
A chevron-type entrainment separator was mounted in a horizontal outlet section
and equipped  with a  spray on the upstream side to remove accumulated solids.
(A  simplified sketch of the overall spray tower assembly is shown in Fig j).

           A set of screening tests (open  loop) was run first to check the
effects of major  operating variables (see Part II of this paper).  An ex-
tended test run then was made, at conditions indicated to be optimum, to
identify any  problems associated with extended plant operation.   The L/G
was  60 gal/Mcf—20 each to the spray banks and the Ventri-Rod unit.  Pressure
drop,  at the  gas  throughput of 2000 acfm,  was about 15 in. H20 across the
venturi and 1 in.  through the rest of the scrubber.  Stoichiometry was 1.5.

           The plant  was operated for 35^ hr with only a total of 20 hr of
downtime for maintenance.  S02 removal averaged 77$ over the entire test
(Fig 8).   The pronounced decrease in S02 removal occurring about 270 hr after
startup and lasting  about two shifts was caused by poor distribution of slurry
in the tower.   A  large hole was eroded in the bottom spray header, resulting
in poor slurry distribution and reduced S02 removal.
                                     452

-------
             ENTRAINMENT
              SEPARATOR
   VENTRI-ROD.
    ELEMENT/
   ADJUSTABLE
   SLIDE GATE
     TO
MECIRCULATION
    TANK
                      FIGURE 6
            Ventri-Rod  - Spray Tower

                           453

-------
                                             MIST ELIMINATOR
01
           INLET
            FLUE
            GAS
         f
                         MAKE-UP
                          WATER
     1
                                                                                   , ,     FLUE GAS
                                                                                            TO
                                                                                          STACK
                                               VENTRI-ROD ELEMENT
                            HUMIDIFICATION  Tuff"*
                               SPRAY 	

       SOLIDS
      DISCARDED
SETTLING
  TANKS
                                     RECIRCULATION
                                         TANK
                                                                                   PULVERIZED
                                                                                   LIMESTONE
                                          RECIRCULATION
                                              TANK
  SLURRY
PREPARATION
   TANK
                                                  FIGURE 7


                          Pilot Plant with Ventri-Rod -  Spray Tower Scrubber Configuration

-------
o

UJ
(E

 N
O
CO
en
en
       o|
       3 <
       -IK
       CO Q.
         01
           >-
           o
       pi
       rffeo
           V)
 80





 70





 60





6.8


6.4


6.0


5.6


5.2





 16


 12


 8


 4
                                      DOWNTIME
                      24   48   72   96   120  144  168   192  216  240  264   288  312   336   360

                                   CUMULATIVE OPERATING  HOURS SINCE  START-UP
                                                   FIGURE 8
                   S05 Removal,  pH, and Solids Concentration in Spray Tower Reclrculation Slurry

-------
           After the test was routinely terminated,  the spray tower was
 inspected.  No  significant scaling had occurred but there were heavy
 deposits  of limestone and reaction products  in the  tower above the upper
 spray bank.   These soft deposits,  which were several inches thick, occurred
 only in areas not wetted by the sprays; more than 50$ of the mist eliminator
 inlet was sealed off.  Although the deposits were heavy, the total pressure
 drop across the scrubber system,  including the mist eliminator, had increased
 by  only 1 in. H20--from 16 to 17 in.  H20.  The accumulation could be removed
 easily with high-velocity jets of water.

           There was some indication in this  run that better operation would
 be  obtained if  the spray banks were raised to a higher level, to give more
 space between the Ventri-Rod and the upper sprays.  Therefore the sprays were
 raised 5  ft and another set of open-loop  screening  tests was carried out.
 (A  third  spray  was also installed in the  top of the scrubber to prevent solids
 accumulation.)   The results indicated that the changes improved operation.
 It  was also indicated (see Part II) that  higher gas velocities through the
 scrubber  could  be used to reduce the scrubber cost and the area requirements
 by  making the cross-sectional area smaller.   The previous screening tests
 had been  made at superficial gas velocities  of k.6 to 6.2 ft/sec; the new
 test series indicated that 7.8 ft/sec might be used without significantly
 reducing  S02  removal.

           An  extended closed-loop run (l68 hr) was made, therefore, at (l)
 7-8 ft/sec, (2) L/G of 55 gal/Mcf (2O to  Ventri-Rod, 5 to first spray bank,
 15  to second, and 15 to the new top spray),  and (3) 1-5 stoichiometry.  The
 S02 removal was only 66%,  however,  as compared with the 76% predicted from
 the screening tests and the j8% obtained  in  the earlier run at 4.6 ft/sec.
 Possible  reasons for the lower removal are discussed in Part II.

           These tests indicate that the spray tower is a reliable scrubber
 for S02 removal, but is not as good a contactor as a fixed-bed scrubber.
 Also,  the limitation in gas velocity may  be  an important cost factor.


Mobile-Bed Scrubber

           The mobile-bed scrubber type tested was the Turbulent Contact
Absorber  (TCA),  manufactured by Universal Oil Products Company.  It consists
 of  three  beds of mobile packing spaced vertically (Fig 9) and is designed to
 operate at 8-13 ft/sec.   Each packing unit consists of a 1-ft static depth of
 1.5-in. hollow  plastic  spheres,  retained  by  wire mesh grids k ft apart.  The
 flow of gas through the bed causes  the spheres to move rapidly in a random
pattern between the two retaining  grids.  The scrubbing liquor is fed above
 the  top retaining grid  through an  open-pipe  distributor and flows counter-
currently to  the gas flow.
                                   456

-------
                                       •*•
      ' LIQUOR OUTLET
         FIGURE 9

Mobile-Bed  (TCA) Scrubber
           457

-------
           A chevron-type mist eliminator manufactured by the Heil Process
 Equipment Company was installed in the horizontal housing at the outlet
 of the scrubber.   The pilot plant configuration with the TCA scrubber is
 shown in Figure 10.

           After the usual series of screening tests (Part II), a 172-hr
 closed-loop run was made at a gas velocity of 12.5 ft/sec, a total L/G of
 *4-8 gal/Mcf, and 1.5 stoichiometry.   The total pressure drop through the
 scrubber was 9 in. H20 during the entire test period.

           After routine termination of the run, inspection of the scrubber
 showed a small accumulation of solids immediately below the bottom retaining
 grid of the first stage.   Otherwise, the scrubber elements were clean with
 no significant evidence of accumulated solids or scale.

           Removal of S02 was excellent, averaging about 92^.  The main problems
 were ball wear, erosion of grids,  and particulate entrainment.

           Further tests of the mobile-bed scrubber are being made.  If the
 indicated resistance to scaling can be verified, and the problems of erosion
 and particulate carryover solved,  the scrubber should be a good candidate
 for large-scale use.


 Erosion

           Severe and rapid erosion by the circulating slurry has been a major
 oroblem with all three of the scrubber types.  For each of three, frequent
 replacement of pipefittings, valves,  and pump components, has been necessary.
 Other problems have been specific to the particular equipment.

           Ventri-Rod;  During preliminary tests with the crossflow scrubber
 assembly,  a hole was eroded in the Ventri-Rod housing.  A 1/2-in.-thick mild
 steel wear plate welded over the hole was cut through after about 200 hr of
 operation.   A test panel coated with Urecal (a urethane-based coating) was
 installed in the housing in the area prone to highest erosion.  After 300 hr
 of  operation,  the welds supporting the panel were eroded through, but the
 coating  on the panel was undamaged.

           Spray Tower;  During the tests with the spray tower system, the
 303 stainless  steel  spray nozzles and the mild steel lower spray header had
 to  be replaced at intervals of about 300 hr.

          Mobile-Bed Scrubber;   In the tests made with the TCA scrubber, the
weight loss of the plastic spheres has averaged about 10$ in 300 hr of operation.
During the  same period,  the weight loss of the bottom retaining grid was about
30$.  In tests now being made,  finer limestone (88$ through 325 mesh) is being
used  in  an  effort to reduce erosion.  A small test loop is also being assembled
in which  test  specimens  can be studied under conditions more closely controlled
than  is  feasible  in  the  scrubber.


                                     458

-------
en

MIST
ELIMINATOR





m
i'V





l^/
....
P.
•





























^



4
HUMIDIFICATION

INLET ^R
PI UF ^ ^
GAS w


PULVtKIZbD ' 	 """'
LIMESTONE
U 9
if " H
SPRAY
r







J









MAKE-UP
WATER — ^
> & T '
r
f
y^pj j->

i ^
t



^
1 ,,
/
/


4
f
i
-WTP-^

i
i





FLUE GAS
TO
i ' STACK
i
.c5
^id








j
/•"<^ 	 •
==-^>^ J r%. - ~ ^~ - j 	 .a^TP-*'
^^S^^t'^dLl^- 2!1
SLURRY'^ RECIRCULATION RECIRCULATION'0" 1 SETTLING
PREPARATION TANK TANK ' TANKS
TANK







SOLIDS
DISCARDED
FIGURE 10
                                         Pilot Plant with TCA Scrubber Configuration

-------
Particulate Loading in Exit Gas

          The objective in the test program has been to remove fly ash from
the gas to the degree that a "clear stack" is obtained.  The degree of dust
removal required to give a clear stack varies with several conditions, but
it seems generally accepted that a level of 0.02 gr/ft3 (dry) or less is a
reasonable objective.

          In the TVA pilot plant tests, the particulate loading in the exit
gas from the scrubbers tested has averaged about 0.0} gr/scfd from the cross-
flow, 0.027 and 0.05 gr/scfd from the spray tower (for k.6 and 7.8 ft/sec gas
velocity, respectively) and 0.07 gr from the TCA scrubber (12-5 ft/sec velocity).
The particulate loading has been a linear function of the apparent gas velocity
in the barrel of the scrubber.

          Petrographic analyses of the particles indicate that the material
in the exit gas is calcium carbonate or calcium sulfite - sulfate rather than
fly ash.  Apparently, the fly ash is readily captured and removed in the first
stage of each scrubber configuration, even at the low pressure drop in the
mobile-bed scrubber  (see Fart II).  The particulate in the exit gas comes from
the mist of scrubber slurry that passes the chevron mist eliminator.   Tests
are currently in progress to develop a more effective mist elimination system.

          Tests of scrubber configurations in the pilot plant are continuing.
Evaluation of the spray tower and TCA scrubber have not been completed.


Summary

          Of the three scrubber types tested, all should give adequate S02
removal if properly designed.  As to reliability, the fixed bed (crossflow
type) is unacceptable because of scaling and plugging.  The spray tower and
mobile-bed types avoid this problem, but each has other problems that must be
solved.  The advantages and disadvantages of each, as indicated by the pilot
plant work, are listed in Table III.

          Further tests are under way to continue evaluation of the spray tower
and mobile-bed types.  Other scrubber configurations may also be tested.  The
main remaining problems, on which special studies are planned, are erosion,
mist carryover, and slow settling of solids.  Ways for avoiding unstable pH
at low limestone stoichiometry will also be explored.
                                   460

-------
                                TABLE III
                       Evaluation of Scrubber Types
             Spray Tower
Advantages

1.  No internals for build up of mud
    or scale

2.  Infinite turndown
    Spray nozzle and distributor
    replacement only routine mainte-
    nance item

    Low pressure drop through tower
    -AP-2.0 in. H20 or less
     Mobile-Bed Scrubber (TCA)


Advantages

1.  Gas velocity of 12-5 ft/sec
    possible with good S0e recovery

2.  Fly ash removal possible in
    same unit

3-  Slurry distribution may require
    only open pipes with splash
    plates

4.  Excellent gas contacting
    efficiency
Disadvantages

1.  Gas velocity limited (however,
    preliminary indications are that
    the velocity may be increased
    significantly without loss of
    efficiency)

2.  Separate fly ash removal system
    required

3-  High velocity (atomizing nozzles)
    required for slurry distribution,
    requring high-pressure pumps

4.  Gas contacting efficiency rela-
    tively low
Disadvantages

1.  Some internals (support grids,
    etc.) required that could promote
    solids accumulation
    Turndown limited
    Bouncing balls and supporting grids
    must be replaced at regular inter-
    vals because of erosion

    Pressure drop across scrubber of
    9 in. or more required

    Mist elimination problem aggravated
    by high gas velocity
                                    461

-------
               II.   EXPERIMENTAL DESIGN AND DATA ANALYSIS
                    FOR SPRAY AND MOBILE-BED SCRUBBERS
                                   By
                               N. D. Moore
                    Division of Power Resource Planning
                        Tennessee Valley Authority
                          Chattanooga, Tennessee
          As described in Part I of this paper, three scrubber types have
been tested in the TVA pilot plant studies on limestone slurry scrubbing of
S02«  The fixed-packing type was abandoned because of scaling, leaving the
spray tower and mobile bed as the candidates for full-scale application.
In the tests of each type, short-term tests (open loop) were made to evaluate
effects of the major operating parameters and to determine the best combination
for extended closed-loop runs aimed at evaluating operational reliability.

     N    The basic experimental plan used for all the screening tests was
the 2 (N = 1, 2, 3, . . . ) factorial or fractional factorial.  At the com-
pletion of each test series, the data were analyzed by analysis of variance
and regression analysis.

          For both scrubber types, preliminary tests indicated that of the
several operating parameters, four had the most significant effect:

     1.  Liquid to gas ratio (L/G)

     2.  Gas velocity (v)

     3.  Pressure drop (AP)

     4.  Distribution of slurry between scrubber units


Spray Tower

          The first test series consisted of 20 tests with the spray tower
as shown in Figure 6 in Part I.  This was a 1/2 replicate of a 2  factorial
experiment, with four center points added.  The five variables studied
and the range of each variable were as follows:
                                  462

-------
                           Variable
                                                        Range
1.  L/G to Ventri-Rod (L/GVR), gal/Mcf

2.  Pressure drop across Ventri-Rod

3-  L/G to first spray above venturi

4.  L/G to second spray above venturi

5.  Gas velocity, cfm
                  ft/sec
                                                 in. H20
  20-40

   7-15

   0-10

   0-10

2000-2700
 4.6-6.2
The other operating conditions are given in Part I.  Detailed test data
are listed in Table IV.
                                TABLE IV

            Test Conditions and Results from Screening Tests

                  of Spray Tower Scrubbing (First Series)
       Gas              AP
Test  flow,  APvR,   overall,  L/G, gal/Mcf    $     S02 removal,
No.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
cfm
270O
2000
2000
2000
2350
2700
2700
2000
2000
2350
2700
2000
2700
2000
2350
2700
2000
2700
2700
2350
in. H20
6.5
6.0
tt.5
15.0
10.0
13.2
13.9
15.6
7-2
11.1
7-4
15.2
15.0
7-4
11.3
7-0
7.0
15.0
7-2
11.6
in. H20
10.4
-
15.0
15-6
10.6
15.9
16.9
17.0
8.5
12.4
8.3
16.7
17.0
8.7
12.7
8.9
8.4
17.0
8.4
14.1
VR
4o
20
20
20
30
20
40
4o
4o
30
20
4o
20
4o
30
20
20
4o
40
30
SI
0
10
0
10
5
0
10
10
10
5
10
0
10
0
5
0
0
0
10
5
S2
10
0
0
10
5
10
10
0
10
5
10
10
0
0
5
0
10
0
0
5
solids
12
25
16
16
16
13

12
16
15
13
16
17
16
14
17
13
16
20
18
% average
64
55
58
68
58
64
71
60
59
59
60
61
60
53
56
50
59
66
57
60
Range
59-69
53-58
57-62

54-63

70.5-71.4
60.0-60.3
57-60
58-61

60-63
59-62
51-55
56-57
49-51
59
66
56-59
58-6!
                                    463

-------
          Although each variable had a statistically significant effect
on percent S02 removal, the two largest effects were the pressure drop
across the venturi and the L/G to the second spray above the venturi.
The regression equation derived from the data is as follows:

     Predicted $ S02 removal =

           59-9 + IT xx + 51 x2 + 15 X3 + 47 x4 + 19 X5 + 31 xs
                  T£      T5T      T&      TS      15"      ib

     where
           X2 =

           X3 =  (L/Gsl-5)/5

           X4 =  (L/GS2-5)/5

           X5 =  (Gas velocity-2350)/350

           X6 = Xi-Xs

           This equation accounts for 90$ of the total variation of the data.
 The  remaining 10$ is composed of 3.0$ due to curvature (i.e., x^2) and J.0$
 error.   The  analysis of variance is tabulated in Table V and the predicted vs
 observed values  are given in Figure 11.  Based upon the analysis of the data,
 95$  of  the actual or observed values will be within 3$ of the predicted percent
 S02  removal.

           Two of the results from  this test series were encouraging.  First,
 the  contribution to S02 removal from the lower spray (si) was small compared
 with other variables.   Since this  spray was located in a position just above
 the  venturi  and was subject to severe erosion, removing it would simplify
 operation  without any  great loss of S02 removal efficiency.  Secondly, in-
 crease  in  gas velocity,  within the range studied, improved S02 removal slightly
 rather  than  reducing it.

           The S02 removal  in the extended run made under closed- loop con-
 ditions  in the venturi  -  spray tower (Part l) checked with the predicted
 removal  from the  open- loop screening tests within experimental error.

           The second spray tower test series was run after the spray banks had
been raised to a  higher  level and  a wash spray (53) installed at the top of the
scrubber.  The series consisted of two blocks of tests composed of a full factorial
for two variables  (22) plus one center point in the first block (Table Vl).
                                     464

-------
                         TABLE V

        Analysis of Variance for Screening Tests
                                               a
        of Spray Tower Scrubbing (First Series)
     Source
Mean
  L/GVR
L/G
     S1
  Velocity
       ; velocity
Residual
  Lack of fit
    Curvature
    Interaction
  Error

     Total
 Sum of squares

71,760.2
    18.06
   162.56
    14.06
   138.06
    22.56
    60.06

    48.42
      39.6750
         13.6
         26.0750
       8-75

72,224
Degrees of
 freedom

   1
   1
   1
   1
   1
   1
   1
    10
      1
      9
                                                 Mean
                                                square
 18.0625
162.5625
 14.0625
138.0625
 22.5625
 60.0625
                                                 2.897)  2.9 (pooled  estimate)
                                                 2.92 )
                                     20
  Correlation coefficient = 0.96; index of determination
b = °'92
  All values significant at a = 0.01, where a indicates
  degree of confidence  (i.e., 0.01 indicates 1% possibility
  of variable not being significant).
                              465

-------
                                                                       S02  REMOVAL,  %
                55          60          65
                OBSERVED  S02  REMOVAL,  %

                                          FIGURE 11

Comparison of Predicted and Observed Results  in Screening Tests of Spray Tower Scrubbing  (First Series)

-------
                                TABLE VI

            Test Conditions and Results from Screening Tests
                                                        a
                of Spray Tower Scrubbing (Second Series)

Test
No.
1
2
3

5
6
7
8
9

APV
in.
7
7
15
15
11
15
15
15
15
b
R>
HgO
•3
.4
.1
.2
.2
.1
.0
.1
.0
AP



overall,
in.
10
10
17
18
14
18
17
18
Hao
•
.
»
,
*
*
*
*
18.
3
6
9
3
5
0
8
0
2
VRb
20
38
20
38
30
20
20
20
20

L/G
SI
10
10
10
10
10
0
20
0
, 20


S2
10
10
10
10
10
18
0
0
17



1°

S02
S3 removal
10
10
10
10
10
9
10
10
10
67
71
75
75
70
64
62
58
72
.1
.6
•9
•3
.1
•3
• 7
.8
.8
            a Gas flow, 2700 cfm  (6.2 ft/sec); L/G to humidi-
            ,  fication  spray, 5-5 gal/Mcf.
              Ventri-Rod  scrubber.


The variables studied and the range of each were as follows:


                           Variable
                1.  L/G  to Ventri-Rod, gal/Mcf     20-38
                2.  L/G  to SI, gal/Mcf              0-20
                3.  L/G  to S2, gal/Mcf              0-l8
                4.  AP across Ventri-Rod, in- H20   7-15

All tests were conducted at the maximum  system velocity of 2700 ft /min
(6.2 ft/sec).  In the first block of  five tests, the AP across the venturi
and the L/G to the venturi were varied.
the following equation:
Analysis of these data resulted in
                       S02 removal = 72 + 3AP + L/G -  1-5(L/G)(AP)
                               (2)
                     where      AP =  (APVR-ll)/4

                               L/G =  (L/GVR-29)/9

Table VII gives the analysis of variance.  The L/G was not  significant  over
the range tested.
                                    467

-------
                                TABLE VII

                     Analysis of Variance for Block 1

                   of Second Test Series on Spray Tower
                           Degrees of  Sum of    Mean
                 Source      freedom   scuares  square  F value
Mean
Ap
L/G
(AP)(L/G)
Residual
Error0
Total
1
1
1
1
1
(12)
5
25,920
36
k
9
5

25,97^
p
36 12
4 i.y
9 ?
5 1.6?
3

                 Significant at a = 0.01.
                 Not significant or significant at ci> 0.10.
               0 Error variance assumed to be 3.0 with 12 degrees
                 of freedom.
          The next block of tests was run with L/G to sprays SI and S2
varied but with venturi AP and L/G held constant.  The resulting equation
was
           % S02 removal = 65 + 3-25L/G1 + 3.75L/G2 + l.asCL/G^L/G.,)   (3)

           where    L/G^ = (L/G to Sl-lo)/10

                    L/G2 = (L/G to s2-9)/9

The analysis of variance is given in Table VIII.1

          This block indicates that the two sprays are both important in
S02 removal but are independent and the greater the L/G to each the greater
the SO  removal.  Thus raising the sprays in the scrubber did increase the
effectiveness of the lower spray (Si).
  From previous error estimates, the error variance is estimated to be 3.0
  with 12 degrees of freedom (see Table v).  This estimate is used as the
  error throughout the Ventri-Rod - spray tower test work.
                                      468

-------
                                TABLE VIII

                      Analysis of Variance for Block  2
                                                        a
                    of Second Test Series on Spray Tower
                       Degrees of    Sum of    Mean
             Source     freedom     squares   square   F value
Mean
L/G!
L/G2
L/G-^L/GS
Total
1
1
1
1
4
16,770.25
42.25
56.25
6.25
16,875.00

42.25
56.25
6.25

b
14.08
18.75C
2.08

            a Error variance assumed to be 3-0 with 12 degrees
            ,  freedom.
              Significant at a = 0.01.
              Not significant or significant at a>0.10


          A third series of tests (Table IX) were run to test  the  system
at the maximum velocity attainable,  3400 ft3/min (7.8 ft/sec).   This  group
consisted of a 1/2 replicate of a 25 factorial experiment with 4 center
points added for a total of 20 tests.   Again, open- loop conditions were used.
Analysis of the test results gave the equation:
  % S02 removal = 68-9 + O-QZL/Gy^ + 0.^6^/G^ + 2.J58L/G2 +

      - i. 55(1/6^) (L/G) + o.95(VGvR)(ApvR) - i-93(L/G XL/G )
                       o                                J.     &
      - 0.85(L/G2) (L/G5) + 2.2(L/G2)(APVR) + l^L/GgXAP^)            (4)

  where

      L/GVR = (L/G to Ventri-Rod-28)/7

       L/GI = (L/G to si-io)/5

       L/G2 = (L/G to S2-io)/5

       L/G3 = (L/G to S3-io)/5

             (AP to Ventri-Rod-12)/3
                                   469

-------
Each of the variables in the above equation is significant at a = 0.2  or
lower.  The analysis of variance is given in Table X, along with the
t - statistic for each variable.
                                TABLE IX
             Test Conditions and Results from Screening Tests
of Spray Tower Scrubbing (Third
Test
No.
1
2
3
4
5
6
T
8
9
10
11
12
13
Ik
15
16
IT
18
19
20
AP,
VR
14.5
9
15
15
12
8.9
15.1
15.1
13-3
13.5
9.0
15.0
14.9
9-2
12.6
11.8
11.0
15.0
13.0
12.1
in. H20
Overall
19.8
19.1
21.6
2J.1
21-5
22.0
23.0
23.1
22.6
23.2
24.5.
l.8b
*'ll
2.0b
20.5
24.5
21.5
20.5
25.9
20.6
a
Series)
L/G
VR
20
35
20
35
28
35
35
35
20
28
20
20
35
35
28
20
35
20
20
28
SI
5
15
15
5
10
5
15
5
15
10
15
15
15
15
10
5
5
5
5
10
S2
15
5
5
15
10
5
5
5
15
10
5
15
15
15
10
5
15
5
15
10
!1
15
15
15
5
10
5
5
15
15
10
5
5
15
5
10
15
15
5
5
10
 S02
removal
76.8
68.9
70.4
77-6
77-3
65.8
68.2
64.3
71-3
68.7
68.9
68.3
75-6
70.7
70.4
65.7
67-1
57-0
70.1
69.5
           a Gas flow, 3400 cfm  (6.8 ft/sec); L/G to humidi-
           b fication  spray, 4.4 gal/Mcf.
             Inch Hg.
                                    470

-------
                                TABLE X


        Analysis of Variance for Third Test Series on Spray Tower*
          Source


        Regression
        Error

             Total
Degrees of
 freedom


    10
    19
Sum of
squares


 367-73
  22.59


 390.32
Mean
square


36.67
 2.51
F ratio
          Variable
t - statistic
   Variable
    t - statistic
L/G to VR
L/G to SI
L/G to S2
AP to VR
(L/GvR)(L/G3)
(L/GVR)(AP)
1-73
2.40
5.30
1.78
-3.69
1.71
(L/GX)(L/G2)
(L/6j)(L/Gg)
(L/G2)(AP)
(L/G3)(AP)


-lf.58
-2.08
Ml
3-21


          Correlation coefficient = 0-97; index of determination
          = °-^'
          Significant at a = 0.01.
          The most significant variables in the third series are pressure
drop across the venturi and the L/G to the upper spray, S2.  These tests
indicate that although increasing the velocity above 2700 ft3/min (6.2 ft/sec)
does reduce the S02 removal, this effect can be offset by judicious choice
of slurry distribution.  The following tabulation illustrates the effect of
velocity on S02 removal when the slurry distribution to the sprays is held
constant.


             Velocity,   APyR>    L/G, gal/Mcf      S02
              ft/sec    in. 1^0  VR  Sll  S2  S3_  removal, %
                k.6
                6.2
                7-8
                k.6
                6.2
                7-8
                k.6
                6.2
                7-8
   15
   15
   15
   15
   15
   15
   15
   15
   15
20
20
20
30
30
30
38
38
38
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
               72
               76
               68
               72
               75
               71
               72
               75
               72
                                 471

-------
          Under  these  conditions,  2700 cfm gives  the maximum  removal.   By
varying  the distribution to  the sprays,  however,  good  results can be ob-
tained at 5^00 cfm--as shown in the following tabulation.
Velocity,
£t/seca
7-8
7-8
7-8
7-8
7-8
APVR,
in. H2(
9
9
15
15
15
L/G, gal/Mcf
) VR
35
20
20
20
35
SI
5
15
5
c;
15
s2
15
5
15
15
15
V i
tt
s
10
15
10
15
S02
removal, %
73
TO
76
74
78
               3400 cfm.
           In view of these results,  an extended  closed- loop  run (165 hr)  was
 made at ^400 cfm, 15&VR,  20L/GVR>  5VGS1>  15L/GS2> and 15L/GSy   The
 average removal,  however,  was only 66$ as compared  with the  predicted 76$.
 The cause of this is unknown; the removal was  about J0% at the  end of the
 run (stopped to install a different type of scrubber), and perhaps would
 have improved further if time had been available for  extending  the run.
Mobile-Bed Scrubber

          In screening tests with the mobile-bed (TCA) scrubber, the variables
were gas velocity, L/G to the main slurry inlet (Si) over the mobile beds,
and L/G to the wash spray (S2) in the top of the scrubber.   Test conditions
and results for the 16 tests are given in Table XI.  In developing the equation,
test 5 was not included and tests 11 and 16 were averaged to give a total of
14 tests.  The equation derived is as follows:1

                                                                          (5)
   % S02 removal = 64-58 + 5.04L/G2 + 0.855V- L/G! - 0.94l(L/Gi

         where V = velocity (cfm) /100, L/G! = (L/G to Sl)/10,

        and L/G2 = (L/G to S2)/10

The analysis of variance and t - statistics are given in Table XII.  The
t - statistics for each variable are significant at ey = 0.05 or lower.
  It should be remembered that all regression equations strictly apply only
  within the experimental region and that extrapolation can be hazardous.
  This is particularly evident in the TCA system as it is very easy to flood
  the beds at certain conditions and in fact too large an L/G may be detri-
  mental.
                                   472

-------
                                TABLE XI

             Test Conditions and Results for Screening Tests

                         on Mobile-Bed Scrubber

Test
No.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Gas
flow,
ft/sec
10.5
16.2
16.2
10.5
16.2
10.5
16.2
10.5
16.2
14.4
13.3
8.3
8-3
8.3
8-3
13-3
AP through
scrubber,
in. H20
7-1
16.5
19.6
5-6
24.0
9-0
16.7
9-0
18.6
22.0
7-7
6.1
5-4
8-5
7.7
7-2
L/G,
gal/Mcf
SI
12
4
12
12
32
52
4
52
12
35
20
15
15
65
65
20
S2
29
10
16
17
15
29
16
17
10
16
17
36
21
36
21
17
% S02
removal
81
77
82
71
74
95
76
95
79
94
84
86
79
89
86
77
               L/G to humidification, 15 gal/Mcf.
          It is interesting to note that velocity was not a significant in-
dependent variable in these tests.  It will appear if the level of significance
is selected at a = 0.25, but the coefficient is positive indicating that
increase in velocity will increase S02 removal.

          At the conclusion of this series, an extended closed-loop run
(172 hr) was carried out at conditions selected on the basis of the screening
tests (12.5 ft/sec, j4 L/G to main slurry inlet, and 14 L/G to the upper
wash spray).  Average removal was 92$, close to that predicted.
                                  473

-------
                                TABLE XII

                Analysis of Variance for Screening Tests
                                                a
                          on Mobile-Bed Scrubber
                           Degrees of  Sum of    Mean
               Source	   freedom    squares  square  F ratio

          Regression            3       621.18   207-06   19-5b
          Error (residual)      10       106.0^    10.6

               Total           13       727-22
                        Variable          t - statistic
                    L/Gj,                     2.9103
                    (velocity)(L/GS)         5-teiu
                                           -2.6095
            Correlation coefficient = 0-92
            Significant at  a = 0.01.
 Particulate Removal

           Since removal of particulate matter as well as S02 is required
 in wet scrubbing with limestone slurry,  small-scale tests were made to
 determine the pressure drop necessary for good removal.  This study was
 conducted at the Widows Creek plant in order to have the same type of dust
 as in the later full-scale scrubbers; no limestone was involved.

           The equipment used to determine required pressure drop was the
 "Dust Difficulty Determinator," a small  orifice-type test scrubber (with
 a  following mobile bed) manufactured by  Environeering, Inc.  Composition
 of the coal burned during the test period is given in Table XIII, particle
 size  of the ash in Table XIV,  and results of the tests in Table XV.

           Data analysis was performed on tests 16 through 22, 2k through 26,
29  through 42,  and ^5.   The first 15 were preliminary, obtained during the
shakedown period;  the others were rejected because equipment difficulties
were believed to have made the results atypical.
                                   474

-------
                                TABLE XIII

   Composition of Coal Burned During Dust Removal Tests at Widows Creek


               Total      	Weight percent,  dry basis
apple
1
2
3
4
5
6
moisture, $
6.6
5-7
5-6
7.4
7-0
7-7
Volatile matter
28.4
31-5
38.2
31.0
31-2
32.2
Ash
20.7
21.0
17.4
18.3
18.1
19-1
Fixed carbon
50.9
47-5
44.4
50.7
50.7
48.7
Sulfur
1.1
2-9
U.I
2.5
3A
3.1
          The analysis indicated that both L/G and AP affect the degree
of dust removal.  The following equation was developed:
              Y = 0.12(-AP)-°-54                                    (6)

              where
                                                   Q
                    Y = outlet grain loading, gr/ft

                    L = liquid pumping rate, gal/min

                    G = gas rate, acfm x 10"

                   AP = pressure drop, in. H20


If the pressure drop across a venturi is defined as

                   AP = L . V1'87                                      (7)
                        G

              where V is gas velocity (ft/min),

then substituting equation 7 in 6 gives
              Y =
0.12G
                              _           _O.54
                       0. 54
                      L0.54
                              (L) .
                                    475

-------
                                    TABLE XIV
        Bahco Particle Size Analysis of Fly Ash During Widows Creek Tests



Test No.   i     2     3     31    32    33     3^    35    36           1


Micron   Inlet (6-28-71)    Inlet (6-29-71)    Inlet (6-50-71)    Outlet (6-28-71)
                                less than        % leas than        % less than
0 J»C*«
3-3
6.5
9.6
1J.2
18.0
23.9
^ *
29.5
J2.4
/" *
3-1
13.0
*.j • **
27-6
48.2
69-5
85.0
89.0
90.3
«V 0 ** *•* A
4.6
20.7
r*w i
42.5
57-1
64-3
69-5
72.7
74.3
**»*
2-7
14.0
31.8
50.6
65-7
79-4
83.8
85.4
/ —
3.8
17-0
36.0
59-6
78.6
89.6
92-3
93-1
5-4
19.5
37.1
47-7
54.0
63.6
70.3
73-5
6.2
31.8
49.7
62.9
70.2
74.5
77.1
78.6
4-3
20.7
42.9
60.0
69.2
77-7
81.3
82.8
I'1?
18.4
41-3
63-7
79-9
87.7
89.7
90.4
4.2
20.6
40.8
51.7
56.1
5M
66.0
68.6
i.
9-
22.
35-
48.
65-
71.
76.
4
5
2
0
4
6

3
 74      94.6  86.7  92.4   94.6  86.1  87.5   88.9  91-6  84.5        91-6
                                                                 (Filter No. 702)

-------
                             TABLE XV
   Dust Removal from Power Plant Stack Gas (Widows Creek Tests)
        &PLS     APHS
L/G, gal/Mcf
 Inlet    Outlet
loading,  loading,
Run No.
1
2
3
4
5
6
7
8
9
10
11
12
13
tt
15
16
17
18
19
20
21
22
23
A
25
26
27
28
29
30
31
32
33
3^
35
36
37
38
39
to
41
42
43
1*
45
in. H20
_
-
-
3-6
3-9
4.0
6-5
3-1

0.8
1.6
3-0
3-o
4.0
4.9
4.0
2.2
1.8
2.1
2.1
2.1
-
-
_
-
-
_
0.2
0.2
0.2
1.2
0.8
0.4
1.7
3-2
3-2
2
2
2
_
_
_
_
_
-
in. H20
33
11
20
10
28
11
11
21
46
10.4
21.4
32
5
10.3
15
-
-
-
-
-
-
5-5
10.9
4.9
3-4
^3
4.4
-
-
-
_
-
-
-
-
_
_
-
_
4.9
4-5
4.2
8.8
10. 1
10.1
in. H20
33
11
20
15.6
31-9
15
17-5
24.1
46
11.2
23
35
8
14.5
19-9
4.0
2.2
1.8
2.1
2.1
2.1
5-5
10.9
4-9
3-4
M
4.4
0.2
0.2
0.2
1.2
0.8
0.4
1-7
3-2
3-2
2
2
2
4.9
4-5
4.2
8.8
10.1
10.1
LS
20
20
24
17
17
19
7.4
18
18
20
19
18
20
24
13
27
27.7
28.,3
20.2
17.5
12.8
-
-
-
-
-
-
2.6
3.9
5.2
5-2
3.9
2.6
2.6
5-2
7.8
4.7
9-4
13
-
-
-
-
-
—
HS
34
19
17
22
22
14
15
15
15
15.1
14.2
14.0
16
14
14
-
-
-
-
-
-
15.2
10.0
7.2
5-0
7-2
10.0
-
-
-
-
-
-
-
-
-
-
-
-
7.6
5.4
2.1
5.1
4.0
3.0
T
54
39
41
39
39
33
22.4
33
33
35-1
33.2
32
36
38
27
27
27-7
28.3
20.2
17.5
12.8
15.2
10.0
7.2
5-0
7-2
10.0
2.6
3.9
5-2
5-2
3.9
2.6
2.6
5.2
7.8
4.7
9-4
13
7-6
5-4
2.1
5-1
4.0
3.0
gr/scfd
1.20
2.00
5-40
2.70
0.62
2.30
2.00
2.00
0.34
1.68
1.40
1.57
4.0
3.0
2.8
5-9
3-2
2-9
2.7
2-9
3.0
1-9
4-5
5-9
6.01
4-53
3-59
6.2
2.2
3-9
6.6
4.0
3-7
3-4
3-5
2.7
2.4
3-7
3-3
2.2
-
5-2
5-5
5-2
4.4
gr/scfd
0.0068
O.0097
0.027
0.12
0.014
0.10
0,16
0.027
0.0032
0.006
0.010
0.010
0.007
0.006
0.006
0.011
0.009
0.009
0.025
0.007
0.031
0.006
0.068
0.016
0.013
0.018
o. 031
2.7
0.11
0.09
o.o4
0.044
0.13
0.075
0.028
O.o4£
o.o64
0.038
0.026
0.02
0.0246
0.0445
0.0743
0.0483
0.03
HS = orifice section of scrubber.
LS = mobile-bed section of scrubber.
 T = total.
                               477

-------
Now let G = pAV,  where A is the cross-sectional area.

Substituting G = pAV in equation 8 and approximating coefficients,


                    Y = 0.12[~Al                                       (9)


Thus the outlet grain loading is a function of the cross-sectional area
and the resistance (i.e., L where L = liquid pumping rate) across the
area.  The analysis of variance, for equation 6 as the regression model,
is given in Table XVI.
                                TABLE XVI

               Analysis of Variance for Dust Removal Tests


                           Sum of    Degrees     Mean
                 Source    squares  of freedom  square  F_

               Mean         362.52       1                a
               Regression   191-97       1      191-97  75
               Residual      62.29      2£        2.56

                    Total   616.78      25


                         (R2= 0.755; R = 0.87)
                 Significant at a = 0.01.


          To  confirm the model (equation 6), data supplied by the Western
Precipitation Corporation from tests on various venturi scrubbers in the
western United States were compared with the Widows Creek data (Fig 12).
Good agreement was obtained.

          In  sampling at the Colbert pilot plant, the particulate in the
gas from the  scrubber has been found to be mainly calcium salts picked up
in the scrubber  (see Part l).  The grain loading attributable to fly ash
(based mainly on petrographic examination of the total particulate and
estimation of the fly ash percentage) has been less than 0.02 gr/scfd.
Correlation with the Widows Creek model is shown in Table XVII.  Agreement
was fairly good, considering the observed values for fly ash were estimated.
                                    478

-------
 .13


 .12


 .11


 .10


 .09


 .08

r
 .07


 .06


 .05


 .04


 .03


 .02


 .01
                                                                      I     i
                       • PREDICTED BY MODEL:

                        Y (DUST IN  OUTLET GAS, GR /SCFD) • 0.12 X ~°'54
                           WHERE X - A P•L/G

                       * OBSERVED DUST IN OUTLET GAS


                       * STANDARD DEVIATION
                                                                •O
                                                                -o

                                                                X
                                                                                   I
             10
15   20   25
60   65  70   75   80  85   90  95  100
               30   35  40   45   50  55

                                  X


                                FIGURE 12


Correlation of Observed and Predicted Dust Removal Data from Widows Creek Tests
              Data from the Western Precipitation Corporation have been added (denoted by No. 1-12)

-------
                    TABLE XVII
Correlation with Widows Creek Model



APyRa Total L/G,b
in. H20
6.5
6.0
11*. 5
15.0
X
10.0
13.2
13.9
^ *
15.6
— ^x • v
7.2
11.1
7.1*
15.2
15.0
7-1*
11.3
7.0
7-0
15.0
7-2
11.6
15.1
gal/Mcf
56
38
28

1*6
36
66
58
68
1*6
1*6
58
36
1*8
1*6
26
38
1*6
56
1*6
68
L/G- Ap
361*
228
1*06
720
1*60
1*75-2
917.^
901*. 8
1*89.6
510.6
31*0.1*
881.6
51*0.0
355.2
519.8
182
266
690
1*03.2
533-6
1026.8
Fly ash in
outlet gas,
gr/scfd
Observed
0.0101*
0.0093
0.0063
0.001*0
0.001*7
0.0033
0.0037
0.0138
0.001*8
0.0031
0.0041
0.0059
0.0030
0.0027
0.001*2
0.0028
0.0032
0.0023
0.0027
0.0012
0.0016
Predicted^
0.0050
0.0063
0.001*7
0.0031*
0.001*1*
0.001*3
0.0030
0.0031
0.0042
0.001*1
0.0051
0.0031
0.001*0
0.0050
0.001*1
0.0072
0.0059
0.0035
0.001*7
0.001*1
0.0028
a Pressure drop in Ventri-Rod scrubber.
b Total slurry flow to venturi and spray tower.
C Based on Widows Creek model:
                  Y (outlet loading, gr/scfm) = 0.12 X'0-54

                     where X = L/G • AP
                           480

-------
SULPHUR DIOXIDE REMOVAL BY LIMESTONE SLURRY
             IN A SPRAY TOWER

                A. Saleem
                D. Harrison
                N.. Sekhar
     The Hydro-Electric Power  Commission
                of Ontario
            T or ont o,  Canada
               Presented At

    Second International Lime/Limestone
          Wet Scrubbing Symposium
             New Orleans, USA

           November 8 - 12, 1971
                    481

-------
              SULPHUR DIOXIDE REMOVAL
        BY LIMESTONE SLURRY IN A SPRAY TOWER

                     A. Saleem
                     D. Harrison
                     N. Sekhar

   The Hydro-Electric Power Commission of Ontario
                  Toronto, Canada
ABSTRACT

A MOOO-cfm pilot plant for removing sulphur dioxide
from power plant flue gases has been operated for one
year.  The sulphur dioxide was removed by contacting
flue gas from a coal-fired boiler with a limestone
slurry containing 12 per cent by weight of solids in
water in a spray tower.  Automatic control was used
to maintain the pH of the slurry between 5*6 and 6.0.
A spray tower was selected because of its simplicity
of operation, and because it has a minimum of solid
surfaces for scale accumulation.  The variables
studied included shell velocity, liquid to gas ratio,
pressure at the spray nozzles and the slurry reten-
tion time.  The main variables affecting the sulphur
dioxide removal efficiency were gas velocity and
liquid to gas ratio.  During a continuous 1000-hour
run, deposits that formed on the gas inlet duct and
on the demister were removed by water jets.  The
results indicate that a spray tower can be operated
with a high degree of reliability and that a sulphur
dioxide removal efficiency between 70 to 80 per cent
can be achieved using from 20 to 30 per cent above
the stoichiometric amount of limestone.
                       482

-------
INTRODUCTION
          Ontario Hydro, like most North American utilities,
is faced with the problem of finding ways to control pollutant
emissions from fossil-fuel power plants.  Considerable research
and development work is being directed toward finding a process
to control sulphur oxides emissions and the limestone slurry
process has been selected as the most promising for immediate
development.  This process is being studied in a ^OO-cfm
pilot plant which draws flue gases from a 300-MW coal~fired
boiler, after the electrostatic precipitator.  This paper des-
cribes the results of recent pilot plant studies.
          The sorption of sulphur dioxide by aqueous limestone
slurry constitutes a three-phase system, namely, gas, liquid
and solid.  When sulphur dioxide is transferred into the liquid
phase, hydrogen and sulphite ions are formed.  Calcium carbonate
is also transferred to the liquid phase where it is partially
ionized to form calcium and carbonate ions.  The calcium and
sulphite ions combine chemically to form sparingly soluble
calcium sulphite which crystallizes as a separate solid phase.
Oxygen in the flue gas is transferred into the liquid together
with the sulphur dioxide and partially oxidizes the sulphite
resulting in the formation of calcium sulphate, which also
crystallizes because of its low solubility.  In the presence
of a high concentration of sulphite ions in solution, the
sparingly soluble calcium sulphite is converted to the more
soluble calcium bisulphite.  The equilibrium between sulphite
and bisulphite ions is reversible and pH controlled.
                              483

-------
          The transfer of sulphur dioxide is believed to be
controlled by the slow rate at which calciuia carbonate dis-
solves.  To overcome this limitation, the slurry contacted
with a unit volume of gas must contain calciuia carbonate sub-
stantially in excess of the stoichiometric requirement.  With
a given per cent solids in the slurry, this can be accomplished
by having either a large liquid hold-up in the scrubber or by
having a large liquid to gas ratio.  The former approach is
less desirable because it not only increases the pressure drop
across the scrubber but can also lead to the formation of
highly supersaturated solutions of calcium sulphate and sul-
phite with the consequent formation of hard scale over the
scrubber surfaces.  It therefore appears more desirable to
achieve  the  desired rate of mass transfer by using a high
liquid to gas ratio and a spray tower is particularly suitable
for  this.
          Reliability is the key factor affected when a
 scrubber plugs due  to the deposition of solids.  Plugging by
 solids can occur in two ways, namely, by settling of solids in
 parts of the equipment where  the slurry is relatively stagnant,
or by  the crystallization of dissolved salts from supersatu-
rated solutions  on  internal surfaces.  Our experience indicates
 that deposit formation due to settled solids can be controlled
by irrigation of the affected surfaces with water.
          The crystallization of salts from supersaturated sol-
utions can be a  much more difficult problem to control.  Both
calcium sulphate and sulphite have strong tendencies to form
supersaturated solutions from which the salts crystallize in an
                             484

-------
unpredictable way.  If crystals form on the internal surfaces
of the scrubber, a very hard, cenacious film is formed which
tends to promote further crystallization.  To control this type
of scale formation, it is necessary to keep the concentrations
of dissolved calcium sulphite and sulphate as low as is practi-
cal which is probably at or near the saturation level.  In
practice, this can be accomplished by maintaining a very high
liquid to gas ratio.
          Scrubbers that are designed for high gas sorption
efficiency at low liquid rates, such as packed towers and
flooded bed and tray types, are therefore considered unsuitable
for use with the limestone slurry process, unless very high
liquid rates are maintained, in which case the main advantage
of these devices is lost.  In the spray tower where very high
liquid rates are needed to increase gas sorption efficiency,
there is very little possibility of forming highly supersatu-
rated solutions with attendant scaling problems.  This has been
verified in pilot plant testing where no hard scale formation
has occurred during a year of operation, which included a 1000-
hour continuous run.  On the other hand, hard scale formation
has frequently been observed in flooded bed and tray-type
scrubbers.
          When large liquid to gas ratios are used, coupled
with low sulphur dioxide concentrations, the equilibrium back
pressure of sulphur dioxide over the slurry surface is negli-
gible.  For this reason, a true countercurrent-type scrubber
is not essential and a single equilibrium stage is adequate to
obtain a high mass transfer rate.
                              485

-------
          A spray  tower offers  further advantages in handling
 large volumes  of flue  gas  such  as  those from power plants.   It
 has  the  highest turn down  ratio and  the lowest pressure drop
 of the alternatives  considered. Further,  it has  a minimum of
 internal surfaces  and  so reduces the possibility  of plugging
 even if  scale  formation does  take  place.   Although its  mass
 transfer characteristics are  not as  good as  those  of the other
 scrubbers,  it  is considered adequate in this respect.
 DESCRIPTION OF PILOT PLANT
          Figure 1 is  a diagram of the  pilot plant,  a detailed
 description of which is contained in Appendix I.   Four  thousand
 cfm  of flue gases  at 260 F from a coal-fired boiler  were drawn
 off  after the  electrostatic precipitator and passed  through a
 spray tower and then to the stack.   The spray tower  was  32  inches
 square,  16  feet high,  and  contained  6 banks  of sprays, each
 fitted with 9  full-cone spray nozzles.  The  slurry,  containing
 about 12 per cent  by weight of  solids, was recirculated  by  cen-
 trifugal pumps from  two interconnected  500-gallon  tanks  contain-
 ing  a total of about 900 gallons of  slurry.  A pH  sensing probe,
 located  near the pump  feed on the reclrculating tank, switched
 on the dry  limestone feeder and the  clarified liquor puinp so  that
 a  mixture of fresh limestone and clarified liquor from the
 settling tank  was added to the  slurry recirculating  tank each
 time  the pH fell below the set  point  of 5.8.  This fresh feed
caused an overflow of  spent slurry from the recirculating tank
to the settling tank.  Fresh make-up water was used  to wash the
demister and to flush  the  pump  seals.  There was no  liquid waste
stream.
                              486

-------
          The sulphur dioxide concentrations were measured by
an infrared analyzer.  The operating variables studied included
the gas velocity, liquid to gas ratio, nozzle pressure and the
slurry retention time.
RESULTS AND DISCUSSION
          The results are expressed in terms of the overall
mass transfer coefficient K  , the number of transfer units N
                           ga7                               og
and the overall efficiency of sulphur dioxide removal.  In the
calculations, it was assumed that the equilibrium back pressure
of sulphur dioxide over the slurry surface was negligible.  The
following formulae were used to calculate K   and N   values:
                                           ga      og
               Nog  =  ln Vy2
               Kga  =  (GVV) ln Vy2
where:  G* is the molar flow rate of flue gas
        V is the effective volume of the spray tower,
        y1 and y~ are the inlet and outlet concentrations of
                  sulphur dioxide.
Effect of Gas Velocity;  The performance of a single bank of
sprays was studied under various gas and liquid flow rates.
The liquid flow rate was varied by changing the nozzle pressure.
Figure 2 shows the effect of gas velocity on the mass transfer
coefficient K   under various liquid flow rates when using the
             ga
spray bank having nine 1/M—inch orifice nozzles.  It can be
seen from Figure 2 that K   increases with increasing gas velo-
                         ga
city.  The strong dependence of K   on gas velocity indicates a
                                 ga
                              487

-------
significant gas phase resistance to mass transfer.  Figure 3
shows similar data for another spray bank containing nine
1/2-inch spray nozzles.  Because of the larger orifices, more
slurry could be sprayed at a given nozzle pressure than with
the lA-inch nozzles.  The marked increase in K_Q with gas
                                               6<*
velocity can also be seen in Figure 3.  It is interesting to
note that when gas velocity is increased at constant liquid
flow rate, the liquid to gas ratio decreases, but in spite of
this decrease, the mass transfer coefficient shows an increase.
Increasing the gas velocity not only decreases the gas phase
resistance, but probably decreases the liquid film resistance
as well by inducing some surface regeneration of the liquid
droplets.
Effect of Nozzle Pressure and Size:  The effect of nozzle
pressure was studied by comparing the performance of a bank
with lA-inch orifice nozzles with that of a bank with
1/2-inch orifice nozzles.  The results, given in Table I, show
that under similar gas and liquid flow rates the mass transfer
coefficient K   is practically unchanged in the nozzle pressure
             ga
range of 10 to 20 psig.  Further, the two-fold change in orifice
size did not affect the mass transfer coefficient.
                             488

-------
                           TABLE I
            EFFECT OF NOZZLE SIZE AND PRESSURE OH
   MASS TRANSFER COEFFICIENT AT CONSTANT LIQUID FLOW RATES
Run
No
1
2
3
*f
5
Nozzle
Orifice
Size (in)
1/2
1A
1/2
1A
1/2
6 1A
Effect of Liauid
PN
(psig)
10
20
10
20
10
20
Flow
L
(IGPM)
^7.5
1*6.5
*f 6.0
W.5
*f 6.0
*f 6. 5
hate: The
VQ
(fDS)
^.79
^.79
7.19
7.19
9A7
8.97
liquid
L/G
( I gal/
1000 cu ft)
23.65
23.15
15.23
15. MO
11.98
12.3^
flow rate
V
(Ib moles/hr
cu ft atra)
2.09
2.06
2.26
2.30
2.75
2.83
at a given
gas velocity was varied by operating various spray banks simul-
taneously.  Figure If shows the effect of liquid flow rate on
mass transfer coefficient K _ at a constant gas velocity of
                           ga
8.6 ft/sec.  It can be seen from Figure *f that the liquid flow
rate has a marked effect on K  .  The increase in 1C   is due to
                             ga                    ga
increased gas-liquid interfacial area and the greater volume of
slurry present to dissolve and react with the gas.  Figure 5
shows the number of transfer units as a function liquid flow
rate.  Extrapolation of the linear part of the plot gives the
order of magnitude of the wetted wall effect.  Figure 6 shows
the overall efficiency of sulphur dioxide removal as a function
of liquid to gas ratio at a gas velocity of 8.6 ft/sec.
                              489

-------
Slurry Retention Time:  Retention time is defined as the time
the slurry is retained in the recirculation tank before being
returned to the scrubber.  Some retention time is necessary to
eliminate any supersaturation which may be present in the liquid
at the scrubber outlet.  The recirculation of supersaturated
solutions increases the probability of scaling.  The length of
retention time also affects the pH recovery of the slurry.  A
plot of typical pH recovery of a sample of slurry taken at the
scrubber outlet is shown in Figure 7.  The exact shape of this
curve is dependent upon the nature of the limestone, its parti-
cle size, the solids concentration in the slurry and the amount
of sulphur dioxide absorbed per unit volume of slurry.  Some
experiments were also conducted to study the effect of pH
recovery on the mass transfer rate.  The results are shown in
Table II.  A decrease in mass transfer coefficient with
decreasing retention time was noted, probably due to the
lowering of the pH of recirculating slurry.
                          TABLE II
       EFFECT OF RETENTION TIME ON MASS TRANSFER RATE
Run
1
2
£
6
7
8
Cmins )
17.2
8.6
|J'5
5.7
2.9
2*2
L
(IGPM)
*iO rt
M-Q f\
99.0
99.0
Hf6*5
196.5
196.5
(fSs)
7.19
7.19
7-19
7.19
7.19
7.19
7.19
7.19
K
Clb moles/hr cu ft atm)
1.^-2
1.20
1.98
1.75
2.98
2.59
3.72
3-18
                             490

-------
General Operating Experience
          Operation of the pilot plant was automatic except for
filling the limestone feeder hopper and discharging the spent
sludge from the settling tank.  The process was controlled by
monitoring the pH of the slurry.  The limestone feed was auto-
matically turned on when the pH in the slurry recirculation
tank dropped below 5.8 and was turned off when the pH was 5.8
or more.  The limestone consumption was from 1,2 to 1.3 times
the stoichiometric amount.  A typical composition for the
slurry is given in Table III.
          During the pilot plant operation, which included a
1000-hour continuous run, deposits accumulated in the demister
and in the flue gas inlet zone.  These deposits were soft and
were easily removed by water jets or mechanical scraping.  The
chemical compositions of the deposits are given in Table III.
The deposit in the demister zone contained mainly calcium sul-
phate along with calcium carbonate and fly ash.  The deposit in
the gas entry zone contained about equal amounts of fly ash and
calcium salts.  The scrubber shell was found to be remarkably
free of scale and deposits.  The spray banks were also free of
deposits except for the top two banks which were used infre-
quently.  Deposits usually occurred on surfaces which were not
adequately irrigated with slurry or water.
          The bulk settling rate of the spent slurry was
approximately three inches per hour and gravity settling gave
a sludge containing about 35-l+0 per cent solids by weight.
This could be dewatered to approximately 80 per cent solids by
                              491

-------
vacuum filtration.  The sludge could be discharged directly
into a settling pond but, if the sludge has to be transported
for disposal, some conditioning might be necessary to improve
its handling properties.  Even after vacuum filtration, the
sludge is serai-plastic and when jarred or vibrated forms a
uniform levelled mass.
          The average pressure drop through the spray tov/er
including the deraister was about 0.5 to 1.0 inch of water.
                           TABLE  III
              CHEMICAL COMPOSITION OF  SOLIDS  IN
              THE SLURRY AND THE  DEPOSIT  SAMPLES
 Typical
 Recirculatj
 Demister Deposit
 Gas Inlet Deposit
; Slurry
sit
lOSit
Calcium
Sulphate
30
7^
30
Calcium
Sulphite
Per Cent
56
—
if
Calcium
Carbonate
bv Weight
12
3
15
Balance
ash, etc

2
23
51
 COirJLUSIONS
 •  A spray tower is suitable for scrubbing large volumes of flue
   gases with limestone slurry.  From 70 to 80 per cent of the
   sulphur dioxide can be removed from the flue gases and,
   although considerable energy is needed for pumping the slurry,
   the gas phase pressure drop across the tower is very low.
 •  The mass transfer coefficient is a direct function of both
   gas velocity and the liquid flow rate.  The increase in mass
                               492

-------
  transfer coefficient with increasing gas velocity is of
  particular interest and may lead to reducing the size of the
  spray tower.
•  Deposit formation in the scrubber can be prevented by irri-
  gating surfaces with slurry or make-up water and by limiting
  the degree of supersaturation in the liquid phase.
•  The consumption of limestone can be kept within 1.2 to 1.3
  times the stoichiometric requirement by maintaining the pH
  of the recirculating slurry between 5.6 and 5.8.
                              493

-------
LIST OK SYMBOLS
a      -  gas-liquid interface, sq ft/cu ft of active
          scrubber volume
G      -  gas flow rate, cu ft/hr at 120 F
G*     -  molar gas flow rate, Ib moles/hr
K      -  mass transfer coefficient, Ib moles of S0p/hr cu ft
 E        of scrubber volume/atm of driving force
L/G    -  liquid to gas ratio, I gal/1000 cu ft of flue gas
          at 120 F
N      -  number of transfer units
Pfl     -  nozzle pressure of the slurry, psig
Rt     -  retention time, defined as the time the slurry is
          retained in the recirculation tank before being
          returned to the scrubber, minutes
VG     -  gas velocity through the scrubber shell, ft/sec
y« ,y2  -  concentration of sulphur dioxide in the gas at the
          scrubber inlet and outlet, ppm by volume
                            494

-------
                         APPENDIX I

                   PILOT PLANT DESCRIPTION

          Figure 1 is a diagram of the pilot plant.  Flue gas
was drawn from the main duct, after the electrostatic precipi-
tator, and through the scrubber shell by an induced draft fan
(F1).  The flue gas entered the scrubber at about 260 F.  After
scrubbing it was cooled to about 120 F and discharged to the
atmosphere from a 25-foot-high chimney.  The gas flow rate was
regulated with a damper (D1) and was monitored with a pitot
tube.  The 32-inch-square scrubber was constructed from 1/16-
inch-thick 316 stainless steel sheet.  The shell was assembled
in 7 flanged sections with a total height of 16 feet.  The
bottom section was conical and a chevron-type demister was
installed in the top section.  The slurry was sprayed into
the tower from 6 spray banks (S1-6).  The top M- banks sprayed
counter-currently to the gas flow and the bottom 2 sprayed
cocurrently.  Each bank was equipped with 9 full-cone spray
nozzles arranged to completely cover the tower cross section
with spray.  The top bank had 1/2-inch orifice nozzles and the
remaining banks had 1/V-inch orifice nozzles.  These nozzles
when operated under 10-20 psig pressure gave a spray angle of
about 80 degrees with an average drop size of 2500 microns.
The slurry flowed out of the tower through the bottom conical
section and a 6-inch discharge pipe which was immersed in the
slurry tank (T1) to provide a liquid seal.
                             495

-------
          The slurry, containing about 12 per cent solids,
was recirculated from 2 interconnected 500-gallon tanks  (T1 &
T2) containing about 900 gallons of slurry.  This capacity was
chosen to provide a minimum of 3 minutes delay time in the tank
for the recirculating slurry.  The slurry in the tanks, kept in
suspension by paddle-type agitators (A1 & A2), was recirculated
by three centrifugal pumps (P1-3).  The slurry flow through the
banks was regulated with diaphragm valves and metered through
orifice plates.
          Limestone was added to the recirculating slurry auto-
matically using a pH control system.  A pH sensing probe, placed
near the pump suction in tank (T2), operated a screw-type lime-
stone feeder (F1) and the clarified liquor recirculating pump
(1%).  The dry limestone feed was mixed with the clarified
liquor in a mixing cone (C1) and introduced near the bottom of
tank (T1).  Spent slurry was discharged from an overflow in
tank (T1) into the settling tank (T3).  The settled sludge
containing 35-l*O per cent solids was pumped into the station's
settling pond.
          The sulphur dioxide in the flue gas before and after
the scrubber was monitored with an infrared analyzer.  The
moisture content of the incoming gas was determined by wet and
dry bulb temperature readings.   The make-up water was used for
deraister washing and flushing pump seals.
                            496

-------
                           APPENDIX II
                            TABLE IV

             SUMMARY OF DATA OBTAINED WITH BANK NO
(NINE 1 A-INCH ORIFICE FULL-CONE SPRAY NOZZLES. COUNTERCURRENT
Ron
No_
1
2
3
it
5
6
7
8
9
10
11
12
PN
(psie)
25
25
25
20
20
20
15
15
15
10
10
10
L
(IGPM)
52.0
52.0
52.0
if6.5
W.5
»f6.5
M-1.5
M.5
M-1.5
33.5
33.5
33.5
VG
+7.7
38.7
36.5
Uo.8
32.7
29.95
V
(Ib moles/hr
cu ft atm)
2.20
2.5^
2.98
2.06
2.30
2.8M-
1.86
2.11
2.*f5
1.51
1.72
1.92
                              497

-------
                      APPENDIX II (Cont'd)
                             TABLE V

             SUMMARY OF DATA OBTAINED WITH BAM NO 6
(NINE 1/2-INCH ORIFICE FULL-CONE SPRAY NOZZLES. COUNTERGURRENT
Run
PN
L
JSfi_ (psie) (IGPM)
1
2
3
4
5
6
7
8
9
10
11
12

25
25
25
20
20
20
15
15
15
10
10
10

86.0
86.0
89.5
78.0
7^.5
78.0
60.0
61.5
64.0
47.5
46.0
46.0

T /r S00 Cone
v VG 2
VG (I eal/ (ppm)
(fps) 1000 cu
4.8
7.2
9.5
if. 8
7.2
9.5
4.8
7.2
9.5
if. 8
7-2
9.5

42.8
28.5
22.4
38.9
24.9
20.3
29.9
20.4
16.7
23.7
15.2
12.0

% so2
ft) In Gas Exit Gas Removed
1500
1600
1525
1500
1560
1620
1500
1560
1620
1450
1560
1600
TABLE VI
SUMMARY OP DATA OBTAINED OPERATING VARIOUS
Run
1
2
3
if
5
6
L
(IGPM)
50
100
136
186
232
282
VG
(fps)
8.6
8.6
8.6
8.6
8.6
8.6
VG
(I gal/
1000 cu ft) ]
13.9
27.7
37A
51.5
64.2
78.1






S02 Cone
(ppm)
In Gas Exit G^g
1020 600
1020 410
1020 340
1030 290
1030 210
1040 170
550
780
725
560
780
810
650
840
900
700
925
970

63.4
51.2
52.5
62.6
50.0
50.0
56.6
46.1
i.i. cr
fT. J
51.7
li^\ *7
39.3

V
(Ib moles/hr
cu ft atm)
2.87
3.10
4.09
2.83
2.99
3.80
2.40
2.66
3.23
2.09
2.26
2.75

BANKS SIMULTANEOUSLY
% so2
Removed

59.8
66.7
71.9
79.6
83.7
N
0.53
0.91
1.10
1.27
1.59
1.81
V
(Ib moles/hr
2.75
4.71
5.68
6.54
8.71
9.36
                              498

-------
<£>
FIGURE 1
ANT FOR FLUE GAS

-------
           3.5
         I-  3.0
         t
         8
         111
         O
         2
         3
            2.5
            2.0
            1.5
  PN          L
O 25PSIG 52  IGPM
A 20 PSIG 46.5 IGPM
o 15 PSIG 41.5 IGPM
•  10 PSIG 33.5 IGPM
                        678
                         VG-FPS
                          10
                       FIGURE 2

EFFECT OF GAS VELOCITY ON MASS TRANSFER COEFFICIENT
AT DIFFERENT LIQUID FLOW RATES AND NOZZLE PRESSURES
             (BANK S-4. COUNTERCURRENT)
             (ORIFICE DIA OF NOZZLE —1/4")
                      500
                                                   125098-RD

-------
        <
        u
        o:
        LJ
        o
        m
        j
           4.0
           3.5
3.0
2.5
           2.0
           1.5
                     PN
                        678
                          VG-FPS
                                 10
                         FIGURE 3
'EFFECT OF GAS VELOCITY ON MASS TRANSFER COEFFICIENT
AT DIFFERENT LIQUID FLOW RATES AND NOZZLE PRESSURES
              (BANK S-6, COUNTERCURRENT)
             (ORIFICE DIA OF NOZZLE = 15/32")
                      501
                                                        125103-RD

-------
                    VG = 8.6 FPS
                    PN ~ 15 PSIG
                               ISO       200
                            L - IGPM
250
300
                           FIGURE 4

EFFECT OF LIQUID FLOW RATE ON MASS TRANSFER COEFFICIENT
       AT CONSTANT GAS VELOCITY AND NOZZLE PRESSURE
                           502
                                                         125097-RD

-------
   1.5
   1.0
 O)
 o
   0.5
                     VG - 8.6 FPS
                         100                 200
                               L- IGPM
300
                             FIGURE 5

EFFECT OF LIQUID FLOW  RATE ON THE NUMBER OF TRANSFER UNITS
        AT CONSTANT GAS VELOCITY AND NOZZLE PRESSURE
                             503
                                                              125096-RD

-------
   75
3
UJ
a:
   50
   25
                 VG = £.6 FPS

                    ^ ISPSIG
                   25             50            75

                    L/G - I. GAL/1000 CU FT (I20°F)
100
                           FIGURES


EFFECT OF L/G ON THE EFFICIENCY OF SULPHUR DIOXIDE REMOVAL

      AT CONSTANT GAS VELOCITY AND NOZZLE PRESSURE
                             504
                                                           I25IOI-RD

-------
I
a
                            SOLIDS - 14.0% BY WT OF SLURRY


                            CaC03  - 1.68% BY WT OF SLURRY
                                 20            30


                            TIME- MINS
                             FIGURE 7


                  pH RECOVERY OF SPENT SLURRY
                              505
                                                         I25IOO-RD

-------
  THE MOHAVE/NAVAJO PILOT FACILITY

                FOR

       SULFUR DIOXIDE REMOVAL
                BY

    J.  L.  SHAPIRO AND W. L.  KUO

      Bechtel Corporation
      Vernon, California
           Presented At

         Second International

Lime/Limestone Wet Scrubbing Symposium

        New Orleans, Louisiana

         November 8-12, 1971




                507

-------
I.   INTRODUCTION




The Mohave/Navajo S02 Removal Research Program is a project for  testing




alkali absorption processes on a pilot-size basis.  The major thrust of




the work  is  to determine  the characteristics of these processes  as applied




to boilers  fired with low-sulfur coal.






The program was initiated in June 1970.  The owners of the Navajo Generat-




ing Station, which  is under construction near Page, Arizona, asked the




Bechtel Corporation (engineers-constructors of the Station) to survey




the status  of S02 removal technology for application to the Station.




Schedules call for  initial operation of each of three 770 Mw units in




1974,  1975,  1976, respectively.






Coal dedicated to Navajo will come from the Black Mesa mine.  Surveys




have indicated that over  the life of the plant, this coal will have




sulfur content varying from about 0.3% S to 0.8% S  with an average of




about 0.5%.  The heating value of this sub-bituminous coal is about




11,000 BTU/lb.







We need not, for this audience, describe the state of technology that




we observed  in the  Summer of 1970.  However, it is worthy of note that




there was considerable surprise evidenced in the industry that utilities




with such very low  sulfur coal were seriously considering sulfur dioxide




removal.  The concerns of the industry had been directed towards reducing




flue-gas concentrations of sulfur dioxide to levels approaching the
                                 508

-------
400 ppm that are the expected values for Navajo at the boiler outlet.




Therefore, relatively little was known concerning the performance that




could be achieved, the operational problems that might be expected




when attempting to reduce concentrations of SO,, from 400 ppm on down,




or the economic trade-offs involved.






Incidentally, events during the past year have shown that while Western




utilities have access to low-sulfur coals, they may still be required by




state emission regulations to apply flue-gas absorption equipment.






Another feature distinguishes the approach adopted for the Navajo Station




from many others.  High efficiency particulate removal equipment will be




located upstream of the S02 absorption equipment, so that the designs




and constraints of the system will be based almost completely on S0?




absorption.






These facts shaped the basic nature of the research program that we




describe here.






The characteristics of the Navajo Station are very similar to those of




the Mohave Generating Station.  The latter consists of two 770 Mw




units which also burn Black Mesa coal.  Both units of the Mohave Station




have been completed.  Therefore, the owners of the two stations have




joined in sponsoring this experimental program located at Mohave.




Southern California Edison Company was selected by the participating




owners as Project Manager of the program.
                                 509

-------
A  flexible, dual-loop  system was  designed  and  installed  at Mohave.



Construction was  completed  in  July,  1971.  As  it was desired  to test



as many  different types  of  absorbers as  possible in a  short time,



two identical  testing  loops were  built,  each with  individual  controls



and instrumentation.   In each  loop provision was made  for installation



of two absorbers  in parallel with ducting  that could pass the flue



gas  through either one.   At the present  time,  the  four absorbers



installed contain a single  stage  venturi,  a turbulent contact absorber,



a  combined Lurgi-impingement tray, and a low pressure-drop egg-crate



packing, respectively.





The Mohave units  have  electrostatic  precipitators  that are operating
                                <


at very  high  efficiency. Ducting inlets to the test loops were



placed both upstream and downstream  of the precipitators so that the



effects  of particulate loading can be tested.  As we shall show, the



process  loops  are designed  to  simulate on a small  scale, the  operations



that  a full-scale unit may  be  required to undergo.





Downstream of  each absorber is an electrically powered flue-gas reheater,



installed mainly  to enable  measurement of the demister efficiencies.




The present program plan provides for testing three absorbents -



limestone slurry,  lime slurry, and soda-ash solution.   Parametric



variations  of  L/G,  gas flow rate, and pH are made,  while measuring



SC>2 absorption efficiency and  particulate removal.   In addition



the possibility of  regenerating the  soda-ash absorbent (using lime)



is being investigated.
                                510

-------
II.  PILOT PLANT DESCRIPTION




The pilot plant facility is constructed on three skid-mounted platforms




as shown in Figures 1 and 2.  It accommodates the four previously men-




tioned scrubbers and the two identical loops which take a flue gas




sidestream of 500 acfm to 4000 acfm from the Mohave Station (Unit 1)




exit duct.  A 10' x 50' trailer equipped with control panels is at




one quadrant of the pilot plant.  The trailer also contains a wet chemistry




laboratory with chemical analysis equipments and an atomic absorption




spectrometer, permitting analyses on-site.






Figure 3 shows the flow diagram.  The sample flue gas is withdrawn, by




the suction generated by a 20" fan, from either upstream or downstream




of the electrostatic precipitator.  Special care was taken in designing




the sampling valve for the gas intake at the electrostatic precipitator




upstream line in an attempt to obtain representative fly ash concentra-




tion at all flow rates.  At the downstream of the precipitator, since




the fly ash distribution consists mainly of small particles, the gas




is withdrawn through a simple 14" collecting elbow parallel to the main




gas flow.  The gas flow rate up to 4000 acfm automatically is controlled




by an orifice flow controller through a butterfly valve in the gas line




to the blower.






Upon entering the gas system of the pilot plant, the flue gas is auto-




matically monitored for gas temperature, pressure at many key points,




and S02, NOX and Oo concentrations at locations in and out of the scrubber.




Dust loadings of the feed stream and the effluent stream of the scrubber




are periodically measured by manually collecting gas samples.
                                 511

-------
The flue gas from the scrubber passes through an electric preheater




which is controlled by a temperature controller.  Heat input of  the




reheater is accurately monitored so that by heat balance, the quantity




of entrained liquid can be calculated.  The gas is then discharged




through a short stack.






For the measurement of S(>2, NOX and 0£ concentrations, the on-line




monitoring  system has its own gas circuit shown in Figure 4.  The gas




is sampled  through a microporous dust filter in the collecting point




and a condensate trap prior to entering into the electrochemical cells.




Shutoff valves and air-back-purge are provided for cleaning the  filter




if necessary.  Electric time and solenoid valves are used to switch




alternately from one gas to the -other so that one set of analyzers can




be used to  monitor both feed and effluent streams.  Standardized gas




mixture is  used to calibrate the electrochemical cell periodically.






For the measurement of dust loading in the gas in and out of scrubber,




fiber glass filters for total mass and inpaction-type particle classi-




fiers for size distribution are employed.






The four types of absorbers to be tested are as follows:




     A.  Turbulent Contact (UOP)




     A 19-inch ID by 27-foot height absorber, with three trays of




     mobile packing (each tray containing 850 of pinch ping-pong




     balls) and a 24-inch-square demister section.  It is rated at




     1500 cfm and requires up to a 60-gpm solution circulation rate




     (stainless steel construction).
                                  512

-------
     B.  Venturi  (Chemico)




     A 30-inch ID Venturi  (6-1/16 inch to 7-1/8 inch  throat)  scrubber,




     with a 30-inch ID centrifugal separating chamber.   It is rated




     at 1500 cfm and requires 30- to 60-gpm solution  circulation




     (stainless steel construction).






     C.  Polygrid Packed  (Heil)




     A 24-inch ID by 15-foot-high Heil absorber, which will be tested




     with a 7-foot depth of polygrid (Ecodyne Corporation, formerly




     Fluor Products Company) "eggcrate" packing.  It  is  rated at 2000




     to 3000 cfm and required 25- to 50-gpm solution  circulation




     (fiberglass scrubber, polypropylene packing).






     D.  Lurgi Impingement (Peabody)




     A 30-inch ID by 14-foot-high scrubber, combining a  Lurgi Venturi




     and three Peabody impingement trays.  It is rated at 1500 to 2000




     cfm and requires 40- to 50-gpm solution circulation (stainless steel



     construction).






The liquor flow rate to each scrubber is controlled by a magnetic flow




recorder-controller in the 3" main PVC line from the hold tank.   A liquid




level controller is  provided in the hold tank to adjust  the tank volume




to allow sufficient  hold time for the supersaturated calcium sulfite




slurry from the scrubber to desupersaturate.   A hold tank of 3-15 minutes




can be achieved with the equipment sizes selected. 'To control solids




content and to maintain a constant alkalinity and pH control in the liquor




to the scrubber,  a continuous withdrawal of liquor is necessary for solids
                                 513

-------
 removal or for regeneration.  For this purpose,  a centrifuge is  provided




 for loop A while a settling tank with a  long winding  channel is  in use




 for loop B.






 The discharge stream  from each  loop is measured  and analyzed before




 respectively flowing  to the centrifuge or settling tank with the treated




 solution being returned continuously to maintain close solids and pH




 controls.  The solids in loop A are processed continuously during the




 test, while the solids in loop  B are accumulated in the bottom of the




 settling tank and measured before being sent to  the centrifuge at the




 end of the test.






 For the testing of both calcium and sodium absorbents in the removal of




 S02 from flue gases somewhat different process systems are required due




 to different solubility of these absorbents in water.   In the calcium




 system, the scrubbing liquor contains  fine particles  of absorbents and




 is circulated through the  scrubbers  as  a  slurry.   On  the other hand,




 the sodium absorbent  is  passed through  the system in  the dissolved




 state, with only a  small percentage  of  fly ash as solids.






 In the System I  mode  of  operation, as shown  in Figure  5,  the  S02  is




 absorbed from the flue gas by  either a  lime  or limestone  slurry that is




 circulated to the scrubbers  from the hold tank, fed from  one  of the two




 lime-limestone feed tanks.   The  rate of product withdrawal  is controlled




 by a specific gravity  controller in the hold tank which activates a




 valve  to  the  centrifuge.  The  liquid level in  the hold tank is adjusted




 by a level controller  connected  to the makeup water line.  This level,




 in combination with the  liquid circulation rate,  is used to maintain



hold time up to 10 minutes or more.





                                 514

-------
In the System II mode of operation, as shown in Figure 6, the SC>2 is




absorbed from the flue gas by a sodium solution circulated to the




scrubbers from the hold tank.  The rate of product withdrawal from




the hold tank is controlled by a pH recorder controller  located in the




liquid line from the hold tank.  The regeneration in the reaction tank




is controlled by a pH recorder-controller.  Liquid in the reaction tank




is pumped with level control into the centrifuge.  The liquid from




the centrifuge (which constitutes the regenerated sodium sulfite absorp-




tion solution) flows by pressure back into the hold tank.
                                  515

-------
III. ANALYSIS PROCEDURES




Test procedures of some of the more complicated data measurements  are




as follows:




     A.  Analyses of Gas Streams




         1.  SOo Concentration




         862 in the inlet and outlet gases is measured continuously by




         a Dynascience monitoring instrument that utilizes an electro-




         chemical transducing cell.  This unit is routinely calibrated




         with a standard gas mixture.   Periodically, the accuracy of




         SC-2 analyzer is checked either by the iodometric method or by




         the barium sulfate turbidimetric method.
         2.  NOX Concentration
             in the inlet and outlet gases is measured continuously by




         another Dynascience monitoring instrument.   This  unit is cali-




         brated routinely with a gas mixture, and its accuracy is checked




         by a modified phenol disulfonic acid method.






         3.  Dust Loading




         Particulates in the inlet and outlet gases  are measured  in dup-




         licate for each test   Isokinetic sampling  is used.   Particulate




         mass is composed of the solids (dried to constant weight at HOC)




         caught on the filter thimbles or filter  pads, including  particulates




         deposited in the filter probe and sample lines.   The  filter pads




         used in the  outlet  gas  have a reported efficiency of  99.7  percent



         for  0.3 particulates.
                                516

-------
    4.  Trace Metals




    Small amounts of metallic elements, such as mercury  and  chromium




    may be present in the inlet and outlet gases and are measured




    by a Varian Techtron Atomic Absorption Spectrophotometer.   These




    elements are measured in both the particulate  filtered from the




    gases and in the impinger liquids through which the  gases pass




    after flowing through the filters.






    5.  SO, Concentration




    A gravimetric method of determining 803 in the flue  gas  is  used.




    The method involves collecting gas samples in  two  fractions:




    the condensate and the isopropy alcohol scrubber.  The absorbed




    SO-J is precipitated as barium sulfate using barium chloride.






    6.  Fluorine Concentration




    Part of the fluorine contained in the coal, present  in the  flue




    gas, is condensed or absorbed in the traps of  the  803 apparatus.




    The contents in these traps is combined and analyzed by  a SPADNS--




    zirconium spectrophotometric method.






B.  Analyses of Liquor Streams




The pH is measured continuously by an industrial pH meter, which is




routinely calibrated with standard buffer solution; density  is  recorded




continuously by a differential pressure transmitter located  in  the




hold tank.  This unit is checked periodically by gravimetric methods.




Na, Ca, Cr, and other cations in the solutions are measured  by  the




atomic absorption unit.  SC>2 concentration (sulfite +  bisulfite) is




measured by iodometric titration.  Spectrophotometric  methods are used




for 804, NC-3, etc.  Sulfate determination is also  confirmed  by  barium




gravimetric method.



                              517

-------
IV.  TEST PLAN


The test plan is divided into four parts


     A.   Startup of pilot plant includes check-out of  facility  and


     equipment and calibration of instrumentation systems, both  manual


     and automatic controls.




     B.   Initial reference experiments using blowdown water  from  cooling


     towers  (8 experiments)




     C.   Parametric Experiments (228 experiments)




     D.   System Performance Tests (2 tests)
                                   •



The test plans are scheduled for testing for two levels of pH  (7,  9),


three  levels of  liquid  to gas ratio (3.3 to 100) and gas flow  rate


(500 acfm -  3000 acfm).  The range over which the gas,  liquid, and


L/G vary differs with each type of scrubber.  For each  scrubber,


maximum allowable ranges of gas, liquid, and L/G are tested.




The system performance  tests involve two extended runs, each  lasting


two weeks in duration,  for testing system performance from the viewpoints


of erosion,  corrosion,  and scaling.




The parametric experiments and system performance data  are to  provide


essential comparative data for each of  the absorbers and absorbents


tested.
                                 518

-------
In order to acquire meaningful experimental data for subsequent analysis,




operating variables and their responses throughout the scrubbing system




are measured, during each run.  The relevant experimental data collected




include:  (1) gas flow rate; temperature; absolute pressure; pressure




differential across the scrubber; S02, 803, and NO  concentrations;




and dust loadings of the feed and effluent gases; (2) absorbent liquid




circulation rate; pH; temperature; and chemical composition; (3) liquid




composition of feed and effluent of the regeneration tank;  (4) slurry




density in solids and desupersaturation time in the hold tank; (5) water




losses and water droplet entrainment; and (6) quantity and composition




of makeup water and absorbent feed.






The experimental data obtained from each scrubber are analyzed to assess




the validity of the assumptions adopted for the mathematic model describing




the behavior in each scrubber.  Further, scrubber performance data such




as S02 and particulates removal efficiencies and pressure drop across




the scrubber are correlated as a function of operating parameters such




as gas flow rate, L/G ratio, pH, etc., for each scrubber absorbent




combination.
                                519

-------
V.   CONCLUSION




The utilities sponsoring this research feel a sense of urgency which




has made  the program highly specific and compressed in time.  The




testing,  initiated in July, 1971, will be completed by March, 1972.






We believe that  the results will add significantly to the body of




information on alkaline wet absorption and will be of particular




value  to  those using low-sulfur fuel.
                                 520

-------
(71
[\5
                                          FIGURE 1
                            GENERAL PLANT ARRANGEMENT
                               S02 REMOVAL PILOT PLANT
                        EQUIPWIENTSKID
                        W/PLATFORWl ABOVE
     D
   480V POWER
      WATER
    110 LB.AIR
-B-M.4V-Q
                              4" ORA»N

-------
                                                                                    FIGURE 2
cn
ro
ro

-------
                                                FIGURE 3
                    GENERAL FLOW DIAGRAM SO2 REMOVAL PILOT PLANT
cn
ro
CO
      LEGEND

      FRC15
      FRC53
      AE 4. 5.6
      DP
                                             FLUE GAS
                                            PRECIPITATOR
                                            \\\\\\\\\\\V
                                                                               T-3
                                                                             MAKEUP
                                                                              TANK
                                SCRUBBER


                                ABSORBER
GAS FLOW RECORDER CONTROL
LIQUID FLOW RECORDER CONTROL
pH CONTROL ANALYSIS ELEMENT
DIFFERENTIAL PRESSURE
PRECIPITATE

-------
                                    FIGURE 4

              PILOT PLANT GAS MONITORING SYSTEM
      ABSORBER OUTLET GAS
                                                                   ABSORBER
                                                                   INLET GAS
PRESSURE
REGULATOR
AND FILTER
                      POROUS STAINLESS STEEL FILTERS
                                     /-40'LONG
                                    / HEATED 1/4"
                                       SS SAMPLE LINES-•
      MANUAL OR
      AUTOMATIC
        SWITCH
        VALVES
                CYLINDERS OF GAS
                FOR CALIBRATION
                        GLYCOL-H20
VACUUM
GAGE
A COOLING SYSTEM
                                         DIAPHRAGM
                                        VACUUM PUMP
                  CONDENSATE
                  AND EXCESS
                   GAS VENT
                                       MANUAL
                                       SELECTOR
                                       SWITCH
                                    524

-------
                                                       FIGURE 5
en
t>o
en
                                    SYSTEM I -CALCIUM  ABSORPTION
         170°F,750MMHG
            39 PPM S02
L	,
                    o
                  H>h
             B-1
         ID BLOWER
        2500 CFM, 30"
1 	 ^ 	
S-l 1
SCRUBBER
1 y\



— < —

                                                                                              Ca(OH)z,
                                                                                              0.075 Ib/min
                                                                                  r-H20,
                                                                                 1 1.8lbs/min
                                              PRECIPITATOR INLET GAS
                                                250°F,750MMHG.
                                                   390 PPM S02
                                                    1500 CFM
                                                 2.8 gr/cf FLY ASH
                                                    H20. 0.9 LB/MIN
                                                                                                  T-3
                                                                                               LIME TANK
                                                                                                200 GAL
                                                                                                0.2 LB/MIN CALCIUM
                                                                                                SULFITE + SULFATE
                                                                                                0.6 LB/MIN FLY ASH
                                                                                                  0.4 LB/MIN H20
                                              P-5
                                             PUMP
                                          100GPW1.20'
                                        P-1        T-1
                                        PUMP   HOLD TANK
                                     100GPM.70'  1200 GAL
                                                                                                F-1
                                                                                            CENTRIFUGE

-------
                                                         FIGURE 6
                                      SYSTEM  II-SODIUM ABSORPTION
         170°F,750MMHG.
            39 PPM S02
                           130°F
         i- 0.09 LB/MIN LIME

           I—- H20 3.5 Ib/min.
01
ro
o>
                                               - PRECIPITATOR OUTLET GAS
                                                     250°F,750MMHG.
                                                       390 PPM S02
                                                        2500 CFM
                                                          0.65 LB/MIN
                                                     ORIGINAL CHARGE
                                                        400 GAL H20
                                                       71.5 LB Na2 C03
                                                                  tl
                                                                             T-3
                                                                         LIME TANK
                                                                           200 gal.
                         H-1
                      HEATER
                       30KW
     B-1
 ID BLOWER
4000 CFM, 20"
riuu urm, &u
S4 *
CRUBBER
p— ^
HI
t=»^r
^



                 NOTE:
                 IF 5% OF S02 OXIDATION
                 IS ASSUMED AND 3000 PPM
                 S04 IN MAKEUP H20 IS
                 ASSUMED ONLY
                 0.012 Ib/min NazCOs IS REQ'D
                 0.118 Ib/min Ca(OH)2 IS REQ'D
            - 0.04 LB/MIN Na2C03
                                                   P-1
                                                  PUMP
                                               100 GPM, 70'
                                                      T-1
                                                   HOLD TANK
                                                    1200 GAL
   T-2       P-3
REACTION   PUMP
  TANK   5 GPM, 20'
 200 GAL
     CALCIUM SULFITE
        0.22 Ib/min.
      H20 0.22 Ib/min.
    F-1
CENTRIFUGE

-------
DETROIT EDISON PILOT PLANT AND FULL-SCALE DEVELOPMENT PROGRAM
     FOR ALKALI SCRUBBING SYSTEMS - A PROGRESS REPORT
                  J. H. McCarthy
                  Project Manager,  Air
                  Quality Control Projects
                  J. J. Roosen, Director
                  Environmental Studies
                  Division, Engineering
                  Research
                  The Detroit Edison Company
                  2000 Second Avenue
                  Detroit, Michigan  48226
        For presentation at the Second International
        Symposium on Lime and Limestone Wet Scrubbing
        Sponsored by Environmental Protection Agency,
        Office of Air Programs, Division of ControT
        Systems
                               527

-------
      Detroit Edison Pilot Plant and Full-Scale Development
         Program for Alkali Scrubbing Systems—
                    A Progress Report
     This paper includes the material which was originally to

be presented in two separate papers—4c, "Detroit Edison Pilot

Plant for Alkali Scrubbing Systems" and 7c, "Detroit Edison

Full-Scale Development Program for Alkali Scrubbing Systems."

The text of Paper No. 7c is not included in this proceedings.
                             528

-------
         DETROIT EDISON PILOT PLANT AND FULL-SCALE DEVELOPMENT PROGRAM
         	FOR ALKALI SCRUBBING SYSTEMS - A PROGRESS REPORT
                                   ABSTRACT

          In September, 1970, The Detroit Edison Company announced its
plans to construct a full-scale limestone scrubbing system for control of
sulfur oxides on its 280 megawatt River Rouge Unit No. 1.  At that time,
and of this date, there is no commercially proven system utilizing limestone
for scrubbing in operation.  Proven commercial sulfur oxide control systems
for the utility industry are ones that:  1) can be ordered from a manufacturer
as we do with a boiler, 2) can be installed in a coal burning plant with a
reasonable construction time and cost, 3) can be started up and operated
with a high degree of service reliability so that sulfur oxide is controlled
the majority of the time that the generating unit is running and, 4) incor-
porate a practical method of handling and disposing of large quantities of
waste materials.  The River Rouge project is considered a research and
development program.  The Detroit Edison Company developed a pilot plant
program which would provide for the detailed examination of the sensitivity
of a sulfur oxide removal system to variations in the different process
parameters.  This paper describes the basis for formulation of the pilot
plant program, how the pilot plant results must relate to the design and
selection of equipment for the full-scale installation, and preliminary
results of several investigations carried out in the pilot plant to date.

          The specific objective of the pilot plant program is to study,
evaluate, and demonstrate the system behavior with respect to:  1) level of
sulfur dioxide removal from the flue gas, 2) level of entrainment and outlet
particulates in the flue gas leaving the scrubbing system, 3) sensitivity of the
system to turn down in the flue gas rates, 4) short-term scaling and solids
deposition problems, 5) mechanical problems related to equipment and operation,
6) the effectiveness of lime, limestone, and other available alkalis.

                                 INTRODUCTION

          The Detroit Edison Company has announced plans to construct a

full-scale limestone scrubbing system for control of sulfur dioxide on its

280 megawatt, River Rouge Unit No. 1.  The steam generator of this unit has a

front-fired pulverized fuel furnace that is equipped with mechanical cyclone

separators followed by electrostatic precipitators.   When constructed

in 1956,, these devices represented the latest in particulate emission

control technology.  However, more restrictive air pollution codes

indicated the need for upgrading the Collection equipment.   Various
                                          529

-------
alternatives  for  upgrading  the  performance of equipment at  the  plant were

explored.  Among  the  alternatives  was the addition of a wet scrubbing, device

after  existing  equipment.

          Impending regulations on emission of sulfur  oxides indicated

that fly ash  control  systems  should not be considered  independent of sulfur

oxide  control,  even though  the  latter had not been  commercially proven.

Therefore, it was decided that  scrubbers  should be  installed since they

would  also provide for  S02  control.   Proven commercial  sulfur oxide  control

systems  for the utility industry are  ones that:  1) can be  ordered from a

manufacturer  as we do with  a  boiler,  2) can be installed in a coal burning

plant  with a  reasonable construction  time and cost, 3)  can  be started  up  and

operated with a high  degree of  service reliability  so  that  sulfur oxide is

controlled the  majority of  the  time that  the generating unit is running and,

4)  incorporate  a  practical  method  of  handling and disposing of large quantities

of  waste materials.   Examination of available information on sulfur  oxide

control  and work  that had been  done from  the 1930's to now, indicated  that

the most promising concept  for  add-on systems to existing units was  the

limestone wet scrubbing system.  The  major factors  favoring the use  of

limestone in  such a system, at  this time,  are the low cost  and wide

availability  of limestone,  apparent simplicity relative to  sulfur recovery

systems, and  minimum  potential  for water  pollution problems.

                         RIVER ROUGE  SCRUBBER PROJECT

          The River Rouge project  is  considered a research and development

program.  Because of  the lack of demonstrated performance of full-scale

systems and the variety of scrubber designs, it was decided that this

project would include a pilot scrubber installation as well as a full-scale

alkali scrubbing  system.  This parallel approach was needed for the  following

reasons.
                                       530

-------
          1.  Specific pilot plant data concerning performance of various




              scrubber designs and the process chemistry of these systems




              were not available.




          2.  System reliability in terms of equipment dependability and




              operating efficiencies were not known.




          3.  Criteria regarding the effectiveness of different limestones




              had not been adequately established.




          4.  Actual costs as quoted by equipment suppliers bidding to




              specifications were not available.




          5.  Physical and chemical characteristics of waste products,




              with regard to water pollution potential and solid waste




              disposal methods, required further identification.




          6.  Operating experience with this type of equipment had to




              be acquired.




          Bechtel Corporation is designing the full-scale limestone scrubbing




system for River Rouge Unit No. 1, and participating in Detroit Edison's




pilot scrubber program.




                              PILOT PLANT PROGRAM




          In order to accomplish the objectives outlined above, a pilot




scrubbing installation capable of handling up to 2500 ACFM of flue gas from




an operating power boiler was constructed at the River Rouge Plant.  The




pilot program was designed to examine the following:




          1.  Level of sulfur dioxide removal from flue gas.




          2.  Levels of entrainment and outlet particulrtes in flue gas




              leaving the scrubbing system.




          3.  The sensitivity of the system to turn down in flue gas rates




          4.  Short-term scaling and solids  deposition problems.







                                       531

-------
          5.  Mechanical  problems related to  equipment  and  operation.




          6.  The  relative effectiveness  of various  alkalis such ag




              limestone,  lime,  and others.




          In order to minimize  the number of  limestones that would  have to




 be evaluated in the pilot plant,  laboratory screening tests were carried out.




 The reactivity  of available stones was compared with that of reagent grade




 calcium carbonate from two different sources.  Results  indicated that  Presque




 Isle and Rogers City limestones were comparable with regard to  reactivity.




 Arrangements were made to have  a  supply of  Fresque Isle limestone ground to




 various degrees of fineness,  bagged, and  shipped to  the pilot plant.   The




 chemical and physical characteristics of  the  limestone  are  given in Table I.




 Included also are typical coal  and ash analyses for  the unit supplying flue




 gas to the  pilot plant.




          The physical arrangement provides the flexibility to  study several




 different types of scrubbers  with minimum system modifications.   The scrubber




 types to be evaluated were:  1) a series  of two Venturis, 2) a  venturi followed




 by a sieve  tray absorber  and, ^3)  a venturi  followed  by  a mobile bed absorber.




 Series of Two Venturis




          A schematic of  the  pilot scrubber system employing two Venturis




 is shown in Figure 1.   The piping system  was  arranged so that gas could be




 taken from  either  before  or after the precipitators  of  the  No.  3 Unit  at




 the River Rouge Plant.  After the two scrubbers, the gas flows  through a




 metering orifice and an induced draft fan to heaters which  reheat the  clean




 gas  to  approximately 200  F before it is emitted to the  atmosphere.  As  can




be  seen in Figure  1,  the  slurry system is the most complex  part of  the  pilot




plant.  The slurry is prepared  by mixing  limestone with  clarified liquid




from the sludge  thickener  in  a  slurry mixing tank.   Bagged  limestone in  various







                                          532

-------
degrees of  fineness are available.  From  the mixing  tank,  the  slurry of




alkali scrubbing material  is  transferred  to a  large  agitated holding tank




from which  it is metered to the  second-stage recycle tank.  Slurry  is




pumped from this tank to the  second-stage scrubber at a  rate to  satisfy  a




specific liquid to gas (L/G)  ratio.  A portion of the solution from the




second-stage recycle tank  is  bled to the  first-stage recycle tank.  Solution




from this tank is circulated  through the  first-stage venturi to  satisfy  a




specific L/G.  A portion of the  slurry from the first-stage recycle tank




is bled to  the sludge thickener.  Clarified liquid is pumped to  the slurry




mixing tank.  Underflow from  the thickener, spent slurry and fly ash, is




removed to  the plant ash handling system.




          A typical operation of this pilot plant arrangement  consisted




of:




          1.  Preparing a mixture of slurry in the large mixing  tank and




              filling the  recycle tanks to a predetermined level.




          2.  Establishing gas flow and monitoring inlet temperatures until




              a constant value is obtained.




          3.  Starting the recycle pumps at a high rate.




          4.  Monitoring ph of the slurry until it reaches a level  of




              approximately 6.0.




          5.  Adjusting slurry recycle to satisfy preselected L/Gs.




          Pressure readings,  temperature, and sulfur  dioxide concentrations




are taken at frequent intervals.  The instrumentation to make these measure-




ments is installed in a small laboratory that has been built adjacent to the




pilot plant.  Samples of limestone slurry are taken  from the slurry mixing




tank and analyzed for calcium carbonate content.   With these measurements,




the amount of limestone slurry needed to meet a specific stoichiometric







                                       533

-------
value  is  determined.  The  slurry  feed  and discharge rate are then  adjusted




as needed.   Similarly,  the L/G is adjusted to meet a predetermined value by




holding the  gas  rate  constant  and monitoring the  liquid flow using magnetic




flow meters.  The ph  of the  recirculating solutions is monitored in both




the  recycle  tanks to  control scaling in the scrubber and to determine  the




correlation  between ph  and SO- removal.  Suspended solids are measured at




the  inlet and outlet  of the  thickener  to determine its efficiency.




           The results of investigations conducted employing a system of




two  venturi  scrubbers indicate that outlet dust loadings of less than  0.02




grains per standard cubic  foot of gas  can be achieved.  This performance,




however,  has yet to be  adequately demonstrated on a full scale unit.   This




system has been operated for up to 120 hours without serious plugging  or




scaling problems.  During  this series  of tests, the system removed from




55 to  65  percent of inlet  S02, independent of the ratio of excess  limestone




to sulfur dioxide. The optimum removal efficiencies during these  investiga-




tions  were obtained with liquid to gas ratios between 30 and 40 gallons per




minute per 1000 cfm.  When the same system was operated with 20 percent excess




hydrated  lime, greater  than  90 percent of sulfur dioxide was removed.   These




results indicated that, in the case of a venturi scrubber, there is little




internal  circulation  within  the drops; consequently, relatively insoluble




alkalis such as  limestone  do not  readily dissolve and react with sulfur




dioxide which comes in  contact with them for short periods of time.  On the




other  hand,  the  large number of small droplets and large amounts of surface




area produced in these  units provide for excellent particulate removal.




Venturi Followed by Sieve  Tray Absorber




          A  schematic of the pilot system employing a venturi followed by




a sieve tray absorber is shown in Figure 2.   in this system, the alkali




scrubbing material  is transferred from the slurry mixing tank to the large





                                       534

-------
agitated holding tank from «hich it is metered to the tray absorber feed




tank.  The slurry is then metered to the absorber section of the scrubber.




The limestone slurry then cascades over several stages of trays and flows




back to a recycle tank.  Solution from the recycle tank is metered to the




venturi section of the scrubber.  A portion of the solution in the recycle




tank is pumped to the sludge thickener.  As with the previous system, the




clarified liquid is returned to the slurry mixing tank for preparation of




new scrubbing solution.  Sludge from the bottom of the thickener is removed




to the plant's ash handling equipment as with the previous system.




          To date, typical operation of this installation has consisted of:




          1.  Preparing a batch of scrubbing slurry in the mixing tank.




          2.  Filling the slurry holding tank, absorber feed tank, and




              recycle tank to a predetermined level.




          3.  Establishing a flow through the venturi section at a




              predetermined L/G.




          4.  Establishing gas flow through the unit, and immediately




              thereafter, a flow of scrubbing liquid to the absorber




              section.




          5.  Adjusting the rate of flow to the absorber section at a




              predetermined L/G.




          Limited experience with this system to date indicates that outlet




dust loadings of less than 0.015 grains per standard cubic foot of gas can




be achieved.  It should be noted, however, that these results represent the




performance of a l/200th scale of the River Rouge unit.  Difficulties have




been encountered to date in establishing a ph level high enough to control




scaling due to precipitation of dissolved solids in the absorber section of




the  scrubber  system.  In order to provide more precise ph control, modifica-




tions  are presently being made to the recycle and alkali feed systems of the




pilot  plant.  As of this writing, significant sulfur oxide removal data have




not been obtained.                       535

-------
          Operating experience to date has indicated that it is extremely




important to have an on-site capability  for performing chemical analyses.




This has been  found necessary because certain parameters must be monitored




and the results  of that monitoring  fed back into the operation of the




scrubber in order to examine such characteristics as scaling and deposition




potential.  This is particularly true in establishing the degree of  super-




saturation  in  the supernatant  from  the sludge thickener.




          For  each  test  series,  the pilot plant is manned for 24-hour  per




day operation.  On  each  12-hour  shift, a crew of three men is employed.




This  crew  is composed  of an operator from the power plant, a chemical




engineer,  and  an engineering technician  who has been trained in chemical




analysis.   Recently,  an  analytical  chemist has been added to the staff during




the day time part of  the operation.




Future Plans




          Beyond the  present test series, it is planned to study the effective-




ness  of other  alkalis  including  certain  chemical process waste products available




 locally.   Of equal  importance  to the effectiveness of the systems in removing




 S02 are solutions to  the waste disposal  problems.  Therefore, a concurrent




program, utilizing  waste materials  generated by the pilot plant, is underway.




This  effort includes  the development of  practical means of handling, trans-




porting, and ultimately  disposing of the sludge.




          As mentioned earlier,  engineering of a full-scale limestone




scrubbing system for River  Rouge No. 1 is proceeding in parallel with  the




pilot  program.   The ultimate configuration of this unit and its operational




schedule is dependent  upon  the outcome of the present pilot program.   The




current schedule for completion  of  the full-scale scrubbing system is




January 1973.
                                       536

-------
                                  TABLE I

       CHEMICAL AND PHYSICAL PROPERTIES OF LIMESTONE,  COAL,  AND ASH
                   DETROIT EDISON PILOT SCRUBBER PROGRAM
Presque Isle
Percent
by Weight
Silica
Si02
0.60
Limestone
Iron Oxide Calcium
Fe203 Oxide, CaO
0.35 54.5
Magnesium Oxide
MgO
0.60
Sulfur
(S03)
0.085
Loss on Ignition
43.82
             Calcium Carbonate, percent   97.5
             Magnesium Carbonate,  percent  2.5
Coal


01
-1


Percent
by Weight
Volatile
Matter

36




Fusion




Percent
by Weight


Loss on
Ignition

3.4
Ash

Silica
Si02

40

Iron
Moisture
As Fired

6.3
temp (fluid)


Oxide Alumina
Fe70

30
3 A1203

17
Total
Ash

12
- 2350


Sulfur Carbon

2.8
F Heating


Calcium Magnesia
Ca 0

2.6
MgO

0.6

72
Value -


Hydrogen

4.8
12000 BTU/lb


Titanium Sodium
Ti02

0.9
N320

0.3
Nitrogen

1.7
•


Potassium
K20

1.8
Oxygen

5.5



Carbon
C

2.9






Sulfur
S03

1.1

-------
GAS INLET
BEFORE
PRECIPITATOR
                                                                     EXHAUST
                 FIRST
               VENTURI
SECOND
VENTURI
 GAS INLET
 AFTER
 PRECIPITATOR
01
to
00
FIRST
STAGE
RECYCLE
TANK
 LIME OR LIMESTONE

 WATER       	
                                            SECOND STAGE
                                            RECYCLE TANK
   SLURRY
   HOLDING
   TANK
                        SLURRY
                        MIXING
                        TANK
                         FIGURE I
            DETROIT EDISON PILOT SCRUBBER
            NlSAJLTtNVBLVa. VIA- <
                                                       SLUDGE
                                                       THICKENER
                                                                  PUMP
                             •ASH
                             HANDLING
                             SYSTEM

-------
                                                       HEATER
                                                                         EXHAUST
 GAS INLET
 BEFORE
 PRECIPITATOR
                                                                              DRAIN
                                        SIEVE
                                        TRAY
                                      ABSORBER
                                                                   SLUDGE
                                                                   THICKENER
              VENTURI
 GAS INLET
 AFTER
 PRECIPITATOR
            ABSORBER
            FEED TANK
 LIME OR
 LIMESTONE
                               RECYCLE\/
                               TANK     I
                        SLURRY
                        MIXING
                        TANK
                                                   ASH
                                                   SETTLING
                                                   BASIN
SLURRY
HOLDING
TANK
            FIGURE 2
DETROIT EDISON PILOT  SCRUBBER
                       « tm
                           stifi/f TO AW An Ort DO CD l?nHICIAI IDATiriM

-------
      RESEARCH & DEVELOPMENT  IN WET SCRUBBER  SYSTEMS
                      A.  L.  PLUMLEY

                      M.  R.  GOGINENI

                 C-E Combustion Division
                  Windsor,  Connecticut
                       Presented at

SECOND INTERNATIONAL LIME/LIMESTONE WET SCRUBBING SYMPOSIUM


                  November 8-12, 1971

     Sheraton-Charles Hotel, New Orleans, Louisiana
                            541

-------
         RESEARCH & DEVELOPMENT IN WET SCRUBBER SYSTEMS

                         A. L. Plumley

                              and

                         M. R. Gogineni
The subject of this paper  is the research and development sup-
porting Combustion Engineering's product line of air pollution
control and by-product recovery systems to meet Federal standards
for air and water pollution.  Major  emphasis of this program
centers on providing  design criteria and engineering support
for commercial systems.

R&PD PROGRAM  OBJECTIVES

The following have been  established  as both short and long range
objectives of C-E Research and Development Program for Air Pol-
lution Control Systems:

A.  Complete  development of limestone furnace injection
    demonstration systems  to meet  contract obligations for
    SO  and particulate  removal.

B.  Establish design  criteria for  future systems at lower
    cost  using the present APCS principles.

C.  Develop non-furnace  injection  systems for both utility
    and industrial boiler  applications.

D.  Develop improved  air pollution control systems to reduce
    other air pollutants and to allow expansion of a company
    product line.

E.  Develop by-product utilization processes to avoid poten-
    tial  water and land  contamination and to offset operating
    costs of  current  and future air  pollution systems.

The R&D program  is divided into two  principle areas of effort:

1.  Air Pollution Control  Systems  Development

2.  Chemical  Process  Research

AIR POLLUTION CONTROL SYSTEMS DEVELOPMENT

Air Pollution Control Systems Development is comprised of performance
work on full  scale systems in the  field and experimental work done
on our test facilities in  the laboratory.  Design criteria are con-
firmed in this area of development.
                                  542

-------
The first area of development involves the APCS contracts which
C-E has obtained.  These contracts are listed and described in
another paper at this conference, entitled "Contract Applications
of the C-E Air Pollution Control Systems".

The second area of Air Pollution Control Systems Development is
comprised of all the experimental work done in the laboratory on
our scrubber test facilities.  There are two operating APCS test
facilities available in the Kreisinger Development Laboratory
for evaluation of wet scrubber technology:  A 12,000 CFM Proto-
type system and a 1,200 Pilot system (Figures 1 to U).

These APCS test facilities are comprised of all system components
utilized by the existing C-E installations and are assembled in a
modular arrangement to allow rapid system modifications.  In this
way, a number of flow arrangements can be studied.

CHEMICAL PROCESS RESEARCH

The Chemical Process Research section is concerned with two major
areas.  The first is the fundamental understanding of reaction
mechanisms associated with our wet scrubber work and the second
is the disposal of waste materials from APCS operations.  The more
fundamental work is carried out in various types of laboratory
bench reactors and is designed primarily to provide supplementary
information to the larger scale test facilities and to screen
variables to be evaluated during pilot and prototype tests.

As the overall program develops, co-ordination of various technolo-
gies related to environmental control, both solid and liquid wastes,
as well as air pollutants, will be carried out.

TEST WORK

Test work is currently underway which is directed at providing
sufficient design information to resolve problems remaining on
the demonstration field installations thus allowing them to meet
contract guarantees as well as federal and local regulations on
S00 and particulate removal.  Field development is expensive and
time consuming since generation of power takes precedence over
the check-out of field modifications and other work necessary in
evaluation of new processes and new process development.  We have
experienced considerable time, months of delay, between modifi-
cations and their evaluation in these field systems.  As a result,
most of our development work is being carried out on the two
operating APCS test facilities in the laboratory.

Prototype (Figures 1 and 2)

The APCS prototype is approximately 1/U the linear scale of C-E's
field units and utilizes a 25 sq. ft. marble bed  contactor to  scrub
the flue gases from the oil fired packaged boiler.  This boiler  has
a flue gas output of 12,500 to 15,000 ACFM measured at  lU.J psia in

                                 543

-------
125°F.  The inlet gas temperature to the scrubber is variable; the
600°F flue gas from the boiler can be reduced to 250°F by means of
a  finned tube heat extractor prior to the scrubber.

Major components of the system are the scrubber, heat extractor,
reheater, demister, clarifier or retention tank, and the slurry
or hold tank vith provisions for additive and fly ash injection.
The piping and pumps can be readily modified to allow study of
several different processing arrangements.

Pilot  (Figures 3 and 1*)

The APCS pilot system features multiple three sq. ft. marble beds,
as well as 1  sq. ft. rod type contacting surfaces which may be
operated in various combinations to determine the most effective
use of additives for SO  removal from flue gas.  Provision has
been made  for evaluation of other types of scrubbing surfaces.
Separate collection of particulate and SO  is also possible in
this system.   The major system components in addition to the
scrubber with its multiple beds are the demister, clarifier,
slurry and delay tanks additive and fly ash injection systems.
Here,  too, piping and pumps can be modified to allow several
different  processing arrangements.  The system is currently being
used to carry out preliminary screening of variables so that only
optimum modes will be demonstrated on the prototype proper.

Bench  Scale Facilities

There  are  three principal bench scale facilities that are utilized
in the APCS studies.  In general, the purpose is not to perfectly
scale  down APCS either from field or laboratory, but to provide
small  flexible laboratory tools in which to study various phenomenon
including  packed-bed mass transfer rates, scaling phenomena and rates
of reaction to provide a basis for correcting certain problem  areas
observed in operation of either the larger scale laboratory or field
facility.  The first of these facilities, the bench scale scrubber,
is a tube  which has been flanged and divided into three sections  (See
Figures  5  and 6).  The lower section contains a gas inlet in the
liquid drain, the middle section contains the bed, the nozzle and
the bed drain, and the upper section contains the mist eliminating
section and the gas outlet.  These sections are removable for modi-
fications  and repairs.  The rest of the apparatus is concerned with
controlling concentration, temperature and flowrates of reactants.
Instrumentation includes temperature controllers, SO  monitors, tem-
perature monitors, and the pH meters.

The next type of facilities are stirred reactors.  We have a CFSTR
or Continuous Flow Stirred Tank Reactor.  This equipment is used to
understand scrubber reaction kinetics by direct determination of dif-
ferential rate data for the various reactions taking place in the
scrubber.  This is important in design of system controls as well as
in sizing of  system components.  In this reactor the contents are
well stirred  and uniform in composition throughout.  Thus, the exit
stream from this reactor has the same composition as the fluid within


                                  544

-------
 the reactor and analytical results  on this  stream  represent the
 behavior of the system.

 The second type of stirred reactors  are batch  type reactors.
 These consist of a pair  of temperature controlled  flasks.  System
 reactions, specifically  those  relating to potential deposit for-
 mation,  and for preparation of scrubbing solutions, delav tanks,
 holding  tanks, etc.  are  being  studied.  (Figure 7)

 The holding effects  of ponds to allow for completion of reactions
 including precipitation  of solids and dissolution  of potential
 reacting materials which are being recirculated in the overall
 system are being simulated in  bench  tanks.

 CHEMICAL PROCESS DEVELOPMENT

 WASTE DISPOSAL

 A  comprehensive program  for utilization of APCS by-products is
 currently underway.  This  six  phase  program covers:

 1.   Nationwide environmental survey  of present waste
     disposal  procedures  and ecological data relating
     to the needs of  potential  APCS customers.

 2.   Determination of physical  and chemical properties
     of waste  products.

 3-   Studies on direct  disposal of both solid and liquid
     wastes.

 U.   Beneficiation and  utilization of liquid and solid
     waste products.

 5-   Pilot plant studies  of promising utilization of
     procedures.

 6.   Economic  and technical  summary and recommendations.

 A significant  part of  the  understanding of physical  and chemical
 properties  is being  carried out with the assistance  of C-E's
 modern X-ray  diffraction and X-ray spectrographic equipment and
 recently  acquired thermal  analysis equipment.  This  instrument
 (Figure 8)  is  capable  of performing  simultaneously  the three
 most widely used thermoanalytical techniques:  Differential Thermal
 Analysis  (DTA),  Thermal Gravimetric  Analysis (TGA)  and Differential
 Thermogravimetric Analysis  (DTGA).

 The  instrument has been utilized so  far in determination of melting
 point, decomposition temperature and  sinterability  of waste sludge
 from APCS operations (Figure 9).  These are all properties needed
 for  determination of appropriate utilization applications of the
material.
                               545

-------
Characterization

Work is underway on development of a sludge "library" by cataloging
of ten different chemical and physical properties of sludges from
various APCS operations in both field and laboratory studies.  This
information is essential since the utilization or disposal process
in each instance will depend on both the additive used and the mode
of APCS operation.  (Figure 9).

A number of areas for utilization of modified and unmodified fly ash
are shown in Figures 10 and 11.  These schemes range from production
of light weight aggregate developed some years ago to compression
and sintering of drain pipes which is under laboratory development.
Large volume usage will be necessary since 1-2 tons/day/MW of sludge
would be produced from a coal-fired boiler firing 3% S coal and
ash with an APCS for emission control.

Beneficiation

A number of small brick and pipes have been successfully produced
from APCS sludge by Industrial Material Technology, Inc.  Samples
formed by high pressure without sintering appeared most promising.
Efforts are continuing to improve durability and optimize fabrica
tion methods for large quantity production.

ACRES, Inc. has submitted a proposal to define the geotechnical
properties of the APCS sludge.  Properties involved are important
in designing structural fill - dams, dikes, pavement and other
load bearing land-fill usage.  ACRES, Inc. has been retained by
Hiagra Mohawk to handle ash disposal problems.

The evaluation of APCS sludge for hydrothermal bricks is contin-
uing at Michigan Technology University.

Environmental Survey

A report summarizing the results of an environmental survey has
been completed.  Compilation of current Federal, State and Local
pollution codes is continuing.

Direct Disposal

A meeting of the National Ash Association in Washington discussed
plans for transportation of fly ash by rail from the utility to
the disposal site.  It was decided to also include modified fly
ash (from lime/limestone wet scrubbing).  Representatives of
leading coal producers were present to consider use of abondoned
mines as disposal sites.  No firm decisions were reached.

The EPA Office of Solid Wastes is continuing to evaluate an
unsolicited C-E proposal for study of direct disposal procedures
for APCS sludge.
                                   546

-------
BENCH STUDIES


An extensive laboratory investigation designed to reduce the liquid
blow-down required to control calcium sulfate (gypsum)  deposit for-
mation during APCS operation is underway at KDL.  This work is essen-
tial to avoid replacing an air pollution problem with water pollution,

The three-point bench s+ -1y is directed at development of a "closed
cycle" system in which th^ only liquid discharge from the APCS is
that conveying solids which can be recycled, thus requiring no blow-
down to the river.

The most promising technique resulting from the bench study will be
followed by a demonstration run on the 1200 cfm APCS pilot facility.

Other bench scale support  studies are underway  to provide aid in
understanding of scrubber  chemistry  and mass transfer within the
APCS  and similar SO   removal system.  Included  are:

A.  Reliability of  extrapolation of  bench  units  to larger
     systems.
B.   Confirmation  of the marble bed  as a  completely back
     mixed  system  as  part of fundamental  studies on mass
     transfer within the bed.
 C.   Basic  studies  on oxidation of  sulfites and the  sub-
     sequent relationship to both SO  absorption in  the
     scrubber and  control of both  sulfite and sulfate
     deposit formation.
 D.   Use of a Continuous Stirred Tank Reactor to aid in
     equipment sizing by determination of rate controlling
     steps in various phases of the APCS operation.   Dis-
     solution of alkali (CaO and CaCO )  as well as preci-
     pitation rate of sulfates and sulfites are under study.

 TEST RESULTS

 Experimental work since 1970 dealt with specific problems in the
 field demonstration  system.  Control drain line scaling and opti-
 mization of slurry recycle (both clarifier and  scrubber underflow)
 have been studied.

 The under bed addition of slurry and dry  injection of  additive  in
 Se SB stream have  given comparable results  in both SO   removal
 and  chemical efficiency of additive.  Both procedures  resuit in
 passage of both additive  and  SO  through  the bed and it may be
 conceded that the  intimate mix?ng  thus  achieved is  required for
 maximum efficiency of operation.

 SO,  removal of 90  - 95 percent have been achieved  vith both dry
 injection  and  slurry addition of  hydrated dolomite   About 10,
 more hydrated  lime is required than hydrated dolomite to give
  comparable  SO  removal.
                                   547

-------
Experimental work in 1970 on the APCS prototype test facility
emphasized optimization of slurry additive or tail-end systems
utilizing calcium carbonate as the additive.

In a series of single marble bed tests at 1000 ppm SO , 50 to
300$ stoichiometric limestone slurry was fed through underbed
spray nozzles with no recycle.  SO  removal ranged from UO to
75$.  The formation of calcium sulrate (CaSO.) deposits in the
scrubber, is being controlled by blowdown tests.

Experimental work in 1971 on the Air Pollution Control System
(APCS) test facilities has emphasized optimization of tail-end
slurry additive systems which can-be installed in Unit No. 1
at the Meramec station of Union Electric and Unit 6 at the
Paddy's Run Station of Louisville Gas & Electric.  Both single
and dual marble bed studies have covered use of both calcium
carbonate  (for Union Electric) and calcium hydroxide (for
Louisville) with inlet SO  concentrations of 1000 and 2000 ppm.

Following  a series of successful pilot and prototype tests
extended runs to demonstrate reliability were carried out with
dual bed operation and recycle using lime in a scale up of
previous pilot studies.

During an  80 hour run with an appropriate recycle of scrubber
effluent sulfur removal in excess of 80$ was maintained without
scaling.   Dust loadings at the outlet were about 0.03 gr/SCF.

Louisville Gas & Electric approved construction of a full scale
APCS at Paddy's Run Station on the basis of these successful tests.

An  extended run to demonstrate reliability was carried out utilizing
dual bed operation with maximum recycle of scrubber effluent through
the slurry tank using calcium carbonate.  This scale up of previous
studies extended for 125 hours without deposit formation and resulted
in  sulfur  removal efficiency in excess of 75$-

Laboratory and pilot studies are continuing to understand the overall
chemical balance as related to blowdown and waste disposal.  Approval
of full scale APCS construction is expected in the near future.

EPA CONTRACT FOR RESEARCH IN SCRUBBER TECHNOLOGY

C-E has received its first prime contract from the Office of Air
Programs - Environmental Protection Agency (EPA).  This $250,000
contract is for the "Optimization of a Lime/Limes tone Wet Scrubbing
Process for SO  and Particulate Removal in a Marble Bed Scrubber.

A general outline of the program is as follows:
                                  548

-------
Impending national performance standards and standards "being set
by state and local air pollution control office authorities re-
quires the rapid development of as many SO  emission control
processes as possible.  Several processes are now being developed,
including the Combustion Engineering (C-E) Lime/Limestone Wet
Scrubbing process, which .has been under development for about
three years.  Two full scale units of this process have been
installed on separate 125 MW power plants.

The purpose of this contract is to conduct research and development
on small pilot scale, large pilot scale and full plant versions of
the c-E SO  scrubbing process in order to accelerate its commercial
development.  Principle capabilities to be developed and demonstrated
include adequate SO  and particulate removal, proven operational
reliability, adequate equipment lifetime, adequate capability for
scale-up, and sufficient estimates of process costs.

The following specific work steps are anticipated:

1.  Engineering Analysis - Existing C-E data and all related
    data from other EPA contractors to develop the final test
    program and calibrate mathematical process model.

2.  Soluble System Studies - KDL Prototype determination of
    the approach to vapor-liquid equilibrium and the mass
    transfer capabilities of marble bed scrubber and valida-
    tion of input to mathematical process model using  Na CO
    solution.

3.  Prototype Test Program - KDL - To provide  optimization of
    the present C-E APCS based  on input  from 1, 2,  and k.

U.  Prototype tests -  KDL - Verification  of prototype as model
    for prediction of  field unit operation.

5.  Prototype Tests -  KDL - Final calibration  to  determine
    optimum operation  for field units.

6.  Field  tests  - to demonstrate on the field units, under a
     separate  contract, the optimum operating conditions deter-
    mined from the KDL Prototype tests and to arrange for the
     final confirmation of mathematical process r.odel.
                                    549

-------
    ft
    c
    X
    m
in
en  —
o

-------
O
c
ro
                    TO STACK
                       A
     MAKE-UP
        REHEATER
        DEM!ST€R
        MARBLE/
         BEDS \.
                >^
        FLUE GAS
         ""r^/^*" *
         r KUUI
       V.P. BOILER
SLURRY
 TANK
                                             LIQUID
                                            SLOWDOWN
      TO
    VACUUM
    FILTER
                  A PCS  PROTOTYPE-TAIL-END  SYSTEM

-------
   o
   c
   ^o
   m
' 11
< n
i
    I
   CM

-------
TO STACK
J A  I
                                         DEMISTER
                VENTRI
                SECTION
CJl
01
CO
  T,
  o
  c
  ;a
  iM
              GARBLE
               BEDS
                                 V V V V

                                 DRAIN
                                  TAWK
                                  •>3
                                       LJ
         0-*
                                V
                               TO
                             SEWER
                                                               WELL
                                                               V-'ATER
                                                                 V,
                                                                 ftrt'tt
                                                                [R.1
                                                                 I
                                  V V V
                                 MIX
                                TANK
                                                           IFfJ
                                                                 V
                                                   CLARIFIER
                                       TO
                                    DISPOSAL
                                 ARCS PILOT PLANT
CYCLONE
FEEDER

                                           «
                                           I
                                          SI
                                          K
          v
          '.N
                                                   i.
                                                   t
                                                  TO
                                                 SE17ER

-------
I
30
ni



A,

-------
01 rn
01
tn
  cn
                    TO STACK
                               -M
              GAS OUTLET SAMPLE
                      SPRAY  NOZZLE
             ROTOMETER


LIQUID INLET TEMP.-
    ADDITIVE
                -^HXH-1
               IMMERSION

                HEATER
                                       ^DEIVISSTER
                                      B

                                    m GARBLE BED



                                       -DOWNCOMER
                                           13
r
                                               GAS IWLET TE^'P.
                                                             ?
                                             FURNACES

                                      GAS INLET SAMPLE
      JRNACES

      CAfl'IDI ST	1
j
                         ROTOMETERS
                                                   LIQUID OUTLET SAMPLE, pH



                                                   -:>TO DRAIN
              SCHEMATIC  OF BENCH  SCALE  SCRUBBER  SYSTEIV1

-------
    o
    c
    a
    m
en
en
o>

-------
     c>
     c
     -43
tjl

-------
 o
 c
 33
 m
n
       •
       r
       u.
       ;
       3
       - >
       c
       :
       ••:
2
„ , , V tjj
0
O OK
Z 2?
o o x
P H-
-0
x-J
L'J
-------
C
P5
01
              WATER TO
               CLARIFIER
            WET
                   VACUUM FILTER
^=>q
   !
                                      PELLETIZER
CEMENT K
ILN
                                      DRYING KILN
                                          u
                                          V
                                        MIXER
                                                        AUTOCLAVE
                                                 =CH
                          SINTERING
                           MACHINE
                                                         
  LIGHT
 WEIGHT
AGGREGATE
                                                                      SULFURIC ACID
                                                                       OR SULFUR
                                       ^C> MINERAL WOOL
                                        o
       ADDITIVE FOP,
         ASPHALT
                                                                          LAND FILL
         SLUDGE DISPOSAL  SCHEMES FOR  A SINGLE SCRUBBER SYSTEM

-------
Q
C
70
m
                  PELLETIZER
SINTERING

MACHINE
  LIGHT
 WEIGHT
AGGREGATE
1

HJ P
RESL IFLYASHJJ
SCRUBBER JFLYASHC"
o
V
S02
REMOVAL 0
CALCIUM S
A A CALCIUM
\} if CALCIUM C/S
! t
! Lr=3 WATER err:
1 	 .. 	 -,,
VACUUM
FILTER

SLUICED
' TO
POND
ULFATE
OXIDE
^RBONATE
VACUUM
FILTER
POZZOLAN
- crtn
^zm?."rr r OK
CONCRETE
1 0
I
C02
.]
Kl
700 °C
1292 °F

MINE
LAND-FILL DRA'NA^,nM
NEUTRALIZATION


S02
PROCESSING
FACILITY
11
SO2
n
LN
1200 °C
2192 °F

aO BACK TO SCRUBBERLL-^— — ru 	 J
      SLUDGE DISPOSAL SCHEMES FOR A DUAL SCRUBBER  SYSTEM

-------
           LIME  SCRUBBING

                 Of

      SIMULATED  ROASTER  OFF-GAS
                by

            D. E. REEDY
       AIR CORRECTION DIVISION
   Universal Oil Products Company

         Darien, Connecticut
                 For

Presentation at the Second International
 Lime/Limestone Wet Scrubbing Symposium

         New Orleans, Louisiana

          November 8-12, 1971

                  561

-------
Introduction





In recent years, the lime/limestone absorption system



has gained considerable impetus as an economically



attractive flue gas SO2 control scheme, over such



alternates as lower sulfur fuels.  There are other



possible uses for the technology being developed for



power plant flue gases, such as metallurgical roast-



ing operations, but the level of understanding of the



lime/limestone/S02 chemistry does not lend itself to



reliable extrapolation to the higher SO2 levels and



S02/CO2 ratios often encountered in industrial



emission sources.





In this study, the absorption of SC>2 from an ambient



air stream containing 1% S02 and 1% CO2 utilizing a



lime slurry was accomplished using a pilot plant


                                (R)
Turbulent Contact Absorber (TCA) .  The main objective



was to demonstrate the capability of high SO2 removal



efficiencies and to determine the chemical efficiency



for lime usage under a specific set of conditions.





Description of Equipment





A schematic diagram of the pilot plant equipment set-



up is shown in Figure 1.
                        562

-------
            Air Correction Division
            Universal Oil Products Company
            Tokeneke Road • Darien. Connecticut 06820 203-655-8711
            uop
                      TO  FAN
   3 STAGE
     TCA®
   TOWER
      ROOM
       AIR
SO2
C02
                     .ENTRAINMENT
                       SEPARATOR

Y
*sx
II ~V^

TANK
1
I
V"


                                             Co (OH)2

         MAKE-UP
           TANK
MAKE-UP
 STREAM
                                              t   i

                                            METERING
                                              TANK
                                                 i
                                            	I	
             RECIRCULATION
                 PUMP
                                 PURGE
                                STREAM
    MAKE-UP
      PUMP
      FIGURE  1- EQUIPMENT  FLOWSHEET
                          563

-------
Vaporized SC>2 and C(>2 were injected into the ambient



air inlet duct approximately four duct diameters



ahead of the absorber.  The absorber itself was one



of our standard, one square foot Turbulent Contact



Absorbers*!* Three contacting stages, each containing



approximately one cubic foot of static packing,



were utilized in this study.  Recirculation liquor



was introduced countercurrent to the gas flow, above



the top contacting stage, through a full cone nozzle.



Figure 2 depicts the unrestrained contacting motion


                                            ©
achieved with the spherical packing in a TCAT  The



entrainment separator used in this study is similar



to that shown in the illustration.





Liquor draining from the  lower stage was collected



in the scrubber sump connected in parallel with an



agitated retention tank that provided 2-3 minutes



holdup.  The slurry purge from the system was with-



drawn from the discharge  of the recirculation pump.





Slurried,  hydrated lime,  with a particle size distri-



bution similar to that shown in Figure 3, was added



to the liquor recirculation line downstream of the



purge withdrawal point.




Liquor samples referred to in subsequent sections were



taken directly from the retention tank, while gas



samples were withdrawn immediately after the entrain-



ment separator.





                        564

-------
     Air Correction Division
     Universal Oil Products Company
     Tokeneke Road • Darien, Connecticut 06820 203-655-8711
     uop
SCRUBBING
 LIQUID
RETAINING
  GRIDS
                          CLEAN AIR OUTLET
                                  MIST
                                  ELIMINATOR
                                   MOBILE
                                   PACKING
                                   SPHERES
                                 HOT GAS
                                 INLET
    SLURRY
    DISCHARGE
                FIGURE  2
   TURBULENT CONTACT  ABSORBER
                   565

-------


ULJ
M
z
X
1-
oe
UJ
i-
Ul
ae,
O
100
90
80
70
60
50
40
30


20


10
  0.1
0.2
                        Air Correction Division
                        Universal Oil Products Company
                        Tokenska Road • Darlen, Connecticut 06820 203-655-8711
                        uop
                                  FIGURE  3

                    LIME  PARTICLE  SIZE  DISTRIBUTION
0.5     1       2345      10     20


          SIZE .MICRONS
                                                                   50
100    200

-------
Discussion of Results



Table I is a summary of the operating data from

pilot plant operations.  A quick analysis of these

data yields some interesting observations.



Gas velocity variations, illustrated by a compara-

sion of Test 2 and Test 5, seemed to have a relatively

small effect on S02 removal (the magnitude and

direction of change in SO2 removal in these two

tests must be attributed to "data scatter").

Although the liquor residence time has been shown

to be essentially a function of recirculation rate
             /B\
only, in aTCA4* a comparison of results from Tests 1

and 4 would indicate that holdup time, or L/G, are

not significant factors in SO2 removal.



The controlling variable in this study is the pH of

the scrubber outlet liquor.  The relationship is

shown in Figure 4.  This is nothing new,  of course,

for many studies of limestone scrubbing of power

plant flue gases have shown the importance of

limestone "quality" when translated into  terms of

scrubber effluent pH.
                        567

-------
                                               TABLE  I

                                      SUMMARY  OF PILOT  PLANT DATA
 Test
 Number
    1


    2


    3


    4
Gas Velocity,       L/G,       Scrubber AP,  Outlet Liquor
 Ft./Min.       Gal/1000 CFM   in W.G.       pH    % Solids
                                           Inlet S02,    S02 Removal,
                                            Vol %           %
en
CTI
oo
    900


    630


    800


    800


    800


    820


    820
58


82


69


82


85


73


73
9.5


7.5


8.5


8.5


9.2


8.9


8.9
 8.0


 5.8


13.0


 8.0


 5.8


11.2


12.7
17.8
17.8
                                                                   7.0
0
1
1
1
1
1
1
.92
.32
.01
.01
.01
.04
.04
82
65
98
78
73
89
98
  DER 11/71

-------
         Air Correction Division
         Universal Oil Products Company
         Tokttwke Road • D«ri«n. Connecticut 06820 203-655-8711
         uop
   100
    90
Q
ui
CD


8   80
CO
o
to
    70
    60
                    FIGURE 4

           SO2 ABSORPTION  VS. pH
                     *      8       10


                      pH  IN SLURRY
12
                          569

-------
Slurry samples from several tests were analyzed



for calcium, carbonate, sulfur and sulfite as



shown in Table II, and then the solids were



filtered from the slurry and analyzed for the



same components.  Assuming that the only sulfur



species present are the sulfite and sulfate these



data show that only a few percent of the hydrated



lime input is being lost with the slurry purge



stream.  A somewhat higher percentage of the



hydrated lime may be lost as carbonate, as shown



by the following mol ratios:





                   Molar Ratios






Test  CO3=/Ca++   S/Ca++     CO3=/S    SO3=/S
3
5
7
0.16
0.09
0.11
0.76
0.97
0.90
0.22
0.09
0.12
0.96
0.93
0.91
These ratios would indicate that 10 to 15 percent



of the calcium values are being lost as carbonate,



but it should be pointed out that spot checks of



various batches of the hydrated lime indicated



that there was some carbonate present prior to its



use in the absorption tower, probably as a result



of exposure to the atmosphere.
                        570

-------
                                TABLE II
                         RETENTION TANK ANALYSES
          Test No.                   11!







          Total Stream,Wt.%




            Calcium                  5.5         5.3        2.2



            Carbonate                1.1         0.7        0.2



            Sulfur                   3.3         4.2        1.7



            Sulfite                  4.2          -         2.4
          Solids,Wt.%
Calcium
Carbonate
Sulfur
Sulfite
34.1
8.3
20.6
49.8
29.8
3.7
23.1
54.0
28.7
4.7
20.7
47.3
DER 11/71                            571

-------
The data are consistent in that operation at a



lower pHf Test 5, resulted in an increased re-



jection of CC>2 and higher utilization of calcium.






In Test 3, the makeup lime slurry concentration



was about 10% while in Test  7 the makeup rate



was approximately doubled and the input lime con-



centration cut in half.  A comparison of the above



mol ratios for these two tests would indicate



that either the makeup slurry concentration or



the recycle slurry concentration is affecting the



loss of calcium values as carbonates.





The sulfite/sulfur ratio would indicate that very



little  sulfate is being formed in this operation,



and that sulfate formation is relatively insentitive



to pH.  These test runs were of relatively short



duration, however, and these numbers may not represent



fully equilibrated operation.






Figure  5 shows that, at pH's above the neutral point,



the liquid in the retention  tank was not completely



in equilibrium with the solids and that a long time



was required for this equilibrium to be attained



as the  residual lime went into solution.  Again,



however, the amount of undissolved lime remaining in



the slurry was very small.
                       572

-------
                                        Air Correction Division
                                        Universal Oil Products Company
                                        Tokeneke Road • Darien, Connecticut 06820 203-655-8711
                                        uop
                       12r
                                                         FIGURE 5

                                                  SLURRY  pH  VS. TIME
                       10
01
•vj
10
I
 a
                        8
                                                                   1
                                                    TIME, HOURS

-------
  Conclusions


  I.  Lime  scrubbing can achieve high SO2 removal

      efficiencies  from gases containing moderately

      "high.502 levels.


 II.  Recirculation liquor pH is a dominant factor

      in this  operation and must be maintained at

      a high level  to achieve high scrubbing

      efficiencies.


 III.  In spite of the high pH required to achieve

      the desired SC>2 removal, chemical efficiencies

      for lime utilization in excess of 80% can be

      realized.
                      References
1.  Chen, B.H.  and  Douglas, J.M.,  Canadian  Journal of
    Chem. Engr.  46,  245  (1968).

2.  Pollock, W.A.,  Tomany, J.P.  and  Frieling,  G.,
    ASME 66-WA/CD-4 (1966) .
it U. S. GOVERNMENT PRINTING OFFICE. IB7X	746761/I4OI
                           574

-------