THE DEVELOPMENT OF  NEW AND/OR  IMPROVED
AQUEOUS  PROCESSES  FOR REMOVING  S02 FROM
FLUE  GASES.  VOLUME  II

A .  F .  Graefe ,  et  al

Envirogenics Company
El  Monte,  California

October  1970
                                                    ...'to foster, serve
                                               and promote the nation's
                                                  economic development
                                                     and technological
                                                        advancement.'
      NATIONAL TECHNICAL INFORiATSOM SERVICE

-------
           ENVIROGENSCSCOMPANY
               A DIVISION OF
    AEROJET-QENtRAL  CORPORATCO!
     THE DEVELOPMENT OF MEW AND/OR IMPROVED
AQUEOUS PROCESSES FOR REMOVING S02 FROM FLUE GASES


                FINAL REPORT

                 VOLUME 11


      PREPARED UNDER CONTRACT PH 86-68-77
               SUBMITTED TO

  NATIONALAIR POLLUTION CONTROL ADMINISTRATION
U.S. DEPARTMENT OF HEALTH, EDUCATION, AND WEiFARE

-------
 STANDARD TITLE PAGE
 FOR TECHNICAL REPORTS
1. Report No.
    APTO-0620
                                                                      3. Recipient's Catalog No.
 STTiue and Subtitle
      The Development of New and/or Improved Aqueous Processes
      for Removing S02  from Flue Gas       Vol\sn© n
                                                5. Report Dace
                                                   October 3,970
                                                6. Pefformiri§.Or>!aniKiuion Code
 7. Aushor(s)
      A. F. Graefe at al.
                                                fl. Performing Organisation Rept.
                                                  No.
   Performing Organization Name and Address
      Air Pollution Control Department,
      Division of Aerojet-General Corporation
      El Monte, California
                                                10. Project/Task/Work Unit No.
                                                11. Contract/Grant No.

                                                    PH 86-68-77
 12. Sponsoring Agency Name and Address
      National Air Pollution Control Administration
      Cincinnati, Ohio  45227
                                                13. Type of Report & Period
                                                   Covered
                                                                      14. Sponsoring Agency Code
 15. Supple me ncary Notes
  i. Abstracts >Efficient  absorption of S02  at  flue gas concentrations can be effected  through
 the  use of dry, fluidized basic materials in the range of  50  to 60 C, if sufficient water
 is incorporated into the gas phase upstream of sorbent contactor. The formation of sul-
 fate can be essentially eliminated in  a fluidized bed absorber, and reduced  to  a  very low
 value in an aqueous  absorber, through  the use of ferrous ion  in an aqueous prescrubber to
 reduce N02 to NO. The thermal decomposition of both zinc and^magnesium sulfites is markedly
 promoted by the presence of steam. A new  process for the removal of S02 from flue gas is t
 described in which dry fluidized zinc  oxide is used as the absorbent. The oxide is recovered
 for  reuse upon thermal decomposition of the resulting sulfite,  and the liberated  S02 is
 recovered as such. Little or no sulfate is formed. NOx (especially N02) is the  major con-
kributor to oxidation of the sorbent in aqueous solution systems.\In general, the inhibitors
 ' id  complexing agents investigated did not lower the level of oxidation in the  presence o;'
 jNOx  in the flue gas.  The level of oxidation is less in sorbent solutions saturated with a|
 inert salt. "The efficiency of S02 removal from flue gas is not affected by the  presence o|
 JNOx.  The economics of the conceptualized  fluidized-bed zinc oxide process appear  to be
  ', Key Words and Document Analysis.  17o. Descriptors
Air pollution control  equipment
'Scrubbers
Sulfur dioxide
 Oxidation
Nitrogen oxides
Expenses
Iron inorganic compounds
jiZinc inorganic compounds
Magnesium inorganic  compounds
Reduction (chemistry)
                      superior  to  other regenerable. processes tor the
                      moval of  S02 from flue gases.(/But the state of
                      development  of this process  is ir& its very early
                      .stage.	;	_
                  Decomposition  reactions
                  Sulfites
                  Oxides
                  Sulfates
                  Nitrogen dioxide
                  Materials recovery
                                                                                               e-
    Identifiers/Open-Ended Terms
 17c. COSATI Field/Group    13/02,  07/01
 18. Distribution Statement

   Unlimited
                                     19. Security Class (This
                                       Report)
                                         UNCLASSIFIED
                                             LAJ
                                             Cl
 21. No. of Pages

     175
                                                          20. Security Class (This
                                                             Page
                                                               UNCLASSIFIED
'22. Price
 IFOKM CFSTI-33 (4-70)

-------
This report was furnished to the Air
Pollution Control Office by the
Aerojet-General Corporation in ful-

-------
      THE DEVELOPMENT OF NEW AND/OR IMPROVED

AQUEOUS PROCESSES FOR REMOVING SO2 FROM FLUE GASES
                      FINAL REPORT

                        VOLUME II
                        October 1970




                            by


      A.  F.  Graefe,  L. E. Gressingh, and F. E. Miller
         Prepared under Contract PH 86-68-77 by the
        Air Pollution Control Department, Envirogenics,
           A  Division of Aerojet-General Corporation,
                     El Monte, California
                         Submitted to
        National Air Pollution Control Administration
     U.S. Department of Health,, Education,,  and Welfare

-------
                                                   1 Octobe? *970



             CONTRACT FULFILLMENT STATEMENT




       This is Volume n of a final report submitted to the Natiooai

Air Pollution Control Administration in partial fulfillment of Contract

No. PH 86-68-77.   This report covers the period 29 December 1967 to

1 September 1970.

                                Aerojet-General Corporation
                                L. E. Gres singh
                                Program Manager
                                Approved:
                                E7 M. Wilson, Manager
                                Air Pollution Control Department

-------
                             TABLE OF CONTENTS
                                  Volume Two
                                   GENERAL
I.     INTRODUCTION
H.    SUMMARY AND CONCLUSIONS
IL    FLUIDIZED ZINC OXIDE AS AN SO2 ABSORBENT
      A.   Volume I _ _      3
      B.   Volume U _ _ _____      6

                                  PART TWO
                          NEW AQUEOUS PROCESSES
L     INTRODUCTION                                                        14
      A.   Introduction	     14
      B.   Apparatus	     15
      C.   Absorption  and Oxidation of SO-	     17
      Do   Absorption  of NO	'_     33

III.   ZINC OXIDE PARTICLE SIZE AND ACTIVITY
      A.   Introduction	     34
      B.   Small Particle Studies (Kadox-15)	     34
      C.   Large Particle Studies (Pelleted Kadox 215)	
           1.    Introduction 	     39
           2.    Initial Fluidization Experiments	     39
           3.    Attrition During Absorption of SO  in a
                 Fluidized System	
                 a.  Apparatus	     40
                 b.  Experimental Work  ^	     41

-------
                                                                           ~J
                      TABLE OF CONTENTS - Coat'd

                                                                    Page
IV.   SCREENING OF SELECTED FLUIDIZED BASIC MATERIALS
     AS SO, ABSORBENTS
A.
B.
C.
D.
E.
Introduction
Alkali and Alkaline Earth Sulfites
Sodium and Calcium Carbonates
Magnesium Oxide
Conclusions
44
45
47
51 J
52
                              PART THREE                                J

                        PROCESS IMPROVEMENT                             \
                                                                           _J
I.
II.
in.
IV.
INTRODUCTION
53
DISPROPORTIONATION OF ZINC SULFITE
A.
B.
C.
Introduction
Results of Experiments Conducted in a Muffle Furnace
Results of Experiments Conducted in a Tube Furnace
DISPROPORTIONATION OF MAGNESIUM SULFITE
THERMAL DECOMPOSITION OF ZINC SULFATE
55
56
63
70
73
                              PART FOUR

                 A ZINC OXIDE FLUIDIZED BED SYSTEM FOR
                THE ABSORPTION AND REGENERATION OF SO2          77


                               PART FIVE

                       OXIDATION AND NO  STUDIES
                                         x

     IN TRODUC TION                                                  8 1

-------
                       TABLE OF CONTENTS - Cont'd
IL   EXPERIMENTAL RESULTS
A.
B.
C.
D.
E.
F.
Laboratory Evaluation of Inhibiting and Completing Agents 	
1. Introduction
2. Screening Tests
3. Experimental
Base Line Tests in Bench Scale Unit
Oxidation and NO_ Experiments
1. Introduction
2. Effect of O0
3. Effect of NO^ and 2. 8% O^
4. Effect of Fly Ash
5. Effect of Ferric Ions in Solution
6. Effect of Inhibitors
7. Effect of High Concentrations of Na-,50,, and NaCl
8. Effect of Oxidation of SO_ on Required Solution Rate
9. Effect of Solution Flow Rate
10. Effect of Type of Scrubber
11. Effect of NO 0 Removal in the Prescrubber
12. Effect of Type of Solution
a. Solution Molality
b. Type of Solution
NO Removal Data
Sulfate Removal with Reverse Osmosis
Other Sulfate Removal Studies
PART SIX
S2
82
84
89

93
93
100
100
102
103
103
	 104
104
108
109

109
110
110
112
113

                  PROCESS AND ECONOMIC EVALUATION



I.    INTRODUCTION                                                  118

-------
                           TABLE OF CONTENTS - Cont'd
 U.   JOHNSTONE ZINC OXIDE PROCESS _ _____   1 18
      A.    Conversion of Sulfur Dioxide to Sulfur                            119
      E.    Conversion of Sulfur Dioxide to Sulfuric Acid                     ^9
      C.    Sulfate Removal with Reverse Osmosis __________________________   125
KI.   FLUIDIZED ZINC OXIDE PROCESS 	
      A.    Initial Approach	
      B.    Final Approach
            1.     Introduction
            2.     Capital Cost Estimate	   130
            3.     Operating Cost Estimate	_____	   134
            4.     Profitability	   138
IV.   MAGNESIUM BASE SLURRY SO  SCRUBBING SYSTEM
                                      .
      A.    Introduction  	    140       •
                         	,	.	__             ^
      B.    Process Engineering 	____________________________________    *^0
      C.    Capital Cost  Estimate	    143
                                 •^^*—^~^^^~n^*~^^**^—^^^^*ii^^^^^^i^m*~—^**~i^i^~^^^**»*i*ii^******»mi**i^i^^^^*i*mi^*^              f
      D.    Operating Cost and Economic Analysis	    143      "*
REFERENCES	    '	    '       145       ;
                                                                                      _J

                                LIST OF TABLES
Number
         Reaction of Fluidized Zinc Oxide (60-200 mesh Kadox-15) with
         Selected Flue Gas Components _ __ _ _    19
         Gaseous Compositions and Space Velocities for Runs Given
         in Table 1 _ _ • _ _    20
         Oxidation of Aqueous Iodide Ion by Selected Flue Gas
         Components Under Various Conditions __ _ _ „_________    ^8
         Rate of Solution of Selected Metallic Oxides and Hydroxides

-------
LIST OF TABLES - Coat8d
Number
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
Attrition Experiments During SO0 Absorption
The System Sodium Carbonate - Water
Reaction of Fluidized Absorbents with Selected
Flue Gas Components
Gaseous Compositions and Space Velocities for Runs
Given in Table 7
Thermal Decomposition of ZnSO • 2-1/2 HyO in a
Muffle Furnace at 545 + 3°C * &
Thermal Decomposition of ZnSO-' 2-1/2 H-O in a
Muffle Furnace at 275 + 3°C *
Thermal Decomposition of ZnSO • 2-1/2 H_O in a
Tube Furnace at 275 + 3°C
Effect of Water Vapor on the Thermal Decomposition of
ZnSO,* 2-1/2 H0O in a Tube Furnace
Effect of Temperature on the Thermal Decomposition of
Magnesium Sulfite in a Current of Nitrogen or Steam
and Nitrogen
Effect of Water Vapor on the Thermal Decomposition of Zinc
Sulfate in a Tube Furnace
Base Line Inhibitor Screening Tests
Inhibitor and Complexing Agent Screening Tests
Base Line Oxidation Studies - Recirculating Scrubber
Base Line Oxidation Studies - Once Through Scrubber
Oxidation Studies - Once Through Scrubber
Oxidation Studies - Recirculating Scrubber
NO Removal with Aqueous Systems
Reverse Osmosis Tests - Solution Composition
Reverse Osmosis Test Results
Capital Investment Summary - Zinc Oxide Process with
Conversion of Sulfur Dioxide to Sulfur
Profitability - Zinc Oxide Process, Plants Operating at 70%
Plant Factor - SO, Converted to S
Profitability - Zinc Oxide Process, Plants Operating at 90%
Plant Factor - SO., Converted to S
Profitability - Zinc Oxide Process, Plants Operating at 70%
Plant Factor - SO Converted to Sulfuric Acid
Profitability - Zinc Oxide Process, Plants Operating at 90%
Plant Factor - SO? Converted to Sulfuric Acid
Page
42
48
49
50
57
62
64
68
72
75
85
87
90
91
94
99
111
114
116
120
121
122
123
124

-------
                        LIST OF TABLES - Conrd

                                                                      Page
30
31
32
33
34
35
36
37
Reverse Osmosis Test Results
Sodium Sulfate Removal from Sulfite - Bisulfite Absorbent
Heat Input Requirements to Regenerate Zinc Oxide -
1400 MW Power Plant
Fluidized Zinc Oxide Process: Capital Cost Estimate
Fluidized Zinc Oxide Process: Operating Cost Estimate
Fluidized Zinc Oxide Process: Raw Materials and Chemicals 	
Fluidized Zinc Oxide Process: Utilities
Fluidized Zinc Oxide Process: Profitability
Magnesium Base Slurry Process: Capital Cost Estimate
Summary
126
127
129
132
135
136
137
139
144
                           LIST OF FIGURES
1
2
3
4
5
6
7
Fluidized Bed Reactor System
Modified Fluidized Bed Reactor System
Screen Analysis of Zinc Sulfite and Zinc Oxide
Block Flow Diagram - Fluidized Bed Zinc Oxide System
% SO2 in Scrubber Outlet Gas vs. S/C of Rich Solution
for Recirculating Scrubber and Once Through Scrubber
Effect of NO on Oxidation of SO to SO." in Once Through
Scrubber x ^4
Theoretical Effect of Oxidation on S/C in Once Through
Scrubber a
16
18
36
78
92
101
105
 8   Oxidation vs.  Solution Rate for Sodium Scrubbing Solution
     in Once Through Column	   106

 9   Oxidation vs.  Solution Rate for Potassium Scrubbing Solution
     in Once Through Column	   107

10   Test Cell Assembly - Flow Diagram	   115

11   Block Diagram - Optimized Fluidized Zinc Oxide Process__	   131

12   Magnesium Base Slurry SO_ Scrubbing System for
     1400 MW Power Plant                                             142

-------
                            APPENDICES




                                                                   EM®.
APPENDIX A.    Oxidation and NO Studies
I.
IL
in.
IV.
x -'"•-"-•—•-'""-—"- 	 — -^.—...1.1.
Description of Equipment
Operating Procedure
Analytical Methods
Calculations
A-l
A-6
A-ll
A- 13

-------
                             PART ONE
                             GENERAL
I.
INTRODUCTION
      The initial objective of Contract No.  PH 86-68-77 was to assay the fea-
sibility of using aqueous systems for removing sulfur dioxide from flue gases.
The period of service of the initial program was from 29  December 1967 to
31 May 1969.  This technical effort is reported in Volume I0 "Applicability of
Aqueous Solutions to the Removal of SO_ from Flue Gases. "  An extension of
this program covered the period 31 May 1969 to 1 September 1970,  This part
of the program is  reported in Volume II, "The Development of New and/or
Improved Aqueous Processes for Removing SO, from Flue Gases. "

      The general discussion, Part One, consisting of the Introduction,  and
Summary and Conclusions, is identical in both volumes of this report and
provides a resume of the entire project.

      The following three phases define the program effort of the  initial
period:

      Phase I.    Assessment of Aqueous Solution Methods

             e  Literature  survey
             ©  Preliminary  economic evaluation for comparative purposes
             «  Selection of Candidate processes

      Phase IL   Laboratory Experimentation Relating to Candidate Processes

             e  Process  simplification and improvement of each candidate
                 existing process
             a  Demonstration of process feasibility of any candidate new

-------
       Pha.   III.   Preliminary Plant-Scale Process Evaluation and
                   Cost Estimates for the Candidate Processes                  -'
             o     Application of processes selected on the basis
                   of Phases I and II to both new and existing power             —'
                   plant facilities
             •     Application of processes selected on the basis
                   of Phases I and II to a new smelter facility.
L~liase I was accomplished during the first five months of 1968 with
Phases II and III conducted concurrently during the remainder of the             —
calendar year.
       The following parts of Volume I are concerned with the results of         -*1
the application of the various tasks listed above.  Parts Two to Four
cover the work conducted under Phases I,  II, and III,  respectively. Part         -J
Five discusses recommendations for future work under  the contract ex-
tension.   Part Six,  "Bibliography" is the result of the extensive literature        __
survey which was carried out at the beginning of Phase I.  Nearly  700
references are listed,  together with an appropriate author index.
                                                                              ~_/
      vThe program extension, deeignated-a* Phase IV, consisted of.-^he
following  tasks:                                              . _.--
       A.    Conceive New Aqueous Scrubbing Processes^'  >,
       B.    Develop Improvements to Previously Conceived Aqueous            —'
             Scrubbing Processes,
       C.    Determine the  Degree to Which Inadvertent Sorbent                 _
             Oxidation Can Be Minimized:  N
       D.    Determine the  Degree of Interference Which inadvertent
             Sorption of NO  Has On SO, Removal Efficiencies; ^ ^             ~^
                          x           &
       E.    Support the Laboratory Investigations with Preliminary
             Process Evaluations and Economic Analyses.                       -'
       The tasks of Phase IV are covered in Parts Two through Six of

-------
            II.     SUMMARY AND CONCLUSIONS
                   A.    VOLUME I
                         Approximately 500 technical documents, selected from the
            bibliography of Part Six,  were collected, catalogued, and reviewed for
"-          the identification and description of various aqueous processes which
            have been used, or are currently being investigated, developed, or used
v__          for the removal of sulfur dioxide from flue gases„   Some thirty processes
            were identifieds and of these sufficient data were available for a prelimi-
v            nary economic evaluation of twenty-two.  As a result of the evaluation
            the following four processes were considered to merit further investiga-
            t ion:
                         ©    Zinc  Oxide Process (Sodium  Sulfite Scrubbing)
^_                       »    Cominco Exorption Process (Ammonium Sulfite
                                    Scrubbing)
                         «    Ammonia-Hydrazine Exorption Process
                                    (Hydrazine Scrubbing)
                         &    Mitsubishi Lime Process (Lime water Scrubbing)
            The Ammonia-Hydrazine Exorption process,  conceived at Aerojet, repre-
            sented a paper study,  subject to an experimental demonstration of process
            feasibility.
                         Following process selection, a laboratory program was con-
V—-'
            ducted relating to process demonstration and/or improvement. Attempts
            to improve the Zinc  Oxide process were mainly concerned with lowering
"~          the calcination temperature required for the release of sulfur dioxide from
            the regeneration feed, zinc sulfite.  No  significant improvements were
^~          effected, but the investigation led to the conception of a new process based
            on the use of zinc oxide, in which a fluidized bed of this material is used
v__          directly for the  low-temperature (50  C) sorption of sulfur dioxide.  The

-------
             r.."-s  Cominco Sxcrption process  suffers from the relatively
'.•  ;_/i jjecr* coutj associated with the desorpticn of sulfur dioxido from
c* ,-iLuv.'. £.:••• w.-ijarati bisulfite oolution.  The uae of acids as yromoters for
this reaction was  therefore investigated.  Although several acids were
found to be partially effective it was found that the coat cf the acid, the
additional complexity to the process caused by the use of the acid,  and
o^ier factors would not be compensated by the limited reduction in steam
requirements which might be attainable in this manner.  It wao concluded,
therefore,  that this process must be considered as  uneconomical.
             The Ammonia-Hydrazine  Exorption process was designed to
combat the high steam requirements of the Cominco Exorption process
through the use of hydrazine as the  absorbent  for sulfur dioxide.  Since
hydrazine salts are highly soluble in aqueous media it appeared that the
desorption of sulfur dioxide from aqueous hydrazine bisulfite might be
effected without the simultaneous volatilization of large quantities of water.
As the result of an experimental program designed to demonstrate  process
feasibility, it was found that concentrated hydrazine sulfite solution readily
absorbed sulfur dioxide under simulated process conditions.   However,  an
unavoidable loss of hydrazine by oxidation occurred during the regeneration
reaction,  so  that any savings in steam  costs through the use  of this method
was nullified.  Therefore, the process was no longer regarded as
economically feasible.
             No experimental work was indicated relating to the Mitsubishi
Lime process.   The process was regarded as  economical, provided that
by-product gypsum could be sold in quantity.   However, a  subsequent mar-
ket survey indicated that  gypsum requirements could readily be filled from
natural deposits and that  no appreciable synthetic gypsum market exists at
the present time.  A simplified version of the  process, in which the gypsum
is discarded  as waste, appeared more  attractive. A laboratory effort

-------
             In addition to the laboratory effort described above, which
was designed to overcome problems associated with specific processes,
some attention was also directed to a problem which is common to all
aqueous scrubbing methods in which sulfur dioxide is  recovered as  such;
namely, the  oxidation of sulfite to sulfate in the absorber.  The  literature
indicates that in some proceoses oxidation can amount to 10 to 14%  or
more (expressed as a percentage of the incoming SO,).  It was planned
to investigate the extent to which oxygen and nitrogen oxides in flue gas .
contribute to oxidation, and to investigate the use of various oxidation
inhibitors, such as hydroquinine,  for its prevention.  Work  in this area
was initiated toward the end of the  contract period,  and was completed
during the second year of the  program (Phase IV).  The results  are re-
ported in Volume II.
             Of the four candidate  processes,  only the Zinc Oxide pro-
cess (Na  scrubbing/ZnO regenerant) was considered for a. complete
evaluation in Phase III.  Thus, process evaluations  and cost estimates
were completed for large new and existing power plants,  a small existing
power plant,  and for an existing smelter facility. An evaluation was  be-
gun on the simplified lime process, but was not completed since the
analysis of limestone systems was being done on another contract.
             The major conclusions which were drawn from the work
reported in Parts Two to Four of Volume I,  are the following:
       &     Of the four candidate  processes which were selected
             for  further study as the result of the Phase I effort,
             the  Zinc Oxide process was considered to merit
             further study, both in,the form of a fluidized bed
             system, as proposed by Aerojet, and in the form
             of the original Na  scrubbing process, as developed
             by johnstone to the small-scale pilot stage.  For
             the  Johnetone process available data in Phase III
             indicated that for a large power plant (2, 5 MMSCFM
             of flue gas) to be operated at break-even conditions,
             product sulfuric acid would have to be  salable at al-

-------
                                                                                I
                                                                               -J
             A. ..J gas concentration of 3000 ppra),  Ifa  howevor,
             the product acid from this plant could be  sold for
             only $10/ton,  the operation of this  "add-on" SO-                    -'
             control process would represent a  net cost to the
             utility of about $1. 23/ton of coal burned.   Applied                  _j
             to a medium-sized smelter effluent (220, 000 SCFM
             of the flue gas), the Johnstone process could be
             operated at break-even conditions, if produce
             sulfuric acid were sold at about $18/ton.
                                                                               ~j
       »     The three remaining candidate processes  (Commco
             Exorption, Ammonia-Hydrazine Exorption,  and
                                                                               —i
             Mitsubishi Lime) are not considered to be as
             economically attractive as the Johnstone  process.                   ,
       •     A major problem confronting any aqueous process
             in which sulfur dioxide is recovered as such is that                  .'
             of oxidation in the scrubber.  Such oxidation in-
             evitably leads to the formation of sulfate,  which
             is  in general less  readily isolated from aqueous                     -'
             solution and less  readily decomposed than the
             corresponding sulfite. As a result it may be                       _J
             anticipated that equipment and operating costs
             will increase, and product yields will decrease,                      '
             in proportion to the extent of oxidation encountered.
       B.    VOLUME II                                                       _j
             In  the area relating to new aqueous processes for the  re-
moval of SO? from flue gas, attention has been focused on the  use of              _^
fluidized solids  as absorbents.  The absorption step, which is conducted
at 50 to 60 C,  requires the presence of appreciable water in the gas
phase, and this  is provided largely through the use of an aqueous pre-
scrubber.  The  prescrubber also serves the function of removing SO.,

-------
              Only basic materials have proved suitable as SO, absor-
                                                            &
 bents.  It was found, for example,, that both alkali (Na, K) and alkaline
 earth (Mg,  Ca) sulfites are too weakly basic to absorb,, but that
 carbonates (Na) and oxides (Zn? Mg)  are good absorbers.  For example,
 when zinc oxide was used it was found that more  than 50 g of SO, was
 absorbed per 100 g of the oxide before SO?  removal efficiency dropped
 below 90%.
              A problem that arises  in all regenerative aqueous SO,
 scrubbing processes is that a portion of the absorbent becomes oxidized
 by the O,, fly ash,  and/or NO components of flue gas.   This is highly
        M                    X
 undesirable,  inasmuch as the  absorbent cannot readily be regenerated
 for reuse from the  oxidized product. The extent of oxidation appears
 to be substantially less for essentially dry fluid!zed bed absorbents
than for bulk water systems, and in particular fly ash, which tends to
 catalyze the oxidation in bulk water  systems,  was determined to be
 without effect when incorporated into a dry fluidized zinc oxide absorber.
              Of the three gaseous oxidizing components of flue gas (O,,
 NO, and NO,),  it was found that NO, is by far the most active, and that
            Ct                      Ci
 fluidized zinc oxide was partially converted to sulfate when both SO,
 aiid NO, were present in the influent gas.  It was discovered,  however,
       Ct
 that oxidation of the bed material could be essentially eliminated through
 the incorporation of ferrous ion into the aqueous  prescrubber.  The main
 function of the ferrous ion is considered to be that of reducing NO, to
 NO.  For an influent gas containing all flue gas components,  less than
 one-half percent of the SO, absorbed by zinc oxide was converted to
                          £*
 sulfate  when the prescrubber contained 1% ferrous sulfate.   In the prac-
 tical case,  ferrous ion would be provided to the prescrubber in the
 form of scrap iron.
              It may be presumed that the use of  fluidized solids as SO,
 absorbents will be attended to some  extent by the attrition of solid
 particles, and consequently some attention was devoted to a study of
 both particle size and particle activity when zinc oxide was used as

-------
v-xicli p«i;/:^ .    are  readily aitrited to fine particle s; but -that if the cxide
i .. :'?lr~i cnxwv. -'Led to the sulfite through SO9 absorption, and the sulfite is
                                  •        £•
thc:i tiiij i.'i.r.
-------
             As a result of the experimental work considered above relat-
ing to the use of fluidized zinc oxide as an absorbent for SO,* & tentative
system was formulated involving the recovery ef the SO, ao auch.  The
overall system  involves the use of an aqueous p re scrubber,  the removal
of SO, from the water-saturated gas by the oxide, and the thermal da com-
     f>
position of the resulting sulfite at 275  G for the regeneration of the osdde
and the recovery of SO,, Any zinc sulfate formed is separately decom-
                      £*
posed at higher  temperatures,  and no waste product results. In an
alternative system,  sulfate is removed by filtration rather than by cal-
cination.  To  accomplish this,  a portion of the  sulfite-sulfate mixture
is dissolved in aqueous SO, and the sulfite is re precipitated with zinc
                         Ct
oxide.  After  filtering the zinc  sulfate  solution,  the  sulfite cake is re-
turned to the process.
             The studies on oxidation and oxides of  nitrogen in aqueous
Solution scrubbing systems were combined due to the contributions of
nitrogen oxides  to sorbent oxidation.  Most of the experiments were
made with sodium sulfite-bisulfite solutions similar to that used in the
Johnstone Zinc  Oxide process.   A  "once through" countercurrent absorp-
tion column was used in most of these  tests.  Fresh absorbent solution
was fed to the top of the column and the spent solution removed from
the bottom.  Another arrangement was used for some tests in which the
absorbent was recirculated through the column.  Other absorbent sys-
tems checked were potassium sulfite-bisulfite, and  magnesium,  calcium,
and sodium hydroxide solutions.
             Commercially available inhibiting and  complexing agents,
widely used in other applications, were screened for their ability to re-
duce oxidation of sulfur dioxide (sorbent) in the scrubber.  Oxidation of
the sorbent due  to oxygen or fly ash in the flue  gas was suppressed by
some of the materials.   When nitrogen oxides were  present in the flue
gas,  however, oxidation was  lowered only by using nitrilotriacetic acid,
and this inhibitor was effective only in a potassium sulfite-bisulfite

-------
              ', was found that, although oxygen in the flue gaei contributes
to the oxia.~t.icn of the sorbent during scrubbingt the high levels  of oxida-
tlui. was pxugj.ee> Lively greater as the concentration of nitrogen dioxide
in the flue gas was increased.  The rate of oxidation was highest in tests
made with 400 ppm each of nitrogen oxide and nitrogen dioxide.
             Fly ash did not significantly increase oxidation.in systems
where fly ash free absorbents were fed to the once through column. 'In
absorbent recirculating systems,,  in which the fly ash accumulated and
some of the iron content was solubilized, a low level of SOy oxidation
was experienced.  The oxidation increased with increasing turbulence  in
the system.  A similar effect was found when ferric ions such as Fe,(SO.)_
                                                                  £>    4 j
were added to the  system.
             Saturating the sodium sulfite-bisulfite scrubbing solution
with sodium sulfate  inhibits oxidation.  This  is  explained by the  limited
solubility of oxygen in high ionic strength aqueous solutions.
             Since oxygen is only  slightly soluble in water, the liquid
phase  is the limiting resistance to the absorption of the  oxygen.,  Thus,
increasing turbulence in the scrubber improves the absorption of oxygen
and the amount of oxidation of the  sorbent increases with the turbulence
of the  system.
             As discussed in Part Three,  a prescrubber circulating a
solution containing ferrous ion removes the nitrogen dioxide from the
flue  gas stream.  Using this prescrubber system in conjunction with
aqueous solution scrubbers also reduced oxidation of the sorbent to a
very low level due to removal of the nitrogen dioxide.
             Although additional investigations would be needed  to verify
the data, it seems that the absorption of nitrogen oxides simultaneously
with sulfur dioxide is about the same quantity as the percent nitrogen
dioxide in the flue gas.  The experiments also indicate that the absorption
of NO  into SO, scrubbing solutions has no effect on SO- removal efficiency.
      X       £•                                       &

-------
             Miscellaneous process and economic evaluations were made
on the Johns tone Zinc Oxide process, the new Fluidiaed Zinc Oxide pro-
cess, and a Magnesium Base Slurry SO, Scrubbing system.
             Evaluations  involving the Johnstone Zinc Oxide process
included an analysis in which sulfur dioxide recovered from the absor-
bent was converted to sulfur using the Asarco process.  If product sulfur
could be sold for $20 per  long ton8  the net cost of operating this SO- re-
moval/sulfur recovery process on a 1400 MW power plant (at a 70% load
factor) would approximate $1. 36 per ton of coal burned.  The economics
of converting the sulfur dioxide to sulfuric acid (see Volume I) was re-
evaluated on the basis of lower sales prices for the  sulfuric acid produced.
An analysis of using reverse osmosis to separate the oxidation product
from the absorbent indicated an uneconomical system based on current
technology.
             The evaluation of the optimized new Fluidized Zinc Oxide
process showed relatively low capital and operating costs for a system
serving a 1400  MW power plant; however, it must be recognized that
this projection  is based on the presumed validity of data that has been
generated on a  very small-scale laboratory equipment.
             The cost study of the Magnesia Base Slurry SOg Scrubbing
system was made only on the absorption system.  An evaluation of the
regeneration system, which was not available, would have to be made to
complete the analysis.
             The major conclusions which have been drawn from the work
reported in Parts Two to  Six of Volume II are the following:
             •      Efficient absorption of SO2 at flue gas con-
                   centrations can be effected through the use
                   of dry, fluidized basic materials in the range
                   of 50  to 60 C,  if sufficient water is incor-
                   porated into the gas phase upstream of
                   sorbent contactor.

-------
c     The formation of sulfate can be eosen.tis.lly
      eliminated in a fluidized bed. absorber, arid
      reduced to a very low value in an aqueous                     __.
      absorber, through the use of ferrous ion in
      an aqueous pre scrubber to reduce NO, to
      NO.
e     The thermal decomposition of both zinc ar.d
      magnesium sulfites is markedly promoted by
      the presence of steam.   The use of steam
      permits the decomposition of zinc sulfite to
      be carried out at a temperature below that
      at which disproportionation occurs.                           -1'
•     A new process for the removal of SO, from
      flue gas is described in which dry fluidized
      zinc oxide is used as the absorbent.  The             '         ,
      oxide  is recovered for  reuse upon thermal                    ~J
      decomposition of the  resulting sulfite,  and
      the liberated SO,  is  recovered as such.                       J
      Little or no sulfate is formed.
•     NO (especially NO,) is the major contributor                 ~/
         X               £
      to oxidation of the sorbent in aqueous solution
                                                                    /
      systems. •-                                                   __,
•     In general,  the inhibitors and complexing agents
      investigated did not lower the level of oxidation               _,•
      in the presence of NO  in the  fiue ga-s.  -
                           x                                        i
•     The level of oxidation is less  in sorbent solutions             _>
      saturated with an inert  salt. ,.
o     The efficiency of SO- removal from flue  gas is                —-'
      not affected by the presence of NO . -
                                       •JC                            ,
o     The economics of the conceptualised fluidized-                —'
      bed zinc oxide procesa  appear to  be superior to
                                                                   /
      other  regenerable processes for the removal of              _j
                                               *
      SO2 from flue gases, but the state of development

-------
One, merely tentative,, "conclusion" bears mentioning:
9     It appears that adding NO- to flue gas to
      obtain an equimolar ratio of NO/NO^
      prior to scrubbing the gas with aqueous
      sulfite-bisulfite solutions or slurries,  for
      SO,/NO  removal will not lower the NO
        £    3C                               X
      content of the gas significantly^ but will
      cause unwanted oxidation of the sulfite to
      sulfate to increase drastically.

-------
                                                                                 .J
                                PART TWO
                         NEW AQUEOUS PROCESSES   .                           -
                                                                                  i
 I.    INTRODUCTION
                                                                                  i
      A major area of interest on the present program has been concerned         -J
 witf i ihe conception of new aqueous processes for the removal of SO., from
                                                                 u
 xlue gas, and with a demonstration in the laboratory of the feasibility of such       _)
 p-ocesses.  The scope of this effort was considered to be bsroac'j in the sense
 that a candidate process  need not require the presence of bulk water to be           '.
 included in the aqueous category.  In particular,  the use of dry fluidized
 zinc oxide as the absorbent, which was briefly investigated earlier in the            \
 program (Reference 1() was considered to fall within the scope of the present
 effort,  since the absorption step will not occur in the absence of water vapor.
 The fact that the envisioned process  in this case involves aqueous  prescrub-        —'
 '•>ing to  remove fly ash and SO,f  and that the absorption of SO- is conducted
           o                                                                      '
 at about 50 C, which is common to all aqueous processes,  would further  in-        _j
 dicate that this type of process should  be included in any general study related
 to aqueous systems.
 II.    FLUIDIZED  ZINC OXIDE AS AN SO2 ABSORBENT
      	.	                            |
      A.    INTRODUCTION                                                      ~>
            The original concept involving direct absorption of SO^ onto              /
 fluidized zinc oxide was formulated on the original contract on the basis that       —'
 the reaction of interest represents a simple neutralization of acidic and basic
                                                                                  i
 reactants,  and that this type of reaction is in general highly dependent upon        _j
 the presence of water:
                                    H?O
            ZnO -I- SO2 + 2-1/2 H2O—	s*.  ZnSO^ 2-1/2 HEO         (1)        ~"
Whether or not water vapor would suffice to promote Reaction 1  to the
desired extent formed the basis for most of the initial experiments on
absorption.

-------
           Early in the program, consideration was given to the use of an
aqueous prescrubber to remove fly ash, and to introduce additional water
vapor into the flue gas.  In the practical case, a prescrubbed gas should
be available for SO? removal at about 50  C, and this temperature was  there-
fore employed as the fluidized bed temperature in most of the early work.
The experiments to be described were conducted in the manner that had
been previously used by the Bureau of Mines in their screening of metal
oxides as SO- absorbents (Reference 2).  Zinc oxide was included in their
study,  but was found to be inactive at 130 C in the  absence of appreciable
water vapor.
           In addition to  the absorption problem per se,  it was considered
necessary to conduct the absorption step, if possible, in such a manner that
little or no oxidation o::' the SO2 occurred:
           ZnS03-2-l/2  H20 + [o]	» ZnSO4* H2
-------
Kote:
  U^psr portion of water spsryar and entire
  rector wr^ped with electrical heating tape
                                          5 28/15 Ball &
                                             Socket
                                              § 18/9 Ball
                                              & Socket
     ercury Buhbler
    lowrrKtor
(D Flexible Tub!ng (Tygon)
         Sparger
     lass
      ter
                                                            ®Reactor, 15" Between
                                                                 (On center) x 1" OD
                                                               Zinc Oxide Bed
                                                               Glass Frit
                                                               Vibrating T&te
                                                               Thermometer Welt and
                                                                TSierznosnster
                                 FLUIDIZED BED REACTOR SYSTEM
                                            Figure 1
                               L .   I
L

-------
the surface of the zinc oxide charge, and was therefore continuously bathed
with fluidized zinc oxide during a run.  Metallic mercury was used for heat
transfer in the thermometer well. Unreacted SO, in the exiting gas was ab-
sorbed in standard 0. 1  N iodines  and the excess iodine titrated with standard
0. 1 N sodium thiosuliate.
            A run was considered to have reached "breakthrough, " in con-
formance with the experimental work conducted by the Bureau of L/une&j v,hen
an analysis of the iodine solution  showed that 10% or more of the influent SO~
was escaping the bed.   In generals the time of breakthrough was sharply de-
fined,  in that the absorption of SO? was normally complete until breakthrough
occurred,  at which time the rate  of absorption fell rapidly to zero.
            Following Run 1,  the  gas inlet system  was modified so that various
gas compositions could be introduced into the bed.  The overall apparatus
(without heating tapes) is shown in Figure 2.  The  panel board shown in the
figure included manometers and flowmeters for the individual gaseous con-
stituents, which included N,! O- (as air), SO,i CO-,  NO, and NO,. Other
changes made at this time included the insertion of a thermowell directly
into the water sparger, and the use of three heating tapes instead of one.
The lower tape was used from the upper part of the sparger to the gas inlet
to prevent premature water condensation. The second tape was placed around
that portion of the absorber that contained the fluidized zinc oxide.  The  third
tape was used to prevent water condensation in the upper part of the absorber.
      C.    ABSORPTION AND OXIDATION OF SO2
            A total of nine runs were made in the absorber,  with progres-
sively complex gas compositions (see  Tables  1 and 2).   Zinc oxide designated
as Kadox-15 (99.6% ZnO), which was obtained from the New Jersey Zinc
Company, was used in all of the runs.   The initial run was conducted on the
previous program, and the results were reported in detail in Volume One of
this report.  An analysis of the bed material at the completion of this first
run was interpreted as indicating  that ZnSO~° 2-1/2 H,O was present to the

-------
                                            b,


                                            §
                                            H
                                            hi

-------
                               TABLE 1

    REACTION OF FLUIDIZED ZINC OXIDE (60-200 Mesh, Kadox-15),
                 WITH SELECTED FLUE GAS COMPONENTS
                          (As Depicted in Table 2)
Run
No.
la
2b
3b«c
3A
3B
4d
5d
6d
?d, e
8d,e,f
gd, e, g
Temp.
35 & 50
50
50
50
50
55
55
55
55
55
55
Reaction
Time
(hrs)
15
27
23
23
23
12
12
12
12
12
12
so2
Absorbed
(g/100 g ZnO)
19.2
55.6
48.1
40. 7
52.4
24.9
24.7
22. 1
31.3
31.7
29.9
Absorbed SO^
Converted to
Zinc Sulfate
2.34
0.41
13.7
13.0
14.2
0.32
2.65
6.55
4.90
4.72
0.42
ZnO Converted to
ZnSO3' 2-1/2 H2O
22.0
67.6
49.0
36.8
56.1
29.9
27.5
22.6
28.6
29.8
28.8
 This run was conducted at 35°during the first 6-1/3 hours,  and carried
 beyond SO2 breakthrough which occurred at 9 hours.
•i          "
 Run terminated at SO? breakthrough.

 Composite of 3A, representing the free -flowing portion of the reaction
 product, and 3B,  which was caked.

 Run terminated arbitrarily.

el% fly ash added to bed.

 5% metallic zinc dust added to bed.
Of
6
 f
 Sparger liquor consisted of 1% iron (as FeSO.) in 5% H9SOA.
                                           4        M   *X

-------
                                TABLE I


CASEOUS
COMPOSITIONS AND SPACE VELOCITIES
FOR RUNS GIVEN IN TABLE I
(Total gas flow approximately 600 ml per minute)
Run
No.
, a
lb
2
3
4
5C
5d
6
7
8
9
e
aAt
N2

95.7
93.9
75.8
72.0
72.2
74.1
70.6
71.5
71.9
72.0
71.9
70.9
35°.
so2

0.25
0.25
0.29
0.28
0.28
0.29
0.28
0.28
0.28
0.28
0.28
0.28
bAt 50°.
H20

4.01
5.90
9.40
11.00
10.60
8.20
12.60
11.38
11.03
10.82
11.00
12. 18

co2
...... fvol-'ftV

^
-
14.3
14.1
14.2
14.6
13.9
14.1
14.1
14.1
14.1
13.9
CFirst
°2

—
-
mt
2.68
2.68
2.76
2.63
2.67
2.68
2.68
2.68
2.70
7 hours.
Space
NO ' NO^ Velocity
£, ,
- - (>•" }
— — J— -**— \».A /
1095
1118
1122
1143
1140
0.058 - 1106
0.055 - 1161
0.056 0.005 1147
0.056 0.005 1141
0.056 0.005 1138
0.056 0.005 1141
0.045 0.005
Last 5 hours.
Theoretical values, based on flue gas leaving aqueous prescrubber at 50 C
with 90 percent of the contained NO  in the form of NO.

-------
extent of 39.7%.  It may well be, however,  that the basic sulfite,  ZnSO-- ZnO,
was actually the major sulfite species formed, especially in view of the limited
conversion of zinc oxide to the sulfite .  This is suggested by the observed mode
of thermal decomposition of the sulfite,  which yields the basic sulfite as an
intermediate (Reference 3):

            2  ZnSO-2 1/2 HO-^-®>ZnSO.ZnO + SO   +  5 HO§         (3)
            ZnSO,. ZnO--»2 ZnO  + SO,                               (4)
                 j                        £»

            Although the gaseous mixture utilized in the preliminary studies
(0. 3 vol-% SO_ in N_) contained no oxygen, an additional analysis of the bed
             M     M
material following absorption showed that 2. 34% of the absorbed SO^ had been
converted to sulfate. Presumably, the formation of sulfate arose from occluded
air on the oxide used,  and from exposure of  the solid to the atmosphere on
several occasions during the absorption period.
            Beginning with Run 2 the Kadox-15 was screened with standard Tyler
screens, and the combined fraction in the range  60 to 200 mesh was used for
absorption.  The data for Run  1, observed on the previous contract, are included
for comparison.  In this run absorption fell to 90% in about 9 hours, but the run
was nevertheless continued for 15 hours, as discussed in  detail in Volume I of
this report.
            The relatively efficient absorption in Run 2, in which CO., was incor-
porated into the influent gas, is attributed to the increased water content in the
gas phase over that present in Run 1, as shown in Table 2.  In the practical case
the flue gas leaving the prescrubber  should be saturated with water vapor at
about 50 C,  and would then exhibit the following composition:

-------
                                 Flue Gas (Excluding Fly Ash';
                         Entering Prescrubber    Lieaving Prescrubber
                                (vol-%)	(vol-%)  50°C
74.9
14.7
7.25
2.8 .
0. 30
0.05
70.9
13.9
12. 18
2.7
0.2S
C.05
            N2
            co2
            H20

            °2
            so2
            NO
              x
From the above data it appears that even in Run 2 the influent gas to the fluid
bed absorber was not saturated with water vapor.  The water sparger was
maintained at 50°C for the run,  so that presumably the contact time was in-
suf'icient to permit the gas to become saturated.
            The fact that rather extensive absorption occurred in Run  2 indicates
that either zinc carbonate does not form, or that if formed it reacts readily with
SO2 to displace CO2<   Some caking of the  solid in the lower portion of the
absorption bed was noted at the conclusion of the run, but  most of the  solid
remained highly fluid throughout the run.  Screening  of the solid showed that
little or no attrition occurred during the 27 hour absorption period:

                               Reactant               Product
                               Kadox-15              of Run 2
           Mesh Size           (wt-%)                (wt-%)
           Over 60                 -                       1.7
           60 to 100              52. 1                    53. 3
            100 to 150            24. 3                    23.0
            150 to 200            23.6                    19. 3
           Under 200              -                       2.7

-------
            Run 3 represents the first run in which oxygen was incorporated
into the gas phase.  The run was complicated by the appearance of liquid
water in the bed, and this resulted in partial defluidization.  Probably for this
reason breakthrough ( >10 percent of the influent SO^ escapee in the exit gas)
occurred before the 27 hours realized in Run 2.  Ths presence of water in the
bed is attributed to the relatively high water content in Run 3(11 vol-% of the
influent gas),  which corresponds to a nearly saturated gas (12. 18%,  see Table
2) at the bed temperature of 50  C.  It may be presumed that either temperature
fluctuations within the bed resulted in the condensation of water from time to
time, or that  the zinc sulfate formed was deliquescent under the conditions
employed.   In the latter case, the process of water condensation would be
further promoted by the fact that liquid water aids in the further formation of
sulfate.  This is indicated by the dramatic decrease in sulfate content which
occurred in Run 4, which was carried out at 55 C, arid in which no water con-
densation occurred.
            At the conclusion of Run 3, only the top  20 vol-% of the bed material
was fluid enough to be poured from the tube; the remainder was loosened with a
spatula and dried overnight in a stream of nitrogen.  The water lost in this
manner represented 31 wt-% of the total bed material after absorption.  The
dry solid was then removed and ground in a mortar  prior to analysis.  Separate
analyses of the fluid portion and residual portion of  the bed material are shown
in Table 1  as  3A and 3B, respectively.   The sulfate content was approximately
the same in both samples, but the sulfite was located largely in the  defluidized
or lower portion of the bed.  This is probably to be  attributed to the insolu-
bility of the sulfite, which remained at the site of formation once the bed had
defluidized, whereas the soluble sulfate  would tend to migrate as  a result of
digestion (solution followed by precipitation).
            In Run 4 the formation of liquid water in the  bed was prevented by
raising the bed  temperature to  55  C, while  maintaining the sparger temperature
at 50 C.  The higher bed temperature could probably be realized in the practical
case through  adjustment of the  amount of heat removed,  if any, from the hot

-------
 rcj^tn..i^.'.c. . ziuc c:;.u3 before returning it to the absorber,  by taking advantage
 (.2 ihe jucotheriTuc nature of Reaction 1, or by providing an esfiernal source of
 heat.  Because of the  relatively long absorption periods required for break-
 through in Runs 2 and 3,  it was decided to conduct subsequent runs for an
 arbitrarily selected 12 hours.  This time period is sufficient for a degree of
 absorption commensurate with reasonable accuracy in the analytics,;, valuee
 shown in the table.
            As can be seen from a comparison of Runs  3 and 4,  a remarkable
 reduction in  sulfate formation occurred when the gas was maintained above its
 dew point.  The bed material from Run 4 was  completely free-flowing.
            In Run 5, NO was incorporated into the gas, with the result that
 an increase in sulfate formation occurred.   It is not known whether the NO
 reacted primarily as such in causing the formation of sulfates or whether it
 wao first partially oxidized to NO2 by the oxygen present in the gas.  If it is
 assumed that no oxidation of NO occurred,  it was calculated that 36. 8% of the
 available NO would be utilized in sulfate formation, according to the following
 equation:
     2 NO + ZnS03 • 2-1/2 H2O	- N2O + ZnSO4 • H2O + 1-1/2 K2O      (5)

 If only NO_ were involved, it would require that 18. 4% of the NO be oxidized:
           £•
     NO2 -I- ZnSO3  • 2-1/2 H2O	»-NO + ZnSO4 •  H20 + 1-1/2  K2O       (6)

 These alternatives can be distinguished by utilizing a gas containing NO,  but no
 oxygen.
           In Run 6,  NO2 was included in the gas,  to the extent that 10% of the
NO  consisted of NO~.  Oxidation was somewhat more  extensive in this case,
   X               c*
as noted in the table.   If it is  assumed that no oxidation of NO to NO2 occurred,
it appears  that the action of NO2 is a catalytic one in the sense that 3. 6 sulfate
moieties were formed for each NO2 molecule present.   This could arise as a

-------
NO • t o O i er-
O,N-O, • -!• 2 SO,= —
£ £> \ J
02N-02.
	 K5&NO,
result of Reaction 6,  in which the NO, is destroyed (non-catalytic) together
with a reaction involving the formation of an activated NO,-oxygen complex
(catalytic):
                                                                      (7)

                                                2 S04=                 (8)

Reaction 7 is considered reasonable in view of the diradical nature of the
oxygen molecule,  and the tendency of NO- to dimerize.  Since the free radi-
cal NO does not dimerize, its tendency to form a complex analogous to that
shown in Reaction 7 would be considerably less.  The occurrence of a catalytic
process, such as that shown in Reactions 7 and 8, can be  distinguished from
non-catalytic oxidation, involving pre-oxidation of NO to NO, followed by oxi-
                                                         • C*
dation of sulfite via Reaction 6, by utilizing  NO, in the gas phase in the ab-
                                             £
sence of NO.
           In Run 7, fly ash was present in the bed for the first time.  In-
asmuch as an efficient prescrubber did not form part of the experimental
apparatus, it was  not possible at this time to introduce the ash directly into
the feed gas, remove the  bulk of the ash  by pr escrubbing, and then conduct
the gas containing residual ash into the bed.  Instead, ash was  introduced
(1% of the bed weight) directly into the bed,  and thoroughly mixed by flui-
dization of the bed with nitrogen  before admission of the flue gas.  This
amount of ash represents approximately  6. 5 times the amount which would
be expected to enter the bed if the prescrubber were 99%  efficient in re-
moving the ash, and if the fluidized bed  run had been conducted to break-
through (i. e. , the point at which 10% or more of the SO, passes completely
through the bed).
            The presence of fly ash in the bed did not result  in increased
oxidation.  This can be explained on the basis that no leaching of iron oxide
from the ash occurs in  the absence of a bulk liquid phase, and that the oxi-
dation of zinc sulfite by oxygen is very slow in the absence of heavy metal
ions (e. g., ferric ion) as catalysts.  Consequently s the need does not arise

-------
 in thy c. ise of •-. fluidized zinc oxide absorber for the use of inhibitorsj, such
 •^i, Iiv^roqi iricne{  as antioxidants.
            It will be noted from Tables 1 and 2 that only 0. 32% oxidation
 occurred in the fluidized bed system for a. gaseous mixture containing O_
                                                                      £*
 but no NO  (Run 4), and it is therefore inferred that the oxidation in Run 7
          JC
 is ai-jribv-.table to the presence of NO  in the  gas.  On this basis, Run 8 was
                                   3C
 conducted with metallic zinc incorporated into the bed. In aqueous  solutions
 of S X, sulfites are reduced by zinc dust to give dithionite ica,  S,O ~   which
 acts as a very strong reducing agent (Reference 4).

                                      2
      Zn + ZnSOj. 2-1/2 H2O + SO2 - -*« ZnS2O4 + ZnO + 2-1/2 H2O  (9)
      ZnO + S02 + 2-1/2 H20      -*.  ZnSO3 • 2-1/2 H2O                (1)

If water vapor were to promote Reaction 9,  the zinc dithionite formed might
then effectively reduce the undesired NO  to NO and/or N,O,  which would be
                                       X               £•
less effective in oxidizing sulfite to sulfate:
                                                                      (10)
      ZnS204 + 2 NO	» ZnS03 + SO   + N2o                       (11)

It was found, however, that Reaction 9 does not occur in the absence of liquid
water.  Evidence for this conclusion was obtained on noting that the addition
of a sample of the bed material to  acid, following the completion of Run 8,
gave no precipitate of free  sulfur.  The formation of sulfur in this manner
is characteristic of the decomposition of dithionites:
                          H+
  2 ZnS2O4 + 3-1/2 H2O  "  •• Zn(HSO3)2 '+ ZnSOy 2-1/2 H2O + Si     (12)

Run 8 may therefore be considered as a duplicate of Run 7, and. it will be
noted from Table 1 that the results for these two runs are largely in agree-
ment.  An interesting observation  that was made in connection with Run 8
was that during the course of the run the zinc metal  tended to accumulate in

-------
the upper part of the tube. At the end of the run most of the gray metal had
separated in this manner.   The residual bed material (60-200 mesh) was
nearly white, although speckled with fly ash. Apparently the small particle
size of the metal (about 4 microns) counteracted its relatively high density
(Zn, 7. 14 g/cc; ZnO, 5.47g/cc).
            Run 9 represents what is considered to be the ultimate solution
to the oxidation problem in the zinc  oxide fluidized bed system.  As a result
of a series of experiments involving the oxidation of iodide ion by NO ,  which
                                                                  Jw
will be discussed below in some detail,  Run 9 was conducted with ferrous sul-
fate present in the sparger.  Under  these conditions NO, is reduced to NO
(Reference 5):
      2 FeSO4 + H2SO4 + NO2 - «• Fe2(SO4)3 + H2O + NO            (13)

Fortunately, the oxidation of ferrous ion by oxygen is inhibited by the presence
of sulfuric acid  (Reference 6),  so that loss of ferrous ion due to this undesired
oxidation reaction is not appreciable.
            In the practical case,  the sulfuric acid required for Reaction 13
would be provided to the prescrubber by the SO- content of the  incoming flue
gas, and the ferrous sulfate through the use of scrap iron:

      Fe + Fe2(S04)3 - •» 3 FeSO4                                 (14)

However, the iron needed for effecting Reaction 14 would not be added directly
to the prescrubber, inasmuch as additional experiments have shown that SO,
is partially reduced to free sulfur by the direct action of metallic iron in acid
solution.  Rather, the iron would be located in a recirculating side  stream, so
that only ferrous iron would be fed to the prescrubber proper.   The optimiza-
tion of this  system will require additional experiments.
            The experiments which ultimately led to the use of ferrous ion
in the sparger are shown in Table 3.   The main purpose of these experiments
was to determine the relative activity as oxidizing agents of the three oxidi-
zing components of flue gas;  i.e. 8 NO, NO~» and O-.  This could have been

-------
                              TABLE 3
OXIDATION OF AQUEOUS IODIDE ION BY SELECTED
FLUE GAS COMPONENTS UNDER VARIOUS CONDITIONS
Run Gas , 12 Formed Reduction
No. a Components (ml Na?S?O^J sf Gas (%)
ld
2d
3d
d
4
5d

6d
7d

8e
f
9
10f'g
llf'g
12f'g

13f,h

°2
NO
N02

NO 4 O2
NO, + O,
2 2
NO2 + NO
NO, + NO + O_
2 2
N0402

NO + O2
NO
N02
NO, + NO + O-
2 2
NO, + NO * O.
2 2
0.00 0.00
0.32 5,25
1.32 97.4

0.41
1.40

1.80
1.87

0.58

0.58
0.06 81.0
0.1 1 91'. 0
0.25 87. 01

0.75 60. 01

Each run was conducted at room temperature for a duration of one hour.
Whenever O_ was present it constituted 2. 68 vcl-% of the gas; NO,  0. 056%;
and NO_, 0.005%.   The diluent was always  N,,  with the total flew main-
       L                                    f,
taiiied at 600 ml/min (cf. Table 2).
The  ly was formed in a bubbler which,  for each  run, contained 250  ml of
a stock solution prepared from 4. 4 g of 47% HI and 32C g of XI diluted to
4 liters with water.  The L, was determined by titration with standard
0.1107 N Na2S2O3 to the starch endpoint.
In this run the 1^ bubbler was located at the position normally occupied by the

-------
                            TABLE 3 (continued)

FOOTNOTES - continued:

eln this run a water sparger was located at the normal position us©d during
 a run, with the I- bubbler immediately downstream.
 In this run both the sparger and bubbler were located in their normal posi-
 tions  (i. e., sparger upstream of the fluid bed and the I- bubbler downstream).
%n this run the sparger contained 1% Fe in 5% H^SO..
 In this run the sparger contained 1% Fe in 1% H,SO4.
 Based on NO + NO_  content only.

-------
 done ty conta.cL.'ng the gases with solutions of sul£ites8 and noting the amounts
 ol suilaL^ fern cd under various conditions.  However8 even the relatively
 simple turbidimetric  method of analysis for  sulfate is tedious for large num-
 bers of samples, and moreover is subject to a relatively large (5%) experi-
 mental error.
            The  approach taken to the problem at hand consisted in utilizing
 aqueous acid solutions of iodide ion as an indicator for the  relative activities
 of ti.e various oxidizing gases.  One advantage of this system was  that an
 immediate color change occurred (formation of iodine) if oxidation occurred.
 Eque.lly important, the liberated iodine could be rapidly and accurately de-
 termined by titration with standard thiosulfate  solution to the starch endpoint.
            The  experimental  conditions used in carrying out the various runs
 shown in Table 3 are largely given in the footnotes to the table.  In Runs 1 to
 3 the water sparger that would normally have been used in carrying out the
 runi, shown in Table 1 was replaced by the iodide bubbler, with no apparatus
 downstream of the bubbler.  The results of these runs  show that oxygen alone
 is relatively ineffective  as an  oxidant,  although the stock solution (see Foot-
 note c of the table) slowly became  slightly yellow in color (liberation of iodine)
 on standing, due to the presence of dissolved oxygen.   This required several
 days,  in contrast to the  one-hour period utilized for each of the runs*  From
 Run 2 it would appear that NO  slowly oxidizes iodide ion.  Howevers  further
 experimental results, discussed below, indicate that the observed oxidation
 may rather be attributable to NQ~, formed from NO and dissolved O2, and
 that this reaction is actually promoted by the presence of liquid water.  The
 extent of reduction of  NO  (Run 3) is especially noteworthy.
            Runs 4 to 7 show that an enhancement of oxidizing power occurs
when at least two of the three oxidizing gases are present simultaneously.
In Run 4,  for example, the observed value of 0.41 is larger by 0.09 than
the sum of the  results obtained from Runs  1 and 2.   If such an enhancement
is assumed to operate chiefly on the less (or least) active oxidant3 rather

-------
than mutually,  Run 4 would then be interpreted as showing an enhancement
for oxygen of 0. 09.  In the same manner, Run 5 shows an enhancement for
oxygen of 0.08, Run 6 shows 0.16 for NO8 and Run 7 shows as one possibility
0. 08 for O, and 0.15 for NO.  The observed enhanced activity can be explained
          b
as resulting from the formation o£ complexes,,  such as ON-O.»» a N-O78 and
                                                          &    & j
O-N-O-.  However,, in the case of Run 4 a more probable explanation is that
NO- formation occurs rather than complex formation.  It will be noted that
the enhancement of NO is approximately twice that of O~.
                                                    £n
           Run 8 is interesting in that the inclusion of a water sparger into
the system upstream of the bubbler results in a further increase in the oxi-
dizing capacity of the gas (0. 58 for Run 8 vs 0. 41  for Run 4).  This is in-
terpreted (compare the above paragraph) as  indicating that the formation of
NO_ is promoted by the presence of liquid water.  That a similar  promotion
   L*
does not occur in the presence of-water vapor is  shown by Run 98  in which
the fluidized bed reactor was interposed between the sparger and  bubbler.
           In Runs 10 to 12 the sparger contained 5% sulfuric acid, and
therefore more closely approximated a true  prescrubber liquor composition
than does pure water, since flue gas normally contains a small amount of SO-.
Powdered iron was added to the sparger just before each run, and hydrogen
was slowly liberated during the run with the  formation of ferrous  sulfate:
      Fs + H2S04	* FeS04 + H^'                                  (15)
As can be seen from the table, a dramatic decrease in oxidizing capacity of
the gas  entering the bubbler occurred in each case.   The effect was some-
what less when a 1% sulfuric acid solution was used (Run 13).
            The direct use of metallic iron in the sparger was subsequently
found to result in the  partial reduction of SO., to elemental sulfur, as pre-
viously noted, and for this reason an additional experiment (not shown in
Table 3) was conducted in order to establish the approximate sparger con-
ditions to be used in Run 9 of Table 1.  In this experiment a 1% solution of
ferrous sulfate was prepared by allowing metallic iron to react with 5%
sulfuric acid (Reaction 15) until all of the iron was consumed-   The resulting

-------
      anl colorless solution was then heated to the normal sparger tem-
        j ox 5C°Cl, the bubbler was placed immediately downstream of the
        in the manner of Run 8, Table 3, and a mixture of NO,,  NO8 and
O2 in the proportions  used in Runs 7, 12, and 13,  Table 3, was passed
through the system at 600 ml per hour.  A heating tape was used to prevent
water condensation between the sparger and bubbler.   The following results
were obtained, in terms of ml Na-S-O,  required for titraticn of the liberated
iodine after the times indicated: 1st hour, 0; 2nd, 0,09; 3rds 0.06; 4th, 0.08;
5th, 0.05;  6th, 0.16; 7th, 0.30; 8th, 0.28.   These results indicate that the
incorporation of ferrous sulfate into the sparger should be highly effective
in diminishing the oxidizing capacity of the  gas exiting from the sparger,
particularly if fresh solution were used  after about the fifth hour  of the run.
In carrying out Run 9  of Table 1,  the solution was replaced after  the sixth
hour.
            That oxygen did not oxidize ferrous ion at an appreciable rate
under the conditions employed in the experiment described above was shown
by noting that if all of the available oxygen had reacted, 10 ml of  Na^S.O-
would have been required for titration of the liberated iodine after only 4. 9
minutes.
            The main conclusions to be derived from the data in Tables 1
and 2 are as follows:
            •     For  a gas nearly saturated with water at 50 C,  more
                 than 50 g of SO- is absorbed per 100  g of zinc oxide
                 before breakthrough occurs.
            •     If liquid water is permitted to condense in the bed,
                 extensive oxidation results.  In the worst cases
                 approximately 14% of the absorbed SO2 was converted
                 to sulfate (Reaction 2)  under conditions such that 31%
                 of the total weight of the bed following absorption con-
                 sisted of condensed water.  Bed defluidization ?lso
                 occurs.

-------
           •     Oxygen does not cause appreciable oxidation in the
                 absence of a condensed phase,,  It was found „ for
                 example,  that with a a pa A- g or temperature of SO C
                 and with the bed maintained at 55 C to prevent
                 water condensation,  the inclusion of oxygen in the
                 gas resulted in only 00 32% of the absorbed SO-
                 being converted to sulfate 0
           a     The presence  of fly ash in the bed does not promote
                 oxidation.
           o     The presence  of NO~ promotes oxidation, but this
                 component of the gas can be removed in the pre-
                 scrubber through the use of ferrous  ion.
           It will be noted from Tables  1 and 2 that the observed extent of
oxidation for Run 2 (no oxidizing component), Run 4 (oxygen present), and
Run 9 (fly ash and all flue gas components present) was  in all cases approxi
mately the same (0. 3-0.4%).  This would suggest that in all probability no
oxidation at all occurred in any  of these  runs.  The small percentages of
oxidation reported were based on experimental turbidimetric readings of 2
to 3 ppm of sulfate on a  scale ranging from 0 to 300 ppm, so that in all
cases the readings were within the range of experimental error.
      D.   ABSORPTION OF
            The bed materials from the various runs described above have
been stored for future studies in which it could be determined whether or
not NO  had been absorbed.  No work was done in this area on this program.
      3t
The experiments of interest involve the thermal decomposition of samples
of the bed materials, and the  collection,  identification, and analysis ox the
liberated gases.  If NO  has been absorbed, various nitrogen oxides should
be evolved on heating, perhaps at relatively low temperatures .

-------
III.   ZTNTC OXID,: PARTICLE SIZE AND ACTIVITY
      A.    INTRODUCTION
            The initial work during Phase IV was in part directed to the
problem of particle size degradation in connection with the use of ainc oxide
as a fluidized absorbent for SO2«  Earlier results (Reference 1)  indicated
that the absorption step is highly dependent on particle sizes and that  ex-
tremely small particles are required for good absorption.  In particular,
New Jersey Zinc's Kadox-15,  of 0. l/i mean particle size,  was much  more
effective than their XX-504, of 99. 6% purity and 1. 5 ft.  The use of such
small particles is considered to be highly undesirable, both because of the
inevitable loss of solids by entrainment from the bed,  and because the cal-
cination of zinc sulfite, as  studied extensively by Johnstone and Singh (Ref-
erence  7),  does not yield an oxide of such a small particle  size for  reuse in
the bed.  In general, the calcination of the sulfite at 450° to 775°C yields
material exhibiting particle sizes in the  range of 60 to 270 mesh, or 250ft
to 53 p, respectively.  Even this material,  however, is undesirably small.
Recent  studies have shown  that it should be possible to utilize relatively
coarse material (1/16 in. to 1/8 in. pellets) for absorption.
                                             /
      B.    SMALL PARTICLE STUDIES (Kadox-15)
            Although Kadox-15 has been found, by microscopic examination,
to exhibit a mean particle size of 0. 1 p.  (data provided by the New Jersey
Zinc Company), it was found that agglomeration of the particles  readily
occurs, so that in fact a much larger particle size is realized.   An analysis
of Kadox-15 with standard Tyler screens gave the following results:
                       Mesh Size           Wt-%
                       Over 65        .      1.6
                       65 to 100
                       100  to 150           18.0                                    |
                       150  to 200
                       Under 200

-------
In view of the above analysis the New Jersey Zinc Company was consulted,,
with the only comment of interest being that a  relatively high humidity would
be expected to contribute to the agglomeration of the oxide.  However,, the
material was  stored in a closed plastic bag which was in tern contained in a
metal can, and the screen analysis was carried out under conditions of low
humidity.  It would appear,, therefore,,  that the material inherently tends to
agglomerate,  and that earlier work at Aerojet with SLuidisad Kadox-15 (Table
1, Run 1} involved particle sizes in the ranges noted above, particularly in
view of the fact that the  gas entering the bed had been passed through a water
sparger.
           Although the effective particle size of Kadox-15 is much larger
than had been supposed, it is somewhat smaller than that used by Johnstone
and Singh in their sodium  scrubbing/zinc oxide process for the removal of
SO2  from flue gas (Reference 7).  Figure 3  shows a screen analysis of dried
zinc suifite,  as employed by Johnstone, of the same material after it had been
passed through a small hammermill, and of the  zinc oxide resulting from the
calcination of the  suifite.  Kadox-15, which has been included for comparison,
shows  roughly the same range of particle sizes as that of Johnstone's milled
suifite.  It is  of interest that an increase in particle size occurs when the sui-
fite is  calcined, even though the process is  a degradative one,  involving the
loss of both water and SO,.  This further attests to the tendency of zinc oxide
to agglomerate, and indicates that any particle attrition which occurs during
the absorption of SC>2 by fluidized zinc  oxide may be effectively offset on cal-
cining the resulting suifite.
            The zinc oxide obtained on calcination of the suifite was described
by Johnstone as being highly active in the sense that it dissolved in a specially
prepared aqueous solution of sodium suifite-sodium bisulfite in a "few seconds."
The particular medium chosen for dissolution of the oxide closely approximates
the spent scrubber liquor  obtained in the Johnstone Zinc Oxide process.  The
regeneration  step in this process involves the conversion of bisulfite to  suifite
through the use of the oxide;
      2 NaHSO, 4- ZnO + 1-1/2 H,O	s* ZnSO,° 2-1/2 H,OA + Na-SO,  (16)
              J                 £,             3        £,  u     £t   J

-------
                   (Data from Johnstone, Reference ?, except for Kadox-15)
        100
8
4)
§
•tf
O
.s
$
V
0)
u
N
6
s)
O
                          Cadox-15

                          Sine Oxide
20
                                  100                  200

                                 Size of Opening in Microns


                   SCREEN ANALYSIS OF ZINC SULFITE AND ZINC OXIDE


                                         Figure 3

-------
The zinc sulfite obtained in Reaction 16 is calcined to regenerate the oxide,
and the sodium sulfite solution is returned to the scrubber.  In the case of
Kadox-15, which is of smaller particle size  than Johnstone's calcined oxide,
it was found by Aerojet that solution in the particular sulfite-bisulfite medium
of interest required at least two minutes, and that a  small amount of material
had failed to dissolve during ten minutes, at which time the  suifite began to
precipitate  in accordance with Reaction 16.  This phenomenon was considered
of sufficient interest to warrant  a further brief study.
            The results  of a series of  qualitative tests involving the  solution
of a variety of zinc oxide samples  in aqueous sodium sulfite-sodium bisulfite
are shown in Table 4.   In order  to determine whether magnesium and/or
calcium might be substituted for zinc in the  Johnstone Zinc Oxide process,
basic compounds of these elements were also included in the study.  With
respect to this point it was found that magnesium might be substituted for
zinc, but that calcium is unsuitable in that the  hydroxide does not dissolve.
Since the medium used was acidic  (pH approximately 5. 5) the surface  of the
hydroxide must have reacted,  but the rate of formation of calcium sulfite on
the hydroxide surface was apparently rapid enough to prevent appreciable
solution of the hydroxide.  The use of  magnesium in  the Johnstone process
is  desirable from an economic standpoint, but  was not further  investigated,
pending preliminary results relating to the possible  elimination of the dis-
proportionation reaction when magnesium sulfite is calcined.
            The results  shown for zinc oxide are interesting in that  they
clearly indicate that calcination greatly improves the ability of the oxide
to  dissolve.  Since the solution of the oxide is accomplished in an aqueous
medium,  the heating of the oxide in the present study must involve the loss
of one or more surface contaminants other than water.  It may also be noted
from the data that  small particle size  favors solution of the oxide, and that
heated material of small particle size gave the best  results, that is, rapid
and complete solution.

-------
                                 TABLE 4



             RATE OF SOLUTION OF SELECTED METALLIC

    OXIDES AND HYDROXIDES IN AQUEOUS SODIUM SULFITE-SODIUM

                           BISULFITE SOLUTION

                  C* = 3. 00, 3** = 2. 50, S/C = 0. 83

            Oxide or Hydroxide added = 3/8 mole per mole of NaHSO,
                              Time Required for
Material Added
ZnO, Kadox-15
same, over 32 mesh
same, 65 to 100 mesh
same, under 200 mesh
same, under 300 mesh
same, calcined*^
same, under 200 mesh,
calcinedJ
same, under 300 mesh,
calcinedJ
ZnO.XX-504 .
ZnO, Allied Chemical
same, calcined^
Zn(OH)2
Calcined ZnSO3- 2 1/2H2O
MgO
Mg(OH)2
Ca(OH)2
Dissolution of Solid (min)
2
2
1
1
3/4
1/2
1/2
1/2
4
2
1 1/2
1 1/2
3/4
1 3/4
1/2
Remarks
a
a
b
b
b
c
d
' d
a, f
b
c
b, f
d
c, g
d, e, g
e, h
  Solution moderately hazy

  Solution slightly hazy

  Solution very slightly hazy

  Solution clear
Q
  Solution slightly yellow

 Solution very slightly yellow
** Precipitate began to form in about
  2 min; heavy crystalline precipitate
  in 20 min.

  Solid did not dissolve.

J  Heated at 550°C for 2 hrs.
*
  moles of Na per  100 moles  of water.

	moles of SO  per 100 moles of water.

-------
            In summary,  it may be stated that the overall results of Table 4
are considered significant in indicating that magnesium, but not calcium,
might be substituted for zinc ia the Johnston© Zinc Oxide process8 and that
calcined zinc oxide might be a better absorber for SO- in a fluid'ized bed ap-
plication than zinc oxide which has been exposed, to ais- for an appreciable
length of time (if, in fact,,  the rate of dissolution of zinc oxide in aqueous
media is a measure of its ability to sorb SO,}.
     C.    LARGE PARTICLE STUDIES (Pelleted Kadox-215)
            1.    Introduction
                 In addition to the studies described above relating to
Kadox-15, a brief experimental program was conducted in which it was
shown that considerably larger particles could be used for absorption, and
that under the proper experimental conditions very little particle attrition
occurred.
            2.    Initial Fluidization Experiments
                 All of the experiments described below were conducted with
1/16 - 1/8 in. particles,  and at superficial gas velocities in the  range of 2. 7
to 3.0 fps.  The fluidization tube, 15 in.  x 7/8 in. ID, was  that designated as
Item 9 in Figure 1.  Air was used for fluidization.
                 Initial experiments were conducted with 30 g samples of
New Jersey Zinc's pelleted zinc oxide, Kadox-215, and with suitably screened
commercially available zinc S'olfite. Considerable attrition occurred in the
case of the oxide during a two-hour fluidization period, with only 33% of the
final material retained on a 10 mesh (1/16 in.) screen.  Of the material
passing the screen,  37% was retained on a 60 mesh screen.  On the other
hand, zinc sulfite was highly resistant to attrition. After two hours,  the
following screen analysis was obtained:
                       Mesh Size          • Wt-%
                          -S- 10
                        -10-1-24
                          - 24

-------
                 Subsequent experiments involved zinc oxide derived from
    calcination of the commercial sulfite at 275 C.  Although this material
•was friable,  it possessed sufficient mechanical strength to permit fluidization
without observable attrition during the 1 -hour test period employed;,  When
mixed with zinc sulfite,  the fluidized mass readily separated into two solid
phases, with the oxide appearing as the  upper phase.  Overall,  the  calcined
oxide was found to be much more  resistant to attrition than the commercially
pelleted Kadox-215.
                 Another experiment with the calcined oxide involved a
preliminary absorption of sulfur dioxide (from gaseous SO^-N^-H-O) prior
to fluidization.  The purpose of this experiment was to determine the resis-
tance of sulfite-coated oxide to  attrition.  The coating process was arbitrarily
terminated after 15% of the theoretical amount of sulfur dioxide had been
absorbed.  Subsequent fluidization with air at 2. 7 fps  for one hour resulted
in material which  exhibited the  following screen analysis:
                       Mesh Size           Wt-%
                          + 10
                        + 10-24
                          - 24
            3»    Attrition During Absorption of SO, in a Fluidized System
                 a.    Apparatus
                       A fluidized bed absorption system was improvised,
largely from equipment that had been used in studies relating to the oxidation
of Na_SO,-NaHSO, solutions.  The resulting system was not considered ade-
quate for long-term experimentation, but did serve to  provide basic fluidized
bed attrition and absorption data.
                       A mixture of N~ and SO- was saturated with water at
45-50  C in a packed column.   The saturated gas flowed to a fluidized bed ab-
sorber made from 43 mm ID glass tubing.  A coarse glass frit was used

-------
as a fluidizing plate and a 300 mesh screen served as an outlet filter on the
absorber.  Zinc oxide was slug fed at an approximate rate of 1 g/min (later
at 1  g/2 min).  The reacted absorbent was continuously withdrawn through a
glass overflow tube positioned 3 inches above the fluidising plate in the initial
tests.  The tube was later raised to 6 inches above the plate.  The fluidized
bed and the water-saturated gas upstream of the saturator were maintained
at temperati
the system.
at temperatures in excess of 55 C in order to prevent water condensation in
                       A muffle furnace was used to regenerate zinc oxide
from zinc sulfite formed from the oxide in the absorber.  The sulfite was
placed in Petri dishes and decomposed at 350-425°C while the furnace was
purged with nitrogen.
                 b.    Experimental Work
                       Table 5 shows the results of the attrition tests.
Kadox-215 zinc oxide pellets were used in Run 1 with a particle  size of 50%
-12+16 mesh and 50% -16+24 mesh.   The bulk density of this mixture was
1.1 g/ml (68. 6 Ib/ft ). An initial charge of 90 g was placed in the absorber
and was  then soaked for 3 hours with a water-saturated mixture of N, and
SO~,  fed at a rate of 26 liters/min N- and 360 ml/min SO2.   After soaking,
the absorbent analyzed 18.4 wt-% SO,.  The particles partially converted to
zinc sulfite in the absorber were hard and firm, whereas the unreacted feed
pellets were soft and could be pulverized with finger pressure.  It was ob-
served that a small increase in pellet diameter occurred during absorption.
                       In Run 1 the unreacted, relatively soft zinc oxide
pellets were fed at a rate of about 1 g/min.  The absorbent overflow was
set at 3 inches  above the fluidizing plate.  A mixture of 50  C-saturated gas
at a rate of 60 liters/min N- and 280 ml/min SO- (0.47% SO,) was fed to
the absorber.  The feed rate was equivalent to a superficial velocity of
2.86 fps.

-------
                                                    TABLE 5
                            ATTRITION EXPERIMENTS DURING SO2 ABSORPTION
to


Fluid, Bed-Depth, in.
ZnO Feed, ;=g/min
ZnO Residence Time,
hr
Product Rate,, g/hr
Feed Rate*, g/hr
Product Sieve
Analysis, wt-%:
+ 12 mesh
-12+16 mesh
-16+24 mesh
-24 mesh
-24+60 mesh
-60+100 mesh
-100 mesh
Run No. 1
(Using As -Received ZnO Pellets)
Time, Hr.
0
3
1
«.
-
-

-
50
50
-
-
-
-
1
3
1
0.92
81.9
60.6

0.1
36.0
57.3
6.6
-
-
-
2
3
1
0.90
83.8
65.8

0.4
38.2
53.1
-
8.2
0.09
0.01
2.5
3
1
0.82
91.8
72.4

0.4
32.5
56.1
-
10.9
0.1
0.02
Run No. 2
(Using Regenerate^ ^nO Pellets)
Time, Hr.
0
3
1
—
-
-

-
50.0
50.0
-
-
-
-
1
3
1
0.89
84.5
70.8

0.2
46.2
49.3
-
4.3
0.01
0.006
2
3
1
0.84
89.6
80.1

0.3
42.4
53.0
-
4.3
0.02
0.003
2.5
3
1
1.4
52.6
47.0

0.1
37.1
58.3
-
4.4
0.02
0.004
3.5
6
1
_
74.0
-

-
-
-
-
-
-
-
4.5
6
1
1.8
71.8
62.5

0.3
43.0
52.9
-
3.8
0.04
0.004
5.5"
6
oa
3.
42.
36.

08
43.
51.
-
4.
0,
0.
    #Calculated from weighed product.
    Fluidized bed temperature was in the range of 63-70°C.
    Product left in bed after Run 2 with 6" high overflow =  130.8 g.

-------
                       In Run 2,  the absorbent feed was prepared as follows:
Kadox-215, of the same particle size distribution that was used in Run 1, was
soaked as a static bed for several hours in a mixture of N^-SO^-H-O.  These
pellets had a bulk density of 1. 44 g/ml (90 Ib/ft ), and were hard and firm.
The pellets were subsequently decomposed by heating to 425 C for  an hour
in shallow Petri dishes.  The material did not go through a noticeable plastic
state,  and did not flow  into an agglomerated mass.   Practically all of the re-
generated pellets  retained their original shape and were quite firm (much
firmer than the original Kadox-215  zinc oxide pellets but softer than  the zinc
sulfite pellets).  About 4 wt-% of the pellets puffed up to almost three times
the original diameter and were easily pulverized with finger pressure.  It is
quite possible that presence of free water caused the change in physical charac-
teristics.  Analysis of both types of regenerated pellets showed 0 wt-% SO,.
                                                                   3
The bulk density of the regenerated pellets was  1.14 g/ml (71.0 Ib/ft ).
                       An initial charge of 90 g of regenerated Kadox-215
with a  particle size of 50% -12+16 mesh and 50% -16+24 mesh was soaked with
N£-SC>2-H2O for about an hour in a static bed.  After soaking, the absorbent
analyzed 11.9 wt-% SO-.   The bed was then fluidized and regenerated zinc
oxide pellets were fed at a rate of about 1  g/min.  The gas rate was 60 liters/
min N, and 180 ml/min SO-,  equivalent to 2.86  fps superficial velocity.
Table  5 shows the data for Runs 1 and 2.
                       The primary purpose of these tests was to determine
the feasibility of using commercially available zinc oxide pellets for  absorp-
tion of SO- in a fluidized bed.  Of particular interest was an indication of the
extent of attrition that could be expected.  Although some absorption  data in
the form of % conversion of the zinc oxide to zinc sulfite were obtained,  no
attempt was made to  optimize the system,  or to determine process conditions
for a practical fluidized zinc oxide system.

-------
                        The following conclusions were drawn from the re-
 sults of the experiments described above:
                        •     Both zinc sulfite,  produced from commercial
                             pellet-size zinc oxide,  and the zinc oxide
                             regenerated from  this sulfite show con-
                             siderable resistance to attrition at particle
                             sizes of -12424 mesh and. at gas velocities
                             of 2. 9 fps.
                        •     The elutriated fines entrained in the absor-
                             ber exit gas amounted to less than 0, 02% of
                             the product rate (reacted absorbent leaving
                             absorber).
                        •     Regeneration can be accomplished without
                             recourse to fluidization if this appears de-
                             sirable.
                        •     Since the regenerated pellets did not ag-
                             glomerate during calcination in a  static
                             system,  it appears that regeneration of
                             pellets could be accomplished in a static
                             or rotary calciner.
IV.   SCREENING OF SELEC TED FLUIDIZED BASIC MATERIALS AS SO?
      ABSORBENTS'
      A.    INTRODUCTION
            The preceding discussion has dealt with the use of fluidized zinc
oxide as an absorbent for SO_.   In principle, any sufficiently basic material
should also be effective for SO-  absorption,  and on this basis a brief screening
program was  conducted involving bases other than zinc oxide.   The  results of
this work are presented in the following sections.

-------
      B.   ALKALI AND ALKALINE EARTH SULFITES
           A magnesium system of interest was considered to be that in-
volving conversion of the sulfite to the bisulfite or pyrosulfite, followed by
thermal decomposition of either of these products to yield SO2 and the
suJLfite:

           MgSOy 3 H2O 4 SO2 -- e» Mg(HSO3)2 + 2 H2O            (17)
      or   MgSCy 3 H2O + SO2 - «* MgS2O5 + 3 H2
-------
 was also maintained at 50 C.  Only 0. 15 g SO? per 100 g of sulfite was ab-           """'
 sorbed before the absorption rate fell rapidly to zero.  Thus, magnesium
                                                                                    f
 sulfite does not appear as a candidate absorbent for SO- under the conditions         —
 employed.
            An additional run was conducted with calcium sulfite, which was
 determined to be 92. 5% pure by iodometry.  Extensive balling of the material
 occurred on attempted screening, and consequently the unscreened sulfite was
 utilized for absorption.  Considerable channeling of the gas occurred during
 the run, even at the highest setting of the vibrator table control (see Figure 1),       _/
 and no absorption of SO, was effected at the 50°C temperature employed.
            Notwithstanding the results  noted above for magnesium and calcium
 sulfites, it was decided to investigate alkali metal sulfites as absorbents, on
 the basis that these compounds are water-soluble,  whereas the magnesium
 and calcium salts are not.  It was considered that high water solubility would
 be effective in promoting the formation  of a mono-layer of water vapor at the
 absorbent surface, which in turn is believed to be an important factor in effec-      ^
 ting the absorption step.  Of some concern, however, was the possibility of
 defluidization of the bed.  This would be expected to occur if the absorption          -j
 were  effected under conditions such that the absorbent became deliquescent.
 For a water sparger temperature of 50  C, the influent gas to the absorber          _
 will contain a partial pressure of about  92 mm of water at saturation, so that
 deliquescence should occur for a given absorbent at all temperatures below
 which a saturated aqueous solution of the absorbent exhibits a vapor pressure
 of 92  mm.
            Sodium sulfite was found to  be ineffective in absorbing  SO, at
 a bed temperature of 60 C.  Fluidization was poor for this system at the
                                     -1                                           —'
 space  velocity employed (about 1140 hr   ), because of the high bulk density
of the  solid.   However, if the sulfite had been an effective absorber some
absorption would have been anticipated,  even under static bed conditions.            —
It was concluded, therefore, that either absorption does not occur, or that
the resulting sodium bisulfite, NaHSO,  or, more likely, sodium pyrosulfite,         _
Na-S-O,.* exhibits an appreciable decomposition pressure of SO- at 60  C.

-------
           Potassium sulfite was found to be deliquescent at 60 C with the
water sparger at 50 C, so that the effectiveness of the dry salt could not be
determined under the conditions employed.  A higher bed temperature  could
be used to counteract deliquescence, but in view of available data relating
to the Wellman-Lord process (Reference 9)8 in which SO, is stripped from
a hot aqueous solution of potassium bisulfite,  it was concluded that at higher
bed temperatures the SO- decomposition pressure of potassium pyrosulfite
                       Lt
would be appreciable.
           From the results of the various experiments described above it
appears that, in general,  sulfites are not sufficiently basic to permit the ab-
sorption of SO, under fluidized bed conditions.
      C.   SODIUM AND CALCIUM CARBONATES
           Attention was subsequently directed to the use of selected car-
bonates for the absorption of SO,.  Solubility data for sodium carbonate are
shown in Table 6 (Reference 10).   The data indicate that with the water spar-
ger at 50 C,  deliquescence would occur at or below about 55 C.  Inasmuch
as sodium sulfite is in general less soluble in water than the carbonate,  it
would be expected that the sulfite would not deliquesce if the carbonate did
not.
           The results obtained for an experiment in which 60 to 200 mesh
sodiurr  carbonate was used to absorb  SO, at 60 C are shown as Run 11 in
                                       £•
Table 7, and the corresponding gas composition and space velocity are shown
in Table 8. For simplicity at this time, NO  was not incorporated into the
gas, nor was CO,, inasmuch as the absorbent was already in the form of the
carbonate.  Run 4, involving zinc oxide with a somewhat similar gas compo-
sition, is also shown in the tables  for comparison.  Both of these runs were
terminated arbitrarily, in view of the long reaction times otherwise involved.
Good fluidization was observed throughout the run involving sodium carbonate,
and no water condensation occurred.

-------
                                                TABLE 6
oo


t(°C)
15
25
32
40
50
55
60
70
80
110
130
THE SYSTEM SODIUM CARBONATE - WATEI
g Na2CO3 per lOOg of
2 Sat. Soln. Solid Phase
16.4 14.1 Na,CO- 10 H,O
£ j £»
29. 4 22. 7 Na2CO3* 10 H2O
45.4 31.2 Na2CO • 10 H2O + Na_CO • 7
48. 8 32. 8 Na-CO • H,O
£• j £»
47. 5 32. 2 Na2CO3« H2O
Na2C03-H20
46.3 31.6 Na,CO -H_O
£» j L*
45.6 31.3 Na2CO3'H2O
45.2 31.1 Na2CO3'H2O
44. 5 30. 8 Na2CO3
40. 9 29. 0 Na2CO3
                                                                                          Vapor
                                                                                         Pressure
                                                                                          (mm Hg)
 12.3


 21.4


 29.0


 43.6


 74.1

 95.0*


121.5


19Z.7


296.2


  1. 19 atm


  2. 25 atm

-------
                                                                         TABLE  7
vO
REACTION OF FLUIDIZED ABSORBENTS (60-200 Mesh)
WITH SELECTED FLUE GAS COMPONENTS
Run
No.
2
4*
11*
14
Run
Temp
Absorbent (°C)
ZnO 50
ZnO 55
Na?CO 60
MgO 55
terminated arbitrarily
Reaction
Time
(hrs)
27
12
19
23-1/2

s°2
Absorbed
(g/lOOg Absorbent)
55.6
24.9
22.0
70.5
S°2
Converted to
Sulfate '(%).
0.41
0. 32
38. 2
14. 1
Absorbent
Converted to
Sulfite (%)
67.6
29.9
25. 0

-------
                              TABLE 8

            GASEOUS COMPOSITIONS AND SPACE VELOCITIES
                      FOR RUNS GIVEN IN TABLE 7

              (Total gas flow approximately 600 ml per minute)
Run
No.
2
4
11
14

N2
75.8
72.2
86.5
71.6
70.9
S02
0.29
0.28
0.28
0.28
0.28
H,O
Vol-%
9.40
10.60
10. 50
11.47
12.18
C°2
14.8
14.2
-
14.0
13.9
°2
.
2.68
2.69
2.68
2.70
                                                        Spaca Velocity
                                                            (hr-1)
                                                             1122

                                                             1140

                                                             1137

                                                             1149
ff
 Theoretical values, based on flue gas leaving aqueous prescrubber at
 50°C.
                                                                            i

-------
V.
                      From the results shown in Run 11, Table 7,  it is evident that
           extensive oxidation occurred.   This is probably to be attributed both to the
           highly basic nature of sodium carbonate and to the appreciable water solu-
           bility of both the carbonate and sulfite, inasmuch as these circumstances
 ~~"         are favorable to both the catalyzed and uncatalyzed oxidation of sulfite at
           high pH (Reference 11).
                      Attempts to absorb SO, with fluidized calcium carbonate were
                                           £3
           unsuccessful.  Inasmuch as sodium and calcium carbonates are both derived
           from strongly basic oxides, it would appear that the lack of absorption in this
           case  may be attributable to the marked water-insolubility of both calcium car-
 _         bonate  and product calcium sulfite.
                D.    MAGNESIUM OXIDE
x~                    In Run 14.of Table 7,  results are  given for 60 to 200 mesh mag-
           nesium oxide.   Neither the oxide  nor  the sulfite are appreciably soluble in
 '--         water in this case, and the run could  therefore be conducted at a lower
           temperature than that used for sodium carbonate.  In order to prevent simple
                                                        i
 ^,         condensation of water in the bed (as opposed to condensation due to deliques-
           cence)  it was necessary, as in previous  runs, to maintain the bed somewhat
           above the sparger temperature of 50  C,  and the run was accordingly conducted
           at 55 C.  It was anticipated that carbonate formation might occur,  in view of
           the fact that magnesium oxide is a strong base, and consequently CO? was in-
 "~         corporated into the gas, as shown in  Table 3.   Run 2, involving zinc oxide,  is
           included for comparison in Table 7,  inasmuch as both Runs 2 and 14 were con-
 —         ducted  to breakthrough  (the point at which 10% of the SO, permeated the bed).
                                                                c»
                      From the results shown in Table 7, it is evident that the mag-
 "~         nesium system  is of interest as a potential absorbent for S'O,.  Absorption
           was extensive in this case  (70. 5 g per 100 g of MgO),  and moreover the extent
 —         of oxidation was considerably less than that observed for sodium carbonate.
           It may  be noted at this point that the weight gain associated with the bed
 ^         material for Run  14  would  indicate that oxide  not converted to  sulfite may
           have  been converted to  carbonate.

-------
                                                                                  J
      E.    CONCLUSIONS                                                         -
            The following conclusions have been drawn from the experiments
described above:                                                                  "~
            •     Sulfites are unsuitable for the absorption of SO, under
                                                              ^                   .J
                 fluidized bed conditions at low temperatures, because
                 of their limited basicity.   This is reflected in the ap-              , ,
                 preciable decomposition pressures of SO- exhibited by             ~
                                                        £»
                 the desired absorption products (such as K-S^Og) at
                 slightly elevated temperatures.                                   J
            •     Weakly basic oxides are suitable absorbents.  The
                                                                                   r
                 only example here is zinc oxide.                                   -^
            •     Strongly basic oxides are also suitable absorbents,
                 but these compounds may be converted to the cor-                 —^
                 responding carbonates as intermediates. If this
                 occurs, the carbonates should rather be considered               ' ^
                 as the absorbents.
                                                                                   r
            •     Magnesium carbonate  and possibly sodium carbonate              •—•'
                 merit further investigation as dry SO., sorbents, par-
                 ticularly in view of their relatively low cost.  As in the             _J
                 case of zinc oxide,  the successful application of these
                 materials will depend largely on the extent to which oxi-
                 dation to sulfate can be avoided.
                                                                                 J

-------
                               PART THREE
                        PROCESS IMPROVEMENT

I.     INTRODUCTION
      One portion of Phase IV was concerned with effecting improvements in
existing aqueous processes for the removal of SO, from flue gas.  An area of
immediate interest was that involving the disproportionation of metallic sul-
fites, inasmuch as this type of  reaction effectively decreases the amount of
SO7 which is obtainable on thermal decomposition of the sulfite.  An example
   £
is provided by the Johnstone Zinc Oxide Process (Reference 7).
      In the Zinc Oxide process the flue gas is scrubbed with an aqueous solu-
tion of sodium sulfite  and  sodium bisulfite.  Zinc oxide is mixed with the ef-
fluent liquor,  forming insoluble zinc sulfite, and regenerating soluble sodium
sulfite which is  returned to the scrubber.  The zinc sulfite is separated by
filtration,  dried, and  calcined to produce zinc oxide, which is returned to
the process, and product sulfur dioxide.
      Inasmuch  as some oxidation occurs in the scrubber to produce sulfate
which cannot be readily calcined, the process includes provisions for its re-
moval.  The effluent scrubber liquor is treated with insoluble calcium sulfite,
and the mixture is passed through a  clarifier.  The underflow from the clarifier,
which contains the calcium sulfite, is acidified with a portion of the product
sulfur dioxide, thereby causing the calcium sulfite to dissolve.   Calcium ion
is thus made available for precipitation as calcium sulfate, which is removed
by filtration and discarded.  The filtrate is treated with lime to precipitate
calcium sulfite, and it is then returned to the clarifier.
      The following represent the important process  reactions:

      Scrubber:
                                       2 NaHSO3                       (19)

                                                                       (20)

-------
      Liming Tank:

            2 NaHSO3 + CaO - » Na2SO3 + CaSO3^ + HgO             (21)

      Gasifier:

            CaS03 + H20 -fr S02 - »  Ca(HS03)2                      (22)

                                      2 NaHSO  + CaSO              (23)
            2 NaHSO3 + ZnO + 1-1/2H2O - » Na2SO3 + ZnSO3«2-l/2
                                                                    (16)
      Calciner:
           ZnSO,• 2-112 U-O-^—~> ZnO 4 SO,*  + 2-1/2 H,O*         (24)
                j         £»                  £t          £t
           4 ZnSO3'2-l/2 H2O A *  3 ZnSO4 +  ZnS + 10 H2O$        (25)

Reaction 25, which represents the disproportionation,  occurs only to the ex-
tent of about 2 to 3 percent (Reference  12).  However,  both zinc sulfate and
sulfide are ineffective in reab sorb ing SO,, and consequently these compounds
slowly build up during processing unless provision is made for their removal.
In the case of the sulfites of the alkali and alkaline earth metals (such as mag-
nesium), the disproportionation reaction is much  more extensive (Reference
13), and represents a serious deterrent to the recovery of SO, by thermal
means.
     As a first undertaking in Phase IV, attention was directed to the problem
of effecting Reaction 1 and  Reaction 24 to the exclusion of Reaction 25.  Work
on the zinc system was followed by a preliminary study involving magnesium.

-------
II.    DISPROPORTIONATION OF ZINC SULFITE
      A.   INTRODUCTION
           The disproportionation of zinc sulfite has been investigated by
several workers, including Johns tone and Singh (Reference 12),  Okabe,
et al. (Reference 14),  Cola and Tarantino (Reference 15), Pechkovskii and
Ketov (Reference 3), Pannetier,  et al.  (Reference  16), and Ingraham and
Kellogg (Reference 17).  The  reaction is thermodynamically favored in the
temperature  range  25  to 800 C (Reference 18),  but the kinetics are such
that the simple decomposition of the sulfite (Reaction 24) is highly pre-
dominating.  The following comments summarize the more important
findings of the studies  considered above:
           •    The precursor to disproportionation is the formation
                 of a basic sulfite having the formula ZnSO,« ZnO.
           •    The oxidation of the precursor can be effected with
                 either atmospheric oxygen or with SO~, forming a
                 basic sulfate.
           •    No disproportionation occurs at or below 300 C, and
                 the rate of disproportionation decreases with increasing
                 temperature above 350 C.
           0    The addition of transition metal oxides increases the
                 rate  of disproportionation at all temperatures at which
                 disproportionation occurs.
           e    The final composition of the basic sulfate which appears
                 as one disproportionation product is probably ZnO- 2 ZnSO..
                 A compound exhibiting the composition ZnO- ZnSO^ may
                 form as a relatively unstable intermediate.
           From the above discussion it is apparent that Reaction 25 repre-
sents an over-simplification of the disproportionation process.  The fact that
the  reaction occurs most readily in the temperature range of 300° to 350 C

-------
 implies that the present thermal decomposition studies should be limited
 to temperatures lying outside this range.  Initial work on the program was
 carried out at relatively high temperatures  and for short reaction times.
 Later work was devoted to low temperatures for extended reaction times.
       B.    RESULTS OF  EXPERIMENTS CONDUC TED IN A MUFFLE
            FURNACE
            Johnstone and Singh (Reference  12) investigated the thermal
 decomposition of zinc sulfite in an electric muffle furnace at 375 , 425  ,
 and 475 C for time periods ranging from 15 minutes to 3 hours. Under
 these  conditions the formation of sulfate varied from about 2 to 5 percent.
 It was considered of interest to investigate the use of higher temperatures
 and shorter reaction times than those employed by Johnstone and Singh, for
 the purpose of determining whether or not the rate of sulfate formation is
 slow relative to the rate of decomposition of the sulfite at the relatively
 high temperatures of interest.
           A series of experiments (delineated in Table 9) was conducted
 with ZnSO,' 2-1/2 H_O which,  by iodometry, was found to be  92. 2% pure.
          j        £••
 From  the weight loss observed on subsequent  calcination of this material
 it was concluded that the principal impurity was probably surface water.
 In carrying out the experiments the sample, contained in a crucible, was
 placed on a preheated  petri dish  in a muffle furnace and was  immediately
 covered with a preheated cover and heated for a specified time, after which
 the petri dish, crucible, and cover were removed and cooled  in a vessel
 through which a stream of nitrogen was passed.  Decomposition runs at
about 545 C were conducted for 5, 15, and 45  minutes,  the decomposition
being complete (i.e., no sulfite present by iodometry)  even after 5 minutes.
 That appreciable sulfide was not present in any of the residues (compare
Reaction 25) was also shown by iodometry,  inasmuch as acidified samples
of the  residues required no iodine, and no detectable sulfur formed:

           H2S + I2  	» 2 HI + St                                   (26)

-------
                             TABLE 9

         THERMAL DECOMPOSITION OF ZnSO3* 2-1/2

              IN A MUFFLE FURNACE AT 545 + 3°C
Reaction
  Time
  (min)

   15

   45
Conversion of
Available SO
 to
     3.92

     4.51
Available SO-
  Released  t
    96.1

    95.5
    Remarks

Covered crucible
    5

   45
    4.94

    4.84
    95.1

    95.2
Uncovered crucible

-------
 It was noted, however, that the acidified solutions exhibited a very faint
 odor of H_S,  and that a very faint positive test was alv/ays obtained with
          i*
 moistened lead acetate paper in the vapor phase above the solutions:

            H,S + Pb(CH-COO), - •» PbS \  + 2 CH.COOH             (27)
              L          3     &       black        .3

            The above series of experiments was repeated, except that the
 preheated crucible cover was left off during the  runs, but was used to cover
 the crucible at the time of removal from the muffle and during subsequent
 cooling.   It was considered that when the sulfite was heated in an open vessel
 the atmosphere present at the solid surface would be  somewhat different
 from that present in the closed vessel.  In the open vessel the decomposing
 basic sulfite (Reaction 4) would be exposed to SO-,  H-O,  and O-, the latter
 being present as part of the  air originally contained in the muffle at the time
 of introduction of the sample.   These compounds would be expected to com-
 pete with the thermal decomposition  of the basic sulfite,  giving rise to sul-
 fate formation (Reference 3):

            ZnSO3« ZnO + 1/2 QZ - » ZnSO4« ZnO                    (28)

            2 (ZnSO3- ZnO) + SO2 — - » 2 ZnSO4« ZnO -f-  1/2 S2          (29)

Although it has been reported (Reference 3)  that the sulfur liberated in Re-
action 29  reacts further  to yield zinc sulfide, the temperature of the muffle
 in the present experiments ( *> 545  C) was above that of the boiling point of
sulfur (444°C), so that it might be  expected  that the reaction involving sulfur
would not be extensive:

            4 ZnO + 3 S2 - » 4 ZnS + 2 SO^                         (30)

            For the series of experiments in which the crucible was covered
during the calcination, the basic sulfite should have been exposed to SO^, but
not to O, or H,O.  The reason for this is that the air and water would have
been displaced many times over with SO? before the  basic sulfite had formed

-------
          appreciably.  It was calculated that in the early part of the decomposition,
          corresponding largely to a loss of water of crystallization, some 54 re-
          placements of the atmosphere contained within the crucible would occur (thus
          completely displacing the air originally present), and that the subsequent
'"-'         liberation of SO? to yield the oxide would result in 22 additional atmosphere
          replacements.  It was considered that the results  obtained under these con-
v_         ditions, insofar as the formation of sulfate is concerned, might differ appre-
          ciably from those observed when the  crucible was  not covered, and that in
^         particular a reaction of the basic sulfite with the dry SO- present when the
          crucible was  covered (Reaction 29) might not readily occur.
i_                    The residues obtained from the calcination experiments  in open
          crucibles were in general quite similar to those obtained earlier. The cal-
          cination was complete within 5 minutes as before,  and only very faint tests
          were observed for the presence of sulfide.
                     During the course of the  experiments in open crucibles described
          above,  it was noted that the percentage weight loss of the  sulfite on calci-
          nation was  the same for both 5 min and 45 min heating periods.  These results
          are in contrast to those reported by Johnstone and Singh (Reference  13), who
, -        indicated that infiltration of air into the muffle, and the subsequent reabsorp-
'"-•         tion of SO., by the zinc oxide,  was probably responsible for the observed oxi-
          dation to sulfate;  i. e.,
""                    ZnO + SO- + 1/2 O, 	» ZnSOA                          (31)
                              £•        £t             *x
,          In order to obtain additional information on this point an experiment was
          conducted in which a covered sample of zinc sulfite was calcined in the
          presence of an uncovered sample of zinc oxide (Kadox-15,  99. 7% pure).
          If it is  assumed that the decomposition of the sulfite resulted in the dis-
          placement of air only from the furnace, the resulting atmosphere within
""         the furnace should have consisted of  9. 1% O,, 15»  5% SO,,  38. 9% H,O, and
                                                    &*           £»          £
          36. 5% N.,.  It was found that Reaction 31  did not occur when the zinc  oxide
          sample was maintained for a one-hour heating period in this  atmosphere,
          and that in fact a  small loss in weight of the oxide  occurred as  the result
\__         of a loss of surface moisture.

-------
                                                                                 J
            Samples of the various decomposition residues from the experi-
 ments discussed above were analyzed for contained sulfate, and in all cases
 an appreciable quantity was found.  The data,  which are shown in  Table 9,
 indicate that sulfate formation occurs primarily during the.actual decompo-
 sition,  and that the subsequent heating of the decomposition residue  in the
 presence of SO, does not lead to an appreciable  increase in sulfate.   The
 somewhat lower sulfate values obtained for the decomposition in closed
 crucibles may indicate a  decreased reaction rate in the absence of water
 vapor,  since under the conditions employed the water of hydration of the
 sulfite would have been displaced from the crucible by the SO, subsequently
 liberated.  The absence of sulfide in the  residues is attributed to the high
 calcination temperatures  used,  in that volatilization of intermediate free
 sulfur (boiling point 444 C) would be expected  (Reference 3):
            2 ZnSO3' 2-1/2 H2O	» ZnSO3« ZnO 4 SO   + 5 H2o     (3)

            2 (ZnSO3- ZnO) + SO2 	» 2 (ZnSO^ ZnO) +1/2 S^       (29)

            4 ZnO + 3 S,   *  •» 4 ZnS + 2 SoJ                        (30)         /
                       22                                     ^
            Inasmuch as sulfate formation is known to occur through reaction
of the basic sulfite, ZnSO,- ZnO, with SO2 (Reaction 29), it would appear that       J,
the rapid removal from the reaction zone of the SO, liberated in Reaction 3
                                                 4*
should tend to inhibit Reaction 29.  The rapid removal of SO, should be most         -
                                                         *•                       —/
effective at relatively low temperatures,  where Reaction 29 is slow.  Accor-
dingly,  experiments were subsequently conducted in which  zinc sulfite was
decomposed at 275 + 3 C but for an extended time.  This temperature was
chosen as lying between the minimum temperature required for the complete
decomposition of the sulfite at a total pressure of one atmosphere (about 260 C,     ->
References 19 and 20),  and the temperature at which disproportionation is
purported to begin (300° to  350°C,  Reference 3).  At 250°C, for example,           _,
decomposition (Reaction 24) was found by Johnstone to occur to the extent of
only about  62% in a muffle  furnace before equilibrium was reached (Reference
21),  whereas at 350 C disproportionation (Reaction 25) occurs to the extent of
about 4% (Reference 3).
                                                                                 —/

-------
                     The results of experiments conducted in the muffle furnace at
          275°C are shown in Table 10.  The decomposition was complete only after
i*^
          6 hours, with the slowness of the reaction resulting in a much greater dif-
          ference in the use of covered versus uncovered crucibles than that which
          was observed in the earlier experiments at 545 C.  In this series of ex-
          periments, the  slower decomposition rate in covered crucibles indicates
^-        that SO_ (which should be the main constituent of the ultimate atmosphere
          within the  crucible) has an inhibiting effect on the decomposition.  Also,
X_        the relatively low sulfate content of the residue under these conditions shows
          that SO2 does not cause appreciable disproportionation at the temperature
          employed.  It was observed, in fact,  that hydrochloric acid solutions of the
          decomposition residues exhibited no odor of H~S,  gave negative tests with
          lead acetate paper,  and gave no precipitate of copper sulfide with copper  ion.
          The residues were completely  soluble in the acid,  and consequently contained
          no free sulfur.  These results  indicate the complete  absence of disproportion-
x~-        ation, so that the observed sulfate  must in all cases  have  resulted from oxi-
          dation by atmospheric oxygen.   This explanation is in keeping with  the greater
          observed oxidation for uncovered crucibles in that the contents of these cru-
          cibles were exposed to the air  present within the furnace, and implies that in
          the case of covered crucibles the oxidation largely occurred during the early
          part of the decomposition, before the displacement of air was complete.   In
          this connection  it may be noted from the table that in general little  or no oxi-
          dation occurred during the final stages of the decomposition.
                     It was considered that the experiments conducted in the muffle
          furnace were of value  in providing  general information relating to the de-
          composition of the sulfite as a  function of temperature,  and to some extent
*~        as a function of atmospheric composition.  However, the employment of a
          much more closely controlled atmosphere is possible through the use of a
^—        tube furnace, and accordingly attention was subsequently directed to the use
          of this  type of furnace.
W
                                              61

-------
                            TABLE  10
      THERMAL DECOMPOSITION OF ZnSO3« 2 1/2H2O IN A

               MUFFLE FURNACE AT 275 + 3°C.
J
Reaction
Time
(hrs)
1 1/2
2
1
2
3
6
S02
Released
(%)
62.7
68.0
70.4
87.8
93.5
96.6
so2
Converted
to Sulfate*
(%)
1.41
1.36
3.44
3.74
3.68
3.36
so2
Total*
(%)
64. 1
69.4
73.8
91.5
97.2
100.0

Remarks

Covered crucible
ii ii
Uncovered crucible
it ii
ii it
ii ii
Zinc sulfide was not observed as a product; therefore,  observed sulfate
formation resulted from oxidation of sulfite by atmospheric oxygen
rather than from disproportionation reaction.

-------
      C.   RESULTS OF EXPERIMENTS CONDUCTED IN A TUBE FURNACE
           The results of a fairly extensive series of experiments which were
conducted at  275 +. 3°C in a 12-inch tube furnace are shown in Table  11.  The
experimental apparatus consisted of the gas metering system shown  in Figure
2, a water sparger which was used in some of the  runs,, the tube furnaces, a
bubbler (filled with either  standard iodine or dilute caustic solution)  for trapping
acid gases, and a  wet test meter, A 1 in.  dia x 18 im  long glass tube was em-
ployed within the furnace,  with the sample  placed in a glass boat at the center
of the tube.  The temperature immediately above the 'sample was monitored
through the use  of a chromel-alumel thermocouple placed in a glass  thermal
well.
           In operation, a gas of predetermined composition was caused to
flow through  the furnace at 3 liters/hr and  an approximately 0. 5 g sample of
the sulfite  was subsequently introduced into the furnace through  the thermal
well port.  At the  completion of the  run the sample boat was moved upstream
to a cold portion of the tube, and allowed to cool in a current of  the gas which
had been used during the decomposition of the  sulfite.  The  decomposition
residue was  then analyzed by iodometry for undecomposed sulfite,, and by
barium sulfate precipitation (turbidimetric method) for contained sulfate.
Excess iodine in the bubbler was back titrated with standard thiosulfate to
determine  the amount of SO- released.  However,  in those runs (see Table 11)
in which oxygen was present as a component of the carrier gas,  the  iodine
bubbler was  not used because of a possible interfering oxidation of HI by
oxygen, and  in these instances the SO-, release was determined by difference,
E,S noted in the table.  Whenever SO_ was deliberately introduced as a gaseous
component, the accuracy of the determination of the SO? released by the
sample was severely limited; in these cases the SO_ released was also de-
                                                 Lt
termined by  difference.
           The most significant result of the experiments conducted in the
tube furnace  is that indicating the marked effect of even small amounts of
water vapor  on  the rate of decomposition of the sulfite.  From the data in

-------
                                 TABLE 11
            THERMAL DECOMPOSITION OF ZnSOy 2 1/2 H2O IN A
                        TUBE FURNACE AT 2751 3°C.
Run
No.
1
Reaction
Time
(hrs)
1
2 6
3 1
i 4 ; 2
i
5 1
6 2
7 ; 1
8 2
9
A
I
10 1
11
12
13
14
15
3
1
2
3
1
16 3
SO2 i
Released
(%)
48.0
61.6
49.2
80.4
77.8
92.1
(54. 0)
(74.4)
(79. 7)

(85. 3)
(93. 3)
(41.4)
(41.2)
(41.1)
(57.3)
(78. 5)
so2
Retained
(%)
49.2
36.4
45.5
15.9
15.8
5. 1
41. 1
20.5
15.4

8.6
0.7
55.9
56.0
56.6
39.0
17.4
so2
Converted
to Sulfate2
0. 76
0.91
1. 16
1.68
1.30
1.35
2.33
2.49
2.25

3.45
3.39
0. 13
-0. 17
-0.26
1.07
1.48
S°23
Total3
98.0
98.9
95.9
98.0
94.9
98.6
97. 45
"
n

"
"
n
n
"
n
"
Gas 4
Composition
0.03 H,O in N-
f, 2
1.01 H-O i.n N2
"
2.27H,Oin N,
£ £
It
21O-+0. 36H,O in N-
* n
n

21O,+2.27H,O in N,
II
5 SO- in N-
^ n ^
n
5 SO- +2. 27 H-O in N-
^n ^ ^
                                                                                    J
                                                                                    J

Values in parentheses were obtained by difference from the total SO-.

The zinc sulfite used in these studies contained 4. 5% sulfate, and a negative value
in this column indicates that less  sulfate was observed in the product than in the
starting material.

Zinc sulfide was not observed as -a product in any of the runs.

All values are given in volume per cent.  The gas flow rate was always 3. 0 1/hr.
The water content of the various gas mixtures was determined by passing the gas
through Desicchlora (anhydrous barium perchlorate).

This value was determined as the average of the values obtained in Runs 1  to 6.

-------
          Table 1 1  it may be noted that when essentially water-free nitrogen was used
          as the carrier gas,  approximately one-half of the SO- was released during
          one hour  (Run 1), and that, even after 6 hours, only 61. 6% of the available
.          SO- had been released (Run 2).  With small,  but  increasing, amounts of water
V
          in the gas phase,  the decomposition rate progressively increased (Runs 4 and
          6), and was essentially complete at the end of 2 hours (Run 6).  That added
L.        water was less effective  during the first hour (compare Run 1 with Run 3)
          can be attributed to the fact that appreciable  water was already available
A
<          during the initial stages of the decomposition as a result of the  release of
          the water of crystallization of the sulfite.
•_,                   As in the case of the experiments carried out in the muffle furnace,
          the residues from the tube furnace contained no sulfide or free  sulfur.  Thus
^          Reaction  29 does not occur at 275  C,  and the formation of sulfate shown in
          Table 11  is presumed to  be due to the presence of oxygen.  It will be noted,
1          in comparing Runs 1 and 2, that the small amount of oxidation would be favored
          at this time by the release of water of crystallization (compare Runs 1-2 with
          Runs 3-4 and 5-6,  and Runs 7-9 with  Runs 10-11, where water  is seen to pro-
^-        mote oxidation),  and presumably is effected  by occluded air on  the surface of
          th.e solid, and/or by air admitted to the tube  during the introduction of the
•_        sample.  That oxidation is not due to  disproportionation is further indicated
          by a comparison of the results obtained in Runs 12-14 with those obtained in
          Runs 1-2.  The data not only indicate that Reaction 29 does not  occur,  but
          further show that the presence of dry SO- is  effective in inhibiting oxidation
          by oxygen.  The  SO-, presumably complexes  with any water present, and the
*""        oxidation of sulfite does not proceed in the dry state:

w                   ZnSO3- 2-1/2 H2O + 1/2 O2 - *— » ZnSO4 + 2-1/2 H2O       (32)

                     ZnSO3' ZnO -t 1/2 O2 — *-«, ZnSO4. ZnO                    (28)
S
•tor*
          Johnstone noted that wet zinc sulfite cake became hot when it stood in con-
*          tact with air in the laboratory,, and that considerable oxidation  took place
          (Reference 22).  On the other hand, in Aerojet laboratories the routine
                                             65

-------
 handling of dry zinc sulfite over several months did not result in an increase
 in the amount of sulfate (about 4. 5%) present at the time of purchase.  The
 effect of water in promoting the oxidation by oxygen can be further seen by          ~~*
 comparing Runs 7-9 with Runs  10-11.
            From a comparison of Runs  1-2 with Runs  12-14 it appears that
 the presence of SO- in the gas phase tends to inhibit the decomposition reac-
                  £t
 tion, and this effect was noted earlier in connection with experiments con-          ~"
 ducted in the muffle furnace.  This indicates that Reaction 1 is reversible.
 However,  the presence of water in the gas phase tends  to counteract this            — '
 effect, as  shown by a  comparison of Runs 15-16 with Runs 12-14.
            In both Tables 10 and 11 the  data relating to the conversion of           — -
 sulfite to sulfate indicate in some instances that a slight decrease in sulfate
 content occurs on prolonged heating.  However, this is considered not to be         _J
 the case,  inasmuch as it was found that the heating of ZnSO^« 7 H,O for 6 hours
 at 275 +_ 3  C in the tube furnace  in a current of nitrogen did not result in de-
 composition.  From the weight of the residue it was calculated that all of the
 water had been lost, but analyses of the  caustic solution through which the           <
 off-gas had been sparged indicated that neither SO2 nor SO, had been evolved.       "^
 It appears, therefore,  that the lack of agreement in the data pertaining to  sul-
 fate  must be attributed to experimental error.  It may be stated that the ex-         •-'
 perimental values of SO- retained and SO- released shown in Tables 10 and 11
 are probably accurate to about 1%, but that the values for SO,  converted to         _^
 sulfate are accurate to no more than 5%.  This arises from the use of the
 turbidimetric method used for sulfate, which is somewhat  less accurate than
 the corresponding gravimetric method (1%).  However,  in  view of the large
 number of sulfate analyses which had to be conducted both  on this and other
 phases of the program,  it was considered that the more tedious gravimetric
 method was impractical.
                                                                                  i
           In summary, it appears that  the decomposition of zinc sulfite can
be carried essentially to completion in about 2 hours (Run  6) at a temperature
 (275  C) which is  sufficiently low so that the disproportionation  of the  sulfite         -//

-------
does not occur.   This is accomplished by introducing a small amount of water
into the gas phase as an aid to the removal of the combined SO,.  The water
                                                           "
of crystallization also  serves th® function  of aiding in the removal of the SO7.
                                                                        £«
Disproportionation (sulfate and sulfide formation) does not occur at the tem-
perature employed,  but oxidation (sulfate formation) will occur whenever
oxygen (air) is present.  It follows that in  the practical caae the calcination
of the sulfite should be  effected in the absence of aire
           Inasmuch as a small amount of water is effective in aiding the
liberation of SO~ from the sulfite,  it was obviously of interest to investigate
the use  of steam for promoting the decomposition at a somewhat lower tem-
perature.   The use of a lower temperature is desirable both from the stand-
point of reducing fuel costs  for effecting th'e  decomposition,  and of decreasing
the rate (and therefore  extent) of oxidation of the sulfite.
            The  effect of steam on the decomposition reaction is shown  in
Table 12.   Some of the  data (Runs 1 to 6) were also given in Table 11, but
are included for comparison. Run 20 was conducted in order to obtain addi-
tional data relating to the manner in which the decomposition occurs. Ac-
cording to Pennetier (Reference 16), the first step involves  partial dehydration
at 9.0° to 100°C  as follows:

            ZnSO3*a-l/2 H2O —^-i» ZnSO3'l/2 H2O  + 2 H2O$          (33)

In view  of the general tendency toward increased SO,  release through the
                                                  £•
incorporation of water vapor  in the gas phase at 275°C (compare Runs 1 to
6), and  because  precisely one-half of the available SO, in the sulfite was
                                                    4*
released during the  first hour when very little water was incorporated in
the gas  phase (Runs 1 and 3), it was  considered  that the further loss of water
from the product of Reaction  33 might be accompanied by a corresponding
loss of SO-S that is,
2 ZnSO,-l/2 H-.O  >10°  C e» ZnSO,-ZnO
                          2
                                                                      (34)

-------
                                           TABLE 12

                  EFFECT OF WATER VAPOR ON THE THERMAL DECOMPOSITION
                             OF ZnSO3- 2 1/2 H2O IN A TUBE FURNACE
Run
No.
19
20
1
2
3
4
5
6
17
18
Temp
(1 3°C)
225
275
ti
it
n
it
n
ii
225
275
Reaction
Time
(hrs)
2
1/4
1
6
1
2
1
2
2
1/2
S02
Released
(%)
17.2
29.6
48.0
61.6
49.2
80.4
77.8
92.1
36.9
81.5
S02
Retained
(%)
78.6
67.3
49.2
36.4
45.5
15.9
15.8
5.1
60.0
14. 1
SO- Converted
to Sulfate
(%)
0.24
0.00
0.76
0.91
1. 16
1.68
1.30
1.35
0.04
0.92
SO,
^*
Total
(%)
96.0
96.9
98.0
98.9
95.9
98.0
94.9
98.6
96.9
96.5
Gas
Flow Rate
(1/hr)
3.0
ii
it
M
n
n
M
n
4.4
3.5
Gas
Compos itioii
(vol-%H_C i
-------
          Although Reaction 34 may occur to some extent,  the results of Run 20 show
          that the loss of water from the hemihydrate is not necessarily accompanied
w*
          by the loss of SO?.  In Run 20 it was observed that all of the available water
          was lost during the 15 minute duration of the run, but only 29. 6% of the
          available SO2 was released.
                      Run 19 shows the marked decrease in the  rate of the decom-
""         position reaction when the temperature is decreased from 275 C  to 225 C
          (compare Run 19 with Runs  20, 1,  and 2).   The use of steam as the carrier
*-~'        gas is effective at this temperature, as  shown in Run 17.  However, the
          complete decomposition of the sulfite at 225  would require several hours,
_^        even in the presence of steam.
                      In Run 18 it will  be noted that the decomposition can be carried
v—        nearly to completion during one-half hour at 275  when steam  is employed
          as the carrier gas  (compare Run 18 with Runs 20 and 1).  The small amount
^_        of sulfate which was observed to  form at this temperature is attributed to the
          presence of air,  as discussed above.  Although the data indicate a release of
          81. 5% of the available SO-,  this result,  as well as the results pertaining to
          SO- release  in all of the other runs, is undoubtedly somewhat low. It was
          noted during the latter stages of the work that a  small amount of SO- was
          escaping through each of two pyrex ball  joints located downstream of the
          furnace, but upstream of the bubbler used to  entrain the  SO-.  This obser-
x—        vation is considered to explain the deviation of the total SO- observed from
          100 percent, and indicates that the data  relating to SO., release are probably
                                                              &
          better derived from the data for the SO- retained.   Thus, in the case of Run
**""*                                             £*
          18, the SO- released is considered to be 100-14. 1-0.9 or 85.0%.
^                    The  results of Run 18 indicate that it should be possible to carry
          out the decomposition of zinc sulfite in the neighborhood of 275 C  in the prac-
          tical case without concurrent disproportionation. Since the latter reaction
          occurs only at temperatures exceeding 300°C (Reference 3), the decomposition
          temperature could be increased somewhat (say to 290 C) if it were desired to
          complete the reaction in a shorter time.  The reaction would perhaps best
          be accomplished in a fluidized bed regenerator,  so designed that the steam

-------
 liberated in Reaction 24 would be available to assist in the final stages of the        """
 decomposition.  The formation of sulfate would be avoided by effecting the
 decomposition in the absence  of air, and the regenerator would preferably           ^
 be maintained at a slight positive pressure to prevent air  leakage.
            No further work was conducted relating to the disproportionation
 of the zinc  sulfite, since it was  considered that the experimental conditions
 required for decomposing the sulfite without attendant disproportionation            -J
 have now been defined.  Accordingly, attention was subsequently directed to
 the effect of steam on the disproportionation of magnesium sulfite.                   _j
 III.   DISPROPORTIONATION OF MAGNESIUM SULFITE
      Magnesium oxide has been found to be effective in removing SO. from          ~~~
 gas streams in fluidized bed  systems (see Table 7),  and it also appears that
 it can be substituted  for ZnO in the Johnstone process (see Table 4 and the           —/
 accompanying  discussion). It was of interest, therefore,  to optimize the
 regeneration of MgO from MgSO,' 6 H^O,  since extensive  disproportionation         _j
 occurs during  the thermal decomposition of the sulfite.  A brief laboratory
 investigation was conducted to determine if steam would lower the effective            '
 decomposition temperature and/or lower the amount of disproportionation
 of MgSO,* 6 H-O.  The 12-inch tube furnace that was utilized in the studies
 relating  to the decomposition of zinc sulfite was also used in this investigation.       u
 An analysis of the magnesium sulfite indicated 90.4 wt-% MgSO,. 6 H-O and
                                                              j     £•                \
 5. 7 wt-% MgSO4.                                                                  ~J
      Elemental  sulfur was produced in small amounts during the decompo-           t
 sition of the sulfite and condensed on the side of the tube downstream of the           —^
 furnace.  The  sulfur  was dissolved in CS, and its  weight determined after
                                        ^                                           /
 evaporation of the solvent.  The residue in the boat was weighed and divided         ___
 into two  parts for analysis.  The first part was analyzed for unreacted SO,"
by adding a  known quantity of- 0. 1 N iodine, then enough 0.  1 N HC1 was added          :
to dissolve the solid phase, and finally the solution was back titrated with
0. 1 N sodium thiosulfate.  If  any thiosulfate was found in the  second analysis,         '
described below, a correction was made for it, since it also consumes iodine.       ~~"

-------
      The second part of the solid residue was analyzed for sulfate and thio-
sulfate by the following  procedure.  Ten ml of concentrated HC1 was added
to the sample, which was evaporated to dxyness at 130°C.  Twenty-five ml
of water was then added, and the solution was heated to boiling.  The hot
solution was filtered, and the residue washed with water.  The filtrate was
cooled to room temperature,, and sulfate determined turbidimetrically.
The residue was dried at 110°CS  cooled, and reweighed as sul£urs  which
is a measure of the thiosulfate in the sample.
      The following equations represent the reactions involved in the second
part of the analysis:

           SO3= + 2 HC1	a. S02 + H2O 4 2 Cl                       (35)

           S2O3= -I- 2 HC1	»  SO2 + s|  + H2O -F 2 Cl                (36)

      Several samples of pure MgS, O, were analyzed in the above  manner and
                                 £• j
sulfur representing 96% of that shown in Reaction 36 was  found.  Since the
temperature of the furnace was above the boiling point of sulfur in  all runs,
no appreciable quantity  of free sulfur could have been left in the solid residue.
      The results of three of the best runs are given in Table 13.   Only those
runs with a sulfur balance of 95-105% are reported.  The results of the Russian
investigators Ketov and Pechkovski (Reference 3)  are reported for comparison.
At 450 C with a nitrogen gas purge (Run 2), the total conversion of S in the sul-
fite was 34. 1% compared with 38. 3% reported by the Russian investigators.
The amount of sulfate formed (14. 7%) also compared well with theirs  (12. 9%).
The amounts of sulfur and thiosulfate formed, however, were appreciably less
than those found by the Russian investigators.  A similar analysis  can be made
of the runs made at 550°C.
      Run No. 9 was  made at 450  C, with the purge gas consisting of 85% steam
and 15% nitrogen.   This  run indicates that the total conversion of sulfur in the
sulfite increased to 57. 9% (compare with Run 2 using nitrogen).  It appears.

-------
                                                 TABLE 13

        EFFECT OF TEMPERATURE ON THE  THERMAL DECOMPOSITION OF MAGNESIUM SULFITE
                         IN A CURRENT OF NITROGEN OR STEAM AND NITROGEN

Run
No.
2
*
9

4
*

Temp.
, c&(-
450
450
450

550
550
leaction
Time
hr
1/3
1/4
1/3

1/3
1/4

Purge
Gas
N2
N2
85% steam
15% N_
b
N2
N2

% Conversion of S in Sulfite to
so2
17.1
16.3
41.4

81.5
68.5
MSS2°3
0.8
5.3
0.0

0.0
0.0
S
1.5
3.8
2.4

5.6
8.4
MgS04
14.7
12.9
14.1

15.1
17.5
Total
Reacted
%
34. 1
38.3
57.9

102.2
94.4
SOj Not
Reacted
%
66.4
-
41.4

1.8
-
S
Balance
%
100. 5
-
99.3

104.0
-
•*Reference 3.
                    L...
L..
L , '   L-. '   L .
U...

-------
         however, that the quantity of sulfate formed was not lowered, due to addition
;         of steam to the  purge gas.  It should be noted that the quantity of sulfate formed
i»*B>'
         did not change significantly at the higher decomposition temperature.
.                This brief investigation is not conclusive,  but it appears that:
                •     Steam will assist decomposition of magnesium sulfite
v  .                   and should reduce the decomposition temperature.
                •     Steam apparently will not lower the percent of mag-
w                   nesium  sulfate formed during disproportionation.
                A more thorough study is needed to verify these assumptions.
W       IV«    THERMAL DECOMPOSITION OF ZINC SULFATE
                From the data given  in Tables 1 and 2,  it appears that the use of ferrous
\_f
         ion in a prescrubber can, under optimum conditions,  essentially eliminate the
.         formation of sulfate.in a fluidized zinc oxide absorber.  It may be presumed,
'v-       however, that in practice,  traces of sulfate probably will form, and it was
         therefore considered of interest to  briefly investigate the thermal decompo-
^
i^,       sition of zinc sulfate.  In particular, it was considered that the  use of steam
         might be effecti\ge in lowering the calcination temperature of the sulfate, in-
•         asmuch as steam has been observed to promote  the decomposition of both
         zinc and magnesium sulfites  to a marked extent.
w              Considerable  discrepancy exists in the  literature relative to the tem-
         perature required to decompose zinc sulfate to the oxide.  According to Margulis
         and Remizov (Reference 24), the decomposition occurs in two stages, as follows:

                     3 ZnSO4  >6l° C a»  ZnO 2 ZnSO4 + SO^                  (37)

~                   ZnO-2 ZnS04  >740°C  a»3 ZnO 4 2 SO^                    (38)

\_       Pannetier (Reference 16) noted that the basic sulfate that formed from  the
         thermal decomposition of zinc sulfite in the presence of oxygen, and which
i__       he formulated as ZnO* 2  ZnSO4 (compare Reaction 38), decomposed at  854 C.

-------
 On the other hand, Pechkovskii and Ketov (Reference 3), reported the for-           "^
 mation cf a basic sulfate of unknown composition from the disproportionation          ;
 of the sulfite, and observed that the  sulfate decomposed in the temperature          --'
 range 993° to 1027°C.
      The  experimental results from this  study are shown in Table 14.  The        ~~
 apparatus  used was that described earlier for the decomposition of the corres-        ,
 pending sulfite in a tube furnace at 275°C, except that the gas entering the          -J
 furnace was preheated through the use of heating tape in those experiments
 in which steam was used in order to prevent water condensation,  and the            J
 exiting gas was similarly heated to prevent the condensation of sulfuric acid.
 The steam was supplied to the furnace by passing a small amount of nitrogen         f
                                                       rt                          —'
 through a water sparger maintained at approximately 100  C.  The effluent
 gas was sparged through a caustic bubbler, which was subsequently neutralized      f
 with acid and analyzed for SO, content. The  values for SO, released shown in
 the table were derived by difference  from  the SO, retained by the partially de-       •
 composed residue.                                                               -J
      The decomposition of the sulfate would be expected to yield SO, as a
primary decomposition product following the loss of water:                         -1'

                                                                     (39)        j

However, at higher temperatures the SO., dissociates:

            SQ3 j—» SO2 + 1/2O2                                  (40)

At 790 C, for example,  the equilibrium constant for Reaction 40 is unity*           -I
The  relatively low values for the dissociation of SO, shown in the table are
therefore attributed to the limited SO, residence time at the furnace tern-           j
perature.
      Runs 1 to  3 in Table 14 are reported merely for completeness, inas-          ._,
much as only dehydration occurred at the relatively low temperatures em-
ployed.   The direct introduction of the heptahydrate into the furnace at               ,

                                                                                J

-------
r
r
r     r
r    r
r
                                                   TABLE 14


                        EFFECT OF WATER VAPOR ON THE  THERMAL, DECOMPOSITION
                                   OF ZINC SULFATE IN A TUBE FURNACE
Run
No.
1
2
3
4
5
6
7
8
Temp
(t 3°C)
275
500
650
750
750
800
850
850
Reaction
Time
(hrs)
6
2
2
3
3
1
1
1
H2°
Released
100
100
100
—
— <
—
so3
Released
0
0
0
21.0
35.4
37.3
48.7
56.0
Dissociation of
SO3 Released
—
—
—
trace
trace
trace
11.52
8.51
Gas
Flow Rate
(1/hr)
3.0
4.5
3.0
3.0
4.6
4.4
3.0
3.7
Gas
C ompo s ition
(vol-%H7O in N_)
&• b
0.03
93.8
0.03
0.03
92.7
93.2
0.03
94.3
  -4
  Ul
           ZnSO4' 7H_O was used in runs 1 to 3; the anhydrous salt was used in the remaining runs.
           SO,

-------
temperatures exceeding 650°C was attended by considerable spattering of
the solid because of the rapidity of the dehydration.  Runs 4 to 8 were there-
fore conducted with sulfate which had been dehydrated at lower temperatures
in the manner shown for Runs 1 to 3.
      A comparison of Runs 4 and 5 indicates that the presence  of water is
beneficial at 750 C in promoting the decomposition of the sulfate.  However,
the rate  of decomposition is very slow at this temperature.  By raising the          —'
temperature to 800  C, the time can be reduced by a factor of three (Run 6),
but the decomposition rate is still slow.  At 850°C (Runs  7 and 8),  the effect
of water is less pronounced,  presumably because any molecular complex in-
volving water and SO, would be largely dissociated at the high temperature           '
involved.  To the extent that such complexing does exist,  the dissociation
of the SO,  is depressed, as can be seen from the table.
      From the data in Table 14 it is concluded that the presence of water
will not materially promote the decomposition of zinc sulfate at the high              i
temperatures required for complete decomposition.  From the time-
temperature data, it would be expected that a. temperature in the neighbor-           ,
hood of 1000 C would be required for complete decomposition in a relatively         ~J
short time (say,  15 to 30 minutes).  This temperature is  essentially that
observed by Pechkovskii and Ketov, as noted in the previous discussion.             J
If equilibrium is  established at 1000°C, all of the liberated sulfur will be
in the form of SO,, rather than SO,.                                                '
                 £t               j                                                '-J
      The calcination of zinc sulfate is considered further in the following
section.                                                                           ,__,
                                                                                  J

-------
                                         PART FOUR
                     A ZINC OXIDE FLUIDIZED BED SYSTEM FOR THE
~                       ABSORPTION AND REGENERATION OF SO2

L.
                From the experimental results discussed in the preceding sections, it
           is now possible to formulate a tentative system for the removal and recovery
*~          of SO- from a gas stream,  such as the flue gas from a power station, through
           the use of fluidized zinc oxide as the absorber. Such a system is shown in
w          Figure 4.  The  following chemical reactions are involved:
                Absorber:
i_
                      ZnO •«• SO2 + 2-1/2 H2O  - » ZnSOj- 2-1/2 H2O           (1)

^                     ZnO + SO2 + H2O 4  1/2 QZ - »  ZnSO4' H2O              (41)

                Regenerator (527°F):

                      ZnSO3- 2-1/2 H2O  steam». ZnO -f SO^  + 2-1/2 H2O$      (24)
                                        ZnSO + Ho                           (42)

                Regenerator (1832°F):

                      ZnSOA - » ZnO + SO,^                                 (39)
                           t                j
                                                                               (40)

           The gas stream is first prescrubbed with water to remove fly ash and SO- to
           cool the gas to a temperature (122 F or 50 C) where absorption of the SO- by
           zinc oxide readily occurs, and to  saturate the gas with water at this temperature
           as an aid in promoting SO- absorption.  The SO? is then absorbed from Stream 1
           at a slightly higher temperature in the zinc oxide fluidized bed absorber, and the
           purified gas, Stream 2, is vented.  Some oxidation to sulfate may occur in the
           absorber, so that Stream 3 will consist of sulfites  sulfate,  and unreacted oxide.
           Sulfate, at this point,  will be in the form of the monohydrate,  ZnSO,- H2O
           (Reference 25).


-------
CXI







Flue gas from \1/
pre scrubber
at 122°F













122°F
132°F




Absorber










Y


4

Cyclone


^ r~
<2>,H.
1






H. E. = Possible heat exchange









"1
E-l
i r
_ j


1
1
1
L





xx ^n

1

527°F


XJ\
v


H. E.









/g\
N/

~1
1
1
J



I Cy<
V

Regenerator









1832°F
<£> i
                                                                                 Regenerator
                                          BLOCK FLOW DIAGRAM
                                   FLUIDIZED BED ZINC OXIDE SYSTEM

                                                  Figure 4

-------
      Stream 3 enters a fluidized bed regenerator at 527 F (275 C).  The
steam required for promoting the decomposition is provided by the water
of crystallization of both the sul£ite8  ZnSO3° 2-1/2 H,O, and the sulfate,
ZnSO4* H,O.  The gas, Stream 8e will consist only of SO, and water vapor.
Most of the solids from the 527°F regenerator are returned to the absorber,
as indicated by Stream 4.
      Depending on the degree of oxidation in the  absorber, a portion of the
material calcined in the 527 F regenerator is further processed in a second
regenerator at 1832°F (1000°C), as indicated by Stream 5.  Both Streams 4
and 5 contain zinc oxide and anhydrous zinc  sulfate,  but little or no zinc
sulfite.
      The 1832 F regenerator may be fired  either directly or indirectly.
In the former case the combustion gases from a gas or oil burner would be
utilized.  At 1832 F,  dissociation of SO, to  SO, is complete, so that the gas,
Stream 95 will consist of SO, and O, (indirect firing) or of SO,, O,0  and
combustion gases (direct firing). In either case, Stream  9 would  probably
be combined with Stream 8 to give an SO,-rich gas suitable for SO, recovery.
      The underflow from the 1832°F regenerator, Stream 6, consists only
of zinc oxide.  This is combined with Stream 4 to yield Stream 7,  which con-
tains zinc oxide and zinc sulfate, but little or no  zinc sulfite.  The zinc sulfate
in the system will build up to an equilibrium value, which will be determined
jointly by the amount of sulfate formed in the absorber  (and the  small amount
possibly formed in the 527 F regenerator) and the amount of sulfate processed
in the 1832  F regenerator.
      The economics of the system shown in Figure  4 will depend  largely on
the amount of sulfate formed in the absorber. If this is small, the 1832 F
regenerator will be relatively small, and the overall heat input to the system
will not be excessive.
      A further discussion of Figure 4 is presented in Part Six, Section III. A,
and relates to the overall heat requirements for the  system.

-------
      It may be noted that zinc is somewhat unique in that the sulfate is
highly water soluble  (ZnSO4, 42 g/100 g H2O at 0°C, 61 g/100 g at 100°;             ,
ZnSO.' H-O reported as soluble at ambient temperature,  89.5 g/100 g at            ^
100  , Reference 26,  whereas the oxide and  sulfite are both insoluble  ZnO,
0.0042  g/100 g at 18°;  ZnSO3. 2-1/2 H2O, 0.16  g/100 g at 25°).  The possi-         ^
bility therefore exists of extracting the sulfate present in Stream 5 of Figure
4 with water.  The extract could be further  processed,  particularly if the           j
flue  gas to be treated is derived from a zinc smelter, where zinc metal is
normally produced through the electrolysis  of aqueous zinc sulfate.   The             ;
water extraction of zinc sulfate would eliminate  the need for the  1832  F
regenerator.                                                                       f
      More recently, a system has been devised in which neither high tem-
perature calcination  nor leaching of the sulfate is required.  The new system         '
includes the use of ferrous ion in the prescrubber,  so that oxidation in the
absorber  will be minimal.  The absorption of SO,  and regeneration of the            /
oxide from the resulting sulfite are conducted in the manner described above,        -1
but any sulfate formed is removed from the  system by simple filtration.  This
                                                                                   <
is  accomplished by dissolving a portion of the  sulfite-sulfate mixture in aqueous     _j
SO-,  and reprecipitating the sulfite with zinc oxide.   The sulfate solution is then
filtered, and the sulfite cake returned to the process.  A discussion of the            '
economics of this system is  given in Part Six, Section III.

-------
                               PART FIVE
                      OXIDATION AND NO  STUDIES
                                          X

I.     INTRODUCTION
       The scope of the work in Phase IV concerning oxidation and NO
                                                                    A>
studies consisted of two separate tasks, one for each study (see Part One,
Section I).
       The purpose of the oxidation studies was to determine the degree to
which inadvertent sorbent oxidation in aqueous scrubbers could be minimized
by the utilization of various oxidation inhibitors and  complexing agents both
with and without fly ash being present in the flue gas being treated.  Pre-
suming that the degree of oxidation could not be economically reduced by the
use of inhibitors  or complexing agents,  the technical feasibility of separating
the oxidation product from the scrubber effluent by chemical or ion-exchange
means was to be  investigated, followed by the thermal, chemical, or electro-
chemical regeneration of the  resultant material.   Finally, the effects that
pre-scrubbing has on the degree of oxidation in the main SO,  scrubber were
to be ascertained.
       The NO studies were to include the determination of the degree of
               X*
interference that inadvertent  sorption of NO  into SO- scrubbing solutions
                                          X        6
had on SO- removal efficiencies both with and without fly ash being present
          &
in the flue gas being tested.  The technical feasibility of achieving high NO
                                                                        Jfc
removal efficiencies in conjunction with high SO, removal efficiencies was
to be assessed.
       As the work progressed, it became evident that these  tasks could be
studied concurrently,  especially since the NO  in the flue gas (in the presence
                                            jt
of O-) was the major contributor to oxidation of SO-  during its absorption.
    Ce                                            Ct
It is for this reason that the tasks have been combined into "Oxidation and
NO  Studies. " The scope of these tasks was such that it was not possible
to investigate  all aspects in the laboratory.
       Section II discusses the  experimental program of the oxidation and NO
                                                                           xC
studies.   Description of equipment, operating procedure, analytical methods
and calculations will be found in Appendix A.

-------
 n«     EXPERIMENTAL RESULTS
        A.    LABORATORY EVALUATION OF INHIBITING AND                      i
              COMPLEXING AGENTS
              1.     Introduction
                    The purpose of this work was the determination of the
 degree to which inadvertent oxidation in aqueous scrubbers can be mini-
                                                                                     I
 mized by the utilization of various oxidation inhibitors and complexing
 agents with and without fly ash being present in the flue gas.                          ,
                    The oxidation of aqueous  solutions of salts of sulfurous
 acid in the presence of oxygen proceeds by a free radical  chain mechanism            '
 (Reference 11); accordingly, compounds capable of breaking the reaction             -J
 chain may serve as suitable inhibitors.  Thus,  the general class of free
 radical oxidation inhibitors were considered that finds wide applications             _/
 in such diverse fields as rubber compounds,  gasoline stabilization, and
 food technology.  Another possible way of preventing oxidation by metal               j
 catalysis is by the  use of complexing  agents, of which there  are many
 commercially available today.  Only those candidates that have been found
 to be effective in various applications, and which are commercially avail-
 able  in production quantities, were investigated.                                     I
              2.    Screening Tests
                   The original plan was to screen the candidate materials
 in simple apparatus using a synthetic  flue gas with fly ash added separately
 and at various levels, which would reflect the approximate degree of ash              \
 removal expected in the prescrubber,  e.g. 95 and 98%.  The candidate
 inhibitors and complexing agents were to be tested in this system at various
 concentrations.   The test solutions were to be analyzed for sulfate  formation
 and compared with the  results obtained from  suitable blank runs  in which no
                                                                                   i
 agent was used.  A reevaluation of this procedure showed that  the desired           _^
 results could  probably be obtained with a simpler approach.  Therefore, a
 test was  substituted that involves the addition of air to the  lean sodium
 sulfite-bisulfite solution (used in Johns'tone's Zinc Oxide process - Refer-
ence 7).   Fly  ash and inhibitor or  complexing agent were added in fixed,

-------
          constant quantities.  This procedure proved adequate, at least as a pre-
^        liminary screening procedure,  for the various inhibitors and completing
          agents tested in the lean sodium sulfite-bioulfits solution with a pH of 6. 5.
i_                          The test procedure consisted of the following:
                                               *
                            100 ml of 0.65 S/C   absorbent (lean solution), which
;           .                                   a
L_        had been prepared using oxygen-free boiled distilled water, was trans-
          ferred to a 3-neck 300 ml flask fitted with a reflux condenser and a
          thermometer.  A nitrogen purge was applied,  using a fritted glass gas-
          dispersion tube.   The nitrogen purge was maintained while the contents
          of the flask were heated to 50 C (expected scrubber temperature). The
          nitrogen purge was stopped and air at the rate of 500 ml per minute was
          introduced for fifteen minutes through the gas-dispersion tube.  The
^~        temperature of the lean solution was maintained at 50 C during the addi-
          tion of air.  The heat was removed and a nitrogen purge was maintained
L.        until the material cooled to room temperature.  A 1 ml sample of the
          cooled solution was added to 10 ml of concentrated hydrochloric acid and
u        the mixture evaporated to dryness on a 130  C oil bath.   This procedure
          removed the sulfite from the  solution.  The residual dry material was
          dissolved in 100 ml of distilled water.  The sulfate content of this solution
          was  determined using the turbidimetric method.
                            Blank tests were made to check the oxidation of the
          solution both with and without the addition of fly ash.  The effect of the
          addition of inhibitors or complexing agents on the oxidation of these
~        systems was compared with the blank test results.  A few tests were
          made using a combination of an inhibitor and a complexing agent.  The
*—        materials  - inhibitor and/or complexing agent and fly ash - were  added
          in quantities of approximately 100 mg each. In the tests containing fly
_        ash and/or insoluble  inhibitor or complexing agent, it was necessary to
          filter the solution before addiag it to the coventrated hydrochloric acid.
         *  S/C  used by Johnstons (Reference 7).  See Appendix A, Section IV
                 for discussion.

-------
                                                                                    J
              3.     Experimental
                    Base line tests were made on the solution, after which
             •     Addition of hydroquinone immediately suppresses
                   the oxidation due to the fly ash addition.
                   The inhibitors and complexing agents checked in the
screening tests are given below.
                   Inhibitors tested were:
                                                                 i
                   Hydroquinone,  N,  N'-dimethylformamide, N, N -
                         i                                    t
d ime t hy lace tarn ide, N,  N -diphenyl-p-phenylenediamine, N, N -di-l-
naphthyl-p-phenylenediamine, N-phenyl-N -cyclohexyl-p-phenylenediamine,
N-phenyl-£-naphthylamine,  butylated hydroxy toluene, butylated hydroxy-
anisole, 1,  l-thiobis-(2 naphthol), and N-phenyl- a -naphthylamine.
                   Complexing agents tested were:
                   Ethylenediaminetetraacetic acid, nitrilotriacetic acid»j,v;vC';v>:'
     i                                                                   ••.••"
N, N -disalicylidene-1, 2-propylendiamine  (80% in toluene), diethylene-
triaminepentaacetic acid pentasodium salt,  and  citric  acid.
 tests with additions of fly ash, hydroquinone, and fly ash plus hydroquinone
 were completed.  Approximately 100 mg quantities of fly ash and hydro-              j
 quinone were added to the 300 ml of absorbent in these tests. The results
 of these tests are given in Table 15.                                                 ;
                                                                                   J
                   The preliminary tests indicate that:
             0     The solution as prepared under oxygen-free                       ,
                   conditions contains a  small amount of sulfate  ion.
             •     Introduction of air will oxidize the sulfite.                        I
                                                                                   -.—/
             •     The addition of fly ash increases the extent of
                   oxidation.                                                        !

-------
                                        TABLE 15
L.
BASE LINE INHIBITOR SCREENING TESTS
Sample
No.
>'f
Blank
1
2
3
Blank*
4
5
6+
7+
Hydroquinone
mg
_
-
100
99
-
100
101
101
Fly Ash SO4~» PPm
in
mg Absorbent
800
2600
1650
1900
900
100 4500
100 1900
101 1850
101 1900
wt-% so2
oxidized
1. 04
3. 38
2. 14
2.47
1. 17
5.84
2.47
2.42
2.47
*  The blank runs were on the absorbent solution as prepared and
   protected with N_ atmosphere.

+  In these tests, the solution containing  hydroquinone and fly ash
   was stirred for 23 hours under N? atmosphere in order to
   determine if it is necessary to allow  the inhibitor extended
   time to react with the iron in the fly ash.

-------
                                                                                      I
                    The test results are shown in Table 16.  Analysis of the
 data indicates the following:
                    The lean solution,  as prepared with boiled distilled                ^
 water and protected with nitrogen atmosphere,  shows a content of 800 -
 950 ppm SO4~.  This  compares with 800  - 900 ppm SO4~ in the blank tests.            -1
                    Introduction of air into this solution increased the SO."             ;
                                                                      4               '
 to 2700 ppm (see blank tests).  The SO.~ content of the freshly prepared              .re-
 solution is due to either SO.~ present in the solid reagents or oxidation of
 SO-  to  SO.  during the solution preparation, or both.       ,                        -J
                    Hydroquinone was the best material tested, inhibiting               ,
 the oxidation both with and without fly ash present (Tests 39 and 40).  These          —'
 results compare favorably with those reported in Table 15.
                    Nitrilotriacetic  acid (Tests 29 and 30) and butylated               ~J
 hydroxyanisole dissolved in toluene  (Tests 15 and 16) were next best in
 activity  (a blank test to check  the effect of toluene in a system containing              _
 fly ash showed no appreciable reduction in oxidation).
                       •                                              •               |
                   N,  N -disalicylidene-2-propylenediamine (80% solution             —'
 in toluene) plus 2 ml toluene,  (Tests 17 and 18),  and ethylene-diaminetetra-
 acetic acid (Tests 27 and 28) both showed some activity in suppressing                _/
 oxidation with fly ash  present.  No noticeable improvement was apparent
 in the; systems not containing fly ash.                                                 j
                   Other materials showing some reduction in oxidation
 with fly ash present but no reduction in fly ash-free solutions were:  N-
         i                                   '      ••                                 —•
 phenyl-N -cyclohexyl-p-phenylenediamine (Tests 9 and 10); diethylenetri-
 aminepentaacetic acid pentasodium salt (Tests 21 and  22); and N-phenyl-              i
 a -naphthylamine (Tests 23 and 24).  Their activities were borderline under
 the test conditions used.
                                                                                   J
                   Two series of tests were made with combinations of an
 inhibitor and a complexing agent in order to determine if their oxidation              i  ,
 suppression characteristics were additive.  Tests 35 and 36,  using hydro-           ~
quinone and nitrilotriacetic acid,  indicate that combining these materials              •
                                                                                   J

-------
               r
r
r '     r""
r
r
r
c       r
r~     r
r
r
r
                                                                 TABLE 16
                                            INHIBITOR AND COMPLEXING AGENT SCREENING TESTS
Test
No.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Blank
Blank
17
18
Blank
Blank
19
20
Additive*
Name
DMF
DMF
DMA
DMA
DPPD
DPPD
DNPD
DNPD
PCPD
PCPD
PNA-2
PNA-2
CaO-3
CaO-3
CaO-7
CaO-7
-
DSPD
D3PD
_
CaO-30
CaO-30
mg
101
98
105
105
100
101
100
101
100
100
99
99
100
102
100
100
_
129
123
_
103
104
Fly Ash
mg
99
101
101
101
102
. 102
98
102
102
99
_
99
SO .", ppm in
Absorbent
3400
4500
4400
5000
4500
5200
5000
5400
3000
3200
2200
4100
2600
5900
2200
2500
950
4000
2500
2900
800
2700
4300
4400
Wt-% SO,t
oxidized
4.42
5.85
5.72
6.50
5.85
6.76
6.50
7.02
3.90
4.16
2.86
5.33
3.38
7.67
2.86
3.25
1.24
5.20
3.25
3.77
1.04
3.51
5.59
5.72
Remarks
DMF = N, N -dimethylformamide
DMA = N, N' -dimethylacetamide
DPPD = N, N -diphenyl-p-phenylene-diamine, insoluble
DNPD = N, N'-di- £-naphthyl-p-phenylenediamine, insoluble
PCPD = N-phenyl-N -cyclohexyl-p-phenylenediamine, insoluble
PNA-2 = N-phenyl-0-naphthylamine
J CaO-3= butylated hydroxy toluene, insoluble. Dissolved in
j 2 ml toluene.
j CaO-7 = butylated hydroxyanisole, insoluble. Dissolved in
j 2 ml toluene.
Lean solution as prepared - no air.
Added 2 ml toluene - added air.
DSPD = 80% sol N, N -disalicylidene-1, 2-propylenediamine.
Added 2 ml toluene.
New lean solution - no air.
New lean solution - added air.
! CaO-30 = 1, l'-thiobis-(2-naphthol) insoluble. Dissolved in
2 ml toluene

-------
                                                                     TABLE 16 (cont'd)
Test
No.
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
Additive
Name
DETPA
DETPA
PNA-1
PNA-1
CA
CA
EDTA
EDTA
•NTA
NTA
Zn
Zn
Zn
Zn
HQ i
NTA»
HQ ,
NTAf
CaO-7i
NTA f
CaO-7i
NTA 1
HQ
HQ
mg
102
99
102
103
103
101
102
101
103
101
103
103
1103
25
103
103
103
102
99
98
103
101
102
99
Fly Ash
mg
101
98
102
103
104
98
102
102
101
102
SO4~, ppm in
Absorbent
2600
3300
2900
3200
2900
4100
2600
2900
2000
2700
'4900
7800
25000
6900
1900
1900
1900
2100
1900
1900
Wt-%. S0,t
oxidized
3.38 I
4.29 '
3.77 J
4.16
3.77
5.33
3.38
3.77
2.60
3.51
6.37
10. 14
32.50
8.97
2.47
2.47
2.47
2.73
2.47
2.47
Remarks
OETPA = diethylenetriaminepentaacetic acid pentasodium salt.
insoluble. Dissolved in 2 ml toluene.
PNA-1 = N-phenyl-oc-naphthylamine. insoluble. Dissolved in
2 ml toluene.
CA = citric acid
EDTA = ethylenediaminetetraacetic acid.
NTA = nitrilotri&cetic acid
Zn = zinc dust
HQ = hydroquinone
CaO-7 in 2 ml toluene
CaO-7 in 2 ml toluene

00
00
            Inhibitor or compleating agent

            Refers to % of SO2 available in the solution.
                                                                I	

-------
   v~_       has no advantage over hydroquinone alone.  In the case of the combination
           of butylated hydroxy anisole and nitrilotriacetic acid (Tests 37 and 38), a
   \__       slight improvement is indicated (compared to the individual compounds)
           especially in the systems containing fly ash.
   l_                          Zinc metal dust was tried on the basis of the electro-
           motive  series of metals, i.e.,  suppress the activity of iron.  Tests 31 -  34
   |^.       indicate,  however, that zinc dissolves in the lean solution which is oxidized
           to zinc  sulfate.
   i
   i^                          The effect of inhibitors on the oxidation of SO,  in the
           standard apparatus used  in the oxidation and NO studies is discussed in
   [        Section II. C.
   *~*>

                  B.    BASE LINE TESTS IN BENCH SCALE  UNIT
   L.
                        A series of runs was made  in both the  recirculating and in the
           once through scrubbers to establish the base line data on SO, removal ef-
           ficiencies in this equipment.  The  data are given in  Tables 17 and  18 and
  ,        Figure 5.
                        The data in Table  17  were obtained in the recirculating scrubber
           by the method outlined in the operating and analytical procedures (see
  t_       Appendix A).  The first run was made with a gas mixture containing CO,
           in addition to SO, and nitrogen.  The second run was made  with SOg in
  L_       nitrogen,  In both runs the overhead gas composition was less than 0. 015%
           SO, when the 0. 85 S/C ratio was reached.  The S/C ratios  were calculated
           assuming there was no appreciable amount of oxidation of SO, to SO.  .
           The actual value of oxidation in the samples was not determined, but from
           all indications it was very low.  The outlet gas composition vs. the S/C
  ^       ratio in the  solution has been plotted in Figure  5 for both runs.  The data
           for both runs fall on a single curve.  This indicates  that the presence  of
  *—       CO, in the flue gas does not have a detrimental effect on the ability of the
           solution to absorb
-  L-
  L_
                                               89

-------
                          TABLE 17
BASE LINE OXIDATION STUDIES - RECIRCULATING SCRUBBER
        Base line data with 0. 9 in ID  glass scrubber
       packed with 32 inches of 0. 5-in porcelain Intalox
          saddles.  Recirculating solution system.

                          Run No.  1

  Inlet Gas: 5/0 liter/min with 0. 3% SO, and 15% CO, in nitrogen
Time

hours

0
0.5
1.0
1.5
2.0
2.5
3.0
Solution
S/C
moles SO,
moles Na+
0.58
0.64
0.69
0.74
0.79
0.85
0.91
Outlet
Gas
mole %
so2
.
0. 0004
0.0012
0. 0022
0. 0048
0. 0087
0.0146
Solution
Temp.
o
C
49
50
49
50
50
50
49
Solution
Rate

ml/min
282
282
282
282
282
318
318
Scrubber
Gas
AP
mm H2O
11
11
11
11
11
12
14
                                                                           _J
0.
0.5
0.8
1.4
2.4
2.3
3.4
0.58
0.64
0.67
0.74
0.84
0.89
0.93
                          Run No. 2
  Inlet Gas: 5. 0 liter/min with 0. 3% SO2 in nitrogen

                                          275
                                          275
                                          275
                                          275
                                          275
                                          275
                                          275
_
0. 0003
0. 0006
0.0017
0.0061
0.0108
0.0220
51
49
50
50
50
50
50
12
12
12
12
12
12
12

-------
                                         TABLE 18
         BASE LINE OXIDATION STUDIES - ONCE THROUGH SCRUBBER

                      Base line data with & 0.2 in ID glass tube 10 ft
                      long packed with a single Teflon helix0  Once
                      through solution system.  Inlet gas: 5. 0 liter/
                      min with 0« 3% SO, in nitrogen.
l_
L.


Time

hours
0
0.7
1.0
1.4
1.8
2. 1
2.3
4.9
5.5
5.8
Outlet
Solution
S/C
moles SO,

moles Na

0.82
—
p.
0.88
•»
0.91
0.98
0.99
0.98
Outlet
Gas
mole %

so2
0.0008
0.0008
0.0012
0.0010
0.0010
0.0012
0.0012
0.0012
0. 0450
0.0218
Solution
Temp.


°C
50
50
50
50
50
50
50
50
50
50
Leah
Solution
Rate

ml/min
2.1
1.8
1.6
1.6
1.5
1.4
1.3
1.0
0.8
0.9
Sc rubbe r
Gas
AP

mm HLO
240
245
240
235
235
300
270
260
265
265

-------
 0. 045
 0.040
 0.035
 0.030
 0.025
 0.020
 0.015
 0.010
 0.005
Notes: 50 C Solution Temp.
       3.5 mole8 Na /100 moles water
       0. 30% SO, in inlet gas
                Legend:
        Run     Type      in
        No.   Scrubber Inlet Gas

         1  O  Recirc.      15
         2  O  Recirc.       0

         3  O  Once through  0
                                        0.9
                                        1.0
       SO, in Rich Solution, S/C  = moles SO9/mole cctive Na
         &t                    cl           £

   % SO, IN SCRUBBER OUTLET GAS VS S/C OF RICH SOLUTION
FOR RECIRCULATING SCRUBBER AND ONCE THROUGH SCRUBBER

-------
 *—                     Table  18 gives the data obtaihed in the once through scrubber.
           Lean solution containing 3. 5 molal sodium with 0.65 S/C ratio wao used in
 L         this run along with 5. 0 liter/min gas  - containing 0. 3% SO, in nitrogen.
           This column turned out  to be very efficient so that with the true counter-
 i
 ^         current operation it was possible to reduce the liquid rate to the point
           where the S/C  ratio in the rich solution approached 0. 98 before the SO-
                        a                                                      £,
 I          content of the outlet gas became 0. 015%.  The relatively small inside
           diameter of the column  combined with the swirling effect of the Teflon
           helix caused a relatively high pressure drop, which in turn was respon-
 v—         sible for the good absorption efficiency.
 '                C.    OXIDATION AND  NO  EXPERIMENTS
 L_                                         X
                       1.    Introduction
 \                            The results of the runs made in the once through scrubber
          are given in Table  19 while the recirculating scrubber test results are pre-
          sented in Table 20.  The tables provide all of the important information
          pertaining to each test.  Oxidation of SO.  to SO4  is reported as a percent
 i         of the SO-, in the entering flue  gas.
 L                 i
                             The absorption system using 3.5 molal sodium sulfite-
          bisulfite with 0.65 S/C was investigated more than any other system.   This
 L_                              a
          absorbent is the same as that used in the John stone Zinc Oxide process.
 ;                      2.     Effect of O2
                             Runs 1 and 2 (Table 19),  which were made without O-
L.        and NO in the feed gas and with a 3. 5 molal,  0. 68 S/C  sodium sulfite-
                 x                                              a
          bisulfite solution,  showed no measurable oxidation.  The presence of CO,
i__        in the feed gas affected neither the oxidation nor the absorption of SO-.
                            Runs 3,  4,  8-1,  8-2,  14, and 16-1 were made with the
L        same solution and with the addition of O^ to the feed gas.  These runs show
          a small but consistent oxidation of about 1% of SO2 to SO^  in the once
(          through scrubber,  which is believed to be due  to the difficulty in preventing
          oxidation of the rich and lean solution samples during sampling and analysis.

-------
                                                                                  TABLE 19
                                                             OXIDATION STUDIES - ONCE THROUGH SCRUBBER
Run
No.
l-la
1-2
1-3
1-4
2-lC
2-2
2-3
2-4
2-5
3-1
3-2
3-3
3-4
4-1
4-2
4-3
4-4
5-1
5-2
5-3
6-1
6-2
7-1
7-2
7-3
Prescrubber
Solution


2



H2°
f.



**2^



H20

H,0
f.
12.8%H2SO4d

12.8%H,SO.
£ V
Inlet Solution
Composition
^
3. 5 molal Na
0.68 S/C
a.

3. 5 molal Na
O£Q c/r"
• OO 3/ V^a


3. 5 molal Na
- 0. 66 S/C
a.

3. 5 molal Na
0.66 S/C
&
3. 5 molal Na
0(\1 C //**
. O7 b/ C
3. 5 molal Na
0.67S/Ca
3. S molal Na
OAA ^/f
• DO O/ \u-
Rate
ml/min
1.72
1.69
1.56
1.45
1.76
1.68
1.63
1.57
1.47
1.70
1.58
1.56
1.47
1.48
1.60
1.70
1.44
1.52
1.26
1.66
1.67
1.84
1.91
1.78
1.68
Moles/liter
soz
1.26
1.26
1.26
1.26
1.25
1.25
1.25
1.25
1.25
1.20
1.20
1.20
1.20
1.20
1.20
1.20
1.20
1.21
1.21
1.21
1..20
1.20
1.21
1.21
1.21
SO4
0.009
0.009
0.009
0.009
0.015
0.015
0.015
0.015
0.015
0.026
0.026
0.026
0.026
0.027
0.027
0.027
0.027
0.032
0.032
0.032
0.052
0.052
0.013
0.013
0.013
Outlet Solution
s/ca
0.88
0.89
0.92
0.93
0.88
0.90
0.91
0.91
0.94
0.90
0.92
0.90
0.92
0.92
0.90
0.89
0.92
1.03
1.08
1.00
0.94
0.91
0.94
0.95
0.98
Moleg/liter
SO2
1.63
1.65
1.71
1.73
1.62
1.66
1.67
1.68
1.72
1.62
1.66
1.64
1.67
1.67
1.62
1.61
1.66
1.36
1.40
1.37
1.61
1.56
1.51
1.52
1.57
so4-
0.010
0.009
0.008
0.009
0.015
0.014
0.015
0.016
0.016
n.d.
0.031
0.026
0.026
0.030
0.030
0.031
0.031
0.275
0.325
0.250
0.076
0.074
0.127
0.138
0.134
Inlet Gas
ppra
NO
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
400
400
400
400
400
0
0
0
NO2
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
400
400
400
0
0
120
120
1ZO
Outlet Gas
mole %
so2
0.0018
0. 0040
0.0022
0. 0022
0.0010
0.0018
0.0010
0.0012
0.0018
0.0014
0.0014
0. 0022
0. 0022
0.0018
0.0018
0.0014
0.0014
n.d.
n.d.
o.d.
n.d.
n.d.
n.d.
n.d.
n.d.
Oxidation
of SO2
to SO f
* *
trace
0
negl.
trace
negl.
negl.
negl.
trace
trace
n.d.
1.5
0.3
0.3
0.8
0.8
1.2
1.0
60.0
60.0
59.0
7.1
6.5
35.3
36.0
33.0
Scrubber
Gas
AP
mm H2O
186
186
185
175
186
186
184
184
188
185
188
186
182
185
185
185
184
187
187
192
185
188
48
48
48
Sulf-ir
Ba2-_ ._>
^*
10 1
i:?.
Ij3
103
101
102
103
102
104
104
104
103
103
103
102
104
102
99
100
101
104
103
106
105
107
NO
          GENERAL NOTES:  The feed gas consisted of S liters/mio of 0. 3 mole-% SOj.  14. 7 mole-% CC«2.  2.8 mole-% Oj in N2,  except as noted, plus NO and NOj as
          indicated.  The gas mixture was saturated with water at 50°C.  The solution temperature w»s maintained at 50°C.   The scrubber was a 0.20 in.  ID x 10 ft tell
          glass tube with a 0.020 in. dia type 304 ss helix in Runs 1 to 6.  The scrubber was changed to a 0. 28 in. ID x 10 ft tall glass tube with a 0.033 in. OD Teflon
          tube helix beginning with Run 7.  See Appendix A, Section TV, for explanation of columns "S/C ," "Oxidation of SO, to SO.", and "Sulfur Balance."
          a.
          b.
          negl.
          c.
          n.d.
          d.
0. 3 mole-% SO2 in NZ in Ron 1.
Na2SO3 - NaHSOj solution.
Negligible
0. 3 mole-% SO2, 14. 7 mole-* CO2 in NZ in Run 2.
Not determined.
See Appendix A, Section n. B.

-------
                                                                                          r
r
r       r
r
                                                                             TABLE 19 (Cont'd)
                                                             OXIDATION STUDIES - ONCE THROUGH SCRUBBER
Run
No.
8-1
8-2
8-3a
8-4a
8-5a
9-1
9-2
9-3
9-4c
10-1
10-2
10-3
10 -4C
11-1
11-2
11-3
11 -4C
1Z-1
12-2
13-1
13-2
14-1
15-1
15-2
15-3
15-4
15-5
!6-ld
l6-2d
l6-3d
l6-4d
Prescrubber
Solution


12.8% H,SO.
2 4


H r*
H2°



H2°



H20


H2°
H,O
2
H20

H,O
2


•
H2°

Inlet Solution
Composition


3. 5 molal Nab
0.67S/C
a

3. 5 molal Na
0.65 S/C
a

3. 5 molal Na
0.65 S/C
a

3. 5 molal Na
0.65 S/C
a
3. 5 molal Na
0.65 S/Ca
3. 5 molal Na
0.65S/Ca
Same as 13

3. 5 molal Na
0.65 S/C


3. 5 molal Na
0.65 S/C,
a
Rate
ml/min
1.78
2.17
2.22
2.08
1.52
1.90
2.30
3.10
2.21
2.94
1.65
2.20
2.10
3.05
1.60
2.20
2.20
2.35
2.20
2.05
2.30
2.14
2.06
2.10
2.07
2.02
1.96
2.07
2.15
2.19
2.15
Moles/liter
so2
1.22
1.22
1.22
1.22
1.22
1.21
1.21
1.21
1.21
1.27
1.27
1.27
1.27
1.25
1.25
1.25
1.25
1.25
1.25
1.22
1.22
1.23
1.23
1.23
1.23
1.23
1.23
1.22
1.22
1.22
1.22
S04
0.027
0.027
0.027
0.027
0.027
0.048
0.048
0.048
0.048
0.009
0.009
0.009
0.009
0.017
0.017
0.017
0.017
0.017
0.017
0.034
0.034
0.040
0.042
0.042
0.042
0.042
0.042
0.038
0.038
0.038
0.038
Outlet Solution
s/ca
0.96
0.93
0.91
0.95
1.04
0.92
0.87
0.83
0.88
0.76
0.88
0.84
0.85
0.81
0.97
0.85
0.86
0.87
0.88
0.76
0.76
0.81
0.87
0.87
0.88
0.91
0.95
0.83
0.84
0.84
0.86
Moles/liter
soz
.74
.63
.64
.71
.87
1.66
1.55
1.46
1.58
1.58
1.81
1.67
1. 70
1.53
1.80
1.59
1.60
1.55
1.57
1.44
1.50
1.67
1.49
1.49
1.47
1.44
1.38
1.47
1.47
1.46
1.46
so4=
0.033
0.035
0.035
0.034
0.037
0.125
0.121
0.094
0.110
0.033
0.049
0.042
0.040
0.045
0.081
0.061
0.059
0.088
0.091
0.082
0.079
0.040
0.057
0.081
0.104
0.144
S.298
0.043
0.066
0.068
0.090
Inlet Gas
ppm
NO
0
0
0
0
0
0
0
0
0
475
475
475
475
475
475
475
475
450
450
0
0
0
450
450
450
450
450
0
0
470
470
NO2
0
0
0
0
0
25
25
25
25
0
0
0
0
25
25
25
25
50
50
25
25
0
0
25
50
100
240
0
25
25
50
Outlet Gas
mole %
S02
0.0017
0. 0038
0.0025
0. 0025
0. 0029
0.0025
0.0025
0.0021
0.0025
0. 001 7
0.0029
0.0017
0. 0021
n-d.
0.0033
0.0025
0. 0025
0.0025
0.0016
0.0033
0.0025
0.0025
0.0029
0.0029
0.0021
0.0021
0.0017
0.0035
0.0035
0.0035
0. 0040
Oxidation
of SO,
to SO =
*4
1.6
2.7
2.6
2.3
2.4
20.0
24.0
21.0
20.0
9.3
8.9
10.0
9.0
12.5
14.8
15.3
13.8
25.6
24.8
13.0
12.7
0.0
5.3
13.3
20.8
33.4
52.5
1.8
9.7
13.8
17.9
Scrubber
Gas
Ap
mm H2O
55
55
55
55
55
44
46
54
45
45
46
44
44
45
45
46
46
46
46
45
45
45
47
46
46
46
46
43
44
44
44
Sulfur
Balance
Out *
-In" *
105
105
102
107
109
102
102
101
101
94
97
99
99
101
103
105
102
103
106
89
90
96
99
101
100
101
100
97
100
100
100
Ol
          See Sheet 1 for General Notes.
          a.      Approximately 24 mg/min fly ash was added to the gas stream entering prescrubber.  Analysis of gas leaving prescrubber indicated that about 95% of
                  the fly ash was removed in the prescrubber.
          b.      Na2SO3 - NaHSOj solution.
          c.      Fly ash approximately 0. 2 wt-% of the feed gas was added to gas entering prescrubber.  About 95% of the fly ash was removed in the prescrubber.
          n.d.    Not determined

-------
                                                                                TABLE 19 (Cont'd)
                                                               OXIDATION STUDIES - ONCE THROUGH SCRUBBER
vO
Run
No.
17-1
17-2a
17-3b
17-4C
18-ld

18-Zd
J
18-3d
19-1

19-2

19-3
ZO-1
20-2
20-3
21-1
21-2
21-3
22-1
22-2
22-3
23-1
23-2
23-3
24-le
24-2e
24-3°
24-4*
Prescrubber
Solution

H,O
2



HO
£



H20
£,

H70
£.
H,0
6
5. 5% FeSO4
and
7. 0% H2S04
5.5% FeSO.
and 4
2. 0% H2SO4
5. 5% FeSO.
and *
2. 0% H2S04

Inlet Solution
Composition
3. 5 molal Na
Oic c //-
• O3 &/ ^*
j a
and
1 3 ppiti iron
as Fe2(S04)3
3. 5 molal Na
0.65S/C
and
30 ppm iron.
as Fe2(S04r
3. 5 molal Na
0. 65 S/Ca
and
250 g NaCl/
liter
7. 0 molal Na
Oftti C If*
. O5 S/C
10.5 molal Na
OAR C //—
. O5 S/C
3. 5 molal Na
0. 65 S/Ca
3. 5 molal Na
OLtL elf*
• 03 o/l*

3. 5 molal Na
0.65 S/C
*
Rate
ml/min
2.06
2.06
2.10
2.12
2.17

2.74

2.19
2.36

2.36

2.60
2.10
1.50
1.03
1.37
1.14
0.53
2.63
2.14
1.65
2.80
2.13
1.51
2.74
2.10
1.53
1.19
Moles
so2
1.26
1.26
1.26
1.26
1.15

1.15

1.15
1.11

1.11

1.11
2.41
2.41
2.41
3.52
3.52
3.52
1.18
1.18
1.18
1.15
1.15
1.15
1.12
1.12
1.12
1.12
liter
so4=
0.018
0.018
0.018
0.018
1.980

1.980

1.980
0.021

0.021

0.021
0.018
0.018
0.018
0.021
0.021
0.021
0.075
0.075
0.075
0.025
0.025
0.025
0.054
0.054
0.054
0.054
Outlet Solution
s/ca
0.87
0.87
0.88
0.87
0.87

0.84

0.87
0.83

0.84

0.82
0.74
0.78
0.85
0.70
0.72
0.86
0.82
0.85
0.93
0.77
0.81
0.90
0.77
0.83
0.90
0.98
Mole*
so2
1.52
1.51
1.53
1.52
1.42

1.37

1.42
1.336

1.356

1.311
2.71
2.86
3.08
3.71
3.87
4.54
1.40
1.44
1.55
1.36
1.43
1.57
1.31
1.41
1.51
1.64
/liter
SOj
0.063
0.063
0.064
0.063
1.980

1.980

1.980
0.048

0.046

0.046
0.036
0.047
0.055
0.172
0.130
0.166
0.086
0.090
0.094
0.038
0.044
0.053
0.071
0.073
0.082
0.085
Inlet Gas
ppm
NO
470
470
470
470
470

470

0
470

470

470
470
470
470
470
470
470
470
470
470
470
470
470
470
470
470
470
N02
30
30
30
30
30

30

0
30

30

30
30
30
30
30
30
30
30
30
30
30
30
30
30
30
30
30
Outlet Gas
mole %
so2
0. 0025
0. 0030
0.0025
0. 0030
0. 0041

0.0040

0. 0055
0.0050

0. 0060

0. 0050
0.0060
0.0080
0.0090
0.0140
0.0140
0. 0280
0.0040
0. 0050
0.0060
0.0030
0.0030
0.0030
0.0040
0.0040
0. 0060
0.0100
Oxidation
of SO,
toso:=
* *
14.8
15.1
15.6
15.5
2.0

2.0

2.0
10.3

9.5

10.4
6.4
6.9
6.3
33..0
20.0
12.0
4.4
5.0
5.0
5.8
6.6
6.9
7.4
6.5
6.9
5.8
Scrubber
Gas
AP
mm H.O
46
45
45
45
45

45

44
45

45

48
47
47
46
46
46
45
47
47
47
47
48
48
34
34
35
34
Sulfur
B^• r.ce
%* «
100
10J
101
101
100

100

too
99

100

100
101
100
104
98
98
100
99
99
101
100
98
98
99
100
101
102
          See Sheet 1 for General Notes.
          a.    Inlet solution also contained 10 ppm hydroqnlnone.
          b.    Inlet solution also contained 100 ppm hydroquinone.
          c.    Inlet solution also contained 1000 ppm hydroquinone.
          d.    Inlet solution also contained 26 g Na,SO4/100 *"'.
          e.  "Scrubber length was only 5 feet.
           L
L


-------
                                                                                                       r
t—
r
r
r
                                                                                 TABLE 19 (Cont'd)
                                                                OXIDATION STUDIES - ONCE THROUGH SCRUBBER
Runa
No.
25-1
25-2
25-2
26-1
26-2
26 -3d
26-4e
27-1
27-2
27-3
27-4f
28-1
28-2
28-3
28-4
28-58
29-18
29-28
29-38
30-lb
30-2b-8
30-3b-8
30-4b'S
31-lb/e
31 -2b
31 -3b
31 -4b« h
32-1
32-28
32-38
32-48
Prescrubber
Solution
5. 5% FeSO4
and
7. 0% H2S04
5. 5% FeSO
and
7. 0% HZS04

, 4
7. 0% H" SO


5. 5% FeSO
and
7. 0% H2S04

5. 5% FeSO.
and
7.0% H2S04

7% H2S04


7
-------
                                                                               TABLE 19 (Cont'd)
                                                              OXIDATION STUDIES - ONCE THROUGH SCRUBBER
Runa
No.
33-1
33-2°
33-3°
33-4b
34-1
34-2?
34-3°
34-4°
35-1
35-2
35-3
36-1
36-2
37-1
37-2c
37-3
37-4
38-2
Prescrubber
Solution

No prescrubber


No prescrubber

No preecrubber

No pre scrubber

No prescrubber

No prescrubber
Inlet Solution
Composition

3. 5 molal Na
0.65S/C
a

7. 0 molal Na
0.65S/C
a
7. 0 molal Na
6C c If
" a
0. 5 N NaOH

5. 0 N NaOH

5. 0 N NaOH
Rate
ml/ min
.94
.90
.97
.88
.20
.61
.56
.54
0.90
1.25
1.40
2.61
3.67
1.36
1.39
1.41
1.41
1.41
Mole*
soz
1.24
1.24
1.24
1.24
2.43
2.43
2.43
2.43
2.38
2.38
2.38
n.8,
n-a.
tua.
n.a.
n.a.
n-a.
n.a.
liter
so;
0.021
0.021
0.021
0.021
0.009
0.009
0.009
0.009
0.064
0.064
0.064
n-a.
n.a.
n.a.
n.a.
n-a.
n.*.
n.a.
Outlet Solution
s/ca
0.92
0.92
0.95
0.86
0.73
0.68
0.69
0.73
0.84
0.78
0.70
0.25
0.14
0.10
0.07
0.06
0.08
0.06
Moles/liter
so2
1.67
1.68
1.64
1.24
2.71
2.51
2.49
2.40
2.94
2.48
2.09
0.13
0.07
0.5!~
0.36
0.29
0.40
0.27
so;
0.021
0.021
0.064
0.208
0.009
0.009
0.048
0.223
0.109
0.271
0.366
0.114
0.073
0.001
0.063
0.137
0.035
0.130
Inlet Gas
ppm
NO
0
0
420
400
0
0
470
400
470
400
400
400
400
0
400
400
470
400
NO,
0
0
30
400
0
0
30
400
30
400
400
400
400
0
400
400
30
400
Outlet Gas
mole %
so2
0.0210
0.0160
0.0160
0.0160
0.0270
0.0120
0.0130
0.0130
0.0180
0.0060
0.0060
0.0000
0.0000
0. 0000
0.0000
0. 0000
0.0000
0.0000
Oxidation
of =G2
toS04=
0.0
0.0
13.5
56.8
0.0
0.2
9.7
53.0
6.6
42.0
68.0
48.0
43.0
0.2
14.1
31.3
8.1
31.0
Scrubber
Gas
AP
mm H2O
36
36
36
37
37
37
36
36
55
55
56
52
57
52
52
53
53
56
Su*?iir
Bs" »-ce
0« „
TiT %
109
1C7
108
92
93
90
91
93
98
94
100
101
85
100
98
96
99
96
00
          See Sheet 1 for General Notes.
          a.    Scrubber height reduced to 5 ft in Runs 33,  34, aad 35.
          b.    Inlet solution contained 100 ppm butylated hydioxy anisole.
          n. a.  Not applicable.

-------
                                                                                   TABLE 20
                                                              OXIDATION STUDIES - RECIRCULATING SCRUBBER
Run
No.
1-4
2-4
3-4
4.2
5-2
6-2
7-2
7-3
8-lC
8-2
9-1
9-2
9-3
10-1
10-2
Circulating Solution
Composition
Fe2(SO4)3 = 28 ppm Fe*
5 g fly ash = 16 ppm Fea
5 ppm iron from fly ash
Fe2(SO4)3 = 100 ppm Fea'b
5 ppm Fe residue
7 ppm Fe residue
2. 5 wt-%
Mg(OH)2
2. 5 wt-%
Mg(OH)2
2. 5 wt-%
Ca(OH)2

5 N NaOH
Rate
ml/min
250
250
250
250
250
250
400
350
350
350
300
300
300
400
400
Solution at Start
Moles/
so2
1.17
1.20
1.16
1.15
1.18
1.14
n.a.
n.a.
nua.
n.a.
rua.
n.a.
n.a.
n,a.
n.a.
liter
so4*
0.018
0.025
0.025
0.033
0.038
0.049
n.a.
n.a.
n.a.
n.a.
n.a.
n.a.
n.a.
n.a.
n.a.
Solution at E
Moles/liter
so2
1.350
1.420
1.350
1.390
1.370
1.270
0.093
0.008
0.070
0.066
0.072
0.041
0,083
0.088
0.104
so4-
0.036
0.042
0.040
0.048
0.044
0.143
0.015
0.120
0.024
0.035
0.003
0.051
0.003
0.017
0.102
nd
s/ca
0.76
0.80
0.76
0.78
0.77
0.87
0.82
0.94
0.67
0.71
0.74
0.82
0.85
0.02
0.02
Inlet Gas
ppm
NO
0
0
0
0
0
480
0
400
400
470
0
400
470
470
400
NO,
0
0
0
0
0
30
0
400
400
30
0
400
30
30
400
Outlet Gas
mole %
soz
0.015
0.013
0.012
0.013
0.013
0.024
0
0
0
0
0
0
0
0
0
Oxidation
of SO-
tosor
% 4
6.9
6.4
5.5
5.8
2.4
38.0
11.8
96.7
18.0
38.0
2.4
38.0
2.4
14.8
68.0
Sulfur
Balance
Out %
TS~
103
100
99
99
100
100
87
102
71d
82d
6ld
72H
70d
93
90
Elapsed
Time
min.
120
120
120
120
120
120
60
60
60
60
60
60
60
60
60
vO
vO
          General Notes:   The feed gas consisted of 5 liters /min of 0. 3 mole-% SO2> 14. 7 tnole-% CO,. 2.8 mole-% O, in N, except as noted,  plus NO and NO, as
          indicated.  The gas mixture was saturated at 50°C in a prescrnbber containing water.   The solution temperature was maintained at 50°C.  The scrubber
          was a 7/8 in.  ID x 3 ft tall glass tube packed with 32 in. of 1/2 in. ceramic Intaloz saddles.  See Appendix A, Section IV, for explanation of columns
          "S/Ca. " "Oxidation of SO2 to SO4=, " and "Sulfur Balance. "
a.
b.
c.
d.
n. a.  Not applicable.
Composition of the circulating solution consisted of 3. 5 mola
1000 ppm hydroquinone added.
No O2 in feed gas.
Some of the solids were held up on the packing* making an exact material balance impossible

-------
              3.    Effect of NO  and 2. 8% O
                               x            Z
                   Runs 5 through 15 were made in order to determine the
 effect of NO  in the presence of O, upon the  extent of oxidation in the once
            X                    6       ,
 through scrubber with the 3.5 molal, 0.65 S/C  sodium sulfite-bisulfite
 solution.  The results of these runs have been plotted in Figure 6.  Two
 series of runs were made in order to cover all ratios of NO/NO,  of interest
 - the first series without NO but with increasing amounts of NO,, and the
 second series with the normal range of NO and increasing quantities of
 NO^.  The first  series of runs (plotted as a solid line on Figure 6) shows
 slightly less oxidation than the second series (dotted line).  In both curves,
 the percent of NO^ in the feed gas was controlled with the percent oxidation
 as the dependent variable.  The data show that the NO^ in the  scrubber feed
 gas caused the large  increase in SO, oxidation rate.  Some of the  NO was
 known to be oxidized to NO- in the prescrubber and in the line  between the
 prescrubber and scrubber; otherwise, the two lines  might have been identi-
 cal.
             4.     Effect of Fly Ash
                   Three runs were made in which the feed gas to the once
 through  scrubber contained appreciable  quantities  of fly ash (Runs 9-4,
 10-4, and 11-4).  During these runs, fly ash, approximating 0.2 wt. -%
 of the feed gas,  was added to the prescrubfyer feed.  The prescrubber for
 these runs had been modified so that only about 95% of the fly ash was re-
 moved in the prescrubber.  The lean solution in these runs was the standard
                                               t
 3. 5 molal sodium solution that had not been in contact with fly ash.  A com-
parison of the amount of oxidation experienced in Samples 9-4,  10-4, and
 11-4 with the oxidation experienced in the same runs  (Samples  1-3) with
no fly ash in the feed  gas,  shows that within the limits of error there was
no change in the level of oxidation.  It should be  assumed,  however, that
in plant operations the solution would be recycled and would gradually ac-
cumulate an increasing concentration of fly ash or  some ingredients from
the fly ash that might greatly affect the extent of oxidation.

-------
o  :
1 ' r f ( < < i [ < i r i f r •• f . r
80 -....-..

70

60
50
CD
ni
O
r° 40
a
• |H
4)
• iH
^) 'in
." Jv;
<3
0^
l£ 20
"3
10


1 1 1 1 |
— Notes: 	 . 	 	 ,_n
1. Glass scrubbing column with 10 ft of Teflon helix packing.
2. Feed gas 5. 0 liter/min of 0. 30% SCfc. 2. 8% 0%, 14. 7% CO,, in N2
3. Column temp' = 50°C; Pressure '- 750 mm"Mg Abs.
4. Le
•










an solution = Na_SO, - NaHSO- of approx.
5 molal sodium, a/CTa = 0. 68.






3
/\
s S
//
*Y/
o ./*/__

//°
x /
X/
Y
cf
Qfv 	 , ...~... »









s
'
^s
s










^•'
^
s'













^*^x"










i
-V--~~~~'~ ~"~~*
—- •*-








'
Legend:
O— — -»=O - °% NO in Feed +• NO,
£*
X-»«.-.e!=.x - 0.040 -.047% NO in Feed -f NO2



i r i
V 0.005 0.010 0.015 0.020 0.025 0.030 0.035 0.040 0.04
                                             Mole % NO2 in Feed Gas




                 EFFECT OF NOx ON OXIDATION OF SO2 TO SO4= IN ONCE THROUGH SCRUBBER







-------
                    One run was made to check this effect in the recircula-            -J
 ting scrubber.   In Rum 2 (Table 20),  the lean solution was shaken for 15
 i.-riuutes with 5 g fly ash per liter and then decanted from the fly ash before            -
 adding the solution to the scrubber.  This resulted in 16 ppm  iron in the
 solution.  Comparing this  run with Run 3, in which the lean solution had               J
 not been in contact with fly ash, but 5 ppm iron was added as  a salt, shows
 that oxidation had increased only slightly due to the higher iron content.                :
 This is an insignificant increase in oxidation when compared to the  3 to
 4-fold increase  found in the inhibitor screening tests (see Table  16).  In               ;
 the latter tests, about 0. 1  g  of fly ash was added directly to 100 ml  of the
 lean solution and the oxidation increased from about 1 to about 5% based                :
                                                                                      |
 on the SO, in solution.  In  this case perhaps it  was the large surface of               —•
 the fly ash and possibly the oxygen adsorbed on the fly ash that caused the
 large increase in oxidation.                                                          _j
              5.    Effect of  Ferric Ions in Solution
                   Runs 3,  4 and 16 (Table 19) made in the once through              "^
 8crubber indicated that 40 ppm iron in the solution did not materially in-
 crease the level of oxidation. This was rather surprising, since many               •—'
 investigators have reported that iron is a mild oxidizing catalyst in sulfite
 solutions.                                                                          _J
                   A series of tests was made in the recirculating scrubber
 to check the  effect of various iron concentrations on oxidation.                        _:
                   Runs 1, 2, 3 and 4 (Table  20) were  made in the recir-
 culating scrubber with increasing quantities of iron added to the lean                 —;
 solution and  no NO  in the feed gas.   Results  showed about 25% increase
                  Ji
 in oxidation as the iron concentration was increased from 5 to 28 ppm.                _
 The accuracy of  the oxidation value obtained for Run 4  (100 ppm iron) is
doubtful,  since a dark solid precipitated from the rich  solution.  These
 results indicate that the presence of iron in the sulfite  solution may in-
crease the oxidation 25% or more when absorbing SO, from gases  con-                  ;
taining O2 but no NO .

-------
              6.    Effect of Inhibitors
                   Several inhibitors, namely hydroquinone, mtrilotri-
 acetic acid (NTA) and butylated hydroxy anisole (CaO-7)#, were found to
 be effective in lowering the extent of sulfite oxidation in the  inhibitor
 screening tests.  These tests were made with O, but with no NO present.
                                               £                X
 Tests made in the once through scrubber showed that these inhibitors were
 not effective in most cases when  NO  was present.  The exceptions to this
                                   2t
 statement were the  runs with  10.5 molal potassium sulfite-bisulfite  solu-
 tion and NTA inhibitor (see Runs 28 and 29,   Table 19).  In  this case, the
 inhibitor lowered the oxidation from a maximum of 69% to 7%, with the
 normal 95/5 ratio of NO/NO,  present in the feed gas.  In these runs, the
 effectiveness of the NTA inhibitor is  believed to be due to  the high pH of
 the solution.  The pH of the 10. 5  molal 0. 5 S/C  potassium  solution  in
                                              2L
 Run 28 was 12. 6 compared to  a 6. 5 pH of the  3. 5 molal 0.  65 S/C  sodium
                          i                                    cL
 solutions.  However, the NTA was not able to lower the level of oxidation
 below 7-9% in Run 29 where the lean  solution pH was 7. 5.
                   Runs 17 and 27 (Table 19) made with 0. 001-0. 10% hy-
 droquinone in the  presence of  NO  failed to suppress the oxidation.   Runs
                                JC
 31 and 34 (Table 19) made with CaO-7 inhibitor also did not lower the
 oxidation rate in the presence  of NO  .
                                   Jt
                   Since the recirculating scrubber was found to produce
 a higher level of oxidation than the once through scrubber  (discussed later)
 it was logical to use this scrubber to  test the effectiveness of hydroquinone
 as an oxidation inhibitor in the absence of NO .  Comparing  Runs 3 and 5
 (Table 20) it can be seen that in the absence of NO  addition  of hydroquinone
 lowered the oxidation rate a little over 50%.
             7.    Effect of High Concentrations of Na^SO. and NaCl
                   Run 18 (Table 19) was conducted with a feed solution that
contained 26 grams of anhydrous sodium sulfate per 100 ml of the normal
stock sulfite-bisulfite solution.   The resulting solution was found to be
* CaO-7 is trade name of Ashland Chemical Co.


-------
                                                                                     —J
 slightly supersaturated at 25 C, but less than saturated at 50°C which is              -*
 tbf? temperature noi-mally employed for absorption.  The data for Run 18               .
 clearly indicate that oxidation is markedly inhibited at high sulfate con-              _,
 centrations both in the absence and presence of NO .   This can be explained
 in terms of the very limited solubility of oxygen in aqueous solutions at               _j
 high ionic strength, and the supposition that little or no oxidation occurs
 in the gas phase.  The observed inhibition of oxidation was tried by em-
                                                                                     •—^
 ploying other soluble salts,  such as sodium chloride.   Here, however,
 with 250 g NaCl/liter  (Run 19) the  effect was not as great.                             I
                                                                                     -J
                    The results obtained with sodium sulfate may be  of
 value in connection with the Johns tone Zinc  Oxide  process, in which oxida-
 tion in the scrubber normally occurs  to the  extent of 10% or more.   It
 would be  required, however, that the precipitation of zinc sulfite  through
 addition of zinc oxide to the rich scrubber solution not  be accompanied by             —>
 co-precipitation of sulfate.
                                                                                      i
              8.     Effect of Oxidation of SO, on Required Solution Rate               —-'
                    In order to aid in selecting the proper lean solution                 •
 rates for experimental runs,  calculations were made to show how oxida-
 tion of SO? to SO ~ in the scrubber affects the rich solution S/C ratio.                <
          f*      4                                             21
 In Johnstone's work, a  0. 85 S/C  was used for design purposes, and in              ~~
                                ct
 a commercial plant it is believed that this may be the highest practical
 ratio that will maintain less than 150 ppm SO, in the outlet gas.                      -'
                   Figure 7 illustrates how the lean solution rate must be              \
 increased at a given rich solution S/C  as the oxidation level increases.              ""'
                                     cL
 For example, at 0. 85 S/C , the solution rate must be  increased from
                          3.
 2. 0 ml/min at no oxidation to 2. 45 ml/min at 30% oxidation.                          —
             9.    Effect of Solution Flow Rate
                   The  effect of solution rate on the extent of oxidation in
 the once through scrubber can be seen from Figures 8 and 9.  With the ex-
 ception of the 10. 5 molal concentrations, there was almost no change in              ~~
 oxidation with solution rate for either the sodium or potassium sulfite -
 bisulfite solution.  The  10. 5 molal solution for both sodium and potassium            —<
 solutions showed a rapid increase in the oxidation as the solution rates
were increased.                                                                     —

-------
o
to
*0
o
 r-j
CQ
     1.00
     0.99
     0.90
     0.85
     0. 80
     0,75
         1.5
                                     Notes:
                          1.
                          2.

                          3.
5. 0 liter/min feed gas containing 0. 30% w.^-
Lean solution contains 1.81 moles active sodium/
liter and 1.22 moles of SO2/liter._
Each mole of SO? oxidized to SO.  lowers the active
sodium content OA the rich solution by two moles.
                     2.0
                                  3.0
3.5
                                          Lean Solution Rate ml/min

                                                NATION O:

                                                 Figure 7
THEORETICAL EFFECT OF SO0 OXIDATION ON RICH S/C  IN A ONCE THROUGH SCRUBBER

-------
CO

n)

O
o>


£
8
 CM
O
CO

t£



i
56
CO
AO
"to
44
40
32
28
24
20
16
12
8
4
0











—n^
Ur









>
^ /
S J
*^S&
^

Symbol
o -c
/•u-,. ....— rf*
	 c
O™ f
LEGEND:
Fe"*"*" in Scrubber
Molality S/Ca Pre scrubber ht, ft
> 3.5 0.65 No 10 !
i 7.0 0.65 No 10
J 10.5 0.65 No 10
* ^ =. n A^ VAB c;

Notes;
1. 0. 28-ii
2. 0. 87-ii
Intal
3. 50°C s
P 4. Tnlet gi
f 0.04



o 	
7. 0 KCol*l


3.5
. Molal with Fe++ in prc
^
i. ID glass scrubbing column Teflon helix.
i. ID glass prescrubber with 32-in. of 1/4
ox saddles.
c rubber temperature.
is contained 0. 3% SO2» 2. 8% Oz> 14. 7% CO,,
7% NO, 0.003% NO2.
fl 3. 5 Molml
Q

^y

scrubber
.
              1.0
                                                           4.0
                                                              5.0
                             2.0                  3.0


               Scrubber solution rate, gallons per 1000 scf flue gas



OXIDATION VS SOLUTION RATE FOR SODIUM SCRUBBING SOLUTION IN ONCE THROUGH SCRUBBER
      I.,
t  _
I...
                                        Figure 8




                                           L-   L  .    I  -,

-------
CO
rt
O
T3
(U
 0)
 N
.f-i
2
4
cf
to
ou
56
52
48
44
40
36
32
28
24
20
16
12
8
4
0







_




/

-Y

1
/
/
LJ




10. 5 Molal with Inhibitor
^


&

LEGEND:
*7> ++ •
h~ " re in
Symbol Molality S/Ca Prescrubber

O— - O 3.5 0.65 Yes
Q— ^ 3.5 0.65 Yes
^&— -«& 7.0 0.65 Yes
A — A 7. 0 0. 57 Yes
E— -B 10.5 0.50 Yes
E£— -0 10.5 0.57 Yes
Notes:
Scrubber
ht, ft
10
5
5
5 (no helix)
5
5
1. 0.28 -in ID glass scrubbing column with Teflon helix.
2. 0. 87-in ID glass prescrubber with 32-in of 1/4 Intalox saddles.
3. 50°C scrubber temperature
4. Inlet gas contained 0. 3% SO2, 2. 8%
CO0, 0. 047% NO, U.UU3yr NO,.
£ £


3. 5 Molal
€^ I ")
(^r^.
O
J ^ , • , , 7. 0 MolsJ

02, 14. 7% C02, 14. 7%
\
©
"O
,
^
             1.0
                          3.0                 4.0
Scrubber solution rate, gallons per 1000 acf flue gas
5.0
        OXIDATION vs SOLUTION RATE FOR POTASSIUM SCRUBBING SOLUTION IN ONCE THROUGH SCRUBBEF

-------
              10.   Effect of Type of Scrubber
                   A comparison of the amount of oxidation under similar
 conditions with no NO and a minimum of iron or fly ash shows that the
                      Jt
 oxidation in the recirculating scrubber is two to five times greater than
 the  oxidation in the once through scrubber.   Compare Runs  3, 4,  8-1,
 and 8-2,  (Table 19) with Run 3 in Table 20.  The much lower level of
 oxidation found in the once through scrubber is due to a lower amount of
 oxygen absorbed in that scrubber solution.
                   Similar runs, which were made with the  5-ft and 10-ft
 once through scrubbers (see Figure 9, O and O points) indicate very little
 difference in the amount of oxidation.  A similar set of runs, made with
 and without the helix in the column (see £ and A points  in Figure 9),
 showed only a small difference in the level of oxidation.  No explanation
 can be offered at this time to account for  these results.   It would  seem
 that the top half of the  10-ft column (where very little SO- is absorbed
                                                       £t
 and  where oxygen from the gas would be absorbed into the lean solution,
 which is high in sulfite) would promote high levels of oxidation. In the
 5-ft column, where there is only one-half of the liquid surface area,
 most of the area is used  in absorbing SO2, so that the average solution
 sulfite concentration and pH are much lower.  Inasmuch as the amounts
 of oxidation were  found to be similar in the two cases, it must be  con-
 cluded that other factors  were involved.  The removal of the Teflon helix
from the column lowered the amount of liquid surface  in the  column, and
also lowered the gas turbulence.  This was reflected by  a relatively large
loss in SO-,  scrubbing efficiency.   Under these circumstances it is logical
to expect a large decrease  in the  extent of oxidation,  but such was not ex-
perienced.  Evidently,  a  better understanding of the oxidation mechanism
is needed in order to be able  to predict the level of oxidation.

-------
               11.   Effect of NO, Removal in the Prescrubber

                    Since it was found that most of the  oxidation is due to
 NO, in the scrubber,  it is assumed that if the NO, could be removed in
 the prescrubber the rate of oxidation would be lowered to about the  same
 rates as that caused by the oxygen in the flue gas.  This was found to be
 substantially correct when an efficient prescrubber similar to that des-
 cribed  in Appendix A, Section I, was used circulating  a solution contain-
 ing Fe   ion.
                    Runs 27-1 and 27-2 (Table 19) with a feed gas contain-
 ing normal amounts of NO  and with 5. 5% FeSO^ in the prescrubber
 resulted in < 2% oxidation or about the  same that would be expected with
, no NO  in the feed gas.  However, when the NO- in the feed gas  was in-
 creased, the oxidation also increased to about 3. 5%, indicating that not
 all of the NO, was being removed in the prescrubber.
             £•
              12.  Effect of Type of Solution
                   a,    Solution Molality
                         In both the sodium and potassium solutions,  the
 7. 0 molal solutions were found to result in the lowest  oxidation.  Also,
 in both the sodium and potassium systems, the 10. 5 molal solutions ex-
 hibited the highest levels of oxidation (see Table  19).   This is attributed
 to the high sulfite  concentrations and the corresponding high  solution pH's,
 both of which tend to produce high levels of oxidation.  Both sodium and
 potassium solutions of 10. 5 molality decrease in oxidation levels as the
 solution circulation rate is decreased,  which reduces the average solution
 pH, and lowers the average quantity of sulfite in the scrubber.  However,
 this does not apply to the 7. 0 molal solutions,  which have higher pH's and
 sulfite concentrations  than the  3.5 molal solutions,  but nevertheless pro-
 duce lower oxidation rates. Evidently more complex and offsetting factors
 are involved.

-------
                    b.     Type of Solution
                          The sodium sulfite-bisulfite solutions seem to
 have rlightly lower levels of oxidation than the potassium sulfite-bisulfite
 solutions at 3. 5 molal concentration about 5% vs  > 6% - see Runs 22 and
 23, (Table 19).
                          Both 5N and 0. 5 N sodium hydroxide solutions
 in the once through scrubber with high NO  in the feed gas (0. 04% NO +
                                         Jt
 ''. 04% NO£) had only slightly lower oxidation  levels than the 3. 5 molal
 sodium sulfite-bisulfite solutions (see Runs 37-3,  36-1, and 33-1,
 Table 19).
                         Runs 7-10  (Table 20) give the oxidation data ob-
 tained on the recirculating column with sodium hydroxide solutions and
 calcium and magnesium hydroxide slurries.  The 5 N NaOH solution with
 equimolar quantities of NO/NO2  in the feed gas produced 68%  oxidation
 in the recirculating system compared to 31%  in the once through column.
 This is in line with the results obtained with the 3. 5 molal sodium sulfite-
 bisulfite systems where the oxidation was found to be several  times  higher
 in the recirculating system.  The oxidation level for Ca(OH)_  was found
 to be the lowest of the three hydroxides (38%) and highest for the MgfOH),
 (96%).
       D.    NOx REMOVAL DATA
             Table  21 summarizes the NO  removal data obtained during
 the project for various solutions  and several types of apparatus.  The data
 show some inaccuracies due to NO  analytical problems.  It is also possible
                                 Jt
 that the desired NO composition was not always maintained exactly due to
 the low gas flow rates and to  the  problems associated with manual control
 of six gas streams. The NO  stream, consisting of 5% NO in Ng» was
especially difficult to control since it frequently fouled the rotameter and
pressure regulator.  This did not make much  difference during most of
the runs,  but when making NO  removal determinations, a small change
                             x
in NO flow rate  during the times when the feed and product samples were
taken could cause a substantial error in the NO measurements.

-------
                                                                     TABLE 21
                                                     NOx REMOVAL WITH AQUEOUS SYSTEMS
Run
No.
17-4*
31 -4*
32-3

32-4
33-3 £
34-3 |
34-4 H
35-2* £
35-3 |
36-1
36-2
38-1*
38-2
6-2
7*3
8-1* S
8-2 £
3
9-2 £
9-3 |
10-1 ,°
fci
10-2
Scrubber
Type



H
a
.£>
•iff
H
•a*
is
^H
3*
5*


h
v -_.
1
2 „
« o
»33
~ |^|
u . o
•l! °!fl
0{ S m
Height
10.0
5.0
5.0

5.0
5.0
5.0
5.0
5.0
5.0
10.0
10.0
10.0
10.0
2.7






Solution
Type
Na
K
K

K
Na
Na
Na
Na
Na
NaOH
NaOH
NaOH
NaOH
Na
Mg(OH)2
Mg(OH)2
Mg(OH)2
Ca(OH)2
Ca(OH)2
NaOH
NaOH
Cone.
3. 5 M
3. 5 M
7. 0 M

7. 0 M
3.5 M
7. 0 M
7. 0 M
7. 0 M
7.0 M
0. 5N
0. 5 N
5. ON
5. ON
3. 5 M
2.5%
2.5%
2.5%
2.5 %
2.5%
5 N
5N
Rate
(xnl/xxiia)
2.1
2.9
1.6

1.5
2.0
1.6
1.5
1.3
1.4
2.6
3.7
1.3
1.4
300
350
350
350
300
300
400
400
Gas Npg Content by
Rotameter Reading
NO
ppm
470
400
470

400
470
470
400
400
400
400
400
400
400
470
400
400
470
400
470
470
400
NO2
ppm
30
400
30

400
30
30
400
400
400
400
400
400
400
30
400
400
30
400
30
30
400
NOx ppm by Gas Analysis
Feed
to Pre-
scrubber
520
807
.

-
-
-
.
880
-
_
-
-
-
550
-
-
.
-
-
-
Feed
to
Scrubber
470
600
434

878
530
496
1000
481
704
695
842
856
740
460
880
786
725
877
702
583
827
Scrubber
Overhead
Gas
430
440
437

520
521
510
598
498
502
691
730
813
640
430
465
307
692
705
653
583
343
% NOx Removal
InPre-
scrubber
10
26
.

-
-
-
.
45
-
.
-
-
-
16
-
-
:
-
-
~
In
Scrubber
8
20
0

41
2
(-3)
40
(-3)
28
0
13
5
13
5
43
61
5
20
7
0
58
NOTES:  K =  potassium sulfite-bisulfite
          Na=  sodium sulfite/bisulfite
          3. 5 M, 7. 0 M  = moles of base per 100 moles water in feed solution
          0. 5 N, 5. 0 N  = normality of feed solution,
          *  Prescrubber used in these runs only
          **No SO2 in the feed to this run - for NO^ measurement without SC«2


-------
              Nevertheless,  some trends are apparent.  The data indicate
 that about 40-60% of the NO was removed when equimolar quantities of NO
 and NO- were present.  The amount of NO  removal depends upon the type
        ™*                                 *t
 of solution as well as the type  of scrubber used.  In the recirculating
 scrubber, 5N NaOH and 2. 5 wt-% Mg(OH), slurry were found to  remove
 about 50% of the equimolar NO/NO2 whereas the 2. 5 wt-% Ca(OH)2 slurry
 was found to remove only about 20% of the equimolar NO/NO, (see Runs
 7-3, 8-1, 9-2, and 10-2).   In the  once through scrubber,  20-40% of the
 equimolar NO/NO, was removed by both sodium and potassium sulfite-
 bisulfite  solutions (see Runs 31-4, 32-4, 34-4 and 35-3).   The removal
 of NO  from flue gases containing a 470/30 ratio of NO/NO,  is in the range
      x                                                   ^
 of 0 to 10% or to about the concentration of the NO9 in  the NO  (about 5%).
                                                 £*          X
              Runs 31,  32, 34 and 35 also show that the absorption of NO
                                                                      x
 into the sulfite-bisulfite scrubbing solutions using the 5-ft once through
 column had no effect on SO2 removal efficiency.
              Although the results  are not conclusive, it appears that adding
 NO- to flue gas to obtain an equimolar ratio of NO/NO- will not lower the
 NO  content of the scrubbed gas significantly.
   X,
        E.    SULFATE REMOVAL WITH REVERSE OSMOSIS
              The possibility of separating sodium sulfate from sodium
 sulfite-bisulfite solutions was investigated.  Calculations indicated that
 the osmotic pressure of these solutions was  in the range of 1150-1200 psi,
 which could be handled with available laboratory test equipment.  Evalua-
 tion of the sulfite-bisulfite-sulfate system showed that  the sulfate would be
 retained and that some  of the sulfite-bisulfite would pass through a high
 flux membrane.
             A brief experimental program was established.  Two solutions
were prepared.   The first was equivalent to a rich,  spent  solution containing
 sodium sulfite, sodium bisulfite, and sodium sulfate.   The quantity of sulfate
added was equivalent to the sulfate formed by 10% oxidation of the SO2 absor-
bed from  a flue gas.  The  second solution contained sodium bisulfite and the

-------
 same quantity of sodium sulfate as in Solution 1.  Sodium sulfite was not
 added in Solution 2 in order to check the difference in permeability, if
 any, between sodium sulfite and bisulfite.  The solutions were prepared
 under oxygen-free conditions  using boiled distilled water.  The solution
 compositions are given in Table 22.
              The test equipment shown in Figure 10 is one of the standard
 systems used for R/O membrane testing and performance.  The membranes
 used in  the test were designated as CAM-38B-80, which were formulated
 for high flux of brackish waters.  The feed solution and the product streams
 were purged with nitrogen to minimize oxidation during the test runs.   The
 test results are presented in Table 23.
             The data indicate that a partial separation of sodium sulfate
 from sodium sulfite and/or sodium bisulfite is feasible, since up to 49%
 of the SCX content of the solution  permeated through the membrane while
 essentially all of the sodium sulfate was retained in the waste.
             The economics of the system are discussed in Part Six,
 Section II. C.
       F.   OTHER SULFATE REMOVAL STUDIES
             Approximate thermal requirements to regenerate zinc oxide
 from zinc sulfate were calculated in conjunction with the fluidized zinc
 oxide process (see Part Six, Section III).  Although the regeneration
 temperature is high, this approach may have merit if oxidation of zinc
 sulfite to sulfate cannot  be controlled at a very low level.
             The investigation of the  use of ion exchange methods for re-
 moval of sulfate from the sulfite-bisulfite absorbent was not pursued due
 to the probable high cost (again compared with the direct addition of lime
 as proposed in the Johnstone Zinc  Oxide process). An  ion exchange sys-
tem would be burdened not only with the initial cost of resins,  but also
with resin replacement costs due to attrition, and with  regenerant chemi-
cal costs.

-------
                          TABLE 22
                 REVERSE OSMOSIS TESTS
                  SOLUTION COMPOSITION
    Component
Sol.  No.  1
Sol. No. 2
Water
NaHSO3
Na2S03
Na2S04
Theoretical content:
SO,
so4=
3000 ml
436.8 g
 87.6 g
 17.4 g
  8. 85 %
  0.332 %
 3000  ml
 606.6 g

  17.4 g
  10.30 %
   0. 324 %

-------
(        r
r       r
r
r       f
       HIGH PRESSURE
          PUMP
                                      3-GAL FEED
                                       RESERVOIR
                                                              THEtHCOSTATlC BATH
                                             arises
                                            PSSSSUSE
                                                                                       BACK PRESSURE
                                                                                       CONTROL VALVE
                                                                                0-2000 PSI GAGE
                                                  0-20OO PSI  BASE
                                                      FLEXIBLE HIGH PRESSURE LINES
                                                       PRODUCT WATER OUTLET
                              •BURST DIAPHRAGM
                               2523 PSI  (OR
                               PRESSURE RELIEF
                               VALVE)
                                                                                                      3 KS TEST CELL
                                           TEST CELL  ASSEMBLY - FLOW DIAGRAM

-------
                                           TABLE 23
                               REVERSE OSMOSIS TEST RESULTS
Solution
No.l
No. 2
No. 1 + Waste
No. 2 + Waste
Product No. 1
Product No. 2
Volume
ml
3335
3416
2890
2921
445
495
Test Time
min
_
—
—
—
75
97
Flux
gfd *




23.7
20.4
pH
5.60
4.50
5.58
4.40
5.20
4.62
S
g/ml
.0885
. 1030
.0968
. 1163
.0378
.0505
°2
% thru
membrane
_
—
—
—
42.7
49.0
SC
ppm
4600
4500
5300
5100
300
300
4=
% thru
membrane
—
—
—
—
6.5
6.7
gfd  = gal/sq ft/day
I. -   L

-------
              There is a direct relationship between the concentration of
 ions to be exchanged and the quantity of regenerant chemical required.
 Deionization of water is usually economical only where total dissolved
 solids are not greater than 500 ppm  (Reference 27); ion exchange pro-
 cesses might be economically justifiable at substantially higher concen-
 trations only when the product has greater value than water.  In the
 aqueous solution processes for removing SO, from flue gases a substan-
 tial quantity of the sulfite in the absorbent is oxidized to sulfate.   For
 example, in the Johnatone Zinc Oxide process it was assumed that the
 equivalent of 10%  of the absorbed SO, was oxidized. This resulted in a
 SO.   concentration exceeding 2. 5 wt-% and a SO,   concentration > 6 wt-
 % in the rich absorbent.
             Even if an anion resin is available which would selectively
 exchange the sulfate ^on from the  absorbent mixture, the resin would have
 to be  regenerated with sodium hydroxide or other base.  The quantity of
base would be theoretically equivalent to the quantity needed for direct
 reaction (such as the lime addition cited above).   It is evident that an ion
exchange system would be more expensive than the  route of lime addition.

-------
                                PART SIX                                        -'
                   PROCESS AND ECONOMIC EVALUATION
                                                                                  —J
 I.      INTRODUCTION
        —«*^MMB^^H^^^B«nW««Ma^BMMW>                                                           (
        The objective of this task, was to support the laboratory invest!-             —
 gations of Tasks A through D (described in Part One herein) with prelimi-
 nary process evaluations, designs and/or economic analyses to illustrate            _,
 the economic feasibility of the concepts undergoing scrutiny in the labora-
 tory.  In addition to these investigations, similar evaluations and analyses
 would be made for aqueous processes under development by other NAPCA
 contractors.                                                                       !
                                                                                  _J
        The evaluations performed during this period covered work on the
 Johnstone Zinc Oxide Process, the Fluidized Zinc Oxide Process and on              [•
                                                                                  _J
 the Magnesium Base Slurry Scrubbing Process. The latter is under de-
 velopment by Babcock and Wilcox Company for NAPCA under Contract               !
 CPA  22-69-162 (and too,  more recently, by the Chemical Construction              ~i
 Corporation for NAPCA under Contract CPA 70-114).                                f
 II.     JOHNSTONE ZINC OXIDE PROCESS                                        ^
       Additional economic evaluations of the  Zinc Oxide process were               [
                                                                                  .—;
made to  supplement the data in the Phase III study (see Volume One) based
on conversion of sulfur dioxide to sulfur.  A re evaluation was also made              f
of the system in which concentrated sulfuric acid is produced but with the            ~"
acid being sold at lower prices than those used in the initial analysis in               <
Phase III.  These evaluations were applied to all four plant cases*.                  ->
Both 90% and 70% plant factors were used for Cases 1 to 3, which apply
to power generating plants.  Case 4, which applies to the smelter facility,            _j
was considered only at a 90% plant factor since it would not be subject to
                                i                                                   •
the load variations of power plants.                                                 f
*  See Appendix B for explanation.

-------
                  A.    CONVERSION OF SULFUR DIOXIDE TO SULFUR
                        The capital and operating costs were based on data developed
           by Allied Chemical Corp. under NAPCA Contract PH 22-68-24.  This
1           information was submitted to Aerojet by NAPCA, and provided data about
*"         the Asarco process, which uses methane to reduce essentially oxygen-
           free  sulfur dioxide to sulfur.   The capital and operating costs submitted
L.         were for a  1400 MW power plant, and were used by Aerojet without change
           for Cases  1 and 2.  The data were factored for Cases 3 and 4.  The capital
|_         investments for complete systems, including the  Asarco process, are
           summarized in Table 24.
,_.                      The profitability was estimated at both 70% and 90% plant
           factors. Selling prices for sulfur were  set at three levels:  $10,  $20,  and
           $30 per long ton.  Tables 25 and 26 summarize the profitability.  A loss
           is indicated for each case.
                                                 1
_                B.    CONVERSION OF SULFUR DIOXIDE TO SULFURIC ACID
                       In the initial Phase III analyses,  sulfuric acid prices per ton
L.         were selected at three levels:  $34 which was the  current price in 1968;
           approximately 70% of the then  current -  $23. 50; and at prices required to
           break even.  These  selections were governed by the fact that similar prices
           were being used by  others in the evaluation of other SO, removal processes,
           both aqueous and nonaqueous.
                       Since these prices were later considered to be optimistic,
           the economics of the Johnstone Zinc Oxide process were reevaluated with
           the sulfuric acid to be sold at $7,  $10, and $13 per ton.  These evaluations
          were made for the four cases,  and are summarized in Table  27, at 70%
~        plant factor, and in  Table 28,  at 90% plant factor.  It is  apparent that the
           data indicate a loss for all systems evaluated.

-------
                                            TABLE 24
                                CAPITAL INVESTMENT SUMMARY
                     JOHNSTONE'S SODIUM SCRUBBING/ZINC OXIDE PROCESS
                       WITH CONVERSION OF SULFUR DIOXIDE TO SULFUR
Capital Investment,  Thousand $
       Process Equipment
       Plume Reheat Equipment
       Sulfur (Asarco Process) Equipment
            Total Plant
Working Capital

Total Investment
Capital Requirements,  $/kw
Case 1
15,085
205
2, 100
17, 390
769^
18, 159
12.97
Case 2
11.925
205
2,100
14, 230
676^
14, 906
8.28
Case 3
3,346
72
700
4,118
5ll
4,169
18.95
Case 4
8,785
46
1, 900
10,731
622i
11,353
-
*  See Table 110 of Volume I for comparison
^  At 70% Plant Factor
2  At 90% Plant Factor
           1,1.
(..„   i..

-------
                     r •"
r.
r
r
r
r
                                                TABLE 25*
                             PROFITABILITY - ZINC OXIDE PROCESS (JOHNSTONE METHOD)
                                PLANTS OPERATING AT 70% PLANT FACTOR
              SO2 CONVERTED TO S (ASARCO PROCESS) -- PLUME REHEAT FROM 112° TO 200°F
     Sales, M long tons sulfur
     Sales, M $

fr^
to    Operating Cost
            Process
            Plume Reheat
            Sulfur (Asarco Process)
                  Total

     Loss, M $/Year

     Loss
            $/ Ton of Coal
            Mill/kwh
     Price at Break-even,
            $/long ton sulfur
                                            Case 1
                   Case 2
                               Case 3
Sulfur Price, $/Long Ton
10
84.5
845
5082
502
934
6518
5673
1.60
.65
20
84.5
1690
5082
502
934
6518
4828
1.36
.55
30
84.5
2535
5082
502
934
6518
3983
1.12
.46
77
10
84.5
845
4473
502
934
5909
5064
1.42
.58
20
84.5
1690
4473
502
934
5909
4219
1.19
.48
30
84.5
2535
4473
502
934
5909
3374
.95
.39
70
10
13.3
133
1225
106
273
1604
1471
2.66
1.09
20
13.3
266
1225
106
273
1604
1338
2.42
.99
30
13.2
39S
1225
106
273
1604
1205
2.18
.89
121

-------
                                                  TABLE 26*
                              PROFITABILITY - ZINC OXIDE PROCESS (JOHNSTONE METHOD)
                                 PLANTS OPERATING AT 90% PLANT FACTOR
                SO2 CONVERTED TO S (ASARCO PROCESS)--PLUME REHEAT FROM 122° TO 172°F
      Sales, M long tons sulfur
      Sales, M $

      Operating Cost
ts>            Process
             Plume Reheat
             Sulfur (Asarco
              Process)
                  Total
      Loss, M $/Year
      Loss
             $/ton of coal
             Mills/kwh
      Price at Break-even,
             $/long ton sulfur
Case 1

10
109
,1090
5645
422
1049
7116
6026
1.31
.54
20
109
2180
5645
422
1049
7116
4936
1.07
.45
30
109
3270
5645
422
1049
7116
3846
.84
.35
65
Case 2
Case 3**
Case 4
Sulfur Price, $/long ton
10
109
1090
5036
422
1049
6507
5417
1.18
.49
20
109
2180
5036
422
1049
6507
4327
.94
.39
30
109
3270
5036
422
1049
6507
3237
.70
.29
60
10
10.5
105
1137
90
291
1518
1413
3.30
1.35
20
10.5
210
1137
90
291
1518
1308
3.06
1.25
30
10.5
315
1137
90
291
1518
1203
2.81
1.15
145
10
9:
93(
39 7(
5(
93J
4958
4028
_
-
20
93
1860
3970
5C
938
495*
309*
•
-
30
93
2790
3970
50
938
4958
2168
_
-
-
      *   See Table III in Volume I for comparison
      **  In Case 3 Operations are 330 Days per Year at 60% Capacity

-------
                          r
                                                    TABLE 27»


                             PROFITABILITY • ZINC OXIDE PROCESS ( JOHNSTONE METHOD)
to
OJ
Sales,  M tons



Sales,  M $


Operating Cost,  M $



Loss,  M $/Year
        Loss
               $/ton of coal


               Mills/kwh
        Price at Break-even,
               $/ton
                                  PLANTS OPERATING AT 70% PLANT FACTOR

                                      SO2 CONVERTED TO SULFURIC ACID

                                     PLUME REHEAT  FROM 122° TO 200°F
Case 1
Case 2
Case 3
Sulfuric Acid Price, $/ton
7
295
2065
7329
5264
1.48
.60
10
295
2950
7329
4379
1.23
.50
13
295
3835
7329
3494
.98
.40
24.85
7
295
2065
6720
4655
1. 31
.53
10
295
2950
6720
3770
1.06
.43
13
295
3835
6720
2885
.81
. 33
22. 78
7
55.4
388
1949
1561
2.82
1. 14
10
55.4
554
1949
1395
2.52
1.02
13
55.4
720
1949
1229
2.22
.90
35. 18

-------
                                            TABLE 28*
                 PROFITABILITY - ZINC-OXIDE PROCESS (JOHNSTONE METHOD)
                          PLANTS OPERATING AT 90% PLANT FACTOR
                              SO2 CONVERTED TO SULFURIC ACID
                              PLUME REHEAT FROM 122° TO 172°F
Sales, M tons
Sales,  M $
Operating Cost, M $

Loss, M $/year

Loss
       $/ton of coal
       Mills /kwh

Price at Break-even,
       $/ton
Case 1
Case 2
Case 3
Case 4
Sulfuric Acid Price, $/ton
7
380
2660
7992
5332
1.50
.61
10
380
3800
7992
4192
1. 18
.48
13
380
4940
7992
3052
.86
.35
21.03
7
380
2660
7383
4723
1.33
• .54
10
380
3800
7383
3583
1.01
.41
13
380
4940
7383
2443
.69
.28
19.43
7
44
308
1879
1571
2.84
1. 16
10
44
440
1879
1439
2.60
1.06
13
44
572
1879
1307
2.36
.96
42.70
7
324
2268
5734
3466
.
-
10
324
3240
5734
2494
^
-
13
324
4212
5734
1522
^
-
17.70
*  See Table 111 of Volume I for comparison
      L .
L ..   L ..

-------
        C.    SULFATE REMOVAL WITH REVERSE OSMOSIS
              .An investigation was made to determine the feasibility of
 using reverse osmosis for separating sodium sulfate from sodium sulfite -
 bisulfite  solutions as an alternate to the lime used in the John at one Zinc
 Oxide process.  The data indicated that the sulfate would not permeate
 through the membrane, and that some of the sulfite and bisulfite would
 pass through the membrane (the low sulfate values in the product stream
 are considered to be due to post-oxidation).
              An analysis of the data indicates that reverse osmosis would
 not be economical.  Table  29 shows that most  of the sulfite and bisulfite
 would remain with th^ sulfate and only 6 - 7% of the SO, in the  solution
 could be  separated from the sulfate.  Dilution of the  solution before  re-
 verse osmosis, however,  would improve recovery.  For example,  if
 Solution 1 is diluted 10:1 and it is  assumed that 90% of the volume can be
 recovered, then                    ,
        33. 35 liter x 8. 85 g/liter » 295 g SO2 in feed
 and    30. 01 liter x 3. 78 g/liter = 114 g SO2 in product
 or      39% of the SO, would be recovered.
 Similar treatment of Solution 2 indicates 44% recovery of the SO,*
              The data from Solution 1,  as applied to Case 1, a 1400 MW
 power plant,  is compared with  the system using lime for removal of the
 sulfate in Table 30.  As in  the Phase III study,  it was assumed that the
 equivalent of 1.0% of the SO, absorbed from the flue gas is oxidized.  The
 cost data  for the lime system was  derived from the Phase III work.  Re-
 verse osmosis costs were based on the  estimated total cost of desalinating
brackish water at $1. 00 per 1000 gal. product.  This cost,  applied to the
 sulfite-bisulfite-sulfate system specified, is probably optimistic.
             It is apparent that removing sulfate with reverse osmosis
would be  substantially more expensive than the lime treatment system.
 The higher sulfur value loss in the R/O system  would also add to the cost
of sulfate  removal.

-------
                             TABLE 29
                 REVERSE OSMOSIS TEST RESULTS
                                               Solution
Feed - volume,  ml
    SO, content, g/ml
    Total SO2,  g
    SO,
ppm
Product - volume, ml
    SO, content, g/ml
    Total SO,, g
    SO, recovered,  %
    SO.
ppm
  3335
 0.0885
   295
  4600

   445
0.0378
  16.82
   5.7
   300
   No.  Z
  3416
0. 1030
   352
  4500

   495
 0.0505
  25.00
   7.1
   300
                                                                          _J

-------
                                 TABLE 30



  SODIUM SULFATE REMOVAL FROM SULFITE/BISULFXTE ABSORBENT
                                       ZnO Process - Case 1 - 1400 MW Plant
SO, Oxidized, Ib/rnin



Na-SO. equivalent, Ib/min
   £•   4


CaSO4* 2 H2O equiv.,  Ib/min



CaO makeup, Ib/min



R/O Feed, 10:1 dil., gpm



Sulfate, Free Recovery, gpm



Waste, gpm



SO. equiv. in waste,  Ib/min



Na-CO, makeup:



   For sulfate loss, Ib/min



   For SO- loss in waste,  Ib/min



SO- Lost from Flue Gas, %





Operating Cost per Ton of Coal



   CaO @ $18. 00/ton



   Na2CO3 @ $1.60/cw



   Utilities

                                  t

   Fixed Costs



   R/O @ $1/1000 gal product




        Total Operating  Cost, $/ton coal
                                       UsingTLim eio r

                                       Sulfate Removal
  121



  268


  325



  106
   10








$0.100






$0.020



$0.045







$0.165
                    using R/O for

                   Sulfate Removal
  121



  268
 6070


 2367



 3703



  180






  200


  300


   24.7
$0.827
 2.370
$3.397

-------
 III.    FLUIDIZED ZINC OXIDE PROCESS
        A.    INITIAL APPROACH
                                                                                   •—*
              The laboratory investigation involving the use of zinc oxide
 in a fluid!zed bed to absorb SO- has been discussed in Part Two. In con-             '
                              £•                                                    ~-i
 junction with this study,  preliminary process and economic evaluations
 were made to determine  the economic feasibility of the system.  These
 evaluations were made on the basis of a new 1400 MW plant and were com-           ~"
 pared with the cost estimates of the Johnstone Zinc Oxide process.                •    ,
                                                                                   -^
              An evaluation was concerned with heat input requirements
 for the thermal regeneration of zinc oxide from the solid absorbent.   At
 the time, absorption rates had not been determined.  Consequently,                  ^
 various levels of absorption and oxidation were assumed  and the heat
 required to regenerate the zinc oxide from zinc sulfite at 527 F  and from            *->
 zinc sulfate at 1832 F was calculated.  The system is illustrated in
                                                                                    i
 Figure 4.  The  study showed that maximum conversion of zinc oxide  to               _j
 sulfite  during absorption with minimum oxidation to sulfate was import-
 ant in order to keep heat requirement costs at a minimum. Order-of-
                                                                                   *—r
 magnitude heat  input requirements are listed  in Table 31. A preliminary
 check of the  heats of reaction in the absorber indicated that the heat evolved          ;
 would maintain or slightly exceed the 122 F absorber temperature.   It
 was noted that heat exchange could reduce the total heat input requirements
 since the  regenerated zinc oxide must be  cooled to absorption temperature           —'
 (approximately 122 F).
             An alternate route was also  considered for the removal and            '-"
 recovery of zinc sulfate from its mixture with zinc oxide, i.e., by leach-
 ing the  soluble zinc sulfate from the mixture of zinc sulfate and insoluble            _
 zinc oxide.  This system was checked for thermal requirements and there
was an indication that a leaching approach to zinc sulfate  removal might be          _
a more economical system than regeneration of zinc oxide from zinc  sulfate
at 1832 F, if there was a market for  zinc sulfate.

-------
                                               TABLE 31

                     HEAT INPUT REQUIREMENTS TO REGENERATE ZINC OXIDE
                                       1400 MW POWER PLANT
Absorption
%~
Oxidation
2
% ~
Regen. @ 1832°F
%3
—
Fluidized-bed ZnO System
15
30
60
30
30
30
30
Johnstons 2
_20
20
20
10
40
20
20
iinc Oxide Pr
50
50
50
50
50
25
100
Dcess
Theor. Heat Input, MM Btu/Min
Regen. @ 527°F

2.5
2.2
2. 1
2. 3
2.0
2.4
2. 1
—
Regen. @ 1832°F

1.5
1.0
0.7
0.8
1.5
0.8
1.5
—
Total

4.0
3.2
2.8
3. 1
3.5
3.2
3.6
2.7
1.      Conversion of zinc oxide to sulfite,  %
2      Sulfite oxidized to sulfate, %

-------
        B.    FINAL APPROACH                                                   -
              1.    Introduction                                                     >
                   It is expected that additional development work would
 determine that oxidation of the zinc sulfite to zinc sulfate would be control-           ;
 led so that recovery of zinc oxide from the sulfate would be minimal.
 Therefore, order-of-magnitude capital and operating costs were estimated
 for the optimized process using fluidized zinc oxide as the absorbent for             '~"
 so2.
                                                                                    i
                   A block flow diagram of the process is shown in Figure           ~*
 11. The process involves:  aqueous prescrubbing of the gas to remove fly
 ash and NO,; the absorption of SO- from the  scrubbed gas by zinc oxide,             -J
 thereby forming  zinc sulfite; the subsequent thermal decomposition of the
 sulfite to re-form SO,, which is recovered and liquefied, and zinc oxide,             _j
 which is returned to the absorber; and an auxiliary  system for the re-
 covery of zinc values from process impurities,  such as zinc sulfate.  The             I
 conditions under which the various process steps should be carried out
 appear  in Part Four.
                                                                                   _J
                   The following cost estimate is based on the  use of the
 process as applied to a new  1400 MW power plant installation.  Profit-                '
 ability has been based on the sale of liquid SO,*  considered at various
 prices.
             2.    Capital Cost Estimate                                           ~J
                   Table  32  summarizes the  capital costs.  The prescrubber         j
 system  selected is  a two-stage mobile bed scrubber with a collection effici-
ency in  excess of 99% for.particles of 2  microns and larger when 5 gallons
of water is circulated per 1000 cubic feet of gas  (Reference 28).  Additional          -1
design and cost data were obtained from a recent TVA study (Reference 29).
The prescrubber system  includes the scrubbers, piping, pumps,  founda-            _j
tions and supports,  and hold tank.
                                                                                   j
                  A recent  report by Tracor, Inc.  was used as a guide  for          —'
the design and cost estimating of fluidized bed systems  (Reference  30).  The

-------
Flue Gas
                            Purified
                              Gas
            CaO
            Fly Ash/Gypsum
                to Pond
                                             ZnO(ZnSO4)
                                         (ZnSO4)
           ZnSO3«2-l/2 H2O
                                                                        Liquefier
               ZnO
                                              (ZnS04)
                                                                                 -a*
                                      ZnSO3« 2-1/2
     Aqueous
     so2
ZnSO4 (to waste
                                    or recovery.
         Prescrubbox-
Fluidized-Bed
  Absorber
Cslciner-
Regenerator
                                                                                    Removal
                        BLOCK DIAGRAM - OPTIMIZED FLUIDIZED ZINC OXIDE PROCESS

-------
                               TABLE 32

                  FLUIDIZED ZINC OXIDE PROCESS -
                      CAPITAL COST ESTIMATE
                   (1400 Megawatt New Power Plant)
Prescrubber System

Fluidized Bed Absorption System

Absorption System Mechanical Separator

Prescrubber/Absorption System Fan (For 13. 5" H2OAP)

Fluidized Bed Regeneration System

Regeneration System Mechanical Separator

Regeneration System Fan

Conveyors

Waste Disposal System

Zinc Sulfite Recovery System


     Sub-total

Sulfur Dioxide Liquefaction System


      Total Fixed Capital

Working  Capital

      Total Investment
Investment,  $

  1,600,000

  1,320,000

  1,110,000

    940,000

    550,000

     70,000

    110,000

    160,000

     90,000

    150,000

  6,100,000

    600,000

  6,700,000

    670,000
  7,370,000


-------
equipment of interest included the absorption system, with cyclonic
mechanical separator,  and the regeneration system, with cyclonic me-
chanical separator and fan.  The purchased equipment cost was based
on Tracer data relating to their fluidized copper oxide process, with the
cost recalculated to conform with the lower gas and solid flow rates and
lower operating temperatures that will be used in the fluidized zinc oxide
process.   The purchased equipment cost was subsequently factored to ob-
tain the fixed capital cost.
                   The absorption system fan cost was based on $50,000
per inch of water pressure drop for 1000 MW plants,  and factored to
1400 MW.  Conveyor costs were based on available data  used in previous
estimates.
                   The waste disposal system cost was factored from
costs shown  in Reference 29 for 10% slurry systems pumped to a pond,
and with no water recirculation from the pond.
                   The recovery system equipment costs include a  gasifier,
filter,  mixer-crystallizer, and centrifuge, at a level capable of handling
the 5% sidestream of the solid absorbent mixture that would be needed if
oxidation occurred to the  extent of 0. 5%.
                   The SO2 liquefaction plant cost was factored from costs
reported in a Bureau of Mines report (Reference 36).  Although the scale-up
from the reference estimate to a 1400 MW plant is considerable, the capital
cost indicated is  believed to be adequate for the liquefaction of all of the re-
covered sulfur dioxide.
                  Plume reheat equipment was not included in the cost
                  ch as the purified gas temperature should approach 1!
(due to the heat of reaction), which may be adequate for most locations.
estimates, inasmuch as the purified gas temperature should approach 155  F

-------
                                                                                   ~J
                   The estimated total fixed capital cost of $6, 700, 000
 is substantially lower than costs appearing in the literature for other
 regenerative SG^ removal processes.   This  is attributed to:
              •    Simplicity of the fluidized zinc oxide
                   process,  which requires  only three major
                   unit operations as a result of the non-
                   occurrence of interfering secondary
                   reactions.
              •    Low operating temperatures.
              •    High SO,  sorption capacity of zinc  oxide
                   which reduces equipment  size.
              3.    Operating Cost Estimate
                   Operating costs were derived from the  schedule shown
 in Table 33, indicating a cost of $0. 64/ton coal and 0.  26 mill/kwh.   Raw
 material and chemical  costs are listed ,in Table  34.  Sixteen operators
 would be required for continuous operation, or four men per shift. A wage
 rate of $4.00 per man-hour was  used.  Utilities costs  are detailed in
 Table 35.  It is assumed that suitable steam would be available at the
 power plant for regeneration  of the  spent absorbent.  The waste disposal
 costs are based on a  10% fly ash slurry being routed to a pond on the  plant
 site  at a cost of $0. 70 per dry ton.  The other operating cost elements are
 self-explanatory.
                   Although Aerojet is aware that NAPCA has standardized
 on a 70% plant factor, a 90% plant factor (330 days operation) was used in
this  estimate in order to provide a level of operation that would be com-
parable with that used in most cost estimates appearing in the literature.
 Under this conditions the process appears to be technically and economically
superior to many other regenerative SO  removal processes.

-------
                                TABLE 33
                  FLUIDIZED ZINC OXIDE PROCESS
                     OPERATING COST ESTIMATE
             (1400 Megawatt New Power Plant, 90% Plant Factor)
                 Fixed Capital Investment:  $6, 700, 000

             ITEM                            TOTAL $
1.     Raw Materials and Chemicals           346,600
2.     Direct Labor                           128,000
3.     Supervision,  15% of 2.                   19,200
4.     Maintenance, 3% of fixed capital cost     201,000
5.     Supplies, 20% of 4.                       40,200
6.     Utilities                              1,034,500
8.     Payroll Burden,  18. 5% of 2 & 3          27,200
9.     Plant Overhead,  50% of 2, 3,  4 & 5      194,200
10.     Waste Disposal                          6,700
11.          TOTAL INDIRECT COST          228,100
12.     Depreciation, 11% fixed capital cost     737,000
13.     Taxes & Insurance,  3% fixed
       capital cost                            201, OOP
14.          TOTAL FDCED COST             938,000
7.            TOTAL DIRECT COST          1,769,500-         60.3
15.          TOTAL OPERATING COST      2,935,600         100.0

16.     Cost: $0. 64/ton coal, 0. 26 mill/kwh

-------
                                                                           J
                            TABLE 34
                FLUID1ZED ZINC OXIDE PROCESS -
               RAW MATERIALS AND CHEMICALS
       (1400 Megawatt New Power Plant, 90% Plant Factor)
                                             Cost per year,  $
ZnOa,  2, 297, 000 Ib @ $0. 15/lb                    344,500

50% NaOHb, 72,000 Ib @ $2. 90/cw                   2,100
                                                                           _J
          Total                                   346,600
 Make-up based on loss of 0. 2% of solids circulated to the regenerator.

 Requirement for recovering zinc values from impurities,
 including sulfate, equivalent to 0. 5% oxidation of the SO-
 absorbed.  Na^SO. formed is discarded.
             L,   *±                                                ••
                                                                           _J



-------
                            TABLE 35

                 FLUIDIZED ZINC OXIDE PROCESS -
                            UTILITIES
        (1400 Megawatt New Power Plant,  90% Plant Factor)
                                             Cost per year, $


Steam,  1, 350, 000, 000 Ib @ $0. 50/1000 Ib           675,000

Power,  48, 660, 000 kwh @ $. 006/kwh               292,000

Process Water, 415, 000, 000 gal @ $. 10/1000 gal     41,500

Cooling Water,  520, 000, 000 gal @ $. 05/1000 gal      26,000
       Total                                    1,034,500

-------
              4.     Profitability
                    The profitability of any SO,  removal system that pro-
 duces a salable product is obviously dependent upon the market conditions
 relating to the product.  The ultimate  demand for elemental sulfur,  SO2,
 and H2SO4 as products is controlled by the demand for sulfuric acid, in-
 asmuch as both elemental sulfur and SO, would eventually be converted
 to the acid to a large extent.  Over 70% of the sulfuric acid produced in
 the  United States  is derived from native  sulfur.  A large new source for
 sulfur, or SO^, such as the sulfur values contained in flue gases, would
 result in a continued drop in sulfur prices.  Sulfuric acid (100%) sold for
 $33. 75 per ton in July 1969,  which is equivalent to $12. 34 per long ton of
 sulfur value (Reference 31).  World-wide sulfur prices in  recent months
 have been fluctuating between $22. 50 and $45. 50 per long ton (References
 32 and 33).  The list price in January  1970 was  $39 to $40 per long ton,
 with prices eroding as  much as $15 per long ton as a result of large  world
 inventories (Reference 34).
                   The profitability of the fluidized zinc oxide process is
 presented in Table 36 with SO, prices  based on  four levels of sulfur  prices,
 as follows:
                                    Sulfur Dioxide Equivalent
                                         $/short ton
                                            4.50
                                            9.00
                                           13.50
                                           17.50
According to the present day market, it would be expected that SO, could
 sell for between $9. 00 and $13. 50 per ton, plus  some additional  revenue due
to value added above the elemental sulfur value.  A break-even situation
would be expected for the fluidized zinc oxide process with SO2 sales at
$10. 87 per ton.  Even at $4. 50  per  ton of SO2,  the cost per ton of coal is
not excessive.
                   This economic study of the fluidized zinc oxide process,
though quite preliminary,  indicates that it should be superior to many other
 regenerable processes for the removal of SO, from flue gases.


-------
                               TABLE 35
                     FLUIDIZED ZINC OXIDE PROCESS:
                             PROFITABILITY
                      (1400 Megawatt New Power Plant)
                              90% Load Factor
        Total Investment:
        Liquid SO2 Sales:
Sales, M $
Operating Cost, M $

Profit (loss),  M $
Profit After Tax, M $
Return on Investment After
  Tax, %
Payout, Years
Net Profit (Cost)*
       $/ton of coal
       mills/kwh
$7, 370,000
270,000 tons/year (95% recovery)

         Liquid SO2 Price, $/ton
                                   4.50
              9.00
      1215
      2936
2430
2936
     (1721)  ( 506)
     (0.37)  (0.12)
     (0.16)  (0.05)
         13.50
3645
2936
          709
          369

          5.0
          6.7

          0.08
          0.03
          17.50
4725
2936
         1789
          930

         12.6
          4.4

         0.20
         0.08
*  Operations would break-even if SO.,
       was sold at $10. 87/ton.

-------
 IV.    MAGNESIUM BASE SLURRY SO2 SCRUBBING SYSTEM

        A.    INTRODUCTION                                                    -1
              The Research and Development Division,  Babcock and                  <
 Wilcox Company, performed development work for NAPCA under Con-              ~"'
 tract CPA 22-69-162 entitled: "Magnesium Base Slurry Scrubbing of
 Pulverized Coal Flue Gas. "                                                      —
              The work was involved with:  (1) the design and construc-
 tion of a pilot plant scrubbing system having a nominal capacity of 1000              -J
 scfm flue gas, and (2) with experimental work concerning the removal
 of fly ash and sulfur dioxide  from flue gas generated by burning pulveri-             _,
 zed coal.  Regeneration of the absorbent was not included in the scope
                                                                                  l
 of the project.                                                                    !
              One of Aerojet1 s tasks in Phase IV of Contract No. PH
                                                                                  i
 86-68-77 was to provide economic evaluation of aqueous processes under           _
 development for NAPCA such as the above-mentioned Babcock & Wilcox
 process.  Unfortunately,  the schedule of the B&W contract precluded
                                                                                •—'
 analysis of this system during the early months of the Aerojet contract.
 However, initial operating data were presented  in B&W's progress report            ;
 for the month of January  1970, making it possible to  start a cost analysis.          ~"
 Several runs appeared in this report in which a  venturi scrubber was used
 as the prescrubber to trap fly ash, and a floating bed absorber was used            —•
 to remove the SO9 from the flue gas.  Run D-047 of the series of runs
                 L*
was selected as a guide for the Aerojet cost analysis; this was an arbitrary         _
choice from several runs that had removed > 95% of the SO, present in
the flue gas.                                                                      ,
        0                                                                       —i
       B.    PROCESS ENGINEERING
             Most of the data of Run D-047 were converted directly to a           '
 1400 MW power plant size by a factor derived from the  ratio of gas flow
in the pilot plant to the gas flow in a 1400 MW plant (Cases 1 and 2 of the
                                                                                —'
Aerojet Phase III study).  The venturi scrubber  product fly ash concentra-
tion was increased from a very dilute to a 10% slurry in order to economize   ,.,.      f
on water consumption.  Temperatures and gas pressures were not changed.         —


-------
              Figure 12 illustrates a simplified flow diagram of the
 magnesium base slurry SO- scrubbing system (similar to the B&W pilot
 plant design) with pertinent operating data for a 1400 MW system.  Ma-
 terial quantities shown are approximate and do not balance since they
 were calculated from Run D-047 stream data and are not necessarily
 theoretical quantities.   No attempt was made to determine the exact
 composition of the circulating slurry to the floating bed absorber, since
 this was not needed for  the cost estimates.  It was assumed that the
 floating bed absorber product liquor would flow to an absorbent regen-
 eration and SO2 recovery system.   No data were given on this system;
 consequently,  the system is not included in this  preliminary study.
             Design and cost information  is limited concerning  venturi
 scrubbers with adequate capacity to handle the large quantities of flue
 gas generated in large power plants.  Data in a recent report (Reference
 35) indicated that four 12-ft dia. x 20-ft high venturi prescrubbers would
 be needed for a 300 MW system handling 671,600 acfm gas.  On this
 basis, seventeen 12-ft dia. units would be needed for a 1400 MW plant.
 The  report (Reference 35) did not provide  cost information on these
 scrubbers; therefore, floating bed scrubbers were substituted to serve
 as prescrubbers in the  cost estimate discussed below.
             Run D-047 indicated a liquid/gas ratio of 18. 5 gal/mcf
flue gas for the floating  bed absorber.  At this rate approximately 55, 000
gpm of circulating liquor would be needed  for the SO- floating bed absor-
bers in a  1400 MW plant.  Scale-up of the  pilot plant floating bed absorber
indicated a total area of about 3900 sq ft of absorber cross-sectional  area
would be needed for SO, removal from a 1400  MW plant.  Presumably,
this could be provided by installing six 29-ft dia.  absorbers in parallel.
It was  as sunned that an equivalent area would also be needed for the float-
ing bed prescrubbers used for fly ash removal.  These areas, converted
to equivalent units, are  similar to those used in  a lime-stone-wet scrubbing
study by TVA (Reference 29).

-------
                              MgO 8<
                              Recovery
                       Pond
                                                                         _g	-
                                                  Floating Bed
                                                  Absorber
                                                                    Mg(OH)z
                                                                   Hold  Tank
                                                              MgO
                                                            Slaking Tank
                                                           STREAM NO.
DESCRIPTION
Components, Ib/min
    Total Gas
    Dry Gas
    H2O
    SOz
    Fly Ash
    MgO
    MgSO3, Solid
    MgS04
H2O, gpm
Temperature, °F
Gas Pressure, in. W.C.
                                                                                                     10     11
197000  209805   Z08605
188335  188335   188272
  8665   21470    20333
  1270     1270       63
   396       6        6
                                                                                                                     1Z
11750  456200    22000  331100    3900    16700   11120    11750
                                                            1090     390
                           928     928
                                           *       2735
                                           *         94
                                  1408    54700     2637   39700     467    2002    1333
   575     145      134    72      94      138       138     140      140      60      60
   5.4    -12.2    -25.7
*Also in circulating slurry.
                       MAGNESIUM BASE SLURRY SOj SCRUBBING SYSTEM FOR 1400 MW POWER PLANT
                                                    Figure 12
                                                          1408
                                                            60
                             L.
                                  L  .      I  ,.     I

-------
      C.    CAPITAL COST ESTIMATE
            Cost data on venturi scrubbers were not available; thus
floating bed scrubbers were substituted for cost estimating purposes.
Since the process was in an early stage of development, only an order-
of-magnitude estimate could be  made at this  time.  Even this provided
only a partial cost since the system did not include regeneration of
absorbent and recovery of SO.,.
            Under these circumstances, cost information appearing in
the TVA Report (Reference 29) was used as a guide to estimate capital
costs.  Table 37 provides a. summary of the  capital costs, with the total
fixed capital estimated at $9, 750, 000.  It is  possible that the cost would
be somewhat less if large venturi  prescrubbers could be substituted.
      D.    OPERATING COST AND ECONOMIC ANALYSIS
            Operating costs and an economic evaluation cannot be made
since the process as described is  incomplete.   The magnesium oxide
regeneration and SO2 recovery systems will add to both the capital and
operating costs.  Credits applied due to reduction in magnesium oxide
make-up requirements and to sale of sulfur values, however, will help
to balance the added costs.

-------
                                 TABLE 37
             MAGNESIUM BASE SLURRY PROCESS:  CAPITAL
                        COST ESTIMATE SUMMARY
                   (Regeneration System Not Included)
                                 CASE  1
        Item .
MgO Storage and Handling Facilities
Slurry Storage and Pumping
Floating Bed Prescrubbing System
Floating Bed Absorption System
Solids Disposal System and Pond
Fan, 30 in. W. C. ap
Control Room and Equipment
Electrical and Water Distribution
Painting and Insulation
Construction Facilities
       Total Direct Cost
Engineering Design
Contractor's Fees and Overhead
Contingency
       Total Fixed Capital Cost
  Cost - $
  480,000
  150,000
2, 000, 000
2,000,000
   90,000
1, 800, 000
  290, 000
  360, 000
  140,000
  300.OOP
7,810,000
  470,000
  950, 000
  520,000
9, 750, 000
                                                                               -J

-------
                             REFERENCES

1.     Volume I of this Final Report.
2.     D.  Bienstock and F. J. Field,  "Bench-Scale  Investigation on
       Removing Sulfur Dioxide from  Flue Gases, " J.  Air Poll. Cont.
       Assn..  1JJ (2).  121-5(1960).
3.     V.  V. Pechkovskii and A.  N. Ketov,  "Study of the Thermal
       Decomposition of Zinc Sulfite, " Zhurna.1 Prikladnoi Khimii,  33
       (8), pp 1724-9  (I960).
4.     F. A. Cotton and G. Wilkinson, "Advanced Inorganic Chemistry, "
       Interscience Publisher, New York, N. Y., 1962, p. 430.
5.     T.  R. Hogness and W. C.  Johnson,  "Qualitative Analysis and
       Chemical Equilibrium, " Henry Holt and Company, New York,
       N. Y.,  1940,  pp.  459-60.
6.     F. A. Cotton and G. Wilkinson, loc.  cit., p.  711.
7.     H.  F. Johnstone and A. D.  Singh, University of Illinois Bulletin,
       38.' No.  19, 31 December 1940.
8.     Ketov, A. N. and Pechkovskii, V. V.,  Zhur, Neorg. Khim.,  4,
       272-6 (1959); CA: 53,  12806h (1959).
9.     J. D.  Terrana  and L.  A.  Miller,  "New Process for Recovery
       of SO2 from Stack Gases, " Wellman-Lord,  Inc.  1967).
10.    K. A. Kobe and T.  M. Sheeky, Ind.  Eng. Chem.,  40_, 99-102  (1948).
11.    L. C. Schroeter, "Sulfur Dioxide, "  Pergamon Press, New York,
       N. Y.,  1966, Ch.  2.
12.    H.  F. Johnstone and A. D.  Singh, loc.  cit., p.  62.
13.    ibid, p. 64.
14.    T.  Okabe,  K.  Kamisawa, and S.  Hori,  Nippon Kagaku Zasshi,  81,
       529 (I960).

-------
 15.     LI.  Cola and S. Tarantino, Gazz. Chim. Ital.,  92,  174 (1962).
 16.     G. Pannetier,  G. Djega-Mariadassou, and J. M. Bregeault,
        "Etude de  Decompositions s'effectuant avec Depart Simultanes
        de Plusieurs Gas.  I - Decomposition Thermique de  Sulfite de
        Zinc Hydrate,  ZnSO?. 5/2H?O, " Bull. Soc. Chim.  France, 8,
        1749-56 (1964).
 17.     T. R. Ingraham and H.  H.  Kellogg,  "Thermodynamic Properties
        of Zinc Salfate, Zinc Basic Sulfate, and the System Zn-S-O, "
        Transactions of the -Metallurgical Society of AIME 227,  1419-
        25 (1963).
 18.     Tracor Report No. TM 004-009-ch 16, 31 December 1968, p. 89.
 19.     H. F. Johnstone and A.  D. Singh, loc. cit., p.  67.
 20.     Z. P. Rozenknop,  "Extraction of Sulfur Dioxide from Gases, "
        Goskimizdat,  Moscow-Leningrad, 1952.
 21.     H. F. Johnstone and A.  D. Singh, loc. cit., p.  63.
 22.     ibid,  p. 88.
 23.     ibid,  p. 68.
 24.    E. V. Margulis and Yu.  S.  Remizov,  "The Chemistry of the
        Thermal Dissociation of Copper, Zinc, Cadmium and  Lead
       Salfates, "  Sbornik Nauch.  Trudov. Vsesoyuz. Nauch -
       Issledovatel Gornomet. Inst. Tsvetn.  Met.  I960 No. 6,  171-82;
       C.A., 56,  3109a (1962).
 25.    K. Kohler  and P. Zaeske,  Z. Anorg.  Allgem. Chem., 331, 1-6
       (1964); C.  A. 6^, 12938h (1964).
26.    N. A.  Lange,  Handbook  of Chemistry, Handbook Publishers,
       Sandusky,  Ohio, 6th Ed.  , 1946pp. 276-7.
27.    A. W. Michalson,  "Ion Exchange, " Chem.  Eng. , TO,  163-82,
       March 18,  1963.
28.    Control Techniques for Particulate Air Pollutants,  Dept. Health,
       Education,  and Welfare,  NAPCA Publ. AP-51,  January 1969.

-------
29.    Sulfur Oxide Removal from Power Plant Stack Gas - Use of Lime-
       stone in Wet-Scrubbing Process,  Prepared for Department of
       Health,  Education, and Welfare by Tennessee Valley Authority,
       1969.
3®'    Tracer, Inc. ,  Applicability of Metal Oxides to the Development
       of, New Processes for Removing 50% from Flue Gases, Final
       Report,  Contract No.  PH 86-68-68, for NAPCA,  Department of
       Health,  Education, and Welfare, 3 July 1969.
31.    Chemical Week, Market Newsletter, 105_ (2),  29,  July 12,  1969.
32.    Chemical Week, Market Newsletter, 1£5(13), 43, Sept. 27, 1969.
33.    Chemical Week, Market Newsletter, 1£5>(19), 37, Nov. 12S  1969.
34.    Chemical Week, Market Newsletter, 106^(2),  53,  Jan. 14, 1970.
35.    Stone and Webster Engineering Corp.,  Sulfur Dioxide  Scrubbers,
       Stone and Webster/Ionics Process,  Final Report, Contract No.
       CPA 22-69-80, for NAPCA, Department of Health, Education,
       and Welfare,  January 1970.
36.    Field, J. H.,  Brunn,  L.  W.,  Haynes,  W. P., and Benson,  H. E. ,
       Cost Estimates of  Liquid Scrubbing Processes for Removing Sulfur
       Dioxide From  Flue Gases,  Bureau of Mines,  RI 5469, 1959.

-------
                              APPENDIX A
                      OXIDATION AND NO  STUDIES
                                          j£
I.    EESCRIPTION OF EQUIPMENT
      The arrangement of equipment used in these studies is shown in the flow
diagram, Figure A-l.
      The equipment was designed as a dynamic bench scale counter cur rent
prescrubber and scrubber, to simulate the SO, removal efficiencies of a
                                             c*
full sized commercial plant treating flue gas from a coal burning boiler.
In doing this it was hoped  to obtain oxidation values of SO, to SO, in the
scrubbers that are similar to oxidation that occurs in a full size plant.
The equipment was made of glass and plastic to eliminate the possible
catalytic effect of iron and other metals on the system.  The equipment was
versatile in that the SO_,  CO?' NO>  NO2' °2'  N2'  and fiy ash concentrations
in the synthetic flue gas could be changed within the desirable limits simply
by turning valves  or switches.
      The flue gas mixtures were made up of gases from the various lines
and cylinders. Each gas stream was reduced to a constant pressure of
260 mm Hg (gage) by a pressure reducing regulator,  metered accurately
with a calibrated rotameter, and controlled with a needle valve.  The gas
streams consisted of:
                    •     4. 8% SO2 in N2
                    •     5. 0% NO in N2
                    •     1.2% N02 in N2
                    •     Pure CO
                                  L*
                    «     Air
                    •     Pure N_

      It was planned to add fly ash from a vibrating hopper at a controlled
rate  through a sharp orifice and into a venturi mounted in the dry,  mixed
flue gas  line.  This  system required revisions that are discussed later.

-------
  VIBRATING  FLY
                                                                              LEGEND
                                                                      Fl  -  FLOW INDICATOR
                                                                      Tl  -  TEMP.  INDICATOR
                                                                      FC  -  FLOW CONTROLLER
                                                                      FCV-  FLOW CONT.  VALVE
                                                                      S  -  SAMPLE VALVE
                                                                      M  -  MANOMETER
                                                                      P  -  PRESSURE INDICATOR
                                                                      V  -  VARIAC
                                                                               *
                                                                                          TRAP
DRY GAS
FLOW 5 STD
LITERS/MIN
                                                                                                      TO
                                                                                     WET
                                                                        ABSORPTION  TEST
                                                                          BOTTLES    METER
                                                                        VACUUM
                                                                         PUMP
 i. ..    L  ._
FLOW DIAGRAM FOR SO2  REMOVAL, STUDIES
      USING ONCE THROUGH SCRUBBER
                     Figure A-l


-------
      The prescrubber consisted of a 22 mm  ID (0.87 in. ) glass column
 packed with 32 in.  of 0. 5-in.  Intalox saddles.  This packing was used so
 that slurries containing fly ash could be circulated without plugging.   During
 some runs late in the program the 0. 5-in. packing was replaced  with 0.25-in.
 saddles in order to obtain more efficient scrubbing.  A calibrated glass
 rotameter was used to measure the rate of liquid flow.   An adjustable
 transformer - controlled electric heater was  installed in the liquid line in
 order to maintain the prescrubber overhead temperature at about 50  C.  The
 column was insulated to minimize heat  losses and a stirrer was provided in
 the liquid receiver to prevent the solid  fly ash from settling out prematurely.
 Thermometers were provided for inlet  and outlet liquid  temperature measure-
 ment.   A manometer was used to measure the pressure drop across the
 column.  Sample connections were provided for liquid and gas  analyses.  A
 small rotameter was added to provide make-up  water for that  carried out
 by the saturated warm  outlet gas.
      An alternate  prescrubber was  also used in some runs to  check fly ash
 and removal efficiency. This system simulated a venturi scrubber by using
 a laboratory aspirator  through which the 50°C prescrubber solution was
 circulated. The synthetic flue gas stream containing the fly ash flowed into
 a small glass gas-liquid separator.  The gas  stream then flowed into the
 scrubber.
      In a few runs  the dry gas mixture by-passed the prescrubber,  and
 the water needed to saturate the gas  at  122 F was metered into the
 preheated line.
      A liquid  recirculating scrubber was used  initially that consisted of
 a packed column of  the same diameter and height as the prescrubber and
was packed with 0.  5 in. porcelain Intalox saddles (see Figure A-2).  The
 solution pump,  rotameter,  liquid heater,  manometer, column  insulation,
 thermometer,  receiver stirrer and sample connections were similar to the
prescrubber equipment. An auxiliary electrical heater was installed on the
 outside  of the column and covered with insulation to compensate for heat
losses and to provide a constant temperature throughout the scrubber. An
electrical heater on the gas line from the prescrubber to the scrubber
prevented the gas from cooling and condensing water in this line.

-------
                                                                                                          t LO'.V INUICATOH
                                                                                                          TEMP. INDICATOR
                                                                                                          FLOW CONTROL VAl V
                                                                                                          •i.AMPLF VALVE

                                                                                                          PRESSURK INOlC<\TOa


                                                                                                          VARIAC
>
I
                                                   FLOW DIAGRAM FOR SO2 REMOVAL STUDIES

                                                      USING RECIRCULATING SCRUBBER
                                                                       Figure A-2
                          L.

-------
      Another scrubber in which fresh absorbent was fed to the top and
spent absorbent removed at the bottom was used in most of the tests
(Figure A-l).  This unit was designated as the "once through scrubber" in
order to distinguish it from the recirculating scrubber.  This scrubber \vas
installed so that data could be  obtained on a countercurrent operation similar
to the full-sized commercial column proposed in the Johnstone Zinc Oxide
Process.   This type of operation had substantially different operating
characteristics than the  recirculating scrubber.
      The original once through scrubber, which was used for the preliminary
base line runs, was made from a 5 mm ID glass tube, 10-ft. long. The once
through scrubber used in most of the oxidation experiments was made of 7 mm
ID glass tubing,  10 ft. long.  A third unit was also 7 mm ID, but only  5 ft.
long.  The 7 mm ID glass tube was contained in a 1 5 mm ID glass tube
which provided a. jacket through which water was circulated in  order to
maintain the desired 50° operating temperature. The 7 mm ID glass tubing
was flared out at each end in order to permit the liquid to enter the top and
leave the bottom without causing flooding.  It was found necessary to add a
helix of 0. 033-in.  thick Teflon inside of the 7 mm  tubing to provide good
liquid-gas  contact.  The lean solution was fed to the top of the  column from
a large reservoir by nitrogen pressure.  A calibrated rotameter was used
to measure the solution rate which was controlled  by a Foxboro flow con-
troller.  The rich solution was collected at the bottom of the column in a
small receiver from which it was withdrawn,  measured and analyzed.
Thermometers indicated  the liquid and gas temperatures,  and a manometer
measured the gas pressure drop across the column.
      The scrubbed  gas from either of the two scrubbers passed through a
water bubbler and was vented to a  hood.  Two sample connections were
provided in the line to the bubbler.  The first connection went to  a gas
absorption bottle,  a wet test meter and a vacuum pump.  The vacuum  pump
was necessary to draw the gas sample through the  bubbler in that the; pres-
sure in the absorption system was not appreciably above atmospheric
pressure.  The second sample connection was used to take gas samples in
glass bulbs for NO  analysis.

-------
      Figure A-3 is a sketch of the final design of the fly ash addition unit,
which had the culpability to deliver a continuous fly ash stream at a rate of
about 15 mg/min to the 5 liter/min to the 5 liter/min flue gas streslm.  It
was found necessary to dilute the fly ash with a relatively large volume
(99%f) of inert free-flowing glass beads (-60 +100 mesh from the 3M Co.)
in order to maintain a steady flow of fly ash.  An  aspirator in the dry gas
line helped to control the pressure drop across the vibrating hopper-orifice        —
feed system.  The hopper discharge line was connected to the side inlet
line of the aspirator.  The aspirator also served to mix the fly ash-glass           _
bead mixture with the high velocity gas stream. A needle valve  in the
gas line to the top of the hopper was adjusted to control the pressure drop
across the fly ash orifice.  The glass beads in the circulating prescruhber
liquid were removed from the system in a settling bulb placed upstream of
the circulating pump.
      Figure A-4 is a photograph of the apparatus and work area.  The once
through scrubbing column is not visible,  but was located at the extreme            ~"
left in the picture.

II.    OPERATING PROCEDURE
      A.    FEED GAS SYSTEM
            The following operating procedure was used in the experimental
work performed in the bench scale equipment:
      The total gas flow was held constant at a rate of 5. 0 liter/min of wet
(7. 25% H2O) gas flowing at 70°F and 760 mm Hg.  This is equivalent to
4. 6 liter/min of dry gas at the same temperature and pressure.

-------
             Glass Hopper
     Glass Beads + Fly Ash
            Vibrator
         Cam
                                                         Fine
                                                         Needle
                                                         Valve
 20 RPM Clock Motor
     Orifice or Venturi Tube
Gas + Ash + Beads
 to Prescrubber
Aspirator
  Tube
Dry Feed Gas Mixture
                       FLY ASH ADDITION SYSTEM
                               Figure A-3

-------
.'v  ...
  APPARATUS FOR OXIDATION AND NO  STUDIES
                                x


                Figure A-4
                                                                      J
                                                                      _J
                                                                     J
                                                                     -J

-------

           Gas compositions used in the base-line runs were:
                                               Volume %

N, -SO,
2 2


N2 -SO, -CO,



N2 -S02 -C02 -02





H,O
2
S°2
Balance N~
H2°
C02
so2
Balance N_
H2°
°2
C°2
so2
Balance N7
Dry
M

0. 32
99.68
-
15.85
0. 32
83.83
-
3.02
15.85
0. 32
80.81
Wet
7.25

0.30
92.45
7.25
14.70
0. 30
77.75
7.25
2.80
14. 70
0. 30
74. 95
           After completion of the base-line experiments,  NO and NO- were
added to give a total of 0. 05% NO at various NO/NO, ratios.  The N,  content
                                X                 £t               £
was adjusted in all runs so that the total gas rate was held at 5 liter/min wet.
In some  runs equimolar quantities of NO and NO- were added equivalent to
0. 04% NO + 0. 04% NO2.
      B.    PRESCRUBBER OPERATION
           In all runs with a prescrubber,  the prescrubber overhead tem-
perature was maintained at 50  C_+ 2.  The gas stream was heated to about
50  C with a line heater in runs made without a prescrubber.  The pre-
scrubber contained water in the base-line runs in order to saturate  the
gases  with water at 50 C.  In the normal runs, the prescrubber was
charged with 7 - 12% sulfuric acid in order to simulate the equilibrium acid
concentration that would build up in the prescrubber due to  the solution of
SO- from the flue gases. In several special runs the prescrubber was
charged with a solution containing 5. 5 wt-% ferrous sulfate plus  7 wt-%
H-SO. in order to remove NO2 from the feed gases.

-------
      C.    SCRUBBER OPERATION
                                                                                .  f
            The majority of the runs were made with a standard sodium           ^
sulfite-bisulfite solution similar to the one recommended for the Johns tone
process.  This solution was made by dissolving reagent grade  sodium             J
sulfite and sodium bisulfite in freshly boiled and cooled distilled water and
                                                                                  I
storing the solutions in nitrogen blanketed bottles.  The standard solution           ''
was made to contain 3. 5 moles  6f Nation in 100 moles of water.  The
quantities of Na2SO3 and NaHSO- used were selected so that the SO,/Na             j
ratio was 0.65.  Since the salts contained some SO~. and since  some oxidation
                                                 4
occurred while the  solutions were mixed,  the SOT content of the lean solutions       i
were always determined just before taking the rich solution samples.  In         ~
                j)C                                              ^
this way the S/C ratios could be corrected for the amount of SOT- present,          ,
and the  actual quantity of SOT formed in the scrubber could be  accurately         >J
calculated.
                                                                                  i
            In the runs using the recirculating scrubber, approximately           Ji
375 ml  of fresh, lean solution was  charged to the scrubber.  The scrubber
was purged with N_  for 5 minutes to remove air, then heated to 50 C with           !
                 &                                                              —i
liquid circulating, after which the desired flue  gas flow was  started at the
normal rate.  The overhead gas was sampled and analyzed every 30 minutes.        {
The solution was sampled during and at the end of the run for sulfite and
sulfate  content.  Similar  runs were made with selected oxidation inhibitors
added to the solution.                                                            ~'
            In the runs using the once through scrubber, the same com-             i
positions of gas and lean  solution were used.   In this case, however, the          '-'
lean solution was fed continuously  to the column and accumulated at the
bottom  of the column.   The column was started up with the desired flue gas        _
flowing at the normal rate and with the heating jacket adjusted  to give  50 C
inlet temperature.   The lean  solution flow was then started at  a predetermined
rate.  Since the liquid hold up in the column was only several minutes and
the gas hold up was only a few seconds, the column came to  equilibrium  in         •
                                                                                _j
15 - 20  minutes. The  column was  then sampled at about 30 minute intervals,
  See Section IV Appendix A

-------
           usually at varying operating conditions such as slightly different liquid or
           gas rates, with inhibitor added,  etc.
           IU.   ANALYTICAL METHODS
                 A.    GAS ANALYSES
                       The scrubber outlet gas was analyzed for SO- by bubbling the gas
           through standard iodine  solution containing excess potassium iodide to pre-
           vent loss  of iodine to the relatively large volume of gas flowing through
           the bubbler  and wet test meter.  The excess iodine was titrated with
           standard sodium thiosulfate solution with starch indicator.  The 2. 8% O,
           and 0. 04% NO were found not to interfere with this SO,  analysis.   The runs
           with 0. 003% NO, were also found not to make detectable errors in the SO,
                          ™*                                                      £*
           measurement; however,  the runs with 0. 04% NO  resulted in low SO,
                                                         2                   "
           analyses in  the outlet gas,  due to the oxidation of the potassium iodide
           by NO2.
                       A relatively simple and accurate colorimetric analytical method
           for NO^ in flue gases was developed that used a Hach  DR-A 1250 colori-
           meter and NitraVer IV   powder pillows.  A calibration curve was prepared
           from the results of tests made on a series  of solutions containing known
           concentrations of potassium nitrate in distilled water.   This curve was checked
           by adding  known quantities of SO~ ion to the standard nitrate samples in
           concentrations equivalent to the SO, present in the flue  gas samples.  These
           sulfate values did not appreciably affect the colorimeter readings for nitrate.
                       The method  is a modification  of the phenoldisulfonic acid  method
           for determination of total nitrogen oxides in the presence of SO, and ammonia.
                                                                        &
                      The oxides of nitrogen were oxidized to NO7, using an oxidizing
           solution containing 5 ml  of 3% H,O  per 100 ml of 0, 1 N H,POA.  Phosphoric
                                         £ £                      j   4
           acid was substituted for  sulfuric  acid specified in the original procedures
           to eliminate interference with analyses for  SO,.  The SO, present in the
           flue gas was oxidized to  SOT at the same time the NO was  oxidized to NO".
            Manufactured by the Hach Chemical Co. , Ames, Iowa.
          C«C
            Atmospheric Emission from Nitric Acid Manufacturing Processes, Department
            Health,  Education and Welfare,  PHS Publ.  989-AP-27,  1966.

-------
            The gas sample was drawn into a 1. 0 liter bottle containing 25
                                                                                  . i
 ml of the oxidizing solution.  After measuring the gas pressure and tern-           _J
 perature the bottle was shaken for 1/2 hour and allowed to stand overnight.
 The liquid was then carefully transferred into a beaker, and the bottle washed       ;
 several times with distilled water which was  also added to the beaker.  The
 sample in the  beaker was neutralized by addition of 1 N KOH and evaporated         i
 to dryness in a 120°C oil bath.  The residue in the beaker was  then dissolved       ~~
 in distilled water and washed into a volumetric flask.   24 ml of the sample           ,
 was transferred to a colorimeter bottle, a NitraVer IV powder pillow added,        -J
 and the bottle  shaken for one minute.  The nitrate content was determined           ,
                                                                                   i
 with the colorimeter after 3  minutes.                                              _j
                                                        *.
       B.    LIQUID ANALYSES
            Liquid samples taken for analysis were  carefully blanketed with
 nitrogen to prevent the normally rapid oxidation of sulfite  to sulfate.  The           \
 sulfate determination was made as soon as the samples were taken.  After
 investigation of existing procedures, a  standard turbidimetric  method, used         i
 for rapid routine tests for sulfate ion in industrial water, was  modified and         —'
 used for these analyses.   The sulfate ion is converted to a barium sulfate
                                                     «                             t
 suspension in the standard method and the resulting turbidity is determined         J
 by a photoelectric  colorimeter  or spectrophotometer.   The turbidity is com-
 pared to a curve prepared from standard sulfate solutions.                           i
                                                                                  —•/
           A major source of interference to this sulfate ion determination
 in the oxidation studies was the presence of sulfite ions which reacted with           \
 the barium reagent and formed BaSO^ precipitate.   It was  necessary, there-
fore, to eliminate  this  interference,  which was accomplished by adding the           (
 sample to concentrated HC1,  evaporating the  mixture to dryness, and dis-
 solving the residue in water.  The HC1  liberated the residual sulfite from            j
 the sample, thus removing this interference to the sulfate  ion determination.        —'
           A Hach AC-DR Colorimeter, used for these tests,  is equipped           !
with direct reading scales for sulfate (and many other tests).  Pre-weighed         U
quantities of  reagents were used.                                                    ,

-------
                        The sulfite  concentrations in both the lean and rich solution
  *•"          samples were determined by adding an excess of 0. 1  N iodine solution and
             back titrating with  0. 1 N sodium thiosulfate  solution to the starch end point.
  v_          Alkaline solutions were acidified with hydrochloric acid.
             IV.   CALCULATIONS
                        The rate of oxidation of SO, to SOT in this work was based on
             the amount of SO- being fed to the scrubber in the gas stream and  not on the
  L-          amount of SOI present in the absorbing solution.  In all sulfite-bisulfite
             scrubbing systems  there was  much more SO- equivalent present in the liquid
  L          than in the gas  phase; however, from the practical standpoint the percentage
             of the SO- in feed gas being oxidized in a commercial scrubber is  of interest.
  L          The percentage SO- oxidized was calculated as follows:

                              moles/min SO?  oxidized to  SOT
  *-                           	1	3_  x 100
                              moles/min SO-  inlet to absorber
  v—.
                  Concentrations and compositions of the sulfite-bisulfite solutions were
  ;           represented by C and S, which indicate, respectively, moles of sodium (or
             other cation) and moles of SO- dissolved in 100 moles of water.  The actual
  |           sulfite concentration is given  by C minus S (C-S) and the bisulfite concen-
             tration is  given by ZS minus C (2S-C).  An S/C ratio of 0. 5 corresponds
  I           to pure sulfite,  whereas a ratio of 1. 0 corresponds to pure bisulfite.  In
  L-          solutions containing sulfate, the concentration of sodium present as
 L
sulfite and bisulfite is'represented by C  , which represents the active base
                                      a,
concentration.   The SOT concentration is subtracted from the total base
concentration to give C .
                      el
 L
                  The sulfur material balance was calculated as:
 L_
            moles _—.  .   .  ,    , •   . moles __. = .    . ,    ,.   . moles _—.   .    ,,  .
r           —:	 SO- in rich sol'n  + —:	 SO .  in rich sol'n + —:	 SO- in outlet gas
 I           mm	2	mm	4	mm	2	°
 i           moles ,,.-  .  ,       ,,   . moles _,,-. = .   ,      .,   , moles ,,._   .  .  .  .
 L-         —:	 SO_ in lean sol'n  + —:	 SO.  m lean sol'n +—:	 SO_ in inlet gas
            mm      2               mm     4                mm      2        °


-------

                                                               J
                                                              -J
This page intentionally left blank.
                                                                I
                                                              J

-------
                                      APPENDIX B
                       DESCRIPTION OF CASES USED IN PHASE III
Li
                  Case  Description

                   I    Large new power
                        plant facility
                        1400 megawatt

                   2    Large existing
                        power plant
                        facility
                        1400 megawatt

                   3    Small existing
                        power plant
                        facility
                        220  megawatt

                   4    New smelter
                        facility (5%SO2
                        to scrubber)
Flue Gas
mmscfm

  2. 5
  2. 5
  0. 5
0. 02
                       Coal
         Exit SO-  Requirement
           ppm      tons/hr
            150
            150
            300
             5,000
580
580
 90
                  ^
                  Case 4 was changed during the program to a smelter generating
                  220,000 scfm (1 atm and 60°F).
L.

-------