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.
------- |