United States Industrial Environmental Research EPA-600/7-78-216
Environmental Protection Laboratory November 1978
Agency Research Triangle Park NC 27711
Advanced Concepts:
SO2 Removal Process
Improvements
Interagency
Energy/Environment
R&D Program Report
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy-Environment Research and Development
8. "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the INTERAGENCY ENERGY-ENVIRONMENT
RESEARCH AND DEVELOPMENT series. Reports in this series result from the
effort funded under the 17-agency Federal Energy/Environment Research and
Development Program. These studies relate to EPA's mission to protect the public
health and welfare from adverse effects of pollutants associated with energy sys-
tems. The goal of the Program is to assure the rapid development of domestic
energy supplies in an environmentally-compatible manner by providing the nec-
essary environmental data and control technology. Investigations include analy-
ses of the transport of energy-related pollutants and their health and ecological
effects; assessments of, and development of, control technologies for energy
systems; and integrated assessments of a wide'range of energy-related environ-
mental issues.
EPA REVIEW NOTICE
This report has been reviewed by the participating Federal Agencies, and approved
for publication. Approval does not signify that the contents necessarily reflect
the views and policies of the Government, nor does mention of trade names or
commercial products constitute endorsemertt or recommendation for use.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/7-78-216
November 1978
TVA Y-139
Advanced Concepts:
SO2 Removal Process
Improvements
by
John M. Potts and John E. Jordan
Tennessee Valley Authority
Division of Chemical Development
Muscle Shoals, Alabama 35660
EPA Interagency Agreement No IAG-D6-E721-BM
Program Element No. EHE624
EPA Project Officer: Robert H. Borgwardt
Industrial Environmental Research Laboratory
Office of Energy, Minerals, and Industry
Research Triangle Park, NC 27711
Prepared for
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Research and Development
Washington, DC 20460
-------
FOREWORD
Man and his environment must "be protected from the adverse effects of
pesticides, radiation, noise, and other forms of pollution, and the unwise
management of solid waste. Efforts to protect the environment require a focus
that recognizes the interplay between the components of our physical environ-
ment—air, water, and land. The Tennessee Valley Authority contributes to
this multidisciplinary focus through programs engaged in investigations of
• chemical and physical characteristics of process emissions and
secondary pollutants formed in the biosphere, and their measurement,
• mechanisms governing the movement of these residuals through the
biosphere, and their potential for impacting human health and
safety,
• effects of process residuals in combination with natural phenomena
on terrestrial and aquatic receptors,
• mitigation alternatives and beneficial uses of residuals, and
a search for economic processes and management alternatives to prevent environ-
mental damages and to reclaim values from process residuals.
This report presents data from a study of a potassium scrubbing system
that will result in the production of useful forms of sulfur from pollutants
while utilizing a low-energy process for regeneration of the absorbing medium.
No solids for disposal are generated in the recovery process. Also included
in the report are studies of the oxidation of some sulfites that might be
formed in sulfur dioxide scrubbing processes. Such oxidation would be benefi-
cial to improve the value of the products, to improve the physical properties
of the products, and/or to reduce the chemical oxygen demand of the products.
Also studied were the use of weak sulfuric acid resulting from some sulfur
dioxide scrubbing processes and the concentration of the products of that use.
Two processes for use of melamine to remove sulfur dioxide from stack gas were
studied.
ii
-------
CONTENTS
Foreword ii
Figures iv
Tables vii
Executive Summary ix
Acknowledgments xi
Potassium Scrubbing 1
Introduction 1
Thermal Decomposition of Potassium MetabisuILfite h
Reduction of Potassium Sulfate with Carbon ik
Scrubbing with Potassium Carbonate Solution 20
Oxidation of Sulfites 29
Prior Work 29
Equipment and Procedure 29
Soluble Salt Oxidation Tests 33
Oxidation of Other Salts Vf
Selective Precipitation of Sulfate with Barium Salts 53
Test Plan 53
Evaluation of Data . 55
Conclusions 68
Utilization of Weak Sulfuric Acid Produced in Pollution
Control Processes 69
Use of Weak Sulfuric Acid for Extraction of Phosphate Rock .... 70
Production of Wet-Process Acid from Unbeneficiated Matrix 79
Concentration of Phosphoric Acid by Freezing 87
Concentration of Ammonium Phosphate Solution by Freezing 96
Melamine Scrubbing 102
Tests with Synthetic Gas 102
Tests with Colbert Steam Plant Gas 108
Minipilot-Plant Tests 120
Sodium Sulf ite-Melamine Process 128
Production of Melamine Sulfite from Sodium Bisulfite
Solution and Melamine 128
Production and Characterization of Melamine-Sulfur
Oxide Adducts in Pure Form 138
Production of Melamine Sulfite from Melamine and Sodium
Bisulfite Scrubber Liquors 1^2
References 153
iii
-------
FIGURES
Number Page
1 Solubility of K2S03 and K2S205 in Water 2
2 Potassium Scrubbing System for Removal of Sulfur Oxides from
Waste Gas Streams—General Flowsheet 3
3 Thermal Decomposition of K2S2Os--Laboratory Test Equipment
Arrangement 5
U Laboratory Apparatus Used in Tests of the Reduction of
Potassium Sulfate with Coke Breeze 16
5 Oxidizer Units Used in Tests of Oxidation of Sulfites
with Air 30
6 Oxidation Vessels Used in Tests of Oxidation of Sulfites
with Air 31
7 Tall Oxidizer Used for Study of Pool Depth 32
8 pH During Reaction of 0.2 N Ammonium and Potassium Hydroxides
with Sulfur Dioxide kO
9 Effect of Time on Proportion of Sulfite Oxidized 10
10 Effect of Time on Moles of Sulf ite Oxidized kk
11 Effect of Pool Depth on Rate of Oxidation 45
12 Effect of Pool Depth on Energy Required Per Mole of Sulfite
Oxidized k6
13 Effect of Pool Depth on Oxygen Utilization k6
14 Laboratory Precipitation of Sulfate with Barium Carbonate—
Test 1 56
15 Laboratory Precipitation of Sulfite with Barium Carbonate--
Test 2 57
16 Laboratory Precipitation of Sulfite with Barium Carbonate—
Test 3 58
17 Laboratory Precipitation of Sulfate and Sulfite with Barium
Carbonate—Test k 59
18 Laboratory Precipitation of Sulfate and Sulfite with Barium
Carbonate—Test 5 60
19 Laboratory Precipitation of Sulfate and Sulfite with Barium
Carbonate—Test 6 6l
20 Laboratory Precipitation of Sulfate and Sulfite with Barium
Carbonate—Test 7 62
21 Laboratory Precipitation of Sulfate and Sulfite with Barium
Carbonate—Test 8 63
22 Laboratory Precipitation of Sulfate and Sulfite with Barium
Carbonate—Test 9 6k
23 Laboratory Precipitation of Sulfate and Sulfite with Barium
Carbonate—Test 10 65
iv
-------
FIGURES (continued)
Number
2k Laboratory Precipitation of Sulf ate and Sulfite with Barium
Carbonate—Test 11 66
25 Effect of Concentration of H2S04 on Dissolution of Metal
Oxides from Phosphate Rock 80
26 Freezing Point of Phosphoric Acid 89
27 Projected Flowsheet for Concentration of Dilute Phosphoric
Acid by Three-Step Batch Freezing 91
28 Cooling Curves for Weak Phosphoric Acid Samples in Freeze-
Concentration Tests 94
29 Utilization of Sulfur Dioxide from Stack Gas in Production of
Ammonium Phosphate Liquid Fertilizer—Flow Diagram 97
30 Laboratory Removal of S02 from Synthetic Gas by Scrubbing with
Melamine Slurry—Typical Analyses of Gases at Inlet and
Outlet of First Scrubber Versus Time 106
31 Laboratory Removal of S02 from Synthetic Gas by Scrubbing with
Melamine Slurry—Effects of Regeneration on Absorption
of S02 10?
32 Equipment Arrangement Used in Tests of Melamine Scrubbing of
Colbert Steam Plant Stack Gas 110
33 Thermal Regeneration of Spent Melamine Slurry Ill
34 Chemical Regeneration of Spent Melamine Slurry Ill
35 Absorption of S02 from Colbert Steam Plant Gas by Melamine
Scrubbers 115
36 Recovery of S02 During Regeneration of Spent Melamine
Scrubbers 116
37 Spray Scrubber Used in Melamine Scrubbing Tests--
Configuration 1 121
38 Photograph of Diverter Screen After Use in Melamine
Scrubbing Test 123
39 Spray Scrubber Used in Melamine Scrubbing Tests—
Configuration 2 124
40 Spray Scrubber Used in Melamine Scrubbing Tests--
Configuration 3 126
4l Effect of Reaction Temperature on Composition of Filter Cake
from Reaction of Melamine with Sodium Bisulfite Solution . . . 133
42 Effect of CsHeNsiNaHSOs Mole Ratio on Composition of Filter
Cake from Reaction of Melamine with Sodium Bisulfite
Solution 13^
43 DSC Trace of Melamine Sulfite Tetrahydrate 137
44 Sodium Sulfite-Melamine Process—Differential Scanning
Calorimeter Analysis of Pure Products 143
45 Sodium Sulfite-Melamine Process—Differential Scanning
Calorimeter Analysis of Mixed or Impure Products 144
46 Production of Melamine Sulfite from Melamine and NaHS03
Scrubber Liquor—Proportion of NaHS03 Reacted Versus Time . . 147
-------
FIGURES (continued)
Number Page
kj Production of Melainine Sulfite from Melamine and NaHS03
Scrubber Solution—Effect of Stoichiometry on Time
Required to Reach Maximum Proportion of NaHS03 Reacted .... 148
48 Production of Melamine Sulf ite from Melamine and NaHS03
Scrubber Liquor—Effect of Stoichiometry on C3HeN6:S Mole
Ratio in the Final Filter Cake 150
49 Production of Melamine Sulfite from Melamine and NaHS03
Scrubber Liquor—Effect of Stoichiometry on Sulfur
Oxidation in the Final Filter Cake 152
vi
-------
TABLES
Number
1 Thermal Decomposition of Potassium Metabisulfite—Factorial
Test 1—Data and Analysis 7
2 Thermal Decomposition of Potassium Metabisulfite—Factorial
Test 2—Data and Analysis 12
3 Reduction of Potassium Sulfate with Coke Breeze (Laboratory
Tests)—Factorial Test—Data and Analysis l8
k Removal of Sulfur Dioxide from Synthetic Stack Gas by Potassium
Scrubbing—Reproducibility Tests 23
5 Removal of Sulfur Dioxide from Synthetic Stack Gas by Potassium
Scrubbing (Laboratory Tests)—Factorial Test—Data and
Analysis 26
6 Oxidation of Sulfites—Laboratory Test Conditions and
Results Jlj.
7 Oxidation of Calcium Sulfite by Air (Laboratory Tests) .... 51
8 Selective Precipitation of Sulfate with Barium Salts—
Test Data 5U
9 Extraction of Phosphate Rock with Weak Sulfuric Acid—
Test Plan 71
10 Extraction of Phosphate Rock with Weak Sulfuric Acid--
Primary Data 73
11 Extraction of Phosphate Rock with Weak Sulfuric Acid—
Calculated Responses jk
12 Extraction of Phosphate Rock with Weak Sulfuric Acid--
Statistical Evaluation of Responses 75
13 Utilization of Low-Grade Phosphate—Test Plan 8l
Ik Utilization of Low-Grade Phosphates—Test Results 83
15 Utilization of Low-Grade Phosphates—Effects of Variables ... 84
16 Laboratory-Scale Batch Tests of Concentration of Dilute
Phosphoric Acid Solutions by Freezing 93
17 Tests of Freeze Concentration of Ammoniated Phosphoric Acid
Solution—Test Data 98
18 Tests of Freeze Concentration of Ammoniated Phosphoric Acid
Solution—Effect of Procedures of Separation on Purity of
Solid (ice) Phase 100
19 Simulated Countercurrent Washing Procedure for Freeze
Concentration of Liquid Fertilizer 101
20 Removal of SuiLPur Oxides from Waste Gases by Melamine Scrubbing
(Laboratory Test Data) 105
vii
-------
TABLES (continued)
Number
21 Removal of Sulfur Oxides from Waste Gases by Melamine
Scrubbing (Data from Laboratory-Scale Tests at Colbert
Steam Plant) 113
22 Chemical Analyses of Final Products from Scrubbing Stack Gas
with Melamine Slurry 118
2J Production of Melamine Sulfite from Melamine and Sodium
Bisulfite Solution--Laboratory Test Data 129
2k Effect of Variables—Differences in Responses Due to
Variables 131
25 Production of Melamine Sulfite from Melamine and Sodium
Bisulfite Solution—Analysis of Products 136
26 Optical Description of Adducts 139
27 Sodium Sulfite-Melamine Process--Preparation of Pure Reaction
Products for Characterization 1^0
28 Production of Melamine Sulfite from Melamine and NaHS03
Scrubber Liquor—Laboratory Tests lk-6
viii
-------
EXECUTIVE SUMMARY
Interest in regenerative potassium scrubbing processes results from the
favorable solubility relationship in this system; potassium pyrosulfite,
heavily loaded with S02, can be precipitated from scrubber effluent. Labora-
tory studies indicated excellent S02 removal from stack gas but indicated
difficulty in precipitating potassium pyrosulfite unless a cooling step is
used to cause precipitation. Antioxidants were helpful in controlling oxida-
tion during scrubbing. Laboratory studies of thermal decomposition of
potassium pyrosulfite indicated disproportionation to potassium sulfate, thio-
sulfate, and sulfide occurred to the extent that appreciable sulfate removal
would need to be provided. Laboratory studies of the reduction of potassium
sulfate indicated that efficient reduction might be difficult but that a
process might be technically feasible which involved potassium scrubbing,
evolution of one-third of the sulfur as S02 by thermal decomposition,
reduction of two-thirds to hydrogen sulfide, and reaction of these to
elemental sulfur by the Glaus process.
In laboratory oxidation studies, a gas-liquor contacting device was
developed that appeared superior to a spinning cup oxidizer developed in
Japan. Conditions were defined for oxidation of sulfites of ammonia,
potassium, zinc, and calcium. Bench-scale studies indicated that increasing
the depth of submergence of the contactor up to 2 feet improved oxidation but
further increases were less helpful.
Laboratory studies indicated that sulfate could be selectively precipi-
tated from soluble alkali scrubbing systems by precipitation with barium salts;
barium sulfate formed could be processed into elemental sulfur by commercially
practiced methods.
Laboratory studies of extraction of phosphate rocks with dilute sulfuric
acid indicated some promise for utilization of acid from stack gas recovery
processes for acidulation of unbeneficiated phosphate matrix to produce dilute
phosphoric acid.
Laboratory studies were made of freeze crystallization of ice from dilute
phosphate systems which might be made by extraction of phosphate rock with
dilute sulfuric acid recovered from stack gas scrubbing systems. Results
indicated that a freeze concentration procedure, with continuous countercurrent
washing of ice, might be used to attain P205 concentrations of 20 to 25$ in
either dilute phosphoric acid or ammonium phosphate systems; these concentra-
tions should be usable. Water removal by freeze crystallization should be
much more energy efficient than evaporation.
IX
-------
EXECUTIVE SUMMARY (continued)
Melamine sulfite is very easy to decompose to melainine and S02 and can
be formed by slurry scrubbing with melamine or by reaction of melamine with
alkali bisulfite solution. Two possibly improved S02 recovery processes were
identified for further development.
-------
ACKNOWLEDGMENTS
The cooperation of the Emission Control Development Projects staff in
their guidance and of the Fundamental Research Branch staff in their advice
and analytical capability are recognized and appreciated. We are particularly
indebted to Dr. John D. Hatfield for his advice and to Mr. Joe Gautney for his
active participation in some of the reported tests.
Most of the test work required for this study was performed by M. C. Nason,
F. D. Nix, W. E. Fowler, J. R. Clemmons, and J. L. Cabler. Their contributions
are gratefully acknowledged.
The cooperation and guidance of Mr. Robert H. Borgwardt, Project Officer,
U.S. Environmental Protection Agency, on this contract is recognized and
appreciated.
We gratefully acknowledge the efforts of the secretarial staff,
particularly Mrs. Martha Witt, who typed this report in accordance with
EPA format.
xi
-------
POTASSIUM SCRUBBING
INTRODUCTION
Several attempts have been made to take advantage of the unique
solubility relationship between potassium sulfite and potassium bisulfite
(potassium metabisulfite or potassium pyrosulfite) in a regenerable S02
recovery process. The more acid bisulfite salt is less soluble than the
sulfite at temperatures below about 8o°C (see Figure l). The attempts largely
have been abandoned for several reasons including unwanted oxidation in the
process, disproportionation in the regeneration step, and high energy require-
ments for recovering the sulfur value in both the regeneration and reduction
steps. A generalized flowsheet is given in Figure 2.
Desired reactions occurring at various points in the process are given
below.
Scrubber
K2S03 + S02 >• K2S205
Decomposer
K2S205 > K2S03 + S02
If these were the only reactions, the process would be very simple and the
reduction step would not be needed. However, some oxidation takes place in
the scrubber. This reaction, K2S03 + 1/202 > K2S04, forms the sulfate
which will not react with S02 in the scrubber and necessitates the removal of
sulfate from the system, either by bleeding a fairly expensive potassium
stream or by reduction to a form suitable for absorption of S02.
In the decomposition step, a complicated disproportionation of the sulfur
species takes place. Instead of the simple dissociation of K2S205, there
occurs a redistribution of oxygen among the sulfur atoms and compounds
containing sulfate, sulfite, sulfide, polysulfides, thionates, etc. are
formed. Of these, the sulfate is predominant.
In the reduction step, the potassium-sulfur salts are reduced to the
sulfides. There is some volatilization of potassium sulfide; however, the
off-gas stream from the reduction furnace may, in part, be used in the
hydration-carbonation reactor where it would be recovered. Reactions
occurring would be:
-------
0 10
20
50 6o 70
Temperature, °C
80 90 100
Figure 1. Solubility of K2S03 and K2S2Os in water, (l)
-------
Stack
gas •
Heat,
Carbon
Heat-
Clean gas
I
SCRUBBER
J
K2S205 Solids
Solution
FILTER
K2S2Os
solids
*
DECOMPOSER
(200°C)
K2S03
K2S04
REDUCTION
FURNACE (950°C)
K2S03-K2S04 solution
Solution
-»-SOg to recovery
Makeup
K2C03
H20 + C02.
.TT
H20«
CO-
KoS
ir
REACTOR
to recovery
solution
DISSOLVER
Figure 2. Potassium scrubbing system for removal
of sulfur oxides from waste gas streams—general flowsheet.
-------
K2S04 + 2C > 2C02 + K2S
K2S + 2H20 * 2KOH + H2S
2KOH + C02 > K2C03
The scale of the reduction loop in the process would "be dependent upon
the amount of sulfate formed in the scrubber and the amount of disproportiona-
tion occurring in the thermal decomposer. Since the reduction step is
included only to recover potassium in a form suitable for scrubbing and is
expensive in both energy and investment costs, it is desired to limit the
amount of material requiring reduction to as small a quantity as is practical.
Means of accomplishing this reduction in quantity include supression of
oxidation in the scrubber, such as limiting excess air in the plant boiler or
including antioxidants in the scrubber and finding some operating procedure
or additive in the decomposition step to limit disproportionation.
The test program was begun with a factorial study of the decomposition
step.
THERMAL DECOMPOSITION OF POTASSIUM METABISULFITE
The problem of disproportionation appeared to be of greatest concern and
to affect most the proportion of the regenerator product that would have to be
directed through the reduction step.
There has been some conjecture that the disproportionation reaction might
be dependent, in part, on the presence of the sulfur dioxide liberated in the
decomposition process. In previous tests of open-air and vacuum heating of
potassium metabisulfite, reported in the December 1968 report of the Applied
Research Branch, Division of Chemical Development, Tennessee Valley Authority,
up to half of the sulfur was volatilized in useful form but an excessive
amount of the sulfur remaining in the residue was in the unwanted sulfate
form. Later (November 1970 report) tests were made to determine whether
sweeping liberated sulfur dioxide away with carrier gases might decrease
sulfate formation. In tests with nitrogen, carbon dioxide, or air as sweep
gas in a fluidized bed, about k5% of the sulfur was volatilized at 200° to
570°C but sulfate formation still was excessive. Lowest sulfate formation was
obtained with humidified nitrogen as the sweep gas, but about 22$ of the input
sulfur was converted to sulfate or to equally undesirable thiosulfate.
In the current work, a first study was made in which reagent-grade
potassium metabisulfite was subjected to a series of factorially designed
tests in which it was heated in a glass vessel in a pot furnace to various
temperatures, for varying times, with varying amounts of a sweep gas (N2)
being directed over or through the bed of material.
Test Plan and Equipment
The equipment used in the tests is shown in Figure J. Test temperature
was controlled by adjusting the temperature of the pot furnace; the tempera-
tures of the furnace and sample were measured and recorded using the thermo-
-------
^>.
V— /
•^ ^{_
=Cx3=7
Dilution
gas
meter (
Sweep gas
inlets
W. .
r=CXj=
p.
w
-n — X - '
T~T
i i
Sweep |
gas .
meter
!
Nitrogen .
source ' 1
i
^ •.•
••"* •
' "^~
. — **
\\
\v
Gas
outlet -7
/ ^ W a*,
^/? — cell
1
1
1 1
1 1
^ Thermocouple
•*^S ' 1
1
/Sample .
...S \ '
'
-------
couples shown. The sweep gas was metered and passed either through the bed
(through the fritted glass disc) or into the gas space above the bed by
opening the appropriate gas inlet line. The alternate gas inlet was sealed.
Off-gases were analyzed for S02 using the UV (ultraviolet) analyzer after
dilution with a measured amount of N2 which was added to keep the S02
concentration within the limits of the analyzer.
Using this equipment, a factorial test series was made in which the
following conditions were imposed.
A -
B
C -
D
Variable
Furnace temperature, °C
Bed condition (gas direction)
Time, min
Sweep gas rate, 1/min
Lower
(-)
168
Static
15
2-5
Limit
Upper
(+)
2(A-
Fluidized
30
3
Responses measured were (l) the temperature attained in the product bed,
(2) the weight loss of the product, (3) the area of the curve of the UV
recorder chart to measure sulfur dioxide in the off-gas, and (k) the chemical
analysis of the product. From the chemical analyses of the raw material and
products, the following indices were calculated, assuming no K20 loss:
(l) the final weight of the product, (2) total sulfur loss, $ of total sulfur
input ($ TS), (3) reduction in suLfite sulfur content, $ TS, (4) dispropor-
tionation, % TS (which is the difference in the apparent losses of sulfite
sulfur and total sulfur species), (5) the increase in sulfide sulfur, % TS,
(6) the increase in thiosulfate suLfur, $ TS, (7) the increase in sulfate
sulfur (by difference), % TS, (8) the total sulfur loss, expressed as S02, %
of input weight, and (9) the weight loss (K20 balance), % of input weight. In
a later series of tests, some P205 was added to determine its effect.
Responses to show P205 recovery were calculated also. For the main body of
data there were, then, 12 responses measured. Results of the tests are given
in Table 1 along with calculations made from the chemical analyses of the
residual solids and the statistical analyses of the several responses
calculated or measured.
Evaluation of Data
The statistical analysis was made by the method suggested by Davies (2)
but modified so that the values shown are the differences in the responses due
to the various treatments; included also are the minimum values for which
these differences are significant at the 90, 95, and 99$ confidence levels.
As an illustration, under the response column headed "Bed temperature, °C,"
it will be noted that there is a 90$ confidence that any difference in
response greater than 2.569°C is a real effect of the variable or interaction,
a 95$ confidence in differences greater than 3.277°C, and a 99$ confidence in
differences greater than 5.1^0°C. For convenience, all differences with 90$
confidence or greater than the effect is real are underlined. Thus, in the
"Bed temperature, °C" column, it is seen that the furnace temperature varia-
tion from about 170°C to 205°C caused a variation in bed temperature of
-------
TABLE 1. THERMAL DECOMPOSITION OF POTASSIUM METABISULFITE
FACTORIAL TEST 1—DATA AND ANALYSIS
Responses
Teat conditions
Plan Furnace
poai. Test taop. ,
tlon No. 'C
Haw Bjaterial analysis
1 1JB 169
2 14D 169
3 15 B 170
4 16B 169
5 2 170
6 1 173
7 4 170
8 3 169
9 10 204
10 11 208
11 12 2O7
12 9 207
13 7 203
14 5 206
15 6 207
16 6 203
Average
Bed
condi-
tion*
5
5
S
S
¥
T
r
F
S
£
S
S
F
P
F
F
Differences^ln responses due
Furnace teapcrature
Bed condition (B)
Tlae (C)
Gas rate (D)
Two- factor Interactl
AB
AC
AP
BC
BD
CD
Least significant diff
90? confidence
95JS confidence
99)1 confidence
(A)
ons
erences
Gas
Tine, rate,
15 2.5
15 3
30 2.5
30 3
15 2-5
15 3
30 2.5
30 5
15 2.5
15 5
30 2.5
30 3
15 2.5
15 3
30 2.5
30 3
to
Bed
tesip. ,
•c
118
107
123
108
146
151
147
145
143
138
152
137
186
183
190
182
147.25
itil
l7*»0
-6.75
*^~^^
4.7;
1^25
-1.00
-2.00
-3.25
2.569
5.277
5.140
Weight
loss
g
(meal.)
0.}
0.3
0.3
0.4
0.1
-0.1
0.25
0.15
0.2
0.25
0.2
0.2
0.05
0
0.35
0.3
0.2031
-0.019
-0.131
p. 131
-0.031
0.094
0.006
0.019
0.119
.0.069
0.019
0.0570
0.0728
0. 1141
Gas analy-
sis chart
area. In*
(UV trans-
oittance)0
0.0012
0.0016
0.0039
0.0041
0.0330
0.0310
0.1210
0.1190
0.0640
0.0230
0.4400
0.}800
0.1420
0.1560
0.7090
0.3360
0.1606
0.2415
0,200*3
-0.0586
0.0185
o. 1617
-0.0564
0.0225
-0.0322
-0.0514
0.0812
0.1036
0.1625
% of Input
Chemical analysis.
Total
2T-5
27.6
27-5
26.9
27.1
27.}
27.4
27.0
27-3
27.4
27.3
26.9
26.9
26.9
27.0
25.8
26.7
Sulfur
SO^' S"
25.6 0.9
24.3 0.8
24.5 0.6
22.6 0.9
23.2 0.6
23.7 0.4
24.5 0.4
23.4 0.6
23.2 0.9
22.3 0.9
22.4 1.0
19.1 1.1
17.6 0.9
21.6 0.9
21.4 1.3
15.2 1.3
18.5 1.3
s 0,
nil
0.1
0.2
0.3
0.7
Nil
Nil
2.8
0.1
Hil
III!
0.1
0.6
Nil
0.}
0.5
0.1
* by «.
50*"
0.8
2.4
2.2
2.9
2.6
3.2
2.5
0.2
3-1
4.2
3-9
6.6
7-8
4.4
4.0
8.8
6.8
42.5
42.9
42.9
42.8
42.9
43-3
42.6
42.7
4}.0
43.0
4.?. 8
43.6
4}.5
4}.l
4}. 6
45.9
44.4
43.3125
0.850
0.525
0.575
-0.200
0.500
0.650
-0.125
0.275
-o. 150
-0.100
0.5br,9
0.7221
1.1324
Final
vt. , g Total
( K20 sulfur
4.9534
4.9534
4.9650
4.9534
4.9076
4.9883
4.9766
4.J419
4.9419
4.9650
4.8739
4.8851
4.9304
4.8739
4.6296
4.7860
4.9073
-0.0943
-0.0571
-0.0626
0.0211
-0.0544
-0.0715
0.0125
-0.02B7
0.0154
0.0093
O.C613
0.0782
0.1226
0.572;:
0.9324
2.8675
2.3734
2.5614
0.5975
2.2780
1.8816
1.5222
1.42}1
4.6497
4.4305
3.5436
4.2952
13.1313
7.0639
3.3827
3.2494
2.0727
2.9055
-1.0160
1.9294
1.719}
-0.3925
0.4557
-0.9030
-0.7782
1.7737
2.2631
3.5492
Reduc-
tion in
sulTlte
6.2784
5.5580
11.4902
10.2412
9.2286
4.9364
9.1256
10.4555
13.6702
12.9346
26.1159
31.2894
16.3662
17.9633
42.6397
29.4242
16.1061
15.5887
2.6177
10.4762
-1.5166
2.7782
6.6554
-0.2786
0.5044
-2.1330
-0.4768
4.3427
5.5410
8.6898
Sulfur
dispropor-
5.7062
4.6256
8.6227
7.8678
6.6672
4.}}89
6.8476
8.5539
12.1480
11.5115
21.4662
26.8589
12.8226
13.6681
29.5084
22.3603
12.7234
12.1395
0.7450
7.5747
-0.5005
0.8487
4.9562
0.1139
-0. 1}14
-1-2507
0.2995
2.7991
3-5715
5.6011
sulfur
]
Sulflde
-0.3908
-1.1113
-0.0229
-1.1113
-1.8451
-1.8216
-1.1011
-0.0381
-0.0381
0.3581
0.6264
-0.0752
-0.0456
1- 335}
1.1044
1.252}
-0.1640
1.4925
oTWB5
0.5267
0.0601
0.6206
-0.1971
0.2407
0.3719
0. 59 57
-0.2049
0. 4476
0.5711
0.8957
H of input weight
Increase 1
Thlo-
aulfate
0.3602
0.7205
1.083}
2.5217
0.0000
0.0000
10.1341
0. 3594
0.0000
0.0000
0.3545
2.1317
0.0000
1.0634
1.6835
0.3481
1.2975
-1.1997
0.8021
2.0590
-0.8089
-0.6499
-1.1954
1.1851
O.B06U
-1.7020
-1.1648
2.5597
( tiS9
',.1218
in
Sulfate
5.7363
5.0165
7.5624
6.4573
8.5123
6.1605
-2.1852
8.2}26
12.1660
11.1735
20.4854
24.8025
12.8682
11. 2694
26.7205
20.7600
11.6099
11.8466
-0. 1352
4.9890
0.2482
0.8780
6.5288
-1.3120
- 1 3O98
-0.1216
1.6692
4.0176
c,. 1202
3.0W2
Total s
loss
as S0?
0.3147
0.5128
1.5771
l.}054
1.4088
0.3286
1.2529
1.0}49
0.8372
0.7827
2-5575
2.4368
1.9490
2.3624
7.2222
3.3851
1.8605
1.7672
1.1400
1.5970
-0.55BO1
1.0612
0.9456
-0.2158
0.2396
-0.4966
-0. 4280
O.^75b
1../447
1.9521
Weight
loss
(K.,0
balance)
0.9324
0.9324
0.7009
0.9324
1.8476
0.2347
0.4684
1.1628
1.1628
0.70QJ
2.5229
S.?963
1.3921
2.5229
T.4074
4.2791
1.84 37
1.1414
1.2559
-0.4213
1.0876
1.4266
-0.2495
0.5742
-0.3077
-0. 1853
1.2208
1.5570
2.4423
S - static bed, awep gao paaocd through space above bed. F » fluldlzed bed, sweep gas paused through bed.
Off-gaB diluted and passed through LO-co flow cell In UV analyzer set at 2&J nonooetero.
Initial weight of K^S^O., was 5<00 grama.
Negative algn Indicates increase in response at the higher Icwl of th-.- variable; underlined valued are significant nt 90Jl confidence level.
-------
33.25°C and that there is greater than 99$ confidence that this is a real
effect of the furnace temperature variation. An even more effective variable,
however, was the variation in bed condition; the tests with the "bed fluidized
averaged 38°C hotter than the static bed tests. This probably was due to the
transfer of heat into the bed more rapidly with the incoming gas, to the
removal of decomposition products (S02 and moisture) from the bed,, and to the
mixing of the bed which exposed more particles to the warmer edges of the bed.
There were smaller, but still at least 90% significant, differences due to the
gas rate (higher gas rate lowered temperature), and to the interactions of
furnace temperature and bed condition (AB) in which the temperature averaged
4.75°C higher when the furnace temperature and bed condition had the same
sign (either + or -) than when the signs of these variables were mixed. Also
effective were the interactions BD and CD.
Weight loss from the samples was measured to the nearest 0.05 gram.
This measurement was not precise enough to give highly significant results
and greater than 0.057-gram differences were required for 90$ confidence in
the effects. Bed condition and time, along with the AB, BC, and BD inter-
actions, were significant. The lack of precision in this measurement was the
result of the relatively small sample weight (5 g) and the large weight of the
decomposition chamber (about 500 g) in which it had to be weighed. In order
to get better material balances, it was decided to calculate the product
weight from the input weight and the K20 analyses of the products and the raw
material. Since there should have been no K20 loss, the weight calculated by
this means should be about as accurate as the analytical method, the variation
of which should give results within .05$ or a variation in product weight of
0.006 gram. The material balances were calculated from the weights determined
in this manner.
Measurement of S02 in the off-gas stream, as reflected by the area under
the UV analyzer curve, responded to variations in temperature, bed condition,
and time, along with the temperature-time (AC) interaction. Temperature and
time were the most effective of the variables, higher temperature and longer
time increasing the amount of S02 in the off-gas stream.
Since the remainder of the responses were to be based on the weight
calculated from the K20 analyses, it was decided to determine what effect the
variables had on the product K20 analyses. These values were affected only by
the temperature and time variables and their interaction as would be expected.
Higher temperatures and longer times caused greater losses of weight and thus
increased the K20 content of the residue. Temperature was the most effective
of the variables in these tests.
The final weight, based on K20 analyses, responded to the same variables
and in the same order as did the K20 analyses but in the reverse direction, as
expected.
Loss of total sulfur content averaged about 3.4$ of the input sulfur or
about 6.8$ of that desired in the process. A maximum value of 13.1$ (26.2$ of
desired) was attained. Total sulfur loss was increased by increases in
temperature and time, by fluidizing the bed, and by the interaction of
temperature and bed condition.
8
-------
Reduction in sulfite content averaged 16.1$ and reached a maximum of
1).2.6$. A value of 50$ is theoretical for pure K2S205. More sulfite was lost
at higher temperatures, longer times, and the interaction between the two.
The proportion of the input sulfur that was changed in oxidation state
(disproportionation) was calculated as the difference in total sulfur loss and
sulfite sulfur loss. This sulfur was still present but in different states of
oxidation. Disproportionation was increased by the same variables that
increased the sulfite sulfur loss and in the same order (temperature > time >
AC interaction). Of the average 16.1$ sulfite sulfur lost, 12.7$ (79% of that
lost) was due to disproportionation. Reduction of this value is highly
desirable but the changing of any of the variables which caused disproportion-
ation would adversely affect sulfite loss which is the purpose of the procedure.
Increases in sulfide, thiosulfate, and sulfate suiLfur content were
measured to determine the forms to which the sul±ite sulfur was altered.
Sulfate sulfur was calculated as the difference between the total sulfur
content and the sum of the values for the other forms. None of the variables
had a significant effect on the thiosulfate sulfur increase which averaged
about 1-3$ of the input sulfur and varied up to 10.1$. Both the sulfide and
sulfate sulfur contents increased with increases in temperature and time.
Sulfide sulfur was increased also by the AB and BD interactions while sulfate
sulfur was increased by the AC interaction. The primary phase produced during
disproportionation was sulfate; the second most abundant was thiosulfate. It
is noted, however, that the oxygen requirement for the indicated amounts of
sulfate and thiosulfate produced (averages) is greater (about 127$) than the
amount of oxygen available from the sulfite disproportionated.
The total weight of the sulfur lost, calculated as S02, averaged about
1.86$ of the input sample weight. This value compares very favorably with the
weight loss of 1.814-$ as calculated by the K20 balance. Both these values were
increased most by increases in temperature and time although the su]jfur loss
value was increased by fluidizing the bed also. Interaction AB increased
sulfur loss and AC increased weight loss.
It was concluded from this part of the study that temperature was the
most significant factor in causing disproportionation and that disproportion-
ation was not decreased by rapid removal of the gaseous decomposition products
from the solid residue. The latter was the primary purpose of the tests.
Generally, the tests resulted in a reduction in sulfite sulfur equivalent to
about five times the value of the total sulfur lost from the sample. Since
most of the sulfur lost was probably lost as sulfur dioxide, disproportiona-
tion probably accounts for about 80$ of the sulfite sulfur reduction. The
disproportionation reaction appears to be one in which sulfur dioxide is
reproportioned, primarily, to sulfate and thiosulfate.
Further Tests, Effects of PP05 and Moisture
Exploratory tests were made to determine the effects of further increases
in furnace temperature, of the addition of phosphate, and of the addition of
antioxidants to the system. The latter two test series were made to attain a
better understanding of the potassium phosphate system and the use of anti-
-------
oxidants to suppress oxidation in the scrubber system. Tests with P205
present, added as monopotassium phosphate were promising.
Pertinent data from those tests, along with comparative tests without
P205, are given below. The addition of P205 decreased the overall reduction
in sulfite content while increasing the total sulfur evolution from the
samples, thus decreasing significantly the proportion of sulfur disproportioned.
P205
content,
%
Furnace
temp. ,
°C
Time,
min
Gas
rate,
l/mina
Reduction
in sulfite
sulfur,
i> of
input S
Total
sulfur
evolution,
% of
input S
Sulfur
dispropor-
tionation,
% of
input S
Tests with 15-Minute Retention Time
5
5
0
202
20k
203
15
15
2.5
2.5
Average
2.5
Improvement
12.28
11.^6
11.87
16.JT
IK 50
11.27
7.99
9.63
3.5k
6.09
l.oo
3.U7
2.2k
12.82
10.58
Tests with 30-Minute Retention Time
5
5
202
20l|-
207
30
30
30
2.5
2-5
Average
2-5
Improvement
27.21
33-91
30.56
U2.61*
12.08
13.99
16.1*3
15.21
13.13
2.08
13.22
17 -k9
15.36
29.51
3k. 15
In all tests the bed was fluidized.
A second factorial test was planned in which the effects of potassium
phosphate addition was the primary area of study. It was possible that, at
the temperatures of the test (168° and 20l)-°C) in the exploratory work, some
water might have been released as the result of conversion of P205 to poly-
phosphate forms and that this water might have inhibited disproportionate.
Also, in earlier work (November 1970 Applied Research Branch Progress Report,
TVA), it was reported that the use of moist sweep gas gave the lowest sulf ate
proportion in the product. The new test series was designed, therefore, to
determine the effects of time, temperature, P205 content, and moisture content
of the sweep gas on the thermal decomposition of K2S205.
Variables and levels tested in the present series are indicated in the
following tabulation.
10
-------
Level
Variable
Temperature, °C
Time, min
P205 content of feed, %
Moisture in sweep gas
a. Sweep gas,
b. Sweep gas,
with water
2.
2.
at
5
5
1/min
1/min
25°C.
-1
168
15
Oa
+1
2<&
30
1(\
Sat.
dry N2.
N2 saturated
The decomposition chamber, pot furnace, and sweep gas metering system
used in these tests were the same as were used in the earlier series. Changes
made in the rest of the system were as follows: (l) the UV cell and dilution
gas system were removed, (2) the gas was directed through the fritted disk and
sample in all the current tests, and (J) a water-filled gas-absorption bottle,
fitted with a fritted sparger, was installed as a saturator between the sweep
gas meter and the decomposition chamber for use in the tests where the sweep
gas was saturated with water vapor.
Evaluation of Data
Results of these tests and the statistical analyses of the several
measured or calculated responses are given in Table 2. Since the addition of
phosphate was a new feature of the system, analyses were made of the residue
after the tests to determine its retention. A material balance, based on
total retention of the K20 in the sample, indicated an average gain in P205 of
0.02 gram or about 5$ of the average input P205 content. This value is very
small, considering the method used to make the calculations. None of the
variables had significant effect on the P205 retention in the sample so any
variation was considered as error.
Total sulfur evolution averaged 10-5$ and was increased by increases in
temperature, time, moisture content of the sweep gas, and the interactions of
temperature with time (AB) and moisture content of the gas (AD). The change
in temperature resulted in the greatest difference but the addition of moisture
was almost as effective.
Reduction in the sulfite sulfur content averaged 16$ and was due to the
same three individual variables with temperature being the most effective,
time was next, and moisture content the least effective; the temperature-time
interaction was significant also.
Disproportionation, the difference between the reduction in sulfite
sulfur content and total sulfur evolution, responded to temperature changes
only.
The forms of the disproportionated sulfur compounds were determined from
analyses for total, sulfite, sulfide, and thiosulfate sulfurs; sulfate sulfur
was calculated as the difference between the total sulfur content and the sum
of the remaining forms.
11
-------
ro
TABLE 2. THERMAL DECOMPOSITION OF POTASSIUM METABISULFITE
FACTORIAL TEST 2—DATA AND ANALYSIS
Plan Furnace P^OS Gas
poll- Test trap., Tla*, content, Kisture
1 13 168 15 5
2 1OA
5 U 10
4 12A
59 30 5
6 14A
7 15 10
8 16A
9 3 204 15 5
10 6A
11 BA 10
12 4A
13 5 30 5
14 7
15 IB 10
16 2B
Average
Raw materials
II 10
Dlfferencescin responses due to
Furnace teoperature (A)
P,0j content (C)
Gas moisture content (D)
Two-factor interactions
AB
AC
Least significant differences
90$ confidence
95% confidence
Dry
Sat.
Dry
Sat.
Bry
Sat.
Bry
Sat.
Bry
Sat.
Dry
Sat.
Bry
Sat.
Dry
Sat.
Bed
te«p.,
153
154
156
151
151
151
«3
154
167
166
lfl9
183
184
184
188
184
Weight
loss,
0.15
-0.15
0.20
0.00
-0.05
0.05
0.05
0.15
0.20
0.25
0.30
0.05
0.55
-0.05
0.60
0.60
25.4
23.8
21.9
20.8
24.9
23.8
21.1
20.9
24.6
22.8
20.6
20.3
23.0
17.7
20.9
15.9
24.9
22.0
—— — ~
—
Chenleal analysis.
21.9
21.0
19.6
18.2
21.4
19.9
18.9
18.5
17.6
17.1
16.7
14.9
15.8
12.6
12.8
10.1
21.9
19.1
0.1
< o.l
0.2
< 0.1
< 0.1
0.1
< 0.1
< 0.1
0.2
0.1
0.1
0.1
0.2
0.2
0.1
0.1
< 0.1
< 0.1
"«""
0.1
< 0.1
0.3
0.2
0.3
0.4
< 0.1
0.3
0.7
0.6
0.3
1.2
0.7
1-3
1.2
0.4
0.3
* by «.
3-35 42.5
2.65 43.3
2.05 41.9
2.25 42.9
3.25 42.6
3.50 44.5
1.75 42.2
2.30 42.8
6.30 43.5
4.90 44.4
3.20 42.5
5.00 45.0
5.80 41.1
4.20 46.0
6.70 45.9
4.50 47.4
2.55 42.5
2.55 41.7
RO SHOD
Final
«., g
(XjO PaO,
5-5
'5.3
10.6
11.3
5-3
5.5
J0.5
11.3
6.0
5.6
12.5
11.0
5.0
5-1
12.2
13-4
5.2
10.35
5.0000
4.9076
4.9761
4.8601
4.9883
4.7753
4.9408
4.8715
4.8851
4.7860
4.9059
4.6333
4.8186
4.6196
I..5425
U-3987
-O.0150
-0.0001
-0.0100
-0.0317
-0.0044
-0.0026
-0.0013
-0.0330
-0.0331
-0.0080
-0.0957
0.0078
0.0191
0.0244
-0.0367
-0.0719
-0.0219
-0.0047
0.0173
-0.0243
0.0004
-0.0013
-0.0255
0.0096
-0.0206
-0.0300
-0.0113
0.0403
0.0514
0.0806
MB
sulfur
evolu-
-2.0080
6.1836
0.9297
8.0992
0.2347
8.7135
5-2273
7.4416
J.4760
12.3521
8.1262
14.4939
10.9818
34.3243
13.6928
36.4183
10.5429
12.3805
8.1727
2.5213
10.9207
Iriro
1*. 1*07 2
-0-5999
3-2695
-1.3015
4. 12J1
5.2608
8.2503
tlon in
•ulflte
0
5.1727
-1.8475
6.4050
2.2098
11.6240
1.9270
•4.8385
18.1092
22,2159
12.3380
24.0576
26.8002
41.1996
33-9602
46.4298
15.9661
24. 5414
8.5620
7.5937
2.0118
2.1118
1.2396
1.2491
0.2888
3-7501
4.7649
7-5040
J of Inwit sulfur 1 at Input wnltfit
Sulfur
dispi-opor-
2.0080
-1.0109
-2.7772
-1.6942
1.9751
2.9105
-3.3003
-2.5531
14.6332
9.8638
4.2118
9.5637
15.8184
6.8753
20.2674
10.0115
5-4252
11.9609
2.1508
-2.4180
-2.3588
1.5242
1.6339
-2.2954
1.6296
-2.0204
1.5902
4-2533
5.4269
8.5107
mo- evolu-
Sulflde sulfate Sulfate tlon
content
0.2008
-0.0037
0.6775
-0.0061)
-0.0005
0.1828
-0.0027
-0.0058
0.5839
0.1836
0.2187
0.1939
0.5733
0.5413
0.1857
0.1726
0.2309
0.2015
-0.0502
-0.1035
-0.1473
0.1234
-0.1743
0.0297
-0.1333
0.1811
-0.0339
0.1631
0.2061
0.3264
consent
-1.4056
-1.2122
-1.1374
-0.0381
-0.8051
-0.4558
0.4330
-1.1422
-0.4293
1.0845
1.3123
-0.1000
3.0380
0.9909
4.0048
3.4350
0.2410
2.8521
1J2099
-0.7707
0.5076
-0.2179
-0.7874
-0. 2192
-1.1193
-0.7731
0.7925
1.0112
1.5858
3.2129 -1.0000
0.2050 3.0794
-2.3172 0.4091
-1.6497 3.5636
2.7806 0.1169
3.1835 4.3393
-5.7306 2.3000
-1.4051 3-2743
14.4786 1.7310
8.5956 6.1514
2.6807 3-5755
9.4697 6.3773
12.2071 5-4689
5.3431 17.0935
16.0769 6.0248
6.4039 16.0241
4.7209 4.9081
9 vpi 5.7955
-yo597 ^5™?
-l.OOH 5-159*
l!s6l5 -TOBIB
., *», 2.0519
1.5175 -0.4l»
-1.546T i'5«5«
1.9327 -0-927"-
4.2047 1-9022
5 5-1,9 2.4270
8.4136 '-8062
lots,
(KjO
0
1.8476
0.4773
2.7972
0.2347
4.4944
1.1848
2.5701
2.2989
4.2795
1.882U
T-3333
3.6281
7.6087
9.1503
12.0253
3.8633
4.3250
J.0154
1.6571
l.?15i>
0.5593
0.6125
0.1127
-0.0046
1.5336
1.9568
3.0683
a Saturated gas prepared by pasting dry gas through fritted glase plug in gu scrubber bottle filled with water.
b Initial weight of charge 5.00 grans.
c Negative aign indicates increase In response at the higher level of the variable; underlined values are significant at ?o£ confidence Level.
-------
The sulfide and thiosulfite sulfur contents increased about 0.2$ each
(average) and the sulfate sulfur content increased about 5$. All three were
increased most severely by increasing the furnace temperature. The thiosulfate
content was increased also by time and P205 content. Interaction AC decreased
sulfide content while BD increased it. Interaction BD decreased thiosulfate
content.
Total sulfur loss (calculated as S02), percent of weight, followed the
same pattern as total sulfur loss, percent of input sulfur. The weight loss
calculated from the K20 balance reponded to all the individual variables and
to the AB interaction. Total sulfur loss, calculated as S02, was greater
(U.9$ vs. 3.9$) than that calculated from the K20 balance. A linear
regression of all the values for sulfur evolution as S02 and weight loss gives
the following equation with a correlation coefficient of 0.85.
Weight loss =0-59 (S02 evolution) + 1
This indicates that the weight loss was not due to the S02 evolution only.
However, if the regression is made on the results from runs with dry gas only,
the equation (correlation coefficient 0.8U) becomes;
Weight loss = 0.99 (S02 evolution) + 0.05
This shows the weight loss to be primarily S02 and indicates that the
samples from tests in which moisture was added may contain a considerably
greater portion of water than those made with dry gas. The moisture addition
assisted in S02 release but the amount of S02 released is not determined by
the weight loss from the sample in this case.
The proportion of sulfur disproportionated in this series of tests was
5.U$ or 3!$ of the sulfite sulfur reduction. This value, calculated for the
previous series, was 79$. Although increasing the P205 content from 5 to
did not significantly affect disproportionation, the addition of 5$ ^2^
have a large depressing effect. Apparently the major part of the improvement
was due to the first 5$ or less. Lowering disproportionation from 79$ to 3^$
would have a large effect on the amount of material that would have to be
processed through a reduction step or would have to be bled from the system.
Conclusions
Although the addition of KH2P04 to the raw material reduced disproportion-
ation and the addition of moisture to the sweep gas increased S02 evolution,
no combination of conditions tested resulted in elimination of disproportiona-
tion. It appears that increased S02 evolution might be attained by increasing
temperature, time, and maintaining a moist sweep gas; but the production of
sulfate, sulfide, and thiosulfate would have to be accommodated. One method
under consideration for accomplishing this would be to decompose metabisulfite
thermally to the extent that one-third of the sulfur was evolved as S02 and
then reduce the remaining two-thirds to the sulfide by reaction with a reducing
agent. The sulfide then would be converted to hydrogen sulfide by reaction
with water and C02, producing the carbonate for return to the scrubber and
13
-------
supplying raw materials for the Claus reaction to produce elemental sulfur.
Equations for the entire process are shown below.
Scrubber
3K2C03 + 6S02 > 3K2S205 + 3C02
Decompos ition
3K2S205 > 2S02 + 2K2S04 + K2S203
Reduction
2K2S04 + K2S203 + K2C03 + 6C > ^K2S + TC02
E^S production
Claus reaction
kE2S + 2S02
It is presumed that the P205 would remain with solids through the H2S
production step and then be recycled to the scrubber.
The present results indicate that enough S02 could be evolved from
potassium metabisulfite by heating to satisfy the needs of the process
described by these equations and that a viable process of this nature could
be possible.
As a result of these two studies of the thermal decomposition of
potassium metabisulfite, it was concluded that sulfur evolution as S02 and
disproportionation to unwanted forms may be improved by the inclusion of
phosphate (added as KH2P04 in these tests) to the solids being decomposed and
by the addition of moisture to the sweep gas. Rapid removal of liberated
gases from, the sample did not affect disproportionation significantly
indicating that reaction of liberated S02 with the remaining solids probably
is not a cause of disproportionation. The other factors (temperature, time,
bed condition, and sweep gas flow), when applied so as to increase sulf ite
loss from the sample, also increased disproportionation.
REDUCTION OF POTASSIUM SULFATE WITH CARBON
Most of the sulfur in the solids, after the scrubbing and thermal
decomposition steps, is in the suLfate form. Recovery of the potassium and
sulfur values of the material is important for economic reasons. Reduction
with carbon at elevated temperatures to yield potassium sulfide and
subsequent reaction with water and carbon dioxide would yield the desired
products as illustrated by the following equations:
-------
K2S04 + C
K2S + 2H20 > 2KOH + H2S
2KOH + C02
Earlier work (October 1969 and March 1970 Applied Research Branch reports,
TVA) with various reducing agents, including coke, on reduction of K2S04 had
indicated loss of both K20 and S from the samples, the production of some
unwanted sulfur species (K2S203 and/or polysulfides), and incomplete reactions.
Some of the results apparently were due to reoxidation of the charge on
removal from the reduction furnace. Some advances have been made in both
analytical and test procedures since the earlier work.
Test Plan and Equipment
The process now has been restudied using a factorial test design in an
attempt to determine the effects of four variables at each of two levels on
the reduction of reagent-grade K2S04. The variables were temperature, sweep
gas rate, time, and carbon proportion. Prior to this series of tests, some
work was done to assist in selecting the most suitable of a list of reducing
agents and the ranges over which the variables were effective. Reducing
agents tested were bituminous coal, graphite, coconut charcoal, and coke
breeze. Coke breeze was selected as the most effective and was used in the
present tests. Levels of the variables tested are given below. The conditions
were chosen to result in less than complete reduction.
Levels
Variable - +
A - Temperature, °C 815 927
B - Sweep gas flow (N2), ml/min 25 75
C - Time subjected to test temperature, min JO 60
D - Carbon proportion, $a 5^- 91
a. $ of stoichiometric for the reaction:
K2S04 + kC > K2S + bCQ. Carbon proportion
values would be twice as great if based on reaction
to form C02 instead of CO.
The tests were made in an electrically heated tube furnace lined with a
fused silica tube which was restricted on both ends, and through which a flow
of nitrogen sweep gas was passed to remove any gaseous reaction products. A
rod and packing gland arrangement in one end of the tube (gas discharge end)
allowed insertion of a sample into the hot zone of the furnace after purging
it with N2 and withdrawal of the sample from the hot zone to allow cooling in
the N2 atmosphere to below 52°C before exposing the reduced product to air.
The method should minimize reoxidation. The apparatus is shown in Figure k.
-------
-53 cm-
Thermocouple
N2
sweep gas
Reaction
tube
(2.5^ cm i.d.
fused
silica)
Electrically
heated tube
furnace
Porcelain
boat con-
taining
charge
Rod to
position
boat
Figure h. Laboratory apparatus used in tests
of the reduction of potassium sulfate with coke breeze.
-------
Evaluation of Data
Losses and gains of chemical species, on which basis the tests were
evaluated, were calculated from the weights and analyses of the raw materials
and products. Analytical determinations were made of the K20, total sulfur,
sulfite sulfur, sulfide sulfur, and thiosulfate sulfur. Sulfate sulfur was
calculated as the difference between total sulfur and the sum of the other
species. Ranges of the losses and gains are listed below.
Low High Average
Losses, % of input
K20 0.66 2k.2h 9-9^
Total s o 20.07 6.59
Sulfate S 20.59 9^.38 55A?
Gains, % of input S
Sulfite S 5.53 20.92 ll.llj-
Sulfide S 4.TO 5^M 32.69
Thiosulfate S 1.26 15.87 5-05
The aim of the process is to achieve essentially complete reduction of
sulfate sulfur to sulfide form. Sulfite sulfur results from incomplete
reduction. Thiosulfate results from disproportionation or a reaction of
sulfite with elemental sulfur or a polysulfide. Losses of 'K20 and total S
result from volatilization or from spatter. The K20:S mole ratio in the
materials lost from the samples averaged about l.U.
Results of the current tests are given in Table 3 in which are shown the
conditions, product weights, product analyses, and indices calculated from the
product weights and analyses. Of the statistical analysis, the differences in
responses for the individual variables and the two factor interactions are
shown, along with the minimum differences for which there is 90, 95 > and 99%
confidence in the reality of the effect. The differences in which there is at
least 90$ confidence are underlined. The three- and four-factor interactions
were considered as error in these calculations.
The potassium sulfate used in the tests was reagent-grade material; the
coke breeze contained 72.6$ carbon. Responses on which the evaluation was
made were calculated from material balances based on the chemical analysis and
weights of the input materials and the residue.
The loss of K20 is a physical loss from the sample due to volatilization
or to spatter. It responded to all the individual variables, being greater at
the more severe condition, and to all the two-factor interactions except that
of time-carbon proportion. Average loss of K20 amounted to about 10$ of the
input. Conditions causing greatest losses were furnace temperature, carbon
proportion, and the interaction of the two. Conditions that restricted the
loss were all the interactions involving nitrogen flow.
17
-------
TABLE 3. REDUCTION OF POTASSIUM SULFATE WITH COKE BREEZEa (LABORATORY TESTS)
FACTORIAL TEST—DATA AND ANALYSIS
Test conditions
Plan Furnace Ns
posi- Test temp., flow,
tion No. °C ml/min
1 1 815 25
2 4 815 25
3 5 815 25
1* 6 815 25
5 7 815 75
6 8 815 75
7 2 815 75
8 3 815 75
9 16 927 2'j
10 13 927 25
11 9 927 25
12 10A 927 25
13 15 927 75
ll* 11* 927 75
15 12 927 75
16 11 927 75
Average
Raw materials
I Reagent -grade KSS04, 5.
Time,
30
30
60
60
30
30
60
60
30
30
60
60
30
30
60
60
000 g
Carbon
propor-
tion15
0.51*
0.91
0.51*
0.91
0.51*
0.91
0.5l*
0.91
0.5U
0.91
0.51*
0.91
0.5!*
0.91
0.5!*
0.91
Chemical analysis, $
Weight, g
Input
6,031*
6.723
6.031*
6.723
6.031*
6.723
6.031*
6.723
6.031*
6.723
6.031*
6.723
6.031*
6.723
6.03!*
6.723
Residue
5.1*125°
5.8297
5.0210
5 .0630
5.1*085
5.7952
5.0060
5.1102
1*. 1*115
1*.2221
1*.0310
3.9751*
1*. 31*15
l*.2137
3-9598
l*.010l*
by wt.
Responses
Loss, % of input total sulfur
Sulfur
Total
17.0
15-1
18.0
17-5
16.2
15.5
17.8
17.6
19-7
19.0
20.9
18.5
19-9
18.8
21.0
19.6
S03=
0.6
1.0
1.1
1.7
0.3
0.7
1.0
2.0
2.1
4.5
3-0
3-6
2.0
4.3
3.0
4.8
s=
2.1
2.8
4.0
7-0
0.8
3-0
U.o
6.2
7.3
10.5
9.1
12.6
7.6
10.2
9-2
12.2
820.,=
0.8
1.0
0.8
0.6
2.7
0.2
0.5
0.5
0.7
2.2
0.6
1.0
0.1*
2.6
0.7
0.5
S04=
13.5
10.3
12.1
8.2
12.1*
11.6
12.3
8.9
9.6
1.8
8.2
1.3
9.9
1.7
8.1
2.1
KgO
1*9.6
1*1*. 7
52.6
51-1
1*6.9
1*1*. o
51.0
50.3
57.2
52.3
59.0
51.5
57-0
51.9
58.2
52.2
jb of input Loss
KsO Sulfur S04
0.659 0.000 20.59
3-573 4-329 34.71*
2.272 1.776 33-97
4.263 3.706 54.88
6.235 4.776 27.11
5.643 2.377 26.94
5.528 3.158 33.08
4.883 2.252 50.57
6.626 5.549 53-97
18.288 12.816 91.74
11.993 8.1*38 64.08
24.241 20.071 94.38
8.428 6.103 53-29
19.076 13.907 92.21
llt.722 9.627 65.11*
22.537 14.572 90.85
9-935 7.091 55-47
so3-
3.53
6.34
6.00
9-35
1.76
4.4l
5-44
11.11
10.07
20.65
13.14
15.55
9.44
19.69
12.91
20.92
10.64
Gain
S=
12.35
17-74
21.83
38.52
4.70
18.89
21.76
34.43
35-00
48.18
39.87
54.44
35-86
46.71
39-59
53.17
32.69
S£03 =
4.71
6.34
4.37
3-30
15.87
1.26
2.72
2.78
3-36
10.09
2.63
4.32
1.89
11.91
3.01
2.18
5-05-
II Coke breeze, 1.03U or 1.723 ft.
Differences'1 in responses due
Furnace temperature (A)
Nitrogen flow (B)
Time (C)
Carbon proportion (D)
Two-factor interactions
AB
AC
AD
BC
BD
CD
Least significant differences
90$ confidence
95$ confidence
99$ confidence
to
11.607 8.589 40.473
1.892 0.011 -1.145
2.739 1.718 10.79S
5-755 4.325 23.135
-0.988 -0.677 0.475
2.530 1.865 -4.985
4.838 3.587 10.040
-0.667 -1.106 -0.773
-1. 1*1*9 -1.964 -2.648
-0.1*03 -0.075 0.468
0.6357 1.4316 5.6823
0.8112 1.8266 7.2502
1.2721 2.8645 IL 3702
q.304
0.131
2.316
5.716
0.756
-1.649
2.096
1.45!*
0.929
-0.856
1.9208
2.4508
3-81*35
22.B25
-1.603
10.523
12.640
1.063
-5.193
0.1*05
0.175
0.183
1.738
2.3647
3.0172
4.7318
-0.21*5
0.313
-3.765
0.453
-0.665
-0.013
3-950
-1.295
-1.793
-0.490
1*. 6946
5-9900
9.3940
a Coke breeze contained only 72.6$ carbon; l.OJl* g required for 0.5!* stoichiometry, 1.723 g for 0.91 stoichiometry.
* Fraction of carbon required to satisfy KaS04 + 1*C » K2S + 1*CO.
c Calculated on total sulfur balance to prevent indicated sulfur gain.
d Negative sign indicates increase in response at the higher level of the variable; underlined values are significant at 90$ confidence
level.
-------
Loss of sulfur is a physical loss from the system also. This loss
averaged about J% of the input sulfur. Losses were increased by increases in
temperature, carbon proportion, time, and the interaction of temperature with
the other two. The carbon proportion-time interaction produced negative
results.
The loss of both K20 and S were increased by increasing the three factors
most likely to result in the production of the sulfide (temperature, time,
carbon proportion). This could logically indicate the loss as the sulfide
(K2S) which melts at 8kO°C and is known to be volatilized in some reduction
processes as is discussed later. The K20:S mole ratio in materials lost was
1.37 indicating a higher loss of potassium than would be accounted for as
potassium sulfide.
Loss of sulfate sulfur is desired, especially if the sulfur is retained
in sulfide form. Retention in sulfite form is not harmful since the sulfite
would be returned to the scrubber where it would form a useful absorbent for
sulfur dioxide. The desirability of retention of sulfur in thiosulfate form
is questionable; however, it may serve to inhibit oxidation in the scrubber.
Loss of sulfate responded, in a positive manner, to all the individual vari-
ables except nitrogen flow; it responded positively also to the temperature-
carbon proportion interaction.
Gain of sulfite sulfur in the product is the result of incomplete reduction
of sulfate. Sulfite formation increased with all the individual variables
except nitrogen flow and with the two-factor interaction temperature-carbon
proportion. Further increase in the severity of the test conditions except
nitrogen flow should eventually result in a decrease in sulfite content because
of reduction to sulfide.
The increase in sulfide sulfur is the measure of the degree of completion
of the desired reaction. Sulfide sulfur was increased by increasing tempera-
ture, time, and carbon proportion. Sulfide sulfur content was decreased by
the interaction of temperature and time.
Thiosulfate analyses were so scattered that none of the responses were
significant at the 90$ confidence level. There was some indication that the
proportion of sulfur in this form was decreased by increasing time and
increased by the temperature-carbon proportion interaction.
In the visualized process, the potassium sulfide product from the reduction
step would be dissolved in water and then carbonated to produce potassium
carbonate for recirculation to the scrubber and hydrogen sulfide (H2S) for feed
to a Glaus su3fur production unit, along with sulfur dioxide from thermal
decompostion of potassium metabisulfite formed in the sulfur dioxide scrubbing
step. The K20 and S volatilized in the reduction step may be recovered if the
off-gas from the reduction step is used to carbonate the potassium sulfide
solution; the volatilized material would be absorbed and any potassium and
sulfur it contained would thus be returned to the system. Any additional
required recovery might be accomplished in the stack gas scrubber to which the
off-gases from the Glaus unit could be fed.
19
-------
In the event that prevention or reduction of volatilization of K20 and S
becomes desirable, however, techniques might be used that are reported in an
Indian publication (3) which claims the addition of excess reducing agents,
sand, and magnesium oxide (MgO) to schoenite (K2S04•MgS04•6H20) for that
purpose. It may be possible to add sand in the envisioned process during the
reduction step; the sand could be removed following the hydration-carbonation
step. A settling or filtration step would be required and the sand recovered
could be recycled to the reduction step, along with any unutilized solid
reductants. The addition of MgO would present some different problems due to
the solubility of its various salts in the several steps of the process. It
might be advantageous to add other forms of Si02 or to add P205. The determi-
nation of the effects of any of these additives will require continued testing.
Conclusions
It was concluded from the present tests that potassium sulfate can be
reduced effectively by heating it in contact with coke breeze. Some volatili-
zation of both K20 and S occurs; this is undesirable but may be accommodated
since off-gas from the reduction step would be used as a source of C02 for the
H2S production step, and the K20 and S should therefore be recovered along
with the C02. It appears that the optimization of the process as related to
the four variables studied herein will involve increasing temperature, time,
and carbon proportion while decreasing or eliminating the nitrogen flow.
Increasing time, particularly, should increase sulfide formation. It also
appears that reduction of nitrogen flow would help to minimize the imbalance
in K20 and S losses.
Sodium sulfate is reduced by a similar method in the kraft paper process
and in a pollution control process developed by Rockwell International. It
appears that the potassium sulfate reduction step could be developed also.
SCRUBBING WITH POTASSIUM CARBONATE SOLUTION
A factorial study was made of the effects of four variables at each of two
levels on scrubbing efficiency and oxidation during laboratory tests of
scrubbing synthetic stack gas with potassium carbonate solution. The variables
were (l) potassium concentration, (2) oxygen content of the gas, (3) P205
presence in the scrubbing solution, and (U) antioxidant presence in the
scrubbing solution. The amount of P205 added was equivalent to about 1.0% of
the total weight of potassium metabisulfate that could be produced from all
the potassium present when the solution contained 1$ potassium carbonate or
5$> when 2% potassium carbonate solution was used.
Test Plan and Equipment
The tests were made in specially designed gas-absorption bottles each of
which consisted of a cylindrical section of 32-millimeter glass tubing, flat-
bottomed and 13 centimeters deep, and containing a 5-millimeter glass thermo-
couple shield fused in the side so as to read the temperature about k centi-
meters from the bottom; an enlarged glass section, 75 millimeters in diameter
and 75 millimeters high was fused to the top of the previously described
20
-------
section to act as a foam breaker. The top of the foam "breaker was reduced and
fused to a 29/^2 ground glass joint to allow closure, entry of the inlet gas
sparger, and directed flow of the exit gas. One scrubber was used per test;
it contained 100 milliliters of the scrubbing solution which filled the 32-
millimeter cylindrical section of the unit. A small magnetic stir bar in the
bottom of the unit served to ensure agitation.
A mixture of synthetic stack gas (3*4-00 ppm S02, 19. ^ C02, balance N2)
and air (either 2 or 5$ 02 in mixture) was sparged into potassium carbonate
solution at approximately 1.5 liters per minute through medium porosity fritted
glass plug spargers (12-mm-dia and 2h-mm-long) mounted so as to clear the stir
bars by about 5 millimeters. The actual gas flow was ikko milliliters per
minute of synthetic stack gas with enough air added to obtain the desired •
oxygen content. A water bath served to maintain the scrubber temperature at
52°C. Off-gases from the scrubber were passed through a water-cooled condenser
and a UV spectrophotometer (set at 287 nm) to monitor their sulfur dioxide
contents. The tests were terminated when the sulfur dioxide content of the
off-gas reached 1500 ppm. A factorial test plan (four variables at two levels
each) was designed for this study; it is outlined below.
Level
Variable - +
A - K2C03 solution concentration, % 1 2
B - 02:S02 mole ratio 6.6? 16.67
C - P205 content, % (Ki.5Hi.5P04) 0 0.183
D - Antioxidant (p_-phenylenedi ami ne)
content, % 0 0.1
The plan requires l6 tests; an additional four tests (two additional
replicates of the tests with all variables at the negative level and two with
all variables at the positive level) were made as a check of the reproduci-
bility of the tests. Problems with reproducibility had hampered completion
of an earlier test series of this nature.
Problems in Earlier Tests
During initiation of the current test series, special effort was placed
on discovering the reasons for lack of reproducibility in the earlier tests.
That earlier experience had resulted in wide variations in the scrubbing time,
the degree of oxidation, and the final solution pH in duplicate tests.
Special care was taken with solution makeup, equipment cleaning, and test
procedure; also, weaker scrubbing solutions were prepared to shorten test
times and eliminate the need for excessive makeup of the synthetic stack gas.
In addition, a gas chromatograph, not available during the earlier tests, was
used to monitor the oxygen content of the scrubber inlet and outlet gases.
It was the gas chromatograph that allowed isolation of at least a major cause
of the inconsistency of the earlier test results. It was found that back
pressure in the inlet gas system increased as the tests proceeded, due to
partial plugging of the fritted glass spargers with products of the reaction.
The meters used to measure the flows of synthetic stack gas and of air were
fitted with ground glass joints to allow easy removal for cleaning or replace-
21
-------
merit, and when the back pressure in the system became sufficient, the ground
glass joint on one or the other of the meters would be forced open (even though
they were mounted under some spring pressure); this allowed part of one of the
gas components to escape. The small flows and very small movement required to
allow the leaks had made detection of such leaks difficult. Rapid oxygen
analysis, made possible by the gas chromatograph, allowed determination of
sudden changes in oxygen content and helped to pinpoint the leaks. The
chromatograph was used throughout the current series of tests, special attention
was directed to the meter joints, and small amounts of water were added to the
inlet gas as required to flush the fritted plug of reaction products; these
precautions improved reproducibility. A table to show reproducibility will be
presented later in the report.
Evaluation of Data
Responses measured during the current tests were (l) scrubbing time before
1500 ppm breakthrough, (2) pH of the final scrubbing solution, and (3) chemical
analysis of the final scrubbing solution. From the chemical analyses, the
total and sulfite sulfur contents, the degree of oxidation, and the degree of
reaction /S:K and (S + P):K mole ratios/ were calculated. Sulfite analyses
were made immediately after each test was completed to prevent errors due to
continued oxidation; other analyses were made later.
On receipt of the chemical analyses for total sulfur, P205, and K20, it
was found that there was a very wide variation between the total sulfur values
reported and those expected, especially in samples from tests containing anti-
oxidant. In the samples from tests with antioxidant, the total sulfur often
was less than half the value for sulfite sulfur determined earlier; total
sulfur values for samples without antioxidant appeared fairly reasonable.
Total sulfur was determined by oxidation and subsequent precipitation of
sulfur with barium, and no interference due to the antioxidant was previously
known. Values for K20 and P205 were as expected.
Since chemical analyses for total sulfur were not reliable (sulfite sulfur
had been analyzed immediately on conclusion of the tests and was considered
reliable), some other means of deriving the total sulfur value was needed. The
total sulfur in the samples should be the total amount fed in the inlet gas
less the amount lost in the exit gas. Oxidation in the inlet gas should be
essentially zero. Thus the total sulfur could be calculated using the
proportional areas of the curves from the recorder of the UV analyzer, which
was used as a continuous monitor of the exit gas sulfur dioxide content. The
curves were integrated and the total sulfur values determined for each test;
these are the values used in this report.
Reproducibility and Accuracy—
Responses from the replicate tests to determine reproducibility are given
in Table k. Values of time required to reach the point of 1500 ppm sulfur
dioxide in the exit gas, of final solution pH, of total and sulfite sulfur
contents, or proportion of sulfur oxidized, and of the sulfur to potassium
mole ratio in the final solution were obtained. Reproducibility was then
determined by calculating the standard deviation (n-1 degrees of freedom), the
22
-------
TABLE 4. REMOVAL OF SULFUR DIOXIDE FROM SYNTHETIC STACK GAS BY
POTASSIUM SCRUBBING—REPRODUCIBILITY TESTS
Response
Test Variable
position level Item
1 All (-) Replicate
Replicate
Replicate
Standard
Mean
Error of
Standard
2 AH (+) Replicate
Replicate
Replicate
Standard
Mean
Error of
Standard
1
2
3
deviation8"
mean
°jo error13
1
2
3
deviation
mean
% error13
Time,
to
min,
1500 "DTDm
50.
50.
47.
1.
49.
1.
3.
153.
163.
153.
5.
156.
3.
3-
75
5
5
809
583
044
64
5
25
75
558
833
209
54
Final
PH
3.6
3.6
3.4
0.115
3.533
0.066
3.26
4.8
4.6
4.6
0.115
4.666
0.066
2.47
Sulfur,
Total
0.3226
0.3092
0.2879
0.0X75
0.3066
0.0101
5.71
1.0563
1.1328
1.0702
O.o4o8
1.0864
0.0235
3-75
.g
Sulfite
0
0
0
0
0
0
18
0
0
0
0
0
0
2
.1068
.1131
.0783
.0185
.0994
.0107
.65
.9235
.8890
.8947
.0184
.9024
.0106
.03
%
S
oxidation
66.
63-
72.
4.
67-
2.
7.
12.
21.
16.
4.
16.
2.
2.
9
4
8
75
70
743
01
6
5
4
47
83
578
65
S:K
mole ratio
0
0
0
0
0
0
5
0
1
1
0
1
0
3
.6688
.6525
.6054
.0329
.6422
.0190
.13
.9825
.0519
.0195
.0347
.0180
.020
.41
a Based on n-1 degrees of freedom.
-------
mean, and the error of the mean. Then, to describe the variability as a single
value, the standard percent error expressed as standard deviation, percent of
the mean was calculated.
Reproducibility in the tests as shown in Table 4 was fairly good,
particularly at the (+) variable levels. In tests where all variables were
held at the (-) level, variability (standard % error) ranged from 1.9 to 18.7$.
Greatest variability occurred in the sulfite sulfur content; the sulfite sulfur
content was very small and a variation of only 0.03*4-8 gram of sulfur accounted
for the large value. The proportion of total sulfur oxidized varied 7%,
largely due to the variation in sulfite sulfur values. The remaining
variability values were less than 6£.
In tests where all variables were held at the (+) level, variability
ranged only from 2.0 to 4.4£. Values for sulfite sulfur in these tests
averaged about nine times as great as in the tests with all variables held at
the (-) level.
Accuracy of solution makeup for the tests was indicated by chemical
analyses of the spent liquor for K20 and P205 content. These results and theii
variability are given below. Maximum variability was 0.48$; that occurred in
the P205 content and was the result of a single measurement. Apparently there
was no problem in solution preparation.
KP0 content of spent liquor,
1% KpCOa 2% KPC03
Item
Alone
Alone + F305
P205 content
of spent
liquor,
Replicates
1
2
3
4
5
6
1
8
Standard
Mean
Error of
Standard
0
0
0
0
deviation8" 0
0
mean 0
% error*5 0
•7025
.7000
.6975
.6975
-
-
_
.0024
.6994
.0012
•34
0
0
0
0
0
0
0
0
.8825
.8800
.877
.8825
_
-
-
.0024
.8806
.0012
• 27
1
1
1
1
0
1
0
0
.4000
.3925
.4050
.3950
-
-
-
—
.0055
.3981
.0028
.40
1
1
1
1
0
1
0
0
.5825
.5800
.5825
• 5700
-
-
-
_
.0060
.5788
.0030
.38
0
0
0
0
0
0
0
0
0
0
0
0
.1825
.1825
.1825
.1825
.1850
.1825
.1825
.1825
.0009
.1828
.0003
.48
a. Based on n-1 degrees of freedom.
b. Standard deviation,
% of
mean.
Analysis of Factorial Design—
The mean values for the replicated tests were used in the statistical
analysis of the data from the test series. All three- and four-factor
interactions (a total of five values) were considered as error in these
calculations. From these error values the least significant differences, at
the 90, 95» and 99$ confidence level, were calculated for the responses due to
the levels of each variable and the two-factor interactions of the variables.
24
-------
The test design, test results, and statistical analysis are given in
Table 5. Although the degree of oxidation of the sulfur and degree of reaction
of the potassium with sulfur, particularly in sulfite form, are the most
important responses in relation to later process steps, other responses were
measured also; these included the time required to neutralize the scrubber
solution to the point that the off-gas contained 1500 ppm sulfur dioxide, the
pH of the final solution, the potassium (K20) content of the final solution,
and the phosphate (P205) content of the final solution. Some of these
measurements (K20 and P205 content) served only to check the accuracy of the
test conditions.
The time of the tests was measured and was found to be most significantly
affected by two individual factors and their interaction. These factors were
the amount of potassium carbonate in the solution and the presence of 0.1$>
antioxidant. Increasing the potassium carbonate proportion from 1 to 2$ of
the solution increased the test time by 58 minutes (67-125 min). A similar
increase in time (53 min) resulted from the addition of the antioxidant.
Increasing the oxygen to sulfur dioxide ratio reduced test time about 3.8
minutes and the presence of P205 increased test time about 3.4 minutes. The
interaction of potassium carbonate content and P205 presence reduced test time
about 3-1 minutes at the higher levels. In these tests, increased time
generally results in increased total sulfur removal and increased K20 reaction
within each potassium carbonate concentration level. The improved performance
with the addition of the antioxidant is the most significant of these results.
The final pH of the solution was lower than the input and was a function
both of increased sulfur dioxide absorption and of oxidation. Values for pH
ranged from 4.8 to 3-2. The statistical analysis showed the presence of the
antioxidant to be the most effective variable; it resulted in the final pH
value being about 1.1 pH points above that of samples without the antioxidant.
Higher potassium carbonate content resulted in slightly higher final pH.
Increased oxygen content decreased pH, probably because of increased oxidation.
A linear regression was made of the values of pH and of percent oxidation;
the equation derived (0.938 correlation coefficient) was
pH = 4.625 0.0148 (% oxidation)
This calculation shows the variation of pH to be almost totally dependent on
the degree of oxidation of the sulfur ions. The sulfate ion is much more
acidic than the sulfite ion.
The total sulfur content of the final scrubber solution responded
significantly and positively to all the individual variables. It was
significant to all two-factor interactions of the potassium carbonate content,
also. The interactions with oxygen proportion and with antioxidant presence
were positive; the interaction with P205 content was negative. Greatest
variations (0-35 and 0.39 g) were due to the presence of antioxidant and to
potassium carbonate content, respectively. The improvement due to the
presence of the antioxidant is the most important of these effects, especially
since the doubling in total sulfur content due to potassium carbonate content
(0.44 g at 1% and 0.83 g at 2$ K2C03) would be expected.
25
-------
ro
TABLE 5. REMOVAL OF SULFUR DIOXIDE FROM SYNTHETIC STACK GAS BY POTASSIUM SCRUBBING (LABORATORY TESTS)
FACTORIAL TEST—DATA AND ANALYSIS
Test conditions
Plan KSC03
posi- Test content,
tlon No. *
1 2A,10,10Ab 1
291
3 3A 1
It 1A 1
5 11 1
6 15 1
7 20 1
8 16 1
9 6A 2
10 7A 2
11 1*B 2
12 5A 2
13 18 2
Ik 17 2
15 1} 2
16 12 ,ll» ,19b 2
Average
Differences0 in responses due
KSC03 content (A)
Oe:SOt mole ratio (B)
PgOs content (C)
Antioxidant content (D)
Two-factor Interactions
AB
AC
AD
BC
BD
CD
Least significant differences
90$ confidence
95$ confidence
99$ confidence
" Values calculated from fe
Anti-
OjtSOj P?0; oxidant
mole
ratio
6.67
6.67
6.67
6.67
16.67
16.67
16.67
16.67
6.67
6.67
6.67
6.67
16.67
16.67
16.67
16.67
to
content , content ,
1
0
0
0.183
0.183
0
0
0.183
0.183
0
0
0.183
0.183
0
0
0.183
0.183
s, time, and
Time,
min before
1500 ppm
pH of
final
$ breakthrough solution
0
0.1
0
0.1
0
0.1
0
0.1
0
0.1
0
0.1
0
0.1
0
0.1
1*9.58
83.6
53-5
92.5
1*3-0
80.3
>*9.5
87.25
91.75
160.2
90.5
165.1
91.0
158.5
90.3
156.8}
96.1*631
58.1187
-3 . 7563
*} .]^4}7
53.11*37
1.O263
-3.1237
\f\ . 1 gh}
-0.6737
-0.8737
1-3263
1.9795
2.5258
3.9611
integration of exit
3.53
It .2
3.65
"*-35
3.2
3.2
l*.l*5
3.8
U.6
3-1*
It .8
3-2
It .6
3-<*
U .66
3.9588
0.1975
-0.1650
0.0600
1.0725
-0.0200
-0.01*50
0.11*25
O.OU25
0.1800
O.OHOO
0. 151U
0.1933
0.3032
gas analysis
Responses
Chemical analysis of final solution. R
KaP
0.7025
0.7000
0.8825
0.8800
0.6975
0.6975
0.8775
0.8825
1.1*000
1.3925
1.5825
1.5800
1.1*050
1.3950
1.5825
1.5700
curves.
PsOc
_
0.1825
0.1825
.
.
0.1825
0.1825
_
_
0.1850
0.1825
.
-
0.1825
0.1825
Total"
0.3066
0.5153
0.3302
0.5679
0.2922
0.55U9
0.3378
0.60}2
0.5850
1.01U8
0.5581
1.0339
0.6337
1.0960
0.6301*
1.0861*
0.631*2
0.3913
0.01*03
0 . 0187
0.31*98
0.0233
-0.0239
0,1061
0.0015
0.0118
0.0089
0.011*2
O.OlBl
0.028U
Sulfur
Sulfite
0.099**
0.1*691
0.19U1
0.1*955
0.0375
0.1*812
0-0395
0.5023
0.1389
0.9291
0.0581
0-91*73
0.011*5
0.8825
0.0902
0.9021*
0.3926
0.2056
-0.01*76
0.0222
0.6172
0.0018
-0.0139
0 . 2228
0.0076
0.0296
-0.0008
0.01*91*
0.0630
O.C988
Sulfate
0.3072
0.01*62
0.1}6l
0.0721*
0.251*7
0.0737
0.298}
0.1000
0.1*1*61
0.0857
0.5000
0.0866
0.6192
0.2135
0.51*02
0.181*0
0. 21*15
0.1858
0.0860
-0.0036
-0.267I*
0.0218
-0.0038
-o.nfili
-0.0050
-0.0178
0.0096
0.1*50
0.0571*
0.0900
Oxidation ,
$ of
Qulfur in
sulfate form
67.6
9.0
1*1.2
12.7
87.2
13-3
88.}
16.7
76.}
8.1*
89.6
8.1*
97-7
19-5
85-7
16.9
1*6.156}
8.3125
Il*.pl25
-2.1*375
-66.0875
-1*.7375
2.1125
-7.9375
-0.0875
-7.0375
3.5625
11.592
lit. 790
23.195
Hole
ratios
Total E:K (Total S + P):K
0.61»2U
1.0835
0.5507
0.91*99
0.6166
1.1710
0.5666
1.0060
0.6150
1.0726
0.5191
0.9631
0.6639
1.156U
0.5863
1.0185
o. 8239
0.0011
0.01*67
-0.1077
0.1*575
0.0152
0.0025
0.0009
0.0001
0.0221
-0.02g>
0.0226
0.0288
0.01*52
0.6U2U
1.0835
0.6879
1.0877
0.6166
1.1710
0.701*6
l.llt}2
0.6150
1.0726
0.5966
1.0398
0.6639
1.1561*
0.6628
1.0956
0.8775
-0.029J
O.oSeo
-O.OOOl*
0.1*575
0.0151
-0.0279
-0.0010
0.0000
0.0221
-0.0289
0.0226
0.0288
0.01*52
Values are averages of three tests.
c Negative sign indicates increase in response at the higher level of the variable; underlined values are significant at 90$ confidence level.
-------
Sulfite sulfur content responded in a positive direction to potassium
carbonate content, antioxidant presence, and to the interaction of these two.
By far the greatest improvement was due to the presence of the antioxidant
(0.70 vs. 0.08 g). Doubling the potassium carbonate content did not double
the sulfite sulfur content (0.29 vs. 0.50 g) probably due to the greater test
length and therefore greater opportunity for oxidation in tests with larger
amounts of potassium carbonate.
Sulfate sulfur was calculated as the difference between the total and
sulfite sulfur values. The amount of sulfate sulfur present varied with
changes in potassium carbonate content, antioxidant presence, and the inter-
action of the two. Oxygen proportion also was an effective variable. Increased
potassium carbonate and oxygen content increased the amount of sulfur in
sulfate form while the presence of the antioxidant and its interaction with
the potassium carbonate content decreased the amount of sulfate sulfur.
The percent sulfur dioxide oxidation is one of the most important of the
responses measured. Thermal stripping of sulfur dioxide from fluid scrubber
discharge, precipitation of potassium metabisulfite (K2S205) for later thermal
decomposition, or reaction of the scrubber discharge with limestone in a
double-alkali process would all be hampered by increased oxidation. The
statistical analysis showed that the proportion of sulfur oxidized to sulfate
form responded, primarily, to one variable, the presence of the antioxidant.
There was a 66 percentage point reduction in oxidation (79% to 13$) due to the
presence of the antioxidant. Increasing the proportion of oxygen in the gas
increased oxidation by Ik%.
The total sulfur to potassium mole ratio was calculated to determine the
degree of reaction of sulfur dioxide and potassium carbonate. The proportion
of oxygen, the presence of P20s, and the presence of the antioxidant were
significant. Oxygen and antioxidant responses were positive; P205 presence
gave a negative response, probably due to the increase in K20 content when
P205 was present (added as Kj..5^.5P04). The interaction of P205 and anti-
oxidant presence was significant and was negative. The most effective variable
was the presence of antioxidant. The reason for increased oxygen content to
cause an increase in the sulfur to potassium ratio is not immediately apparent.
The total sulfur-plus-phosphorus to potassium mole ratio was calculated
to remove, to some extent, the effect of the increased potassium content when
P205 was present. Results were significant and positive for variations in
oxygen proportion and antioxidant presence. Potassium carbonate concentration
and the P205 antioxidant interaction were significant and negative. This ratio
almost entirely eliminated the effect of the P205 presence and probably means
that, in effect, the potassium phosphate in the product is in the form of
dissolved monopotassium phosphate (KH2P04). In all probability, the addition
°f PgOs as a 1:1 molar mixture of monopotassium and dipotassium phosphate
(Ki.5Hi.5P04) increased the scrubbing potential slightly. This was shown in
the increased total sulfur content of 0.0187 gram because of the presence of
P205; it is noted that the increase in sulfite sulfur was not significant
because of this variable.
-------
Conclusions
It is concluded from this work that the inclusion of an antioxidant in
potassium carbonate scrubbing solution results in significant improvement in
the process by reducing oxidation and increasing total sulfur and sulfite
sulfur absorption. The antioxidant probably would be destroyed in the
reduction step, but the concentration reaching this step should be extremely
low because of very low concentration in the scrubber solution. The addition
of P205 as potassium phosphate increased total sulfur content by providing
some additional scrubbing cation but otherwise did not affect the scrubbing
process significantly. Increasing the oxygen content of the input gas from
an oxygen to sulfur dioxide mole ratio of 6.67 to 16.67 (02 content variation
from 2-5$) affected the proportion of sulfur oxidized only slightly.
28
-------
OXIDATION OF SULFITES
PRIOR WORK
A few laboratory-scale tests had been made earlier (August 1970, August
1972, and October 1972 Applied Research Branch Reports, TVA) using a spinning
cup patterned after a device developed and patented in Japan (k). Some of
these tests had shown that weak solutions (about 1$) could be oxidized almost
completely with about ko times stoichiometric air; stronger solutions were not
tested. Oxidation of slurries of suJLfites having lower solubility (ZnS03)
proceeded somewhat more slowly. Oxidation of slurries of calcium sulfite
proceeded very slowly unless the pH of the slurry was lowered to increase the
solubility of the sulfite. Oxidation of calcium sul±ite was showi to be
catalyzed by the presence of iron compounds. All the above work was done with
the spinning cup model operating in a glass vessel as described in the equip-
ment section of this report.
EQUIPMENT AND PROCEDURE
The equipment used in the work included two rotary devices to disperse
air into a pool of liquid, a variable speed motor with torque measuring
controller to drive these devices, and a series of five tanks in which the
oxidation reaction occurred. Aside from the various support stands, thermom-
eter clamps, and so forth required, various scrubbers were used to measure the
amounts of ammonia and S02 in the off-gases or to dry the off-gas prior to
injection of samples into a gas chromatograph set up to measure the 02:N2
ratio and thus determine oxygen utilization.
Figures 5 through 7 show the various oxidizers and oxidizing vessels.
The spinning cup, as noted earlier, is a model of a similar device patented in
Japan; this model was constructed from a description of the unit furnished by
A. V. Slack (formerly Chief, Applied Research Branch, Division of Chemical
Development, TVA; currently president of SAS, a consulting firm) on his return
from a visit to sulfur dioxide control facilities in Asia in 1970. The
centrifugal gas-liquor contactor was designed and constructed at TVA and was
later found to be similar to but more simply constructed and operated than a
mixer patented by Max L. Bard (5).
In operation, the fluid to be oxidized was placed in the oxidizing tank
and the oxidizer unit and air inlet tube were immersed to the predetermined
depths and the unit was rotated slowly as the fluid temperature was brought to
the desired level (52°C). Once the temperature was attained, the unit rpm was
adjusted to that desired for the test, and the proper amount of air (1|-120 ml/
as measured by rotameter) was directed into the unit through the air inlet
29
-------
(spinning cup)
c
3.8 cm
UNIT "CGLC"
(centrifugal gas-liquor
contactor)
mm opening
Figure 5. Oxidizer units used in tests of oxidation of sulfites with air.
30
-------
A
(glass)
23 cm
\j
\r
A
T
8.9 cm
v
(stainless steel)
11
.k cm H
r
8.9 cm
\
12.7 cm
T\
23 cm
12.7 cm
Figure 6. Oxidation vessels used in tests of oxidation of sulfites with air.
-------
Thermocouple
shield
A
L
l!
Off-gas sample line
122 cm
HI
J
>0<
U,!^
<** w\ •
Air inlet
Loose fitting cover
Mixing
baffles
Section A-A
CGLC
Sample line
Figure 7. Tall oxidizer used for study of pool depth.
-------
tube. A continuous measured sample of the off-gas from the oxidizer was
pumped through a series of scrubbers and/or dryers to measure the components
of the gas or to condition it to allow analysis by gas chromatography of the
oxygen content. The torque required to turn the oxidizer units was measured
and calculations were made to determine the power requirements for the test.
Samples of the fluid being oxidized were withdrawn at timed intervals and
analyzed immediately by iodine titration to determine the amount of residual
sulfites. The gas chromatograph gave very early indication of the completion
of the oxidation process. The pH of the fluid was measured on an intermittent
basis.
Calculations then were made of oxidation rates, air utilization, time
requirements, power requirements, and power consumption per unit of sulfite
oxidized.
SOLUBLE SALT OXIDATION TESTS
The first factor to be studied was the comparison of the two oxidizer
units. Comparative tests were planned of the oxidation of ammonium sulfite
solutions containing about 60 grams of sulfite sulfur per liter. Ammonium
sulfite oxidizes fairly rapidly when stored dry in reagent bottles and is not
generally available from chemical supply houses. The ammonium sulfite solu-
tions used in this series of tests, therefore, were made in the laboratory by
adding gaseous ammonia to the fairly stable and available ammonium bisulfite
solution (U5$ NI^HSOs) to a pH of 7-5 and diluting to the desired concentra-
tion. At this pH the solution is essentially ammonium sulfite. Results of
these tests are given in Table 6. In the first of these tests (No. 8), the
spinning cup (SC) was used to oxidize 1 liter of solution containing 65.2
grams of sulfite sulfur. The air rate, as in all tests, was 4120 milliliters
per minute and the cup was turned at JOOO rpm or at a peripheral speed of 8
meters per second (26.2 ft/sec). The torque required to turn the cup at this
rate resulted in a 29 mV reading and amounted to 9^-92 g-cm. Using the 5-08-
centimeter (2-inch) cup, rotated at 3000 rpm, the calculated energy require-
ment for driving the cup was 29.2k- watt-hours per hour of operation. The
solution was sampled at intervals and when analysis indicated that oxidation
was nearly complete, the test was stopped. More than 12 hours (728 min) was
required to accomplish 97•5% oxidation.
Some oxidation apparently occurred during the pretest warmup period in
this and the remainder of the tests. In order to eliminate such unintentional
oxidation from calculations, oxidation was measured from the first timed
sample until the end of the test or until depletion of the sulfite species
caused a noticeable change in the oxidation rate. Oxidation appeared to
proceed at a nearly constant rate over the periods for which the calculations
were made.
In test 8, the period of essentially constant oxidation extended from
108.5 to 728 minutes. During this period, the energy requirement was l8l
watt-hours per mole of sulfur oxidized (l8l Wh/mole); oxygen utilization
amounted to 3.5$ of that fed as air into the unit, and the oxidation rate was
0.0027 mole sulfur per minute.
33
-------
TABLE 6. OXIDATION OF SULFITES—LABORATORY TEST CONDITIONS AND RESULTS
Test Ho.
Salt oxidized
Cone . , g S0a- S/l
Volume, 1
Oxidizer
Rotary unit
Peripheral speed, m/sec
Torque, indication, mV
Vessel
Samples
1
2
5
5
6
7
8
9
10
Constant
Energy
Oxygen
8
NH^SOa
1
sc
8
29
0-1
Time, min pH SOf-S,g/l
oxidation period, nin
used, Wh/nf&e oxidized
utilization, 4
108.5
180
250.6
360
513
590
665
728
Oxidation rate, mole/join
7.25
7.05
7.00
6.65
6.25
5-75
3-95
3-90
108.5-728
181
3-5
0.0027
55-5
1*9.8
1*1
36.9
21.6
13.8
6.6
2.2
9
61.7
1
CGLC
5-6
G-2
I
Time, min pH SO->-S, K/l
60
120
180
2>*0
300
360
1*20
1*80
51*0
600
7-35
7.15
7.00
6.70
6.1*2
6.70
5^5
3-95
1*.10
U.20
60-1*80
121*
1*.0
0.0031*
1*1*. B
39.1*
3^.5
32.2
23-7
15.8
9-3
3-2
2.1*5
1.1
11A
NH^S03
1
CGLC
8
95
SS-1
I
Time, min pH SO^-S, g/1
30
60
90
120
150
180
210
7.T
7-1
6.6
3-1
3-2
3-6
-
30-120
100
13.8
0.0106
35- *
26.7
1U.8
U.9
1-3
1-7
-
12
(mOsSCv
55-0
i
CGLC
5.6
50
SS-1
i
Time, min pH SO-.-S, g/1
30
60
90
120
150
180
210
7.6
7.1*
7.2
6.8
5-1
3-2
3-7
50-210
60
8.1*
0.0065
1*2.8
27.8
20.9
11*. 8
8.5
5-9
5-5
(continued)
-------
TABLE 6 (continued)
Test No.
Salt oxidized
Cone., g SOa-S/l
Volume, 1
Oxidizer
Rotary unit
Peripheral speed, in/sec
Torque, indication, mV
VJ.J
VJl
Vessel
Samples
1
2
3
1*
5
6
7
8
9
10
Constant
Energy
Oxygen
oxidation period, min
used, Wh/mole oxidized
utilization, 4
Time,
30
60
90
120
150
Oxidation rate, moie/min
13
(HHjaSO;
63.95
1.5
CGLC
8
95
SS-l
min pH £
7-3
6.7
6.65
5-95
3.8
30-150
73
18.8
o.oU*5
i
50p-S, g/1
U6.0
30.3
19.8
13. U
8.9
Time,
30
60
90
120
150
180
210
21*0
1*0.1*
1.5
CGLC
8
95
SS-l
min pH
5755
6.1*5
3-95
3-6
3.1*
3-8
3-8
3-8
30-120
152
9.1
0.0070
3i3
SOp-S, g/1
18.3
9.8
6.1*
l*.9
>*-5
1-5
1.1*5
0.96
Time,
60
120
180
210
270
360
1*20
1*80
15
(NHjaSi
60.9
1
SC
8
29
SS-l
min pH
60-U80
210
3-0
0.0023
Qs
SO,-S. g/1
55-9
1*7.2
1*6.8
1*7.2
31*. 0
33-6
29-5
25-1*
Time,
30
60
90
120
150
210
21*0
16
(HH»)2S
63.0
1
CGLC
8
95
G-3
min pH
775"
7-5
7-35
7.15
6.95
6.2
3.65
30-210
131
10.5
0.0081
Oa
50,,-S, g/1
56.1
1*8.1*
37-1
30.8
2U. 5
9A
5-7
(continued)
-------
TABLE 6 (continued)
Teat No.
Salt oxidized
Cone., g S02-S/1
Volume, 1
Cocidizer
Rotary unit
Peripheral speed, m/sec
Torque, indication, mV
Vessel
Samples
1
Z
3
k
5
6
7
8
9
10
Constant oxidation period, mln
Energy used, Wh/mole oxidized
Oxygen utilization, %
Oxidation rate, mole/min
* pH 8.5.
D pH 10.8.
17
i
CGLC
8.8
105
G-3
Time, mln pH S02-Sj g/l
30 7-7 53 •&
60 7.8 ^5-3
90 7-6 36.6
120 7.^5 27-6
30-120
jj^o
12.0
0.0091
'
Time, nrin
30
60
90
120
180
18
56.7
1
CGLC
8
95
SS-1
PH SOp-S, g/1 Time, ml
fTo" 47.0 30
32.9 60
7.0 20.6 90
6.2 6.9 120
3-7 1-7
3O-12O
76
18.1
0.0139
a
1.075
CGLC
8
95
SS-1
Jl pH
8.4
8.15
6.55
30-120
67
20.5
0.0158
50,,-S, g/l
£3.5
35.9
19.7
1.6
Time, mln
5°
60
90
120
150
180
20
1.0
CGLC
8
95
SS-1
i^r
10.1
9.9
9.8
9.6
9-55
30-180
91
15.1
0.0117
SO*-S, g/l
58.0
1*7-5
38.9
28.3
13-1
2.0
(continued)
-------
TABLE 6 (continued)
Test No.
Salt oxidized
Cone., g SOg-S/1
Volume, 1
Oxidizer
Rotary unit
Peripheral speed, in/sec
Torque, indication, mV
Vessel
Samples
1
2
3
1*
5
6
7
8
9
10
Constant oxidation period, min
Energy used, Wh/mole oxidized
Oxygen utilization, %
Oxidation rate, mole/min
Time,
10
20
30
1*0
50
60
70
80
21*
19.6
2C
CGLC
7
100
SS-2
min pH £
10. S6
..
_
_
_
-
.
9-65
1O-70
53
21*. 1
0.0186
50,,-S, g/1
10.2
1U.8
11.8
9.0
5-5
3-5
0.3
0.05
25
i!^
CGLC
7
100
SS-2
Time, m-ln pH
20 10. B»
1*0
60
80
100 9-95
20-100
37
31*. 5
0.0267
30,,-S, g/1
17.3
13.1*
9-1
l*.6
0.23
Time,
20
1*0
60
80
100
120
ll*0
26
KeS03
21.1*9
CGLC
7
100
SS-2
min pH
10. 8S
_
_
_
„
9.8
20-120
31
1*1.3
0.0319
SO-.-S, g/1
18.1*
ll*.9
11.5
8.1*
5.3
l.l*
o.i*
Time,
60
120
180
220
2l*0
38
o|.o
CGLC
7
100
SS-2
iirln pH
10.1*«
_
_
.
10.3
60-220
27
1*7.0
0.0362
SOa-S, g/1
15.9
ll.lfl*
6.17
2.63
1.3
*: Pool depth 15.25 cm (6 in).
Pool depth 31.75 cm (12-5 in).
5 Pool depth 1*8.25 cm (19 in).
Pool depth 111.25 cm (1*3.8 in).
8 pH of solution prior to start Df test.
-------
During test 8, off-gases from the oxidizer were passed through scrubbers
to determine the amounts of ammonia and sulfur dioxide being lost and the
portions of the test over which the losses occurred. Ammonia loss was detected
as soon as the feed of air was begun and continued at a decreasing rate until
a pH of about 6 was reached (about 9 hr). The total ammonia loss amounted to
about 0.02 mole or 0.5$ of the input amount. Sulfur dioxide loss "began at a
detectable rate after the pH reached 6.65 (6 hr) and continued at an increasing
rate until the test ended (pH 3.9, 12 hr). The total loss of S02 detected
amounted to 3.9$ of that present at the beginning of the test. The losses
were fairly low (0.0^ g NH3 and 0.83 g S02/hr) and were not measured in later
tests where time was much shorter. There was, however, no time during the
test when one of the species was not being volatilized and between the pH's
of 6.65 and 6; some of both were detected in the off-gases.
A test was attempted in the same vessel using the centrifugal gas-liquor
contactor (CGLC) but the clearance between the rotary unit and the baffles was
so small that the unit bumped the baffles when rotated, even at 5«6 meters per
second (llt-00 rpm), and the test was stopped. A larger glass vessel (G-2) was
built in which the clearance was the same as in the smaller unit when the
spinning cup was in use. Test 9 was made in the larger glass unit using the
CGLC with a rotational speed of ll<-00 rpm or 5.6 meters per second peripheral
speed and an original sulfite concentration about the same as in test 8.
Oxidation was essentially constant over the period 60 to k&0 minutes after
which about 95$ of the sulfur had been oxidized. The energy requirement over
the period was 12k watt-hours per mole oxidized, oxygen utilization was h% of
that fed, and the oxidation rate was 0.003^ mole per minute. There was an
improvement in air and energy utilization using the CGLC although the
peripheral speed was not the same.
A stainless steel tank with sharp-edged, full height baffles was fabri-
cated (SS-l) in which the CGLC could be operated at high speeds. Test HA
was made in this vessel using the CGLC at a peripheral speed of 8 meters per
second (2000 rpm) and 1 liter of ammonium sulfite solution containing 56.9
grams sulfite sulfur per liter. Oxidation occurred much more rapidly than in
tests 8 and 9. Even though a large part of the sulfite was oxidized in the
warmup period, oxidation during the period of constant oxidation occurred at
the rate of 0.0106 mole per minute, an increase of about 300$ over test 9.
Energy consumption was reduced to 100 watt-hours per mole and oxygen
utilization increased to 13.8$.
Test 15 was made to check the results of test 8 to determine if the very
slow oxidation with the spinning cup was reproducible. Oxidation rates,
oxygen utilization values, and energy requirements were comparable.
In order to determine whether the improvements in test HA were due to
the peripheral speed of the CGLC or to the material of construction and/or
configuration of the tank, test 12 was made at a peripheral speed of 5.6 meters
per second as in test 9. The result was a two-fold increase in oxidation rate
and oxygen utilization coupled with a 50$ decrease in energy utilization
(test 12 vs. test 9). A large part of the increased effectiveness was due to
the configuration or material of construction of the tank.
-------
A tank was constructed of glass and plastic (G-3) to simulate the configu-
ration of the stainless steel tank. Test 16 was made in this tank using
conditions otherwise the same as in test 11A. Energy requirement was greater
and both oxidation rate and oxygen utilization values were lower when the
nonmetallic vessel was used. The use of the metallic vessel apparently
catalyzed oxidation.
Test 1? was made using the same equipment as test 16 (CGLC in nonmetallic
tank) but with the highest peripheral speed allowable (88 m/sec) with the
drive system employed. Oxidation rate and oxygen utilization were increased
"but the energy rate per mole of sulfur oxidized was increased slightly.
Tests 13 and Ik were made in the stainless steel tank using the CGLC at
8 meters per second; the pool volume and concentration were varied. Increasing
volume (1.5 1, test 13 vs. 1.0 1, test 11A) decreased energy requirement and
increased oxidation rate and oxygen utilization values. Apparently the
increased pool depth and consequent greater retention time of the air in the
pool made these differences. Decreasing concentration (Ij-O.lj- vs. 63.95 g
sulfite sulfur/I, test ih- vs. test 13) increased the energy requirement and
decreased both oxidation rate and oxygen utilization values. Apparently the
total salt concentration rather than the sulfite concentration made the
difference since, in all cases, the rate of oxidation remained essentially
constant until the sulfite sulfur concentration had decreased to values well
"below the initial value of this test.
Test 18 was made to duplicate test HA. Results were fairly consistent,
the primary differences being caused by the indicated proportion of sulfite
oxidized before the first sampling time. After l8o minutes in both cases, the
residual sulfite sulfur analysis was 1.7 grams per liter and the pH was about
3'7- Indicated oxidation over the period 30 to 120 minutes varied from about
30 grams to about kO grains because of the variation of analysis in the first
samples. This caused the variations in power requirements (100 vs. j6 Wh/mole)
and in oxidation rate (0.011 vs. 0.014 mole/min).
Two tests were made in which 2 molar potassium sulfite solutions were
oxidized. Reagent potassium sulfite is available and was used in these tests.
The potassium sulfite solution used in test 19 had a pH of 8.5 and contained
considerably less than the calculated sulfite sulfur content; that used in test
20 had a pH of 10.8 and was closer to the calculated composition. It was
probable that the material used for test 19 had oxidized in storage; such
would account for the lower pH and reduced sulfite sulfur content. However,
since the pH of potassium sulfite was not known, it was decided to determine
the value for both the sulfite and the bisulfite. Similar determinations were
made for ammonium sulfite and bisulfite. To determine the desired pH values,
tests were made in which 0.2 N potassium and ammonium hydroxide solutions were
reacted continuously with pure" sulfur dioxide gas in a closed vessel and the
pH values were recorded continuously. The pH values were measured using a
Broadley-James sealed combination pH electrode, calibrated with Beckman pH 7
buffer solution, and read with a Beckman Model 1055 recording pH meter. The
results of these two titrations are shown in Figure 8 and are tabulated below.
-------
10 —
8 -
O Ammonium system
A Potassium system
0
0.25 0.5 0.75 1
Moles S02 added per mole hydroxide
Figure 8. pH during reaction of 0.2 N ammonium and
potassium hydroxides with sulfur dioxide.
1.25
-------
_ PH _
Range _ Center point
Salt
K2S03 8.0-11.2 9.6
KHS03 3.5-5.5 4-5
7- 4-8.0 7-7
3.4-5.0 4.2
In test 19, with the initial pH at 8.5, oxidation proceeded at the rate
of 0.0158 mole per minute, oxygen utilization was 20. 5$, and the energy require-
ment was 67 watt- hours per mole. In test 20, with the initial pH at 10.8,
oxidation was less rapid, 0.0117 mole per minute, oxygen utilization was 15.1$,
and the energy requirement was 91 watt-hours per mole. Apparently oxidation
proceeded more rapidly at the lower initial pH although no reason for this is
obvious .
A comparison of the results of test 20 using the better potassium sulfite
with test 18 in which ammonium sulfite was used indicates that the oxidation
rate of the potassium sulfite might "be slightly lower than for the ammonium
sulfite; however, the indicated difference was small.
In the preceding tests, the volume of material being oxidized was limited
to about 1.5 liters by the size of the oxidation vessel and the static depth
of the pool was limited to about 2J centimeters (9 in). In most tests, the
pool depth was 8.25 centimeters (3-1/4 in) and the CGLC was mounted so that
its discharge was at the midpoint of the pool or about 4.76 centimeters
(1-7/8 in) from the bottom of the tank. The CGLC was mounted on a shaft that
entered the tank from the top; a "J" shaped air tube also entered from the top
of the tank and directed air to an inlet in the bottom of the CGLC where mixing
with fluid was initiated.
A study of the effects of greater pool depth required a new oxidizer tank
and the mounting of the CGLC on a shaft extending through the bottom of the
tank in order to shorten the shaft and to eliminate the need for a line bearing
on the shaft; see Figure 7. This arrangement required a seal on the shaft and
therefore introduced some friction that did not exist in the shorter version
of the oxidizer tank. The new tank was built to the same diameter (12.7 cm
or 5 in) and cross section configuration as the smaller tank but was 122
centimeters (48 in) deep and the four mixing baffles (l.J cm or 1/2 in wide)
extended to the full depth. The off-gas from the pool was allowed to escape
through a partially closed cover to minimize backmixing with outside air;
off-gas samples were withdrawn through an inverted funnel suspended just
above the pool surface to minimize mixing with the air in the chamber above
the sampling point. The air fed to the unit was metered by a rotameter and
introduced through a tube extending through the bottom of the tank to a point
below the inlet of the CGLC which was mounted so that its discharge was 5.72
centimeters (2-1/4 in) from the tank bottom. The temperature of the solution
was measured by a thermocouple placed near the wall of the tank and at the
level of the CGLC outlet and was controlled at about 52°C (125°F) by addition
of heat from a heating tape wrapped on the tank or by the removal of heat with
a stream of compressed air directed on the outside surface of the tank.
-------
A series of four tests was made in which approximately 0.6 molar potassium
sulfite solution (about 19 g sulfite sulfur/l) was treated with ^120 cubic
centimeters of air per minute through the CGLC unit operated at 1750 rpm.
Operation at 2000 rpm could not "be maintained because of the increased power
required to overcome the friction of the shaft seal and the limited power of
the drive system. Samples were taken by withdrawing solution through a valve
at the bottom of the tank, pipetting a known volume from the sample for
analysis, and returning the remainder to the oxidizer; the tests were not
stopped for sampling. Sampling frequency was varied with the solution volume;
samples were taken at 10-minute intervals when the tank contained 2 liters
of solution, 20-minute intervals for U or 6 liters, and 60-minute intervals
for lU liters. The air rate was sufficient to supply the oxygen equivalent
of 0.07725 mole of sulfur dioxide per minute or about 12.5$ of the oxygen
equivalent of 1 liter of the solution. Thus, in the 2-liter test, the
approximate stoichiometric amount of oxygen was supplied each 16 minutes,
4 liters required 32 minutes, 6 liters required kQ minutes, and Ih liters
required 122 minutes. The times are approximate because-of variations in
input solution analysis.
Data from the tests are given in Table 6 and Figures 9-13- Figures 9 and
10 show, versus time, sulfite sulfur oxidation as percent of input sulfite
sulfur and as moles, respectively. The response was essentially linear in
both cases (Figures 9 and 10) and shows a continuous increase in the amount
oxidized until oxidation reached about 95$ or more (Figure 9). The average
rate of oxidation (moles/min) increased with pool depth (Figure 11). Since
the oxidation rates did not decrease perceptably until a high degree of
oxidation (probably > 95$) had been accomplished, production of sulfates
essentially free of sulfites should be attainable.
Figure 12 shows the energy requirements per mole of sulfite oxidized as
measured with a torque meter versus pool depth. The energy required to over-
come the friction of the seal has been eliminated from these values. Since
approximately the same torque was required to turn the CGLC regardless of the
pool depth, the energy requirements decreased as the pool depth and oxidation
rates increased.
Figure 13 shows average oxygen utilization over the period of steady
oxidation rates versus pool depth. The curve is similar to that for oxidation
rate, moles per minute, since the air rate was the same in all tests. Best
utilization occurred at the greatest depth as might be expected but improve-
ment leveled off at volumes greater than 6 liters (k8 cm or 19 in depth).
Perhaps the most optimistic note is the very high oxygen utilization attained
in these tests with the CGLC unit. Almost ^2$ oxygen utilization was attained
in about 1*8 centimeters of pool depth. It is reported that as much as l^
meters or 1*5 feet of pool depth is used to attain similar utilization with the
spinning cup. However, the effects of solution concentration, cation identity,
oxidation promoters or retarders, and tank geometry are not well known with
either oxidizer and may affect the comparison. Continuous tests have not been
made with the CGLC.
-------
100 f—
1
"8
N
•H
5
-p
•H
-------
•d
0)
(U
-P
0}
0)
3
0
0
Symbol
O
A
D
t
Test
No.
2k
25
26
38
No. of
liters
2
^
6
3A
I
1
30 60 90
120 150 180 210
Time, min
270
Figure 10. Effect of time on moles of sulfite oxidized.
-------
0)
N
a
0)
-p
•H
0.03
0.02
g o.oi
CO
0
I
I
^0 80
Pool depth, cm
120
Figure 11. Effect of pool depth on rate of oxidation.
-------
O)
H
•H
T3
-H
0) -H
fc
-------
Results of these tests indicate that the CGLC is a very effective
oxidizing device and that,, at the conditions employed, a shallow tank can be
used effectively. Therefore, the use of such a device should offer means of
reducing energy requirements and costs for pumping and storing liquids, for
compressing air, and for turning the oxidizer unit. Also, shaft seals would
not be required to withstand as much pressure or might be eliminated in large
units by using top-entry shafts.
OXIDATION OF OTHER SALTS
Oxidation of slurries of other salts having lower solubilities than ammonium
or potassium sulfites were made using the CGLC in the short stainless steel tank.
Zinc Sulfite
Oxidation of zinc sulfite is of interest because of the relatively low
temperature at which zinc sulfate decomposes to give zinc oxide (for recycle
to a scrubber) and sulfur trioxide (which may be absorbed in water to form
sulfuric acid). A process has been described (Slack, A. V., to Petrusek, R. A.
"Removal of Sulfur Dioxide from Stack Gas by Scrubbing with Zinc Oxide Slurry,"
memorandum, January 27, 1971) in which this series of stages is included; it
would allow the production of sulfuric acid from stack gas sulfur dioxide and
eliminate the vanadium-catalyzed oxidation step required in the usual sulfuric
acid process.
The oxidation of zinc sulfite is unusual in that the sulfite is a fairly
insoluble material (O.l6 g ZnS03•2H20/100 g H20) while zinc sulfate is highly
soluble (57-^5 g ZnS04/100 g H20 at 25°C). This, of course, means that oxida-
tion tests will start with a slurry of zinc sulfite and will end with a solu-
tion of zinc sulfate; any zinc sulfate available for thermal decomposition
would have to be obtained by crystallization from the final solution.
Obtaining zinc sulfite was a problem. The material is not generally
available in the chemical supply catalogs. A supply, therefore, was prepared
in the laboratory by reacting a zinc acetate solution with sulfur dioxide gas
and recovering the zinc sulfite by vacuum evaporation of water and acetic acid.
In preparation of this material, 1 pound of zinc acetate dihydrate was dissolved
in 9 liters of water and pure gaseous sulfur dioxide was bubbled into the
solution until the pH decreased to b.15- The reaction frees acetic acid and
forms zinc sulfite dihydrate as follows:
Zn(CH3COO)2'2H20 + S02 > 2CH3COOH + ZnS03'2H20
The zinc sulfite is soluble in the weak acetic acid produced. Water and acetic
acid therefore were boiled off under vacuum to recover the zinc sulfite in the
dihydrate form in which it is stable at room temperature. A vacuum of about
63 to 68 centimeters (25-27 in) mercury was maintained during the evaporation
step. As zinc sulfite dihydrate (ZnS03'2H20) precipitated, the slurry in the
evaporator was filtered (without allowing the solution to contact room air)
and the solid product was washed with alcohol and dried with air. Samples were
analyzed by infrared analysis and by the petrographic microscope. About 111
-------
grams of ZnS03-2H20 was collected before any significant amount of foreign
material appeared in the precipitate. Chemical analysis of the ZnS03'2H20
showed it to contain 2.6.7% total sulfur, 1.6.7% sulfite sulfur, and 34.2$ zinc.
The analyses are 9k.7% of theoretical and the zinc to sulfur mole ratio is
1.002. Apparently there was a small amount of moisture or alcohol left in the
sample. The average length of the particles was about 10 microns; this length
generally ranged from about 1 micron to 50 microns, the width was generally
about 60% of the length, and the thickness about 20$ of the length.
Oxidation tests were made with 1 and 5$ slurries of the ZnS03'2H20 in
water at 52°C. One liter of slurry was used in each test; the pH was not
adjusted. Air was introduced into the CGLC at 4120 milliliters per minute.
Samples were withdrawn each 10 minutes and analyzed iodometrically for sulfite
sulfur. Evaluation was made on the basis of these analyses.
In the first test, the slurry was sampled as soon as the 1% slurry was
formed and the analysis (l.6l8 g sulfite sulfur/l) was close to that expected
(1.67 g/l); it was indicated, however, that as much as 3% of the sulfur had
been oxidized in the short time required to suspend the solids. Suspension
was attained by rotating the CGLC without air feed for a short time, probably
less than half a minute. The pH of the slurry at the beginning of the test
was 6.1. Oxidation proceeded as indicated in the following tabulation, which
also includes data for test 2, described later.
Time, Sulfite sulfur Sulfite sulfur oxidized
min pH content, g/l % of input Mole/min
Oxygen
utilization,
Test 1
Input
0
10
20
30
Test 2
Input
0
10
20
30
6.1
4.8
5.0
5.0
.67
,62
,26
•IT
5-1
5-1
4.9
0.08
8.35
3-29
0.58
0.27
o.i4
84
90
95
61
93
97
98
0.00425
0.00028
0.00028
5-5
0.4
0.4
0.00846
0.00096
0.00040
11.0
1.2
0.5
The air feed per minute was sufficient to oxidize 0.0773 mole of sulfite
sulfur, which would be equivalent to 153$ of the sulfite sulfur present. Thus,
in 10 minutes, the time of the first sample, 15.3 times stoichiometric air
had been fed and 84$ of the sulfur dioxide had been oxidized; oxygen utiliza-
tion thus was about 5.5$.
In test 2, preparation of the more concentrated (5$ vs. 1$) ZnS03«2H20
suspension required more time, and oxidation apparently proceeded to about 6l$
completion during this time; however, errors in sampling the slurry could have
48
-------
teen involved. After 10 minutes of operation, the material was 93$ oxidized.
Since the air feed per minute now was equivalent to only about 30.6$ of that
required to oxidize all the sulfite sulfur present, the oxygen utilization
(0-10 min, 6l-93$ oxidation) was about 11$. The affinity for oxygen is
indicated by the very rapid oxidation rate, even in the suspending step.
Initial oxidation in test 2 was indicated also by the low pH of 5.14- at the
beginning of the test.
It is concluded that the oxidation of zinc sulfite was accomplished
fairly rapidly and easily with the CGLC and that its solubility is great
enough that no adjustment of pH would be required to maintain a suitable rate.
Recovery of the sulfate salt in solid form for disposal or use will require
removal of significant amounts of water; the solubility of zinc sulfate is
fairly high and it may precipitate with as many as 7 moles of water of hydra-
tion. No data were developed related to the maximum zinc sulfite slurry
concentration allowable in the oxidizer.
Calcium Sulfite
Oxidation of calcium sulfite (CaS03) produced in lime or limes tone-sulfur
dioxide scrubbers is desirable for several reasons. Major benefits of oxida-
tion would be (l) the formation of a material suitable for landfill that will
dewater more readily and completely than will CaS03 and (2) the obviation of
any continued chemical oxygen demand in the discarded waste material. In
prior work (August 1972 Applied Research Branch report, TVA), the oxidation of
CaS03 was shown to respond to initial pH and to the presence of several iron
salts as catalysts. In those tests, made with the spinning cup oxidizer and
in Pyrex, the oxidation rate of 1$ CaS03 slurries was low when the pH was at
5 or above; about 1 hour and 20 times stoichiometric air were required to
accomplish 87$ oxidation at pH 5« A series of laboratory-scale tests now has
been made in which CaS03 slurries were oxidized with the CGLC oxidizer in a
stainless steel tank.
Calcium sulfite for the tests was produced in the laboratory by reacting
a solution of calcium acetate with sulfur dioxide gas. Aqueous acetic acid
was a byproduct. Precipitated CaS03 was filtered from the aqueous acetic acid,
washed with water, and dried. Chemical and petrographic analyses were made.
The material contained k2.8f> calcium oxide and 2k.3$ sulfur (theoretical values
for CaS03*l/2H20 are Uj.4 and 2k.8$, respectively); about k% of the sulfur was
oxidized to sulfate form. Petrographically, the material was described as
agglomerates of small (less than 1 micron) crystals of CaS03'l/2H20 having a
mean index of refraction of 1.604. The crystals are oriented within the
aggregates indicating that they have an elongated direction along which they
have collected during assemblage. The aggregates range up to 200 microns in
diameter.
Oxidation tests were made with 1 and 5$ slurries of the laboratory-
prepared CaS03*l/2H20 using the CGLC at 2000 rpm in the 12.7-centimeter-
diameter by 23-centimeter-tall (5-in-dia by 9-in-tall) stainless steel tank.
The air rate was hI20 milliliters per minute and the slurry temperature was
52°C in all tests. Slurry samples were analyzed iodometrieally to determine
residual sulfite content and to allow calculation of the extent of oxidation.
U9
-------
The energy requirement was calculated from measurement of the torque required
to turn the oxidizer. Variables in the tests, in addition to the slurry
concentration, were initial pH (adjusted by sulfuric acid addition) and
addition of ferrous carbonate (FeC03) as a catalyst. Sulfuric acid was used
for pH adjustment to eliminate the effect of other ions in the test and to
maintain sulfur dioxide at a known input level. Data from the tests are given
in Table 7.
The first three tests (not in table) were made with 1$ slurries. Oxida-
tion was much more rapid than in the tests reported in the August 1972 Applied
Research Branch report (TVA), reaching 67$ in JO minutes and 96% in 1 hour
when the initial pH was 6. When the pH was reduced to 5, oxidation was nearly
complete (99$) when the first sample was taken at 50 minutes; the addition of
O.lf FeC03 at a pH of 5 resulted in 95$ oxidation within 10 minutes. These
tests were too short and were not sampled frequently enough to determine the
relative oxidation rates.
Tests then were made with 5$ slurries in order to lengthen the tests
enough to obtain meaningful oxidation rates for comparison. In the first of
these tests (test 27; Table 7), the pH was adjusted to 6 prior to start of
oxidation, and the test was continued for 4 hours. There was some difficulty
in sampling due to settling during sample collection as is indicated by a wide
scattering of data. In the 1% slurry tests, sampling by pipette had been
suitable, but in these tests with 5% slurry, the solids tended to stop up the
pipette tip and samples tended to be low in solids. Values obtained early in
the test were in error due to this problem and, since most of the sulfite was
in solid form, analysis for sulfite sulfur indicated high oxidation values.
Analyses were made during the test, however, and the discrepancy was noted;
in samples taken after about J>Q minutes, special care was taken in the sampling
procedure to try to prevent the problem. Subsequent tests were made with a
dipper-type volumetric sampler to further reduce the inaccuracy in slurry
sampling. Data for the latter part of test 27 appear to be fairly accurate.
Although not considered completely accurate, the probable final degree of
oxidation was less than about 10$, and the energy requirement was over 5000
watt-hours per mole (Wh/mole) of sulfur dioxide oxidized. The oxidation rate
was only about 0.0001 mole per minute. Calculated oxygen utilization was less
than 0.2$.
Test 28A was made with the pH adjusted to 5.0 prior to the test and using
the revised sampling method. Essentially complete oxidation was achieved in
50 minutes; almost £0$ of the sulfite sulfur had either been lost or oxidized
prior to start of the air feed. The procedure here was to add the CaS03 to
the water and adjust the pH prior to heating the slurry to test temperature.
Slow circulation of the slurry during heating was accomplished by running the
CGLC at about 200 rpm. Contact with air at the surface caused some oxidation,
and loss of some sulfur dioxide probably occurred prior to the beginning of
air feed and high speed rotation of the CGLC; the combination of these factors
amounted to kO% of the input sulfite. Test results were evaluated on the
basis of incremental oxidation only, however, and are considered valid. The
oxidation rate for the test period before 90$ oxidation was reached was 0.006
mole of sulfite per minute. During that period, the energy requirement was 176
watt-hours per mole of sulfite oxidized, and oxygen utilization, based on slurry
analysis, averaged 7.9$.
50
-------
TABLE 7. OXIDATION OF CALCIUM SULFITEa BY AIR (LABORATORY TESTS)
Test
No.
27
28A
30AC
29A
Time,
min
0
30
60
90
120
240
0
10
20
30
40
50
0
10
20
30
40
50
60
0
10
20
30
Cumulative values
S02 Time period, Energy required, Oxidation rate,
pH oxidized, % min Wh/mDle oxidized mole/min
6.0
6.1
6.05
6.05
6.1
6.4
5-0
5.1
5-2
5-1
4.8
4.7
5.0
4.9
4.9
4.9
4.9
4.9
4.9
4.00
4.10
2.75
2.70
3
1
3
6 90-240 824o o.oooi
4o
59
74
90 0-30 176 0.0061
98
99
14
24
42
56
72
82 0-50 217 0.0050
93
18
82 0-10 46 0.0234
98 0-20 73 0.0147
99
Oxygen utilized, %
0.16
7-9
6.5
30.3
19.0
slurry of laboratory-prepared CaS03'l/2H20.
Values erroneous, see text.
Ferrous carbonate added (0.1$ by wt.).
-------
In test 30A, 0.1% FeC03 was added to the 5# CaS03'l/2H20 slurry, and the
pH again was adjusted to 5 with suliXiric acid. The analysis of the initial
solution indicated that about lh% of the sulfite had been oxidized or lost
prior to the beginning of the test. A shorter time was used to heat the
slurry than in the preceding test. Oxidation proceeded at about 0.005 mole
of sulfite per minute until the test was stopped at 60 minutes. Measurements
taken during the first 50 minutes of the test showed that oxidation had
reached about 82$, the energy requirement was 217 watt-hours per mole of
sulfite oxidized, and average oxygen utilization was 6.5$.
Test 29-A was made to duplicate test 28A except that the initial pH was
adjusted to h. About l8$ of the sulfite was shown to be oxidized or lost
prior to the test; thereafter, the proportion oxidized increased to 82/6 at
10 minutes and 98$ at 20 minutes. During the first 10 minutes, the oxidation
rate was 0.023 mole per minute, the energy requirement was 1+6 watt-hours per
mole of sulfite oxidized, and average oxygen utilization ( calculated from
slurry analysis) was about 30$. On extending the test period to 20 minutes,
the oxidation rate (during the 20-min period) was 0.015 mole per minute, the
energy requirement was 73 watt-hours per mole of sulfite oxidized, and oxygen
utilization was about
The results of these tests indicate that the CGLC device is suitable for
oxidizing CaS03'l/2H20, that the oxidation rate is very slow at pH 6 but
becomes more rapid as the pH is decreased to 5 or below, that oxidation
proceeds about four times as rapidly at pH it- as at pH 5> and that the addition
of FeC03 at pH 5 does not increase the oxidation rate when the process is
carried out in a stainless steel vessel (presumably the stainless steel surface
provides the catalyst that was evidenced by the FeC03 in a glass vessel in
the earlier tests).
It appears evident that oxidation of actual solids from lime or limestone
scrubbing of sulfur dioxide will require some accommodation of any residual
calcium carbonate in order to oxidize the calcium sulfite to calcium sulfate,
otherwise the pH would be too high for a reasonable rate. It is possible,
however, to gain the required acidity from the stack gas through staged
scrubbing; such was reported by R. H. Borgwardt, U.S. Environmental Protection
Agency, at the Flue Gas Desulfurization Symposium at New Orleans in March 1976.
Tests using actual scrubber solids should be made to study the effects of
impurities and the larger CaS03*l/2H20 particle size in the actual slurry.
-------
SELECTIVE PRECIPITATION OF SULFATE WITH BARIUM SALTS
Regenerative soluble salt scrubbing processes for removal of sulfur
oxides from waste gases usually are affected adversely by sulfates in the
system. In most double-alkali systems, the regenerant reacts with the soluble
bisulfite from the scrubber in preference to the sulfate, which then builds
in the system. Sulfite will react with sulfur dioxide in the scrubber, but
sulfate will not. The same type of problem exists in systems where the
bisulfite scrubber product is stripped of sulfur dioxide by thermal means.
Sulfate, whether formed by disproportionation in the thermal processes or by
normal oxidation, builds in the system and eventually will reach saturation;
the sulfate then will precipitate and may cause fouling at some point in the
system. Some method of removing sulfate is required to prevent its buildup.
During some of the previous small-scale scrubbing tests, attempts were
made to monitor the total sulfur content of the solutions by precipitation
with barium salts and to determine the degree of oxidation by infrared analysis
of the precipitates. In this work, it was noted that barium salts exhibited
some selectivity in precipitation and that sulfates were precipitated more
readily than sulfites. From these observations, a process was visualized for
removal of sulfates from scrubber solutions by the addition of barium carbonate
to take advantage of that selectivity. The precipitated barium sulfate, along
with any coprecipitated barium sulfite, would be reduced to barium sulfide and
subsequently reacted with water and carbon dioxide to form barium carbonate
and hydrogen sulfide. Reduction of barium sulfate and conversion to barium
carbonate is a commercially practiced process. The barium carbonate would be
recycled; the hydrogen sulfide would be used to produce sulfur by the Glaus
process. Tests of the selective precipitation of sulfate with barium carbonate
now have been made and are reported herein.
TEST PLAN
In these tests, simulated scrubbing solutions of known composition first
were prepared from the sulfates, sulfites, and bisulfites of sodium or
potassium; then a measured amount of barium carbonate (BaC03) was added as a
finely divided solid and the mixture was stirred continuously until the test
had been completed (usually 3 hr). During the entire test, the solution was
maintained at the desired test temperature. A timed series of samples was
taken and filtered immediately. The solids were water washed, acetone washed,
and air dried prior to infrared analysis to determine the proportions of C03~,
S03=, and S04= present. The barium salts of these three anions have very low
solubility. Precipitated proportions of the individual species of sulfur
oxide and the residual BaC03 then were calculated from the infrared analysis
and the input proportions. Data from the tests are given in Table 8 and in
53
-------
TABLE 8. SELECTIVE PRECIPITATION OF SULFATE WITH
BARIUM SALTS—TEST DATA
Test No.
Soluble cation (X)
Input proportions , nolea
X-SOt
XjSOa
XHSOa
BaCOa
Ba:S04" nole ratio
Ba:S nole ratio
Teisperature > "C
C*
Ca
St
Results at tine indicated
30 minutes
£ of input S precipitated
i of individual ion precipitated
1 hour
$ of input S precipitated
$ of individual ion precipitated
3 hours
it of input S precipitated
i, of Individual ion precipitated
Test No.
Soluble cation
Input proportions, Boles
XaS04
XHSOa
BaCOs
Ba:S04- mole ratio
Ba:S nole ratio
Tenperature, "C
Ct
g.
s«
Results at time indicated
30 minutes
% of Input S precipitated
% of individual ion precipitated
1 hour
t of Input S precipitated
£ of individual ion precipitated
3 hours
$ of Input S precipitated
% of individual ion precipitated
«
S04-
37
37
42
42
51
51
<
sc4-
21
63
23
70
23
69
— 1 »
Na
1
-
-
1
1
1
52
6.0
0
3.0
0
SOa" Ba
- 37
- 42
_
- 51
-7 >•
Ha
1
2
2
2
0.67
52
7.1
3-5
5.3
3-5
SOa" B&
12
18 50
13
19 53
21
31 65
4 2 »
Na
-
1
-
1
-
1
49
5-7
5-7
2.9
2.9
S04- SOa" Ba
51
51 51
57
57 57
71
71 71
< 8 »
Na
1
2
3
3
1
52
6.9
3-4
5-2
3.4
S04" SOa" Ba
32 54 -
95 51 66
32 33 -
97 50 66
30 44
90 66 74
< 5 >
Na
-
-
1
1
-
1
52
3-1
3-1
3-1
3.1
SO*" SOa* ^i
3 52
52 55
7 53 -
53 60
2 71
71 73
« 9 »
Ha
1
2
S
2
0.67
49
0.7
0.4
0.5
0.4
S04* 303* Ba
16 13 -
48 19 44
16 16
48 24 48
19a 18" -
57* 27a 56"
< 1, »
Na
1
1
-
1
1
0-5
49
7.2
3-6
3-6
1.8
SO** SOg* BB
21 7
42 14 56
23 7 -
46 14 60
26 9 -
52 18 71
•« 10 ».
Na
0.1
0.9
1
10
1
52 „
3-8
3-1
3-4
3-1
SO4- SOa' Ba
l£ 36
160 40 52
14 48 -
140 53 52
9 66 -
90 73 75
« 5 >
K
1
1
-
1
1
0.5
7.2
3-6
3-6
1.8
S04' S03' Ba
9 10 -
18 20 37
14 6 -
29 12 41
16 9
32 19 51
< 11 »
Na
0.1
0.9
1
10
1
77
3.8
3-1
3-1
S04- S0a- Ba
7 67 -
70 74 75
8 67
80 74 75
9 70
90 78 80
« 6 >
Ma
1
_
2
1.5
1-5
0.5
52
7.1
3.5
5-3
3-5
S04- 803- Ba
15 7
47 10 U5
16 9
48 1} 50
17 11
50 j£ 54
0 Test duration, 2-1/2 hours.
-------
Figures ih through 2k. Only a portion of the data are given in the table; the
portion given was selected to illustrate the rate and degree of completion of
the reactions. More complete data are given in the figures.
In the table and figures, the expressions Ct, Ca, St, and Sa represent
the moles of sodium (or potassium) or of sulfur per 100 moles of total water.
The "C" values are moles of cation and the "S" values are moles of sulfur.
Subscript "t" indicates total ion present; subscript "a" indicates active ion
present (sulfite or cation associated with the sulfite). Total water includes
free water plus 1/2 mole per mole of bisulfite.
EVALUATION OF DATA
Tests 1, 2, and 3 were made with solutions of the individual salts,
sodium sulfate, sulfite, and bisulfite, to which the stoichiometric amount of
BaC03 was added to determine the rate and extent of the reaction in each
system. In these tests St and the temperature were held at the same levels,
3 moles per 100 moles of water and k9° to 52°C, respectively. Sulfite and
bisulfite reacted about equally rapidly and to the same degree, reaching about
50$ in 1 hour and 70$ in 3 hours; sulfate reacted more slowly, the values
being only k-0 and 50$, respectively. This comparison would indicate that the
sulfate is not preferentially precipitated.
Test k was made with a mixture of sulfate and sulfite, with a slightly
higher St (3«6) than in tests 1, 2, and J, but with half the sulfur in sulfite
form (Sa = 1.8). Only enough BaC03 was added to react with half the sulfur;
thus the preferential reaction, if it existed, could be demonstrated. Prefer-
ential precipitation of sulfate was shown in each sample; sulfate precipitation
was about three times sulfite precipitation. Only about 70$ of the barium was
utilized, as in tests 2 and 3, probably indicating some blinding of the BaC03.
Test 5 was made at the same conditions as test h except that potassium
salts were substituted for the sodium sulfate and sulfite. During the initial
15-minute period, sulfate was preferentially precipitated; at 30 minutes,
sulfate and sulfite precipitation were about equal. After 1 hour and 3 hours,
however, there was a 2:1 preference for sulfate exhibited. The reaction with
potassium salts was preferential with the sulfate, but not to the extent
exhibited in the sodium system.
In tests 6, 7, and 8, mixtures of 1 mole sodium sulfate and 2 moles
sodium bisulfite, dissolved and diluted to an St of 5.3 and an Sa of 3.5, were
mixed with amounts of BaC03 to result in input stoichiometries (Ba:S mole
ratios) of 0.5, 0.67, and 1.0, respectively. The temperature was held at 52°C.
In test 6 (Ba:S = 0.5), the precipitate contained about two-thirds of its
sulfur in sulfate form even though the solution contained two-thirds in sulfite
form. Of the individual species, about three times as great a proportion of
the sulfate was precipitated (50$ at 3 hr) as was the sulfite (16$ at 3 hr).
Total barium utilization was only 5^$ at 3 hours.
55
-------
100
w
0)
•H
o
80
60
Molecular proportions,
Wa2S04 + BaC03
Concentrations,
St = 3
Temperature, 52°C
= BaC03
•p
c!
0)
PH
0
I
0
60
90
Time, min
120
150
180
Figure
Laboratory precipitation of sulfate with barium carbonate — test 1.
-------
100
8o
-p
3
(U
fn
03
O
CO
•d
•p
"in
O
6o
20
Molecular proportions,
Fa2S03 + BaC03
Concentration,
St = 2.9
Temperatiire, lt-9°C
I
60
90
Time, min
120
150
180
Figure 15. Laboratory precipitation of sulfite with barium carbonate--test 2.
-------
100
I
CD
1 80
o
0}
Molecular proportions,
NaHSOg + BaCQa
Concentration,
St = 3.1
Temperature, 52°C
0>
20
S04~ in precipitate
13—0-
0.
IT
I
T
60
90
Time, min
120
150
180
Figure 16. Laboratory precipitation of sulfite with barium carbonate—test 3.
-------
VD
100
•CJ 80
V
•p
t
w
s
6o
Q)
Pi
W
•d
-p
-------
100
•s
-p
0)
S-t
80
•rl
O
CO
•d
•p
a
0)
s>
f^
6o
Molecular proportions,
K2SQ3 + K2S04
Concentration,
St = 3.6
Temperature, 49°C
BaC03 utilisation
" in precipitate
Figure 18. Laboratory precipitation of sulfate and sulfite with barium carbonate—test
-------
100
•8
801-
o
Oj
8
w
5 60
o
CD
Pi
CO
•d
o
+5
-P
fl
0)
o
20
0
Molecular proportions,
2NaHS03 + Na2S04 + l.
Concentration,
St = 50
Temperature, 52°C
1
BaC03 utilization
S04~ in precipitate
S03~ in precipitate
0
60
90
Time, min
120
150
180
Figure 19- Laboratory precipitation of sulfate and sulfite with barium carbonate—test 6.
-------
ON
ro
100
80
0)
Jn
W 60
o>
•H
o
3 + NasS04 + 2BaC03
Concentration,
S* - 5-3
Temperature, 52 C
60
S04=: In precipitate
Q-
BaC03 utilization
S03= in precipitate
90
Time, min
120
150
180
Figure 2O. Laboratory precipitation of sulfate and suli"ite with barium carbonate--test 7-
-------
ON
1001
15-5-°-°-
80
20
0
0
T
T
D
S04- in precipitate
BaC03 utilization
Molecular proportions,
2NaHS03 + Na2S04 + 3BaC03
Concentration,
St = 5-2
Temperature, 52°C
I
60
90
Time, min
120
150
180
Figure 21. Laboratory precipitation of sulfate and sulfite with barium carbonate—test 8.
-------
100
5
CO
o
ra
1
o
fc
-------
o\
100
•p
3
o
fH
0)
Molecular proportions,
0.9NaHS03 + 0.1Na2S04
Concentration,
st = 3.1*
Temperature, 52 C
BaC03
20
0
60
90
Time, min
120
150
180
Figure 2J. Laboratory precipitation of sulfate and sulfite with barium carbonate—test 10.
-------
100
80
w
(U
•H
U
&
CO
(U
-------
Increasing the proportion of BaC03 in test 7 to Ba:S mole ratio of 0.6?
resulted in an early (30 min and 1 hr) precipitation of about three times
greater proportion of the sulfate than of the sulfite. These were about the
same proportions as precipitated in test 6. After 3 hours, however, the
preference was less noted and the proportion of the sulfate precipitated was
only twice that of the sulfite. Final BaC03 utilization was 65$. Final
sulfate precipitation was 69$ of that present.
A further increase of BaC03 proportion to a Ba:S mole ratio of 1 in test
8 resulted in a furtherance of the trend seen in test 7 compared with test 6.
The final proportion of sulfate precipitated was about 1-1/2 times that of the
sulfite, and BaC03 utilization increased to 70$. After 3 hours, 90$ of the
sulfate had been precipitated.
Test 9 was the same as test 7 except that the system was diluted from an
St of 5-3 to a less practical St of 0.5, a tenfold dilution; and the test was
carried out for only 2-1/2 hours instead of 3 hours. Results at 30 minutes
indicated that a 2-1/2 times greater proportion of sulfate was precipitated
than of the sulfite. Thereafter, the preference was reduced to about twofold.
Final results were very similar to those of test 7; the final sample contained
about twice the proportion of the sulfate in precipitated form as it did the
sulfite (57$ vs. 27$) and BaC03 utilization was about 60$.
In regenerative systems the sulfate will build, unless controlled, until
the physical losses from the system remove sulfate at the rate it is being
formed. Sulfate sulfur could, conceivably, reach 90$ or more of the total
sulfur present. In nonregenerative systems, however, suILfate level would be
much less. Tests 10 and 11 were made with proportions of sulfate and bisulfite
at about the levels expected in the effluent of a nonregenerative scrubber
system on a conventional coal-fired boiler; about 10$ of the sulfur was in
sulfate form. The St was held at 3-^« Enough BaC03 was added to provide the
stoichiometric requirement of all the sulfur. The differences in the tests
were the temperatures at which the reaction was carried out, 52°C in test 10
and 77°C in test 11. The different temperatures were tested because the
solubility relationship of barium sulfite (BaS03) to barium sulfate (BaS04)
is reported to reverse at and above the higher temperature. Early (30 min and
1 hr) results from test 10 indicated precipitation of more sulfate than
existed in the sample. This may have been due to the low sulfate proportion
and to incomplete drying of these samples. However, the sulfate to sulfite
ratio probably is correct and the results show that a four times greater
proportion of the sulfate than the sulfite was precipitated at 30 minutes. By
1 hour, the proportion was only about three times greater and at 3 hours the
proportion was reduced to about 1.25- About 75$ of the BaC03 was utilized.
Test 11 results, although about the same after 3 hours, were quite different
during the early part of the test when the proportions of sulfate and sulfite
precipitated were almost equal at 30 minutes and 1 hour. This would support
the fact that at 77°C the solubilities of BaS04 and BaS03 are more nearly
equal than at 52°C at which temperature BaS04 is the least soluble.
67
-------
CONCLUSIONS
These tests show the selective precipitation of sulfate from a solution
of sodium sulfate and sodium sulfites by the addition of BaC03. In tests
where the addition of BaC03 was 1 to 1.5 times the equivalent of the sulfate
content of the solution and the temperature was held at 52°C or lower (tests
4 and 6), sulfate precipitation was more selective than in tests at similar
conditions and higher BaC03 ratios (tests 8 and 10). Test results indicate
that the sodium system might be more amenable to selective sulfate precipita-
tion by this means than would be the potassium system. It is concluded that
selective precipitation of sulfates by reaction with BaC03 is a technically
feasible means of removing soluble sulfates from regenerative scrubbing
systems and thereby obviating problems associated with sulfate buildup. The
selectivity of sulfate precipitation by barium is important since other
cations, such as calcium, are very selective in precipitating sulfite. Any
sulfite precipitated with sulfate is undesirable but not so much so as to
make the process unviable; it would mean that a larger charge of material
would need to be subjected to the reduction, hydrolysis, and carbonation steps,
however. Low barium utilization would add to this load also. Barium carbonate
is poisonous when taken by mouth and this might be a deterrent to its use;
barium sulfate is a nontoxic, insoluble salt.
68
-------
UTILIZATION OF WEAK SULFURIC ACID
PRODUCED IN POLLUTION CONTROL PROCESSES
Several sulfur dioxide (S02) scrubbing processes have been tested from
which a product stream of sulfuric acid is recovered. Some of the more direct
of these processes depend on the following sequence: (l) adsorption of S02
from a dilute gas stream in a bed of activated carbon, (2) oxidation of the
S02 to S03 by the oxygen present in the gas, (j) hydration to the acid form,
and (U) removal of the product sulfuric acid from the carbon by elutriation
with water; the "Sulfacid" process (6) is an example. The processes differ,
primarily, in the means of separating the steps of the procedure. Some
processes also include a sulfuric acid concentration step which utilizes heat
in the gas stream for the purpose. The concentration of the sulfuric acid
produced varies with the design of the process, with the concentration of S02
in the waste gas, and with the degree of SOP recovery attained. Generally,
sulfuric acid concentration may vary over the range 5 to 30%. Sulfuric acid
purity may vary as a result of its contact with impurities from the gas stream;
impurities may include any of the materials in the fly ash, the amount being,
generally, a function of the degree to which the fly ash is removed prior to
the S02 removal step. The most common and most concentrated of the dissolved
contaminants probably are calcium and iron sulfates.
In order for such an S02 removal process to be viable, some use must be
made of the weak sulfuric acid. In Japan, some plants simply neutralize the
acid with lime or limestone and produce gypsum which then is used in production
of the construction materials, cement and wallboard. Such practice does not
appear feasible on a large scale within the United States because of the
competition of natural gypsum and the production processes already installed
to utilize that material. There is an experimental installation to make gypsum
from dilute sulfuric acid in Florida (7)> however.
Other uses of weak sulfuric acid might be found. For example, at about
15$ concentration, it is used in pickling metal surfaces, and at other
concentrations it might be used for water treatment, pH control, etc. A major
power plant installation of such a process would, however, produce so much
sulfuric acid that a large portion might have to be transported over long
distances for disposition. A 500-MW power plant burning 3.336 sulfur coal would
produce about 5000 tons of 10$ sulfuric acid per day. The corrosiveness of the
weak acid, compared with that of the highly concentrated product (> 90$)
usually shipped, would add to the problems and cost of transport. A method of
utilization or concentration of the acid at or near the power plant site is
needed.
69
-------
Concentration of the acid from the 500-MW plant, above, from 10 to 92$
would require the removal of about 185.7 tons of water per hour; if this were
done by heating, the energy required for evaporation alone would be about 92
million kilogram-calories (56^ million Btu) per hour. This is equivalent to
106 MW or about 20% of the power output of the plant.
Some method of utilization of weak sulfuric acid is needed that will
reduce the energy requirement for removal of water from either the acid or its
product. Methods under consideration include the following:
1. React weak sulfuric acid with melamine to produce solid melamine
sulfate; remove solid by filtration or decantation; neutralize
with ammonium, potassium, or sodium hydroxide (or carbonate);
wash the suUFate salt from the melamine as a solution more
concentrated than would be possible from the original acid and
process it to the dry salt; and recycle the melamine.
2. Utilize the weak sulfuric acid to extract phosphate rock and
subject the product phosphoric acid to the same procedure as
in item 1. Such a process should produce phosphate salts of
very low impurity content and might allow formation of clear
liquids from acid produced from uncalcined Florida phosphate
rock.
3. Remove water from the weak sulfuric acid, the product phosphoric
acid (item 2), or from one of the salt solutions from the melamine
separation and neutralization steps (items 1 and 2) by freeze
drying or concentration instead of normal evaporative drying
methods. The energy requirements might be much less if freeze
concentration were employed.
Some work has been done in which freeze concentration of phosphoric acid
and of ammonium phosphate solutions have been tested. Results are reported
later. In other work, reported below, weak sulfuric acid has been used to
extract phosphate rock to produce phosphoric acid.
USE OF WEAK SULFURIC ACID FOR EXTRACTION OF PHOSPHATE ROCK
Test Flan and Equipment
A factorial experiment was carried out for study of the extraction of
phosphate rock with weak sulfuric acid; its parameters are shown below.
Level
Variable - +
A - Rock particle size Coarse Fine
B - Reaction temperature, °C 2k 52
C - Sulfuric acid concentration, % 10 JO
D - Acid addition time, hr 12
The full test plan is shown in Table 9.
TO
-------
TABLE 9. EXTRACTION OF PHOSPHATE ROCK WITH
WEAK SULFURIC ACID--TEST PLAN
Variable
Test
position
1
2
3
k
5
6
7
8
9
10
11
12
13
Ik
15
16
Test
No.
2
5
15A
10
k
6
ik
9A
3
8
-&
12
1A
7
13
11
A,
rock
grinda
Coarse
Coarse
Coarse
Coarse
Coarse
Coarse
Coarse
Coarse
Fine
Fine
Fine
Fine
Fine
Fine
Fine
Fine
B,
temperature ,
c,
0 C concentration,
-------
Prior to making the batch tests, calculations were made of the maximum
phosphoric acid concentration that could be expected from each concentration
of sulfuric acid and the grade of phosphate rock in use. The rock (100 g) was
prewet in a 9.5-centimeter-diameter (3-75-in-dia) glass vessel 25 centimeters
(10 in) high with 1JO grams of phosphoric acid of the calculated concentration
and was held at test temperature for a period of JO minutes; this is the
approximate amount of acid that would be available for recycle if the cake
were washed countercurrently with 1.3 grams of water per gram of rock. At the
end of the prewet period, feed of sulfuric acid to the prewet rock was begun.
The acid was fed from a constant-head feeder (Mariotte's bottle) with the tip
of the delivery line touching the moving slurry surface; this tended to
eliminate dripping and allowed for more nearly continuous feed, particularly
when feeding the stronger acid over the longer time. Agitation was maintained,
using a 7.6-centimeter-diameter (j-in-dia), two-blade propeller-type agitator,
at a constant rate (280 rpm) throughout the prewet, acidulation, and a 15-
minute postacidulation period. The product slurry was transferred immediately
to a stainless steel filter frame with a square 8.33-centimeter (3.28!-in)
100-mesh stainless steel screen /J59.5 cm2 (10.767 in2_)J as the filter medium.
Filtrations were timed to the first appearance of solid cake above the fluid
surface. Vacuum was maintained at 38 centimeters mercury (-0.75 cm Hg) by
installation of a vacuum surge tank and a vacuum control valve on the labora-
tory vacuum system. A plug cock mounted on the filter drain served to apply
and to isolate suction rapidly and as required. Both the volume and the
weight of the filtrates were measured. The filter cake, typically 1.25
centimeters (1/2 in) thick, was allowed to dewater after the second wash by
continued application of vacuum until no more liquid ran from the filter and
then was transferred as quantitatively as possible to trays for oven drying
to essentially constant weight at 57°C (the temperature was limited to prevent
dehydration of the gypsum); both wet and dry weights of the cakes were recorded,
The dried filter cake, product acid filtrate, and both wash solutions were
submitted to the general analytical laboratory for analysis. The data
collected are given in Table 10.
Evaluation of Data
Evaluation was made of the responses measured and of calculations made
from those responses. Calculated values included those to show the proportions
of the rock constituents in the acid and the proportions and forms of P205 in
the filter cake. Also calculated were the filtration rates expressed as kilo-
grams of rock, kilograms of P205, and cubic meters of liquid per hour per
square meter of filter area. Filtration rates were calculated for the initial
(product acid) filtration step alone and for the total process including both
wash steps. The calculated responses are given in Table 11. The statistical
analysis of these responses is given in Table 12.
The statistical analysis shows the average of each calculated response
and the difference in response due to each variable. Also shown are the
averages and differences for the interactions of each pair of variables. The
three- and four-variable interactions were considered to be highly unlikely
and were considered to be error; the error in the data thus was calculated
from the differences in the averages of the three- and four-variable
interactions. From the error, so calculated, the values of the least
72
-------
TABLE 10. EXTRACTION OF PHOSPHATE ROCK WITH WEAK SULFURIC ACID—PRIMARY DATA
Test position
Filtration time , sec
Product acid
First wash
Second wash
Total
Filtrate volume , ml
Product acid
First wash
Second wash
Total
Filtrate weight, g
Product acid
First wash
Second wash
Total
Dry cake weight, g
Chemical analysis , %
Dry cake
CaO
Total
N.A.C. insoluble8
Water soluble
A.C. soluble13
SOa
Product acid
CaO
MgO
Fee03
F 2
Pa05
First wash
CaO
P205
Second wash
CaO
P£0S
1
138
45
45
228
685
131*
191
1010
704
136
191
1031
135-9
32-5
4.56
2.70
o.4o
0.64
37-4
0.09
0.02
0.04
0.07
0.25
1.25
0.12
3.60
0.12
1.03
2
147
52
52
251
700
144
207
1051
723
150
209
1082
139.6
31.6
2-75
1.56
0.39
0.71
36.1
0.10
0.02
0.04
0.08
0.30
3-49
1.18
0.12
3.60
0.11
1.69
3
107
136
93
336
186
138
175
499
213
156
178
547
132.4
33-1
7-50
4.91
1.14
0.70
34.0
0.08
0.04
0.11
0.22
0.54
10.7
5-76
0.08
9.92
0.09
2.07
4
66
133
91
290
116
132
182
430
133
152
190
475
153-1
30.5
4.34
1.50
1.45
1.55
37-0
0.08
0.05
0.15
0.26
0.70
12.0
3-74
0.08
12.12
0.10
5.24
5
46
18
23
87
697
154
195
1133
721
160
198
1079
140.3
31.6
2.05
1.02
0.32
0.62
40.6
0.12
0.02
0.06
0.13
0.34
3-70
1.05
0.15
3-78
0.14
1.43
6
50
24
28
102
64o
176
210
1026
658
184
212
1054
i4o.6
31.2
1.60
0.54
0.37
0.71
41.5
0.12
0.03
0.06
0.15
0.35
3.80
0.91
0.15
5-93
0.13
1.71
7
24
70
63
157
91
150
174
415
106
173
183
462
141.2
31-3
2.40
0.38
1.38
1.49
39.5
0.10
0.06
0.25
O.J9
0.8)
13.4
2.54
0.11
13.82
0.13
5-37
8
34
158
107
299
66
120
196
382
77
138
209
424
132.9
31.0
2.57
0.56
1.33
1.11
39.4
0.11
0.06
0.29
0.40
0.90
13.9
2.64
0.12
14.13
0.13
6.97
9
60
17
16
93
815
126
202
1143
847
130
201
1178
i4o.7
30.8
1.96
0-33
0.62
1.32
4o.7
0.12
o.o4
0.05
o.o4
0.33
3-45
0.92
0.14
2.94
0.10
0.06
10
58
16
17
91
810
127
204
1141
838
132
203
1173
141.6
30.3
2.05
0.25
0.69
1.51
39.1
0.12
o.o4
0.05
0.05
0.33
3.44
0.96
0.15
3.o4
0.10
0.09
11
61
89
65
215
136
131*
165
1+35
156
151
174
481
146.6
30.3
3-13
0.07
1.31
1.91
37-8
0.10
0.09
0.14
o.i4
o.So
12.1
2.67
0.12
12.25
o . rj.
4.59
12
68
141
104
313
109
136
412
125
154
177
456
151.1
29-9
2.55
0.05
1.44
1.94
38.2
0.09
0.09
0.15
o.o4
0.50
12.4
2.67
0.09
12.64
0.10
6.02
13
52
18
18
88
820
122
152
1134
846
126
193
1165
142.3
30.5
1.92
0.24
0.57
1.38
39-9
0.12
0.04
0.06
0.06
0.34
3.52
0.93
0.15
2.67
0.09
0.07
14
64
26
33
123
810
125
190
1125
835
128
189
1152
141.8
31.0
2.10
0.29
0.57
1.49
4o.6
0.13
0.04
o.c6
0.07
0.36
3-52
0.98
0.14
2.70
0.09
0.07
15
25
65
47
137
94
144
169
407
107
164
178
449
142.0
30.3
2.01
0.17
1-33
1.29
39-3
0.10
0.10
0.24
0.20
0.93
13-?
2.15
0.12
13.62
0.15
5.51
16
19
53
55
127
84
132
175
391
96
150
186
432
138.7
27.9
1.85
0.17
1.29
0.92
37-7
0.10
0.10
0.28
0.22
0.92
13-5
2.40
0.12
12.73
0.1=;
6.91
a tl.A.C. = neutral ammonium citrate insoluble (A.O.A.C. method).
" A.C. = alkaline citrate soluble (exclusive of water soluble).
-------
TABLE 11. EXTRACTION OF PHOSPHATE ROCK WITH WEAK SULFURIC ACID—CALCULATED RESPONSES
-j
•P-
Response calculated
Filtration rates
Product acid
Kg rock/ (hr Km*)
ms acid/(hr)(m2)
Kg rock paos/(hr)(m*)
Total filtrate
Kg rockAhrHm*)
ms acid/(hr)(ma)
Chemical responses
Rode components in cake, %
CaO
=Total
N.A.C. insoluble0
Hater soluble
A.C. soluble*
Fluids analysis, *.9
75-2
N.A.C. = neutral amncnium citrate insoluble (A.O.A.C. method).
A.C. = alkaline citrate soluble (exclusive of water soluble).
-------
TABLE 12. EXTRACTION OF PHOSPHATE ROCK WITH WEAK SULFURIC ACID—STATISTICAL EVALUATION OF RESPONSES
Response measured
Filtration rates
Product acid
Kg rock/(hr)(ms)
MS acid/(hr)(m2)
Kg rock PgOs/ChrXm2)
Total filtrate
Kg rock/(hr)(m2j
tf* acid/(hr)(m2)
Chemical responses
Percent of rock constituent
remaining in cake
CaO
PS0S
Total
N.A.C. insoluble11
Water soluble
A.C. soluble13
Fluids analysis, %
pe°s
Product acid
First wash
Second wash
CaO
Product acid
First wash
Second wash
Other rock components
solubilized, %
MgO
Overall
average
1114
3-7
352
357
3-1
93-62
12.50
3-92
4.10
5.48
8.10
7-97
3-05
0.105
0.123
0.115
101.26
47-37
59-37
Difference0 in responses due to variable indicated
A
266.5
1.36
5oTT
119-6
1.14
0.20
-4.38
-5-99
0.97
2.87
0.10
-0.29
-0.27
0.010
0.013
-0.008
5-57
2.26
B
886.2
2BTT
138.2
0^9"
-3-22
-6.81
-o'.31
-0.93
0.95
0.91
0.015
0.020
0.018
4.23
18.31
25T56"
c
613.7
193T3
-205.5
-0.77
405.
0.41
3.81
1-51
9.12
9T37
4^1
-0.020
0.010
-7-47
18.46
6.81
D
-4.4
-0.46
-2.9
-55-2
-0.56
-1.11
-2.64
riT5B-
0.34
0.41
0.29
i.2L
0.003
-0.003
-0.008
4.57
2.96
AB
-76.7
-0.99
-34.2
-72-3
-1.00
-0.45
4.38
4.35
0.21
-1.15
-0.34
-0.70
-0.010
-0.013
-0.010
-4.83
-1.31
1.34
AC
98.1
-1.01
24.4
-30.3
-0.82
-0.47
-2.11
-0.29
-0.89
0.23
0.60
1.12
-0.005
0.003
0.022
-3.03
-2.66
T571*T
AD
107.4
o.n
34.2
4.9
0.03
-0.77
2.12
2-32
-0.13
-o.4i
-0.18
-0.38
-0.34
-0.003
-0.005
-0.005
-5-37
-O.C6
-3.24
BC
515.1
-0.81
160.1
-0.5
-0.52
-3-79
-3.11
-1.32
-0.04
-1.00
Q.77
0.000
0.005
0.012
-1.23
9.34
BD
-59.6
-o.4o
-21.0
-4o.o
-0.47
-2.78
2.14
-0.50
-1.12
-0.11
-0.38
-0.25
0.003
0.003
0.000
4.47
1.14
5.86
CD
63.0
0.22
18.5
10.7
0.31
-0.90
-0.68
-0.70
0.20
-0.13
0.27
0.22
0.83
-0.003
-0.003
-0.002
-5.03
Least significant
differences
90
377.9
1.21
116.7
125-0
1.19
3-17
1.85
2.92
0-53
1.65
0.29
0.63
0.44
0.004
0.010
0.011
9.72
1.82
12.07
95
481.9
1-55
148.9
159.2
1.52
4.05
2-37
3-73
0.67
2.10
0.37
0.81
0.56
0.006
0.012
0.013
12.40
2-33
15.41
99
755-8
2.43
233-4
250.0
2.38
6.35
3-71
5.85
1.C6
3-29
0.59
1.26
0.88
O.OC9
0.019
0.021
19.45
3-65
24.16
Negative sign indicates increase in response at the higher level of the variable; underlined values are significant at 90Jt confidence level.
A = rock particle size, coarse versus fine; B = reaction temperature, 24°C versus 52°C; C = sulfuric acid concentration, 10$ versus 30$;
D = acid addition time, 1 hour versus 2 hours.
Jj N.A.C. = neutral ammonium citrate insoluble (A.O.A.C. method).
A.C. = alkaline citrate soluble (exclusive of water soluble).
-------
significant differences in averages were derived (90, 95, and 99% confidence
levels) and the responses then were evaluated on the basis of these least
significant differences.
There are several two-variable interactions that are significant (90$ or
greater confidence level) and these values are duly listed and underlined in
Table 12. However, they will be eliminated from the discussion for the sake
of brevity. Only responses to the four individual variables will be discussed.
Filtration Rates—
Filtration rates were calculated as units per hour per square meter of
filter area. The units were kilograms of rock, kilograms of rock P205, and
cubic meters of acid. Filtration rates expressed as weight of rock and weight
of rock P205 per hour per square meter responded to temperature and to acid
concentration; both higher temperature and stronger acid resulted in higher
rates. Neither of the other variables, rock grind or acidulation time, caused
significant differences. The slight difference in P205 content of the two
rocks did not cause a significant change in results. When the filtration rates
were expressed as volume of acid obtained, response was significant with all
variables except acidulation time. Tests with finely ground rock and those
made at the higher temperature filtered more rapidly. Tests made with stronger
acid filtered more slowly.
Since filtration rates expressed as kilograms of rock and kilograms of
rock P205 gave identical significance of responses, of these two, only the
results of calculations showing kilograms of rock per hour per square meter
are given. Values for volume of acid obtained are given also. Filtration
rate based on weight of rock was significantly responsive only to temperature
and acid concentration. Higher temperature increased filtration rates and
higher acid concentration decreased rates. Filtration rate based on volume
of acid was responsive only to acid concentration. Higher acid concentration
decreased rates.
Chemical Responses, Solids—
The solids, after washing and drying as described above, were analyzed
chemically, and calculations were made of the proportions of the rock
components CaO and P205 in the solids. Proportions of rock P205 in neutral
ammonium citrate-insoluble, water-soluble, and alkaline citrate-soluble forms
were determined. Statistical analyses showed that the proportion of CaO in
the solids responded significantly only to variations in temperature; higher
temperature resulted in lower recovery of CaO in the solids.
Total PS05 proportion in the solids (cake) responded significantly to all
the variables; lower P205 proportion resulted from the use of the more finely
divided rock and from higher temperatures and longer acid addition times. The
use of stronger acid increased the proportion of P205 in the cake. Total P205
averaged about 12.5$ of that present in the rock. The P205 in the cake was
made up of several forms of P205, and differentiation between these forms is
essential to understanding the process. Such differentiation is discussed
below.
-------
The proportion of P205 not extracted by the sulfuric acid may be assumed
to be that which appears in the cake as neutral ammonium citrate-insoluble
P205; an average of about *4 of the rock P205 was insoluble in neutral
ammonium citrate. Only two of the variables caused significant responses; the
more finely divided rock and the higher temperature resulted in lower citrate-
insoluble P205 content. Neutral ammonium citrate-insoluble P205 contents as
low as 0.8$ of the rock P205 were obtained in tests at the higher temperature
with the finer rock.
The proportion of the P205 in the solids in water-soluble form indicates
the effectiveness of the washing procedure. An average of about k% of the
rock P205 was left in the solids in soluble form; this indicates the need for
better washing. The use of the coarser rock and of the lower acid concentra-
tion decreased the proportion of water-soluble P205 in the solids. When using
these conditions, as little as 1.7$ of the rock P205 was left in the cake in
water-soluble form. Both the more finely divided solids and the increased
acid concentration made washing a more significant problem.
The proportion of P205 in alkaline citrate-soluble form (exclusive of
water-soluble portion) represents, to a fair degree of accuracy, that portion
present as dicalcium phosphate. Presumably, this represents P205 that was
extracted by the sulfuric acid but was reprecipitated in a water-insoluble
form not recoverable from the cake by water washing. This proportion averaged
about 5-5$ of the P205 in the rock and varied significantly only with varia-
tions in the fineness of the rock; the use of the more finely divided rock
resulted in the formation of greater proportions of alkaline citrate-soluble
P205. No reason for this response is immediately evident.
The proportion of P205 soluble in neutral ammonium citrate solution but
insoluble in alkaline citrate solution or in water should represent the
reprecipitated P205 more basic than dicalcium phosphate. This value may be
obtained by subtracting the neutral ammonium citrate-insoluble P205, the
alkaline citrate-soluble P205, and the water-soluble P205 from the total P20s'
The error of all the P205 analyses contributes to the variation in this value.
In the case of the current test series, 12 of the l6 values, so calculated,
gave negative results ranging up to 0.88$. Because of this indicated
inaccuracy of the data, no statistical analysis of these results was made.
Chemical Responses, Liquids—
True countercurrent washing was not used in these tests; the use of weak
phosphoric acid as the first wash solution would have complicated calculations
and would have extended the number of tests required. As described earlier,
the test procedure included a product acid filtration step in which liquid was
collected over the period prior to appearance of solid cake at the slurry
surface. At this point, filtration was stopped, the filtrate was isolated,
and the first portion of wash water was added to the surface of the cake.
Filtration then was resumed until solid cake again appeared at the surface at
which time filtration was stopped again and the liquid collected during the
first wash period was isolated. The second portion of wash water then was
added to the cake surface and filtration was resumed; the time of the first
appearance of solid cake was noted but the application of vacuum was continued
until "weeping" of the cake had ceased. The liquid from this step then was
isolated.
77
-------
All three liquid samples were submitted for chemical analyses to determine
their CaO and P205 contents. In addition, the MgO, Fe203, A1203, F, and S03
contents of the first filtrate were determined, also.
The P205 content of the liquid samples varied within and between the
process steps represented. The P205 contents of all three filtrates varied
with the strength of the siilfuric acid used to acidulate the rock; higher
sulfuric acid concentrations resulted in higher P205 contents as expected.
Higher P205 contents in all three filtrates resulted from the use of the higher
extraction temperature, also. Higher P205 contents in the product acid and in
the second wash filtrate were obtained as a result of the longer retention
time; it is not obvious why this variable did not affect the P205 content of
the first wash filtrate as significantly. The average P205 content of the
product acids was 12.66$ for the 30$ sulfuric acid tests and 3.5^$ for the 10%
sulfuric acid tests. The average P205 content for the first wash filtrates
was 12.65$ (30$ H2S04) and 3.28$ (10$ H2S04). Second wash filtrate P205
analyses averaged 5.3^$ (30$ H2S04) and 0.77$ (10$ H2S04).
The CaO analyses of all the liquid samples were very near the same value.
All values were within the.range 0.115 ± 0.035$. Consequently, a high degree
of significance was attributed to variations in average analyses as low as
O.OCA-5$, a value smaller than the precision of the analyses. The indicated
responses, therefore, may not be as significant as indicated. Greatest varia-
tions were caused by variations in sulfuric acid concentration and test
temperature. Proportion of CaO was greatest at the lower acid concentration
and highest temperature. The proportion of the rock CaO dissolved in these
tests is much greater than would be expected from processes producing stronger
acids. A correlation of CaO content and acid concentration will be given later.
Evaluation of the proportions of MgO, Fe203, and A1203 extracted were
based on comparisons of the ratio of those components to P205 content in the
rocks with the ratio of those components to P205 derived from the rocks in the
strong acid filtrates. Essentially all the MgO was extracted; the average
value for all the tests was 101.26$, the range was 90 to 130$. Analytical
values were so small that precision was a major problem. None of the variables
affected the proportion of MgO extracted at the 90$ confidence level.
An average of about *4-7$ of the Fe203 in the rocks was extracted. The
proportion of Fe203 extracted was responsive to changes in temperature,
sulfuric acid concentration, and time used in the tests. Increasing each
increased Fe203 extraction. Response to rock grind, although significant
statistically, was based on variations in analyses less than the precision of
the analyses.
An average of about 60$ of the A1203 in the rocks was extracted. The
proportion of A1203 was varied by changes in rock grind and temperature. Fine
grinding of rock resulted in reduced extraction. This does not appear reason-
able and may indicate an erroneous rock analysis although the reported A1203
content of the rock (1.1$) appears reasonable. The most effective of the
variables was temperature; higher temperatures resulted in greater A1203
extraction.
-------
In addition to the tests performed as a part of this factorial experiment,
individual tests were made using 5, 15, and 60$ H2S04 at the same conditions
as the tests reported in positions 1^ (lO°i) and 16 (30$). From the results of
these tests, a chart was prepared (Figure 25) to show the effects of acid
concentration on extraction of the various components of the rock. The values
plotted are the weight ratio of the extracted metal oxide to extracted P205
versus the H2S04 concentration used to extract the rock. The plot shows little
effect on the proportions of MgO or A1203 extracted. Extraction of Fe203 was
minimized by reducing H2S04 concentration; this same condition maximized CaO
solubilization.
Conclusions
It is concluded that beneficiated phosphate rock may be extracted to a
reasonably high degree (above 99% of P205 in citrate-soluble form) with
sulfuric acid of low (5-30$) concentration and that extraction is improved
with increases in temperature and fineness of rock grind. Increasing acidula-
tion time from 1 to 2 hours did not improve P205 extraction or filtration
rates significantly. Overall filtration rates, expressed as kilograms of rock
per hour per square meter, were improved by increasing the temperature and by
decreasing acid concentration. The average filtration rate was 357 kilograms
of rock per hour per square meter of filter surface; the range was I^k to 595
kilograms of rock per hour per square meter. Extraction of Fe203 was minimized
without increase in extraction of MgO or A1203 by use of weaker sulfuric acids.
However, the solubility of CaO in the acid increased as the acid concentration
was decreased, and any subsequent thermal concentration probably would result
in troublesome precipitation of calcium sulfate. The increased solubility of
CaO emphasizes the need for methods of concentration not dependent on evapora-
tion of water.
PRODUCTION OF WET-PROCESS ACID FROM UWBENEFICIATED MATRIX
During the beneficiation of phosphate rock matrix to a salable grade of
raw material for the production of wet-process phosphoric acid, as much as
of the phosphate in the matrix may be lost to slime waste. This loss of
valuable material and the associated problems of slime disposal are areas of
increased concern for the phosphate industry. The use of unbeneficiated
phosphate rock matrix for the production of wet-process acid should reduce the
loss of phosphate and eliminate the necessity for slime disposal ponds. Weak
sulfuric acid (< 30$ H2S04) might be available from some pollution control
processes, including those utilizing carbon adsorption-water elutriation.
Test Plan and Equipment
A factorially designed experiment has been made in which two different
samples of Florida phosphate matrix were extracted with dilute sulfuric acid
to produce phosphoric acid by the wet process. Acid proportion used was 1 mole
H2S04 per mole of CaO in the matrix. The experiment was made to determine the
effects of different matrix samples, matrix treatment, acid concentration, and
temperature on the extraction of various matrix components. The test parameters
are shown below; the full test plan is shown in Table 13.
79
-------
0.1C
0
10
20 30
Concentration of H2S04 in extractant,
Figure 25. Effect of concentration of H2S04 on dissolution
of metal oxides from phosphate rock.
80
-------
TABLE 13. UTILIZATION OF LOW-GRADE PHOSPHATES--TEST PLAN
Test Test
position No.
1 1
2 3
3 4
4 6
5 7
6 5
7 2
8 8
9 IB
10 3A
11 4B
12 6A
13 7A
14 5A
15 2B
16 8A
Matrix
sourcea
FM30
FM30
FM30
FM30
FM30
FM30
FM30
FM30
FM20
FM20
FM20
FM20
FM20
FM20
FM20
FM20
Matrix
treatment
Raw
Raw-
Raw
Raw
Calcined
Calcined
Calcined
Calcined
Raw
Raw
Raw
Raw
Calcined
Calcined
Calcined
Calcined
Variables
H2S04
concentration ,_
10
10
25
25
10
10
25
25
10
10
25
25
10
10
25
25
Temperature .
?oc °C
Ambient
52
Ambient
52
Ambient
52
Ambient
52
Ambient
52
Ambient
52
Ambient
52
Ambient
52
Q
Phosphate matrix samples from two Florida fields:
FM30--raw
ppo5
11.8
FM30 — calcined 13-0
FM20--raw
18.1
FM20- -calcined l8.6
Chemical
CaO Fe-aQ
18.0 1.48
19-7 1.64
27-4 1.56
28.3 1.64
analysis , $
a Al^Ott MgO
2.6 0.44
2.9 0.48
2.2 0.67
2.4 0.68
F
1.4
1-3
2.1
2.0
Raw matrix--Florida phosphate matrix as received.
Calcined matrix--raw matrix calcined at 982°C for 90 minutes.
Actual concentration used:
FM20 matrix, 9-77 and
FM30 matrix,9.34 and 25-15^ H2S04;
H2S04.
81
-------
Level
Matrix sourcea
Matrix treatment
H2S04 concentration, $
Temperature, °C
-
FM30
Raw
10
Ambient
+
FM20
Calcinedb
25
52
a. Chemical analyses of matrices given in
Table 1J.
b. Calcined at 982°C for 90 minutes.
The equipment used was the same as that used in previous extraction tests
with weak acid described earlier. Also, the test procedure was the same, with
the following exceptions. The agitator speed was increased to 500 rpm for all
tests and the volume of the prewet acid was increased to 260 milliliters for
the tests in which 25% sulfuric acid was used. These changes were required
because of the consistency of the reaction mixture during the acidulation
period resulting from the larger amount of clay in the matrix as compared with
beneficiated rock. Also, because of the long filtration times encountered in
preliminary tests (as long as 1 hr for acid separation and as long as k hr for
wash separation), each acidulated mixture was poured into a 1-liter graduated
cylinder immediately after the final 15-minute agitation period and allowed to
settle until separated. After the volume of the solid layer became essentially
constant, the clear liquor was siphoned off. The solid layer then was elutri-
ated with water twice, each time filling the graduate to a total volume of 1
liter and allowing the solids to settle to approximately the same volume as in
the initial separation. The solids were then transferred to the 100-mesh
screen filter and allowed to pull dry under a 38 centimeter mercury (15 in Hg)
vacuum before being oven dried at 52°C. The weighed and dried filter cakes
were submitted for chemical analysis. The observed test data, the analyses of
the dried cakes, and the calculated proportions of the matrix components
extracted or left in the cakes (chemical responses) are given in Table Ik. A
large proportion of the product acid was derived from recycled reagent-grade
acid rather than from the matrix; for this reason, the evaluation was made
from cake analyses.
Evaluation of Data
Evaluation of the test results was made using the statistical analysis of
the responses. The values given in Table 15 are the overall averages of the
responses for each variable and the differences in responses due to changing
the level of the indicated independent variable. Originally, it was assumed
that the three- and four-factor interactions would be highly unlikely and
therefore could be combined to form an estimate of the error; however, it was
found that the three-factor interaction of matrix source, matrix treatment,
and acid concentration (ABC interaction) was significant. Thus, that inter-
action was not combined with the error in the final analysis.
82
-------
TABLE Ik. UTILIZATION OF LOW-GRADE PHOSPHATES—TEST RESULTS
00
Test position
Primary data
Separation to recover HgPOt
Tine allowed to settle , hr
Volume clear liquor, ml
Volume solid layer, ml
Weight dry cake, g
Chemical analysis of dry cake , $
CaO
MgO
Fe^
AljOa
SOg
P205
Total
N.A-C. insoluble8
Water soluble
Chemical responses
Cake components , $ of that
present in matrix
CaO
MgO
Fee03
AljOg
PJ.OS
Total
N.A.C. insoluble0
Water soluble
Reprecipitatedb
P20s extracted,0 *5
210 205
103.1 102.1*
13.0 12.0
0.03 0.02
0.1*1 0.52
0.27 0.18
18.1 16.8
0.9 0.7
< 0.1 < 0.1
0.5 o.i*
7U. 5 68.3
7.0 l*.7
28.6 36.0
10.7 7-1
7.9 6.1
0.1* O-1*
U.It 3-5
J.I 2.2
99.6 99.6
5
27
l*8o
385
UU.I*
ll*.6
0.08
0.69
0.26
20.0
1.1
0.2
0.5
81* .6
19.2
1*8.1
10.2
9.7
1.8
U.lt
3-5
98.2
6
94
105
275
109.9
15-7
0.07
0.5!*
0.17
21.6
0.7
0.2
0.3
87.6
16.2
36.2
6.4
5-9
1.7
2.5
1.7
98.3
7 8
1*1* 51
87 375
320 275
113.1* 112.1
li*.i* l"*.8
0.08 0.05
0.65 0.57
0.1*1 0.20
19.8 20.2
1.5 1.1
< 0.1 < 0.1
0.7 0.7
82.9 8i*.2
19.1 11.8
1*5.0 39.0
16.0 7-7
13-1 9.5
0.1* 0.1*
6.1 6.0
6.6 J.I
99.6 99.6
9
69
355
235
117.0
20.9
0.05
0.57
O.ltO
28.0
1.6
0.3
0.8
89-1*
8.7
1.1.7
21.3
10.3
1.9
5-2
3-2
98.1
10
1*6
290
290
118.7
20.2
o.oi*
0.67
0.32
27.9
1-3
0.05
0.7
87.7
7-1
1*9-7
17-3
8.5
0.3
l*.6
3.6
99-7
11
55
10
i»35
112.8
21.1
0.02
0.13
0.17
30.1
2.1
0
1.8
87.0
3-1*
9.2
8.7
13.1
0
11.2
1.9
100
12
76
80
330
109.2
21.5
0.03
0.15
0.20
29.7
2-7
0
2.3
85.8
k.9
10.2
9.9
16.3
0
13-9
2.1*
100
13
28
120
500
128.7
21.0
0.22
0.59
0.39
28.5
2.7
0.3
1.8
95-5
1*1.6
1*7-5
20.9
18.7
2.1
12.1*
l*.2
97.9
11*
25
230
385
124.6
21.7
0.23
0.69
0.36
28.6
1.6
0.8
0-5
95-5
1*2.1
53-7
18.7
10.7
5-1*
3-3
2.0
94.6
15
50
1*0
1*00
1U5.9
19.7
0.03
0.16
0.20
26.5
2.1*
0
2.1
81.1*
5-2
11.7
9-7
15-1
0
13.2
1.9
100
16
1*1*
U5
36o
120.8
18.9
0.10
0.37
0.22
25.9
2.2
0.05
1.9
80.7
17.8
27.9
11.1
I1*. 3
0.3
12.3
1.7
99.7
N.A.C. insoluble = neutral ammonium citrate insoluble (AOAC method).
Extracted P-,05 calculated as 100 - (N.A.C. insoluble P205).
-------
TABLE 15. UTILIZATION OF LOW-GRADE PHOSPHATES—EFFECTS OF VARIABLES
Differencea in average of responses due to variable*1 indicated
Matrix Matrix H2S04 Tern-
Overall source treatment oonc'n perature
average (A) (Bj (C) (D)
(AB) (AC) (AD) (BC) (BD) (CD) (ABC)
Chemical responses
Matrix components in cake
% of input
CaO
Pa05
Total
N.A.C. insoluble0
Water soluble
Reprecipitated
Matrix components solu-
bilized,6 % of input
MgO
Fe203
A12Q3
P205 extracted, % of
input matrix P205
a Negative sign indicates
>
84.1
11.2
1.1
6.9
3-2
84.3
62.0
85.9
98.9
7-46
4.40
0.23
5.26
-1.09
-1.20
13.18
-1.11
-0.23
4.81
1.90
0.75
1.29
-O.l4
-11.75
-1.20
3-11
-0.75
a decrease in response at
-7.09
1.50
-1.80
3-89
-0.59
13.03
24~lH
~B~^g
1.8
0.04
-2-55
-0.05
-1.36
-1.14
2.38
-1.13
4.54
0.05
the higher level of
-4.01
0.75
0.65
0.29
-0.19
-8.90
ziii
-0.65
-1.21 -0.94
1.15 0.70
-0.55 0-55
2.39 -0.61
-0.69 0.76
4.03 -5.63
9.23 ^75
T^f -3.64
0.55 -0-55
-1.41 0
0.25 -1
-0.68 0
-0.14 -1
1.06 -0
3.28 -3
-8.70 0
0.68 -o
the variable; underlined values
.86
•50
•93
.64
.79
.08
.00
• 31
•93
are
-1.74
1.80
0.13
1.56
0.11
-3-50
-3-53
-2.21
-0.13
-4.74
-2.90
^x
-0.58
-1.24
-1.09
10.03
6.10
4.84
0.58
significant
at 90$ confidence level.
A = rock source
FM30(-) vs. FM20(+)
B = matrix treatment
raw(-) vs.
calcined( H
C = H2S04 concentration
i-) 10!
N.A.C. insoluble = neutral ammonium citrate insoluble (AOAC
d Reprecipitated apatite
P20s calculated t
is (total I
t(-) vs. 2'.
method) .
D = reaction
temperature
$(+) ambient(-) vs. 52°C
'2°5) - (water-soluble PoOg)
(+)
- (N.A.C. insoluble PoCt).
e Calculated from cake analyses.
-------
Thus, using this esimate of error, the responses for three of the
dependent variables were rejected because of a lack of significant difference
between the error and the model. Those dependent variables are the neutral
ammonium citrate-insoluble P205 and the water-soluble P205 in the cake and the
P205 extracted from the matrix. The models for all other dependent variables
were found to be significantly different from the error, and therefore their
responses which produced significant differences (at least 90% confidence)
between the levels of each of the independent variables and their interactions
are duly underlined in Table 15.
There are several two-factor and one three-factor interactions which are
significant and those results are listed in Table 15. However, in the interest
of brevity they will be omitted from the discussion. Only the results of the
four individual independent variables will be discussed.
Chemical Responses, Solids--
The proportion of CaO left in the cake (an average of 8lul$ of input) was
affected by three of the independent variables. More CaO was left in the cake
(a desirable condition) when the matrix FM20 was used and when the matrix was
calcined before extraction. Less CaO was left in the cake when the more
concentrated acid was used.
The quantity of total P205 remaining in the cake averaged 11.2$ of input.
Its removal was enhanced with the use of matrix FMJO, with the use of
uncalcined matrix, and with the higher level of temperature. The greatest of
these effects was due to the matrix source. It is noted, however, that the
total P205 content of the cake includes all forms of P205 and therefore is
subject to variations in extraction, washing, and reprecipitation. In the
tests conducted (Table ih) lowest retention in the cake was 5-9$ in test
position 6 (-A, +B, -C, +D).
As previously stated, the responses for neutral ammonium citrate-insoluble,
water-soluble, and the extracted (citrate-soluble) forms of P205 were rejected
because of their lack of significance. The acidulation procedure extracted
the P205 so completely (average 98.9$) that the differences between the tests
were insignificant. In the separation of the solids, however, the removal of
solubilized P205 from the cake varied widely; solubilized P^s1 not removed
averaged about 10$ of the matrix P205. The values were so distributed, however,
that the model did not define the causes adequately. However, values for the
reprecipitated P205 (citrate-soluble but water-insoluble P205) did show a high
probability of being significant (91»-.8$ confidence in the model). If this
degree of confidence is accepted, the amount of P205 precipitated was minimized
both with the use of matrix FM20 and with the higher temperature. In the tests
made (Table 1*0, the average amount reprecipitated was 3.2$ of the P205 in the
matrix; minimum precipitation (l.7$) was in test positions 6 and 16 (-A, +B,
-C, +D and +A, +B, +C, +D).
Solubilized P205 not removed = (total P205) - (N.A.C. insoluble P205),
85
-------
Further evaluation, based on solids analysis, was made to determine the
proportions of the other rock components solubilized. The solubilization of
MgO, a contaminant in product acids, averaged 84$ of input for all the tests.
Its solubilization was reduced with the use of the weaker acid and also with
calcination of the matrix. In the tests made (Table 1*0, minimum solubiliza-
tion (57.9$) was in test position 14 (+A, +B, -C, +D). The other variables
did not produce changes significant at the 90$ confidence level; the three
significant two- and three-factor interactions are noted.
The solubilization of Fe203, another contaminant in product acid,
averaged 62$ of input for all the tests. The use of matrix FM30 reduced the
solubilization of this species. A greater effect was caused by the sulfuric
acid concentration. An average of 50$ of the input Fe203 was solubilized when
the more dilute (10$) acid was used while 74$ was found in the product made
with the more concentrated (25$) sulfuric acid.
A third common contaminant of wet-process acid, A1203, was removed from
the matrix at an average of 86$ of input for all the tests. Its solubilization
was decreased by using the weaker acid and by the use of the lower (ambient)
temperature with the greater effect due to acid concentration. In the tests
made (Table 14), the minimum solubilization (66$) was in test position 1
(-A, -B, -C, -D).
In the production of wet-process phosphoric acid, the product acid
concentration depends on the amount of P205 extracted and on the water balance
in the system. If good extraction is attained and if washing is good (nearly
all interstitial liquor in the cake is water and little or no wash water in
the recycle acid), the product acid concentration will be dependent on the
concentration of the sulfuric acid, the amount of water evaporated, and the
amount of water removed by hydration of the calcium sulfate.
For these tests, a product acid concentration was calculated assuming
complete extraction and perfect washing. As mentioned earlier, when 10$ H2S04
was used, the phosphoric acid concentration was calculated to be about 5.7$ or
14-.!$ P205; with 25$ H2S04, the values would be 21.4 and 15-5$, respectively.
In tests made with 10$ H2S04, the actual product acid averaged 6.39$ H3P04 and
with 25$ H2S04, 23.87$ H3P04. A large portion of these products, however, was
derived from the starting recycle acids.
The product acids would contain an average of about 16$ of the CaO, 62$
of the Fe203, 86$ of the A1203, and 84$ of the MgO originally present in the
rock. These values are based on the observed proportions of these components
in the filter cake. An acid produced from a blend of the matrix samples and
the average sulfuric acid concentration used in the current tests would have
the following analysis: P205, 9.8$; CaO, 2.3$;Fe203, 0.62$; A1203, 1.38$; and
MgO, 0.3$. The ratios of impurities to P205 are greater than in commercial
acid because of the greater solubility in the weak acid, because of the higher
ratios in the matrix, or both.
86
-------
These current tests may be compared with the previously reported tests
which were also made in a factorially designed experiment but which utilized
a normal grade rock. Two of the independent variables studied, temperature
and acid concentration, were the same in both series of tests (rock grind and
acid addition time were the other variables in the earlier tests). Previously,
it was found that temperature produced significant responses in more of the
dependent variables than acid concentration; in the present tests the opposite
was true. However, where significant responses coincide for both experiments,
the directions of the responses were the same. In both experiments, decreasing
the sulfuric acid concentration decreased the solubilization of Fe203. Like-
wise, decreasing the reaction temperature decreased the solubilization of
A1203 and the recovery of total P205.
Conclusions
The results of this study indicate that a large increase in the recovery
of P20s from phosphate rock would result from the acidulation of the raw matrix
rather than flotation concentrate. Large amounts of P205 may be lost in the
flotation process. An average of about 99% of the P205 in the matrix was
solubilized in these tests. There was an increase in sulfuric acid requirement
but that acid would be recovered from flue gas desulfurization processes.
Advantages accruing to the use of weak sulfuric acid to acidulate raw
matrix include lower solubilization of MgO, Fe203, and A1203 from the rock
which is highly desirable. Calcination of the rock further reduced MgO
solubilization. Maintaining low temperatures during extraction reduced A1203
solubilization. The proportion of Fe203 solubilized was a function of the
source of the rock also.
Disadvantages are concerned with the poor filtration rates and the need
to dispose of large amounts of solids. A settling system ahead of the filter
might be required.
It was concluded that the production of phosphoric acid from raw phosphate
matrix and weak sulfuric acids at or near the source of the weak acid would be
technically feasible.
CONCENTRATION OF PHOSPHORIC ACID BY FREEZING
The use of weak sulfuric acid to extract phosphoric acid from phosphate
rock will result in the production of relatively weak phosphoric acid
(generally < J0$ H3P04). The weak phosphoric acid would be a more useful
product if it were concentrated to filter-grade acid (about kO$> H3P04) or,
even better, to a shipping-grade level (about T0$ H3P04).
Experimental Work
A short literature search indicated that some energy savings over usual
heating methods of concentration might result from freezing of water from the
acid and mechanically removing the ice from the acid mother liquor. Laboratory
investigations have been made of the concentration of industrial wastes (8) by
87
-------
freezing and a process for desalination of seawater (9, 10) by freezing is
close to commercialization. Both procedures are reported to function well.
Some energy might be saved since freezing water requires removal of 80 calories
per gram while evaporation requires the addition of 539 calories per gram.
Published data on the freezing point of aqueous solutions of phosphoric
acid are given in Figure 26. A brief investigative program was started to
study the problems which might develop from attempts to concentrate acid by
freezing. Initial attempts indicated that a primary problem would be the
recovery of large amounts of H3P04 left in the ice as interstitial liquor.
The removal of this liquor, probably the strongest acid in the system at the
time of ice formation, would be necessary. Early tests indicated that the
amount of interstitial liquor retained was approximately equal to the amount
of ice formed. A practical approach appeared to be to attempt concentration
by 1.5 times (from 10-15$ H3P04); this should result in approximately one-
third each of ice, interstitial liquor, and product acid.
The interstitial liquor problem might be minimized by washing ice on a
filter with a minimum amount of water at 0°C to remove acid without excessive
dilution of the acid. A few exploratory tests were conducted to check this
procedure.
An 8-liter cold bath capable of operation to -50°C was filled with a
commercial antifreeze solution of ethylene glycol and water and was set at
-U°C. A closed vessel containing about 100 grams of 10$ H3P04 (prepared from
reagent H3P04) was placed in the cold bath and was cooled to -^°C. No
crystallization was apparent until the sample was seeded with a few small
crystals of ice; it then was allowed to equilibrate overnight before it was
filtered using the laboratory vacuum system. There was 50.5 grams of filtrate
obtained indicating that warm air leaking through the vacuum filter probably
caused some melting of ice and washing of interstitial acid into the product.
Titration with 0.2 N NaOH indicated the product contained about lk.Q% H3P04.
The remaining ice was melted, weighed, and analyzed. Results indicated that
about kk.l grams was obtained which contained about 5.1$ H3P04. A calculated
9k.k% of the acid was recovered with 76.9$ of the recovered acid being in the
filtrate.
A second sample was placed in the bath with a small glass propeller-type
stirrer in the sample (stirrer operated at about 100 rpm). The intent was to
cool the sample rapidly to about -1°C and then at a rate of about 0.5°C every
5 minutes to -U°C. The intended rate was attained to about -2°C when the
sample was seeded (with ice) and equilibration begun. A difference in
temperature between the cold bath and sample of only about 0.5°C was sufficient
for the initial cooling. Afterward, the time required for the sample to
equilibrate and cool an additional 0.5°C was over 30 minutes in spite of an
increase in difference between the sample and bath of over 2.75°C. At least
an equal period was required before the sample reached -U°C with the tempera-
ture differential increased to as much as 8°C at times. The ice clung to the
sample container and could not be poured onto the filter. Only 12-5 grams of
filtrate was obtained. No analyses were made.
-------
-90
0
10 20
30 kO 50
Percent H3P04
TO 80
Figure 26. Freezing point of phosphoric acid, (n)
89
-------
The test was repeated with the exception that the ice was scraped into
the filter, the filter was maintained at test temperature by immersion in the
cold bath, and a single wash with 30 grams of ice water was used. The glass
stirrer was operated at 500 rpm until crystallization started and then the
speed was increased to 1000 rpm. The sample was cooled to -1°C rather rapidly.
The bath was then lowered to -2.5°C (0.5°C below the expected ice point) in
about 5 minutes and was held there until ice remained frozen after seeding; the
batch temperature was then kept at -3.3°C until the sample appeared to equili-
brate at about -2°C (an additional 20 minutes was required). The bath tempera-
ture was then decreased to maintain a maximum differential temperature of less
than 2°C in an attempt to increase crystal size. Movement of ice in the sample
became extremely slow when the sample reached -2.8°C and cooling was stopped
when the sample reached -3°C (essentially no movement of ice). The sample and
ice were transferred to a filter and a slight vacuum was applied; filtrate
weight was 43.0 grams. A 30-gram ice-water wash was then applied from which
apparently some water was frozen since only 16.2 grams of wash solution was
recovered. The ice was then allowed to melt and was transferred to a sample
bottle; 68.1)- grains of solution was obtained. Titrations of the portions indi-
cated the filtrate contained 14.9$ H3P04, the wash contained Ik.3%, and the ice
contained 2.4$ H3P04. Thus, 6l.8$ of the acid was in the filtrate, 22.4$ in
the wash, and 15.8$ in the ice.
The test was repeated using three 30-milliliter portions of ice water as
wash. The cooling and stirring rates were approximately the same as in the
test above. The results obtained are shown below.
Filtrate Wash 1 Wash 2 Wash 3 Ice
Weight, g 41.9 16.4 3^-3 27-6 64.7
% H3P04 14.2 12.4 4.2 1.2 0.5
$ of recovered H3P04 59-2 20.2 14.2 3-2 3-2
Indicated % of input H3P04 60.6 20.7 l4.6 3.2 3-3
Again a portion of the first wash may have been frozen by the ice (recovered
only 16.4 g of 30-g wash). It might be possible to wash the ice in a counter-
current procedure and add the first wash filtrate to the product acid. A proc-
ess can be envisioned which includes a series of such countercurrent washes so
that the overall acid recovery might be acceptable. The recovery in each step
would diminish rapidly, however, and further washing would soon become -uneconom-
ical. In the above example with three water washes, the recovery of acid as
product and wash solution was nearly 97$; this recovery might be approached if
the washing were done with decreasing strengths of acid rather than with water.
A projected flowsheet for a three-step concentration process is shown
in Figure 27. Concentrations of the wash liquors shown in this figure are
estimated; actual values depend on the washing efficiency, the amount of ice
melted or water frozen, and the concentrations of the incoming streams. The
concentration of the liquor in the freezing steps would depend on the tempera-
ture attained, the degree of supercooling resulting from the procedure, and
the effects of impurities on the freezing point of the mixture. First-stage
concentration would be from 10 to 15$, as studied in the earlier tests.
Second- and third-stage concentrations would be from 15 to 22$ and 22 to 33$,
90
-------
237-5 g H20
A = AGIO
I = Ice
WF = Wash filtrate
337.5 g
10$ H3po4
STEP 1
10 to
H3P04
1
'
— ~
112.
225
5 g l^
g A
V
112.5
112.5
Rl.t
K A |
STEP 2
15 to 22.%
H3P04
STEP 3
22.5 to
33.75$
H3P04
112.5 g A
112.5 fi WF
75 g I
150 g A
75 g A
75 g I
T
75 g A
75 g WF
L50g
50 c I
100 g A
•75$ Jj
50
50 g A
237.5 g
%
H3P04
237.5 g I
237-5 g WF
237.5 g I
237.5 g W
237.5 g
H3P04
237.5 g I
237.5 g W
237.5 g
H3P04
g
125 g I
125 g WF
125 g
22.5$
H3P04
, 50 g
50 g I
50 g WF
33
H3P04
50 g A 50 g WF
"75$
H3P04
PRODUCT
100 g 33$
Figure 27. Projected flowsheet for concentration of
dilute phosphoric acid by three-step batch freezing.
91
-------
respectively. Concentration higher than 33$ would require temperatures lower
than are considered practical. One test was made of concentration over each
of the three ranges. For each test, fresh acid feed was made to the desired
concentration by diluting reagent-grade phosphoric acid. The filter cake
washing procedure was varied somewhat from that indicated in Figure 27, as will
be discussed.
The procedure in each of the present three tests comprised the following
steps. Phosphoric acid at the initial concentration was stirred constantly
and cooled rapidly in a bath to a point near its predicted freezing point.
The bath then was cooled slowly until the acid reached the predicted freezing
point, at which time it was seeded with a small amount of ice. After ice
formation was initiated, the agitation rate was increased and the bath was
cooled further until the acid-ice mixture reached the predicted freezing
point of the desired product acid or until the mixture became too viscous to
stir. Care was taken to prevent a large temperature differential between the
bath and the acid-ice mixture; a 3°C differential was the maximum allowed.
Usually, about 1 hour was utilized in cooling the acid sufficiently to result
in a 50;: increase in concentration. The acid-ice mixture then was filtered on
a medium fritted glass funnel which had been cooled in the same bath in which
the ice was formed. The ice cake subsequently was washed with ice water or a
series of weak phosphoric acid solutions as described later. After washing,
the ice was allowed to melt and the strong filtrate, washes, and melted ice
were separately weighed and analyzed titrimetrically to determine the
concentration and distribution of the phosphoric acid.
Data from the tests are given in Table 16 and Figure 28. The figure
shows the temperature-time relationships of the bath and the acid for the
three tests in which the aim was to concentrate acid from 10 to 15$, to 22-5$,
and finally to 33*75$ H3P04. The freezing points of the acid at each concen-
tration and the point at which seed ice was added are shown also. The table
gives the initial concentration (intended) and weights of the acids and the
temperatures at which they were seeded and to which they were cooled after
seeding. Also listed are the compositions and weights of the wash solutions
and of the various products. A percentage distribution of phosphoric acid
in the final products is given also.
In test 1, approximately 10$ H3P04 was cooled rapidly to -1°C and then
cooled slowly to -1.94°C before being seeded with a few crystals of ice
scraped from the surface of equipment extending above the liquid level in the
cold bath. Ice began to form in the solution and then to grow in quantity
until the test was stopped with a sample temperature of -3.06°C and a bath
temperature of -4.17°C. At this time, the ice-acid slurry was transferred to
the cold filter (immersed in the same cold bath) and filtered under low vacuum
(estimated at 71 cm Hg absolute) to recover 43 grams of acid having a concen-
tration of 14.9$ H3P04. The remaining ice (67 g, including interstitial
liquor) then was washed with 30 grams of water at 0°C on the same filter and
16.2 grams of filtrate was recovered which analyzed 14.3$ H3P04. The remainder
of the ice-liquor cake was allowed to melt; 68.4 grams of liquor was recovered
which contained 2.4$ H3P04. In the test, 127.6 of the 130 grams of input
weight was accounted for (98.15$ of total). The accounted for phosphoric acid
was distributed so that 6l.8$ was in the first filtrate, 22.4$ was in the wash,
92
-------
TABLE 16. LABORATORY-SCALE BATCH TESTS OF CONCENTRATION
OF DILUTE PHOSPHORIC ACID SOLUTIONS BY FREEZING
rppt-t Mn
Freezing
Feed acid
Cone . , % H3P04
Weight, g
Temperature , °F
Precool
When seeded
Final
Filtration
Step
Wash
Cone . ,
Weight
Products
Cone . ,
Weight
% H3P04
> g
recovered
% H3P04
, g
% of input8
^ i
9.85
100
20.2
28.5
26.5
Filtrate Wash
0.0
30.0
14.9 14-3
43 16.2
61.8 22.4
0
15.23
100
28.8
25-3
19.0
"Ice"
-
-
2.4
68.4
15-8
Filtrate
.
-
22.9
18.4
18.5
Wash
15.23
30.0
22-5
24.4
24.1
Wash
9.85
30.0
19.2
28.0
23-7
"Ice"
_
-
9.0
_
33- 7b
2
22.45
100
20.8
18
15.0
Filtrate Wash
22.45
30.0
29.1 29-1
?1.1 27.3
15.2 19.6
Wash
22.45
30.0
25-8
38.5
24.6
Wash
15.23
30.0
21.8
28.7
15.5
"Ice"
-
-
12.5
25- lb
Total input based on H3P04 in initial acid and washes.
Calculation based on H3P04 input less that recovered as product and wash liquors.
-------
Ice addition
Sample
-15.0
o
Ice addition
r
Freezing point, 10$ H3P04
20
TEST 2
• 22.5$ H3P04
Ice addition
Sample F.P.,
-•—22.5$ H3P04
TEST
22.5—
H3P0
bo 60/0 20 IK) V o 20 IK> 60
Time, min
Figure 28. Cooling curves for weak phosphoric acid samples in freeze-concentration tests
H3P04
-------
and 15•8$ was in the final cake. Obviously, a more strenuous washing procedure
would be required, possibly one in which the cold water would be used in two
or three countercurrent washing steps. The projected flowsheet (Figure 27)
indicates two countercurrent wash steps. The low recovery of wash solution
(l6.2 g of JO g input water) probably was due to the freezing of the wash
water when it contacted the colder ice (-3°C). This illustrates a problem of
operating the process in a batch manner; such problem should be nonexistent
or greatly reduced with use of a continuous countercurrent washing procedure
as discussed later.
Test 2 involved planned concentration from 15 to 22-5$ H3P04 using fresh
acid feed. In this test, the acid was seeded at -3.72°C and was cooled to
the planned -7.22°C. Filtration recovered l8A grams of 22.9$ H3P04; this
low recovery (18.5$) may have been due to the higher viscosity of the acid
which resulted in an increased interstitial liquor content. The ice cake then
was washed, first with a solution containing 15-23$ H3P04 (the same concentra-
tion of acid as was the beginning material) and then with a wash containing
9.85$ H3P04 (the beginning material from the previous step) to simulate a
countercurrent washing procedure as outlined in Figure 27- Filtrate from the
first wash contained 2k.k grams of 22-5$ H3P04 and that from the second wash
contained 28 grams of 19.2$ H3P04; both recoveries indicate the freezing of
some water from the wash solution due to the low temperature of the ice. The
final cake analyzed 9$ H3P04; it is noted that this cake was not water washed
as it would be in the projected procedure outlined in Figure 27- Recoveries
of phosphoric acid, based on total feed including that supplied in the wash
solutions, were 18.5$ in the strong acid filtrate, 2^.1$ in the first wash,
and 23.7$ in the second wash. Proportion remaining in the final cake was
33-7$.
Test 3 "was made in a manner similar to test 2. Fresh input acid containing
H3P04 (intended) was cooled rapidly to -6.22°C then seeded at -8.75°C
and cooled slowly to -9.hk°C. The large rise in temperature (Figure 28) on
seeding may indicate that the input acid was more dilute than thought. Concen-
tration to the intended 33-75$ final acid concentration would have required
cooling to about -15°C; however, the slurry became very thick at -9-^°C and
concentration had to be stopped. Increased viscosity of the acid probably
helped thicken the slurry. Filtration allowed collection of 21.1 grams of
29.1$ H3P04, washing with 30 grams of 22A5$ H3P04 allowed collection of an
additional 27.3 grams of 29.1$ H3P04, washing again with 30 grams of 22.^5$
H3P04 allowed collection of 38.5 grams of 25.8$ H3P04, and a third wash with
30 grams of 15-23$ H3P04 allowed collection of 28.7 grams of 21.8$ H3P04. The
final cake contained 12.5$ H3P04; in the projected procedure (Figure 27) this
cake would be further washed. Recovery of phosphoric acid, based on total
feed including that supplied as wash solution, was as follows: 15-2$ in
strong acid filtrate, 19.6$ in first wash, 24.6$ in second wash, and 15-5$ in
third wash. Proportion remaining in the final cake was 25.1$.
Conclusions
Results of the present series of tests indicate generally the workability
of a three-stage freeze concentration as outlined in Figure 27- Indications
are, however, that final acid concentration might have to be limited to 30$ or
95
-------
less, because of excessive thickening. Thickening might be greater also with
wet-process acid containing impurities. The data are not sufficient to predict
accurately the amount of acid that would remain in the ice discharged from the
three-stage process. For application to acid from a steam-plant scrubber
system, it might be feasible to recycle melted ice to the scrubber as makeup
water and thus recover all the acid. It is noted that good insulation would
have to be maintained to minimize the cost of refrigeration and to minimize
loss of ice due to melting when washing with solutions that have been allowed
to warm. Each solution would have to be maintained at a different temperature.
Consideration should be given to application of a fully continuous
procedure. Such a process might suffice to concentrate wet-process acid to
about 30$ H3P04 while eliminating the problem of transfer of ice counter-
currently to the preceding stage for washing and control of multiple wash
solutions at different temperatures. With the closer control possible with
continuous countercurrent washing, it is likely that the P205 lost in the ice
might be decreased significantly.
CONCENTRATION OF AMMONIUM PHOSPHATE SOLUTION BY FREEZING
In an alternative approach, it would be possible to neutralize dilute
phosphoric acid with ammonia before concentration by freezing. One advantage
that might result would be the removal of unwanted impurities from the acid
by partial ammoniation. The impurities would be precipitated and removed by
filtration or centrifuging; they would be utilized in the production of solid
or suspension fertilizers. Concentration of the filtrate would result in
liquid ammonium phosphate fertilizer of 7-21-0 or 8-2k— 0 grade. Another
advantage would be lower corrosion rates in the concentration equipment as
compared with the conventional evaporation process due to the lower temperature
and neutralization of the acid. A diagram of such a process is shown in
Figure 29.
Experimental Work
This approach was studied in the laboratory. The ammonium phosphate
solution used in the tests was k-lJ-0 grade made by ammoniation of diluted
reagent-grade phosphoric acid (about lii-^ P205). Initial tests were made by
.cooling the liquid sufficiently to obtain about equal proportions (by volume)
of solid and liquid phases, separating the mixture by filtration, and then
repeating the operation with the liquid phase until a total of five separations
had been made. These tests were intended to determine whether the nutrient
concentration in commercial fertilizers could be reached by freeze concentra-
tion of dilute ammonium phosphate solution.
Results (Table 17) indicated that freeze concentration of ammonium phos-
phate liquid to commercial grade might be feasible; the grade attained was
higher (8.5-27.8-0) than commercial ammonium orthophosphate solution (8-24-0).
However, the ice purity was relatively poor and it decreased as the concentra-
tion (N + P205) of the liquid fraction from the separations increased. For
example, the P205 content of the ice from the first separation (test l) was 6$
96
-------
Clean gas
1
Waste gas
containing
S02 ,
Activated carbon
adsorption unit
10-30$ H2S04
Impure
water
to S02
adsorption
unit
Wet-process acid
production
(5-15*
Freeze concentration
by countercurrent
washing process
Product (clear)
7-21-0 or
8-2^-0 grade
Acid purification
by ammoniation
(N:P205 wt.
ratio <^> O.j)
Separation
of impurities
by filtration
Impurities to
granulation or
suspension
fertilizer plant
Figure 29. Utilization of sulfur dioxide from stack gas in production
of ammonium phosphate liquid fertilizer--flow diagram.
-------
TABLE IT. TESTS OF FREEZE CONCENTRATION OF AMMONIATED
PHOSPHORIC ACID SOLUTION13--TEST DATA
00
Separation,
Test
No.
1
2
3
k
5
Crystallization _Jo by
temp., -°C
5.6
5-9
7-2
9.3
11.8
Solid
16
10
27
38
ko
wt.
Liquid
8U
90
73
62
60
Analysis, % in
Solid
N
1.9
1.8
2-5
3.2
h.6
PP05
6.0
5.8
8.0
10.6
15.0
Liquid
N P?05
h.V 1U.UC
U.7C 15-3C
5.5° I8.lc
7-0C 22. 7C
8.5° 27. 8C
Recovery, %
N + Pa05 in
liquid fraction
92.8
95-7
86.0
77-9
73.6
a Procedure: A 1000-gram batch of solution was placed in a closed container and
then cooled to the temperature required to obtain an appreciable amount of solids
which were separated by filtration. The procedure was repeated with the filtrate
portion until five separations had been made.
Starting liquid was ^-13-0 grade made by ammoniation of diluted reagent-grade
phosphoric acid (lU$ P205).
0 Calculated values.
-------
as compared with 15$ from the fifth separation (test 5). Guided by these
results, tests were made to study procedures for increasing ice purity.
A series of exploratory tests was made to study the effect of water
washing and of crushing the ice fraction prior to washing on the purity of
recovered ice. The results of these tests are given in Table 18.
Water washing of the ice fraction was beneficial to the ice purity, and
purity was further increased by crushing the ice prior to washing. Without
water washing, the ice fraction obtained in crystallization at -5.6°C contained
5.2$ P205 as compared with 3.7$ with water washing of uncrushed ice; crushing
of the ice resulted in a further increase in ice purity (contained only 2.8$
P205)-
In another series of tests, simulation of a continuous countercurrent
washing procedure was studied. In this series of tests a batch of It-lj-O
grade liquid was cooled to the temperature required to obtain about equal
proportions of solid (ice) and liquid phases, and then separated. The ice
fraction from each separation was saved but allowed to melt and the procedure
was repeated on the liquid phase for four additional times. Then the melted
ice sample from the fifth separation was combined with the melted ice from the
fourth separation, cooled to freeze out ice and separated; this ice fraction
then was melted and combined with the melted ice from the third separation and
the procedure was repeated until all of the previously saved ice samples had
been processed. The ice from each separation was washed with a small amount
of water (2$> by wt. of total sample used in each separation). The results
of these tests are given in Table 19.
The ice purity was increased appreciably with use of the simulated
countercurrent procedure. In the four consecutive freeze-out operations, the
purity of ice was increased from 2.7-7.8-0 to 0.31-0.95-0 grade. The liquid
fraction from this ice (test 10 was 2.2-6.5-0 grade, which is half the concen-
tration of the original starting solution (U-13-0 grade).
Conclusions
These results indicate that dilute ammonium phosphate solutions made and
purified by ammoniation of dilute phosphoric acid produced from dilute sulfuric
acid might be concentrated to usable grade by freeze concentration. Continuous
countercurrent washing would be necessary, and more than one stage of this type
of freeze concentration might be needed to attain the desired concentration
gradient between the ice and the fertilizer product. However, the ice purity
required in freeze desalination of water would not be needed since the melted
ice could be used for recovery of the dilute sulfuric acid used in the
proposed process.
99
-------
TABLE 18. TESTSa OF FREEZE CONCENTRATION OF AMMONIATED PHOSPHORIC ACID SOLUTION--
EFFECT OF PROCEDURES OF SEPARATION ON PURITY OF SOLID (ICE) PHASE
Freeze concentration data
Solid (ice) phase
Test
No.
Tests
1-1
1-2
1-3
1-4
1-5
Tests
2-1
2-2
2-3
2-5
Tests
3-1
3-2
3-3
3-4
3-5
Melt
Crystallization (at 24°c),
temp.
°r
, - ^
sp. gr.
Made Without Water Washing of
5
6
7
9
11
Made With
5
6
7
9
11
Made With
5
6
7
9
11
.6
•7
.7
.7
.7
Water
.6
.7
.7
-7
.7
Water
.6
.7
•7
•7
•7
1.040
1.056
1.076
1.112
1.192
Wt. %
% by wt.
N
P205
Sp. gr.
at 24° C
Liquid phase
Wt. %
ft by wt.
N
PP05
P205
recovery,
%
Solid (ice) Phase
15-3
22.9
20.1
34.1
27.8
l.6b
2.2?
3.0
3.9?
6.ob
5.2
7.0
9.8
12.7
19.4
1.156
1.179
1.198
1.223
1.252
84.7
77-1
79.5
65.9
72.2
4.8
5.2b
6.ob
6.5b
7.0b
15. 5b
17. ob
19- 5b
21.1
22.8
94.3
89.1
88.7
76.2
75-3
Washing of Solid (ice) Phase
1.025
1.044
1.048
1.100
1.132
11.1
25.6
18.2
35.0
20.8
Washing of Crushed (ice)
1.025
1.024
1.040
1.116
1.155
11.2
20.9
15.5
31.0
22.2
i.l
1.8
2.4
3.6
4.4
Phase
0.9
1.3
2.0
3.4
3.1
3-7
6.0
7.8
11.7
14.4
2.8
4.3
6.6
11.0
10.1
1.152
1.174
1.194
1.222
1.248
1.154
1.178
1.196
1.223
1.251
88.9
74.4
81.8
65.0
79.2
88.8
79.1
84.5
69.0
77.8
4.4b
4.6b
5 5b
6Al
7.0b
4.5
5.0
5.8
6.5
7.1
14. 3b
15. ob
18. ob
20.7
22.8
14. 5b
16. 3b
18. 8b
21.2
23.1
96.9
87.9
91.2
76.7
85.8
97-7
93.5
93-9
8l.l
88.9
a Procedure: Crystallization and separation were by the same procedure as described in Table 17.
After separation the variations in washing of solid phases described were made (amount of wash was
2$ of the total weight of the sample).
"^ "EB-fciTOEfbeii analysis "based, on specific gravity of soliition.
-------
TABLE 19. SIMULATED COUNTERCURRENT WASHING PROCEDURE9"
FOR FREEZE CONCENTRATION OF LIQUID FERTILIZER
Freeze concentration data
Solid (ice) phase
Test
No.
1
2
3
k
Temp. ,
Charge composition -°r
I-U (27.1)b'c
No. le (21.6)
No. 26 (32.3)
No. 36 (73-3)
+ i-5 (38.9)b'°
+ i-3 (65.U)
+ i-2 (132.8)
+ i-1 (235-3)
3-3
2.8
2.2
2.8
Wt. %
32.7
36.3
^5-2
36.6
Sp. gr. <$,
at 2k°C % N P^05
_d _d _d
_d _d _d
_d _d _d
1.005 0-31 0.95
Wt. %
67.3
63.7
54.8
63 A
Liquid phase
Sp. gr.
at 24° C
d
1.080
1.075
1.070
%
% N PP05
d d
_d _d
.d .d
2.2 6-5
a A batch of nominal ^-13-0 grade liquid was cooled to the temperature required to obtain about equal
proportions (by vol.) of solids and liquid, and then separated on a cooled, fritted glass filter. The
solid phase (ice) was saved and the procedure was repeated with the liquid phase for four additional
times to obtain a total of five ice samples (i-1, i-2, i-3» i-^9 i-5) then countercurrent operation
was carried out by combining i-5 and i-k, freezing, and .separating. Ice from this separation was
combined with i-3 j cooled, and separated; this procedure was repeated until all of the previously
saved ice samples were processed. Each of the ice separations was washed with a small amount of
water (2$ by wt. of total sample).
The numbers shown in parentheses indicate the weight of mix, grams.
|j Estimated grade of starting mixture (i-5 + i-*0 was 2.7-7-8-0.
Not determined.
e Ice from indicated test.
-------
MELAMINE SCRUBBING
The Applied Research Branch assisted the Fundamental Research Branch in
the study of a melamine scrubbing process (12) for the removal of sulfur
dioxide from waste flue gases. Our work included tests with actual stack gas
performed at the Colbert Steam Plant. As conceived, the process included the
following steps:
1. Scrubbing S02- and S03-laden waste gas with an aqueous slurry of
melamine.
2. Removing the spent slurry solids (primarily melamine sulfite and
melamine sulfate adducts) from the liquid.
J. Recovering melamine from the above adducts by thermal decomposition
of the melamine sulfite adduct.
k. Recovering melamine from the melamine sulfate adduct by chemical
regeneration means as necessary. Use of ammonia as the
regenerating chemical is suggested.
5- Return of the recovered melamine to the scrubber in slurry form.
6. Use of the sulfur dioxide released in the thermal regeneration
step to produce sulfuric acid or other valuable sulfur-containing
compounds in external units.
7- Use the sulfur trioxide recovered by chemical regeneration
(when necessary) as a fertilizer or in fertilizer processes.
Two Fundamental Research Branch progress reports (August and December
1976) describe development work done on the project. The work done on the
scrubbing step did not include the reuse of regenerated melamine, although
both the scrubbing and regeneration steps were studied.
TESTS WITH SYNTHETIC GAS
Exploratory Tests
Initially, a few exploratory tests were made to assist in defining the
test parameters and procedures to be used at Colbert Steam Plant. In these
tests the concentration of the melamine slurries was varied over the range
2 to 6f>; use of 6ff> was adopted to increase the amount of filterable solids at
the end of each test. The slurries were used to scrub simulated stack gas
102
-------
(0-3$ SOs* 3$ °2> 16$ C02, and balance N2) at a flow of 1.5 liters per minute.
At first, the scrubbing was done in standard 250-milliliter gas scrubbing
bottles equipped with coarse, fritted-glass spargers. However, plugging due
to the buildup of melamine sulfate was found to be a problem. Therefore, a
7-millimeter open-end glass tube was used as a sparger and this later was
changed to a 10-millimeter open-end tube. Also, a 125-milliliter scrubber
bottle was substituted to increase depth of submersion and contact time.
However, when using 125-milliliter scrubbers, there was a tendency for the
slurry to flow over to the next unit; therefore, for later tests the 250-
milliliter scrubbers were used.
After completion of each test, the slurry was regenerated chemically by
adding sodium hydroxide until the pH reached about 7. However, it was found
that a near-neutral pH resulted in the formation of melamine sulfite adducts
with melamine to sulfurous acid mole ratio greater than 2, indicating less
than complete regeneration, so in later tests an excess of base was added
during chemical regeneration. After regeneration, the solids were removed by
filtration and the scrubber mother liquor was used to reslurry the regenerated
solids. This was done to minimize melamine losses due to solubility in the
scrubbing liquor.
Process Tests
Planned startup of the Colbert limestone scrubber pilot plant (on which
availability of stack gas for the melamine scrubbing tests depended) was
delayed; therefore, on completion of the exploratory tests discussed above,
it was decided to test the melamine process in its full cyclic mode using
synthetic stack gas, prior to the tests at Colbert.
Four 250-milliliter gas-scrubbing bottles numbered 1 through h were used,
each with 10-millimeter open-tube gas inlet. Each scrubber bottle was charged
with a slurry of melamine that consisted of 6.0 grams of reagent-grade
melamine and 100 milliliters of scrubber liquor (approximately saturated with
melamine sulfite) from previous exploratory tests. The scrubbing train
consisted of three of these scrubbing bottles in series (positions I, II, III),
and the unused bottle was regenerated while the other three were in use. Gas,
simulating stack gas except for oxygen content, was delivered to the scrubber
bottle in position I at a rate of 1.5 liters per minute. The gas contained
0.32$ S02, 19.k% C02, and 80.3$ N2; oxygen was omitted to avoid possible
oxidation problems in these preliminary tests. Immediately following the
scrubber bottle in position I, a glass tee and stopcock were used to divert
100 milliliters per minute of the off-gases through a calcium sulfate drying
tower and a Manostat flowmeter to an ultraviolet spectrophotometer. The
remaining l.k liters per minute of off-gases was delivered to the bottle in
position II, and thence to the last scrubber (position III). The fourth
scrubber was held in reserve for addition to the system later, as discussed
below.
In operation, when S02 content of the off-gases from scrubber bottle in
Position I reached 50$ of the level in the inlet gas (50$ breakthrough) as
measured by the UV spectrophotometer, that scrubber bottle was removed for
regeneration and the other bottles were advanced from position II to position I
103
-------
and from position III to position II. The reserve scrubber bottle then was
added to the system in position III. At the next 50$ breakthrough, the same
procedure was followed, with the regenerated scrubber bottle being placed back
in the system in position III. These cycles were repeated until each scrubber
bottle had been regenerated and returned to the system four times.
Two methods of regeneration were used--thermal and chemical. Scrubber
bottles 1, 2, and J were always regenerated by filtration and heating (200°C)
of the solids to volatilize the absorbed S02 and H20. This thermal regenera-
tion was carried out in a glass-enclosed vessel which allowed the use of N2
(100 ml/min) as a sweep gas over the surface of the solids. Scrubber bottle
k always was regenerated chemically by filtration and leaching of the solids
with an excess of ammonium hydroxide (25 ml of 12$ NE^OH solution) to remove
both S02 and S03 in addition to the absorbed water. After regeneration
(either thermal or chemical), the solids were recombined with their filtrates
and adjusted to their initial slurry volume using water.
It was observed that after filtering the spent slurry a precipitate
formed in the clear liquor. In one case, this precipitate was removed by
filtration; its dry weight was about 0.6 gram. On further standing of the
filtrate at ambient temperature, more of the precipitate formed. Microscopic
examination later showed that the separated portion of the precipitate had the
optical properties of a compound which was reported to be (C3HSNS)2'H2S04.
Because of the amount of time required for removal and the relatively small
amount of precipitate (0.6 g), it was decided that only one filtration would
be made on each slurry and that any later precipitate would be left in the
liquor.
The results of the tests are shown in Table 20 and Figures J>0 and Jl.
Typical data on the concentration of S02 in the exit gas from the position I
(first stage) scrubber (Figure 50) show that the melamine absorbed all of the
inlet S02 for an average of 87 minutes each cycle or 82$ of the total scrubbing
time. At that time, scrubbing efficiency dropped quite abruptly, and S02
concentration in the off-gas rose quickly to 50$ of the inlet concentration.
The weight of S02 absorbed by each scrubber bottle for each of the four
cycles is shown in Figure 31. The average amounts of S02 removed per scrubber
bottle were 1.60, 1.48, 1.36, and 1.37 grams for cycles 1 through I*,
respectively. The decrease in scrubbing capacity for additional cycles most
likely is attributable to oxidation of melamine sulfite to the sulfate form
(which cannot be regenerated thermally) and to mechanical losses in the
regeneration procedure. The high value for scrubber bottle k in the second
cycle was a result of the greater scrubbing capacity of a slurry of 6 grams
of fresh melamine in distilled water which replaced the original scrubber
slurry which was inadvertently destroyed during regeneration after the first
cycle. The low values for scrubber bottles 1 and 2 in the second cycls
probably were due to shorter times used in regeneration after the first cycle.
Regeneration time at 200°C for bottles 1 and 2 after the first cycle was 30
and kO minutes, respectively, but was increased to 80 minutes after additional
cycles.
-------
TABLE 20. REMOVAL OF SULFUR OXIDES FROM WASTE GASES'
BY MELAMINE SCRUBBING (LABORATORY TEST DATA)
Scrubber bottle No.
Cycleb 123
Initial slurry
Wt. melamine, g 6.0
Wt. liquor, gc 100.15
Final slurry
Time to 50* breakthrough, min 1}1 96.5 100
pH 3.1 2.8 3.1
Wt. slurry, g 103.85 103-95 102.65
Wt. solids, ga 10.135 10.19 9-43
Wt. loss, % of input 2.2 2.1 2.8
Regeneration
Time at 200°C, min 30 80 90
Volume NH4OH (1230, ml
Wt. regenerated solids, g 5-70 5-50 5.25
Wt. regenerated slurry, g 94.25 95-35 117.45
Wt. HS0 added, ge 11.9 10.3 10.2
Average S02 concentration of
inlet gas, ppm 3174 3207 3213
SOS removal, % of inputf 97-5 95. 4 98.2
Weight S05 removed, g 1.7!* 1.32 1.38
Mole ratio
Moles SO, scrubbed6 „, , _, ,
Moles input melamine 0>569 °A56 °'U95
? Simulated span gas: 0.32* S02, 19.4* C02, 80.3* N2 at
Cycle consisted of bottle being operated successively
1
4
101
3-1
102.75
10-7
2.7
80
5-07
3213
97-8
1-39
0-522
a flow
Scrubber bottle No. 2
1 2
6.0
99.9
122 86
3-0 3-1
103-5 104.4
10.1*8 10.35
.2.3 1.1*
1*0 80
5-97 5-66
95-1 95-4
10.8 10.2
3174 3233
97-2 93-2
1.65 1.17
0.542 0.385
of 1.5 1/min.
3
105
3-1
103-1
8.9
2.4
80
5-43
85-7
20.2
3279
97-2
1.46
0.508
in scrubber positions III, II,
4
92
3-1
104.1
11.6
1.7
80
5-43
3246
97-5
1.28
0.462
and I
Scrubber bottle No. 3
1
6.0
99-7
105
3-0
103.7
9.66
1.9
50
5-71
96.1
9-3
3266
93-5
1.42
0.466
(until 50*
2
109
2-9
103-2
10.96
2.1
80
5-50
93.3
11.9
3226
97-6
1-55
0-534
3 4
98 102
2.2
103.0 103.8
8.74 10.88
2.1 1.8
80 80
5-31 5.12
97-8
7-9
3220 3193
97-4 95.9
1.36 1-37
0.486 0.508
breakthrough) , then
Liquor was from previous exploratory tests, approximately saturated with melamine sulfite (slurry from scrubber bottle No
with 6 g melamine in 100 g water after first cycle).
Ethanol -washed and air -dried.
Added to maintain volume of initial slurry.
•m T ffjvct o4-oiT*a
,1 r-o-
1 f*ll1 at-ft
rf Frnm TTV •
CT**sn+-.T-n
tnhnt.nmpt.pT* pHj
Scrubber bottle No
1
6.0
99-1
109-5
3-1
103-0
.
2.0
3266
96.5
1-57
0.515
2
6.0
100°
138
2-7
103-5
8.54
2.4
25.0
4.98
96.2
9-8
3207
98.3
1.90
0.622
3
91
3-0
104.0
9.06
1-9
25.0
5.19
95-0
10.1
3180
97-5
1.24
0.491
. 4
4
104
3-1
102-9
8-55
2.1
25-0
5-12
3161
97.6
1-43
0.541
regenerated.
. 4 was lost and
iT-h flypn*; .
replaced
Assuming all of regenerated scrubber solids were melamine.
-------
W
•H
m
id
bD
a
•H
3200
3000
2800
2600
2400
2200
2000
1800
1600
1400
1200
1000
800
600
400
200
0
Inlet gas to scrubber in position I
—O— Outlet gas from scrubber in
position I
Gas composition
C02, 19.W
N2,
S02,
10
90 100
Figure 30. Laboratory removal of S02 from synthetic gas by
scrubbing with melamine slurry—typical analyses
of gases at inlet and outlet of first scrubber versus time.
106
-------
2.0
a
-p
*
0>
1.5 -
LT\
0)
(1)
O
W
•8
CM
O
02
i.o -
0.5 -
1
t)
cT^*-*^
0
—
_
6 i I
O _,
A " g> — -, y
o
D D
Each scrubber consisted of 250 ml
gas scrubbing bottle containing 6 g
melamine and 100 ml liquor. Each
cycle consisted in
bottle successively
positions III, II,
by regeneration.
1
operation of the
in scrubber
and I followed
Scrubber Regeneration
Symbol bottle No. method
A 1 Thermal
D 2 Thermal
O 3 Thermal
0 h Chemical
X Av
1 1 1 1
0
Cycle
Figure Jl« Laboratory removal of S02 from synthetic gas by scrubbing
with melamine slurry—effects of regeneration on absorption of S02.
-------
Chemical analysis of the regenerated scrubber solids showed that bottles
1, 2, and 3 (thermal regeneration) were regenerated to 86.7$ of their theoret-
ical scrubbing capacity /assuming formation of the (C3H6N6)2•H2S03•nH20 salt/
after four cycles. In these bottles, 89.6$ of the sulfur that was not
regenerated was in the oxidized (sulfate) form, which is equivalent to an
average of 0.019 gram of sulfur oxidized per regeneration. This is equivalent
to 1.3$ of the sulfur absorbed. Bottle k- (chemical regeneration) was
regenerated to 99.2$ of its theoretical scrubbing capacity with none of the
remaining sulfur in the sulfate form after the fourth cycle. After the test,
recovery of melamine used in all four of the bottles amounted to only an
average of 82.8$ of the input. This probably was due to mechanical losses
from the small quantities of materials that had to be handled.
It was concluded from this test that 100 milliliters of a 6$ slurry of
melamine used in each of four scrubbers would allow sufficient scrubbing
capacity so that any one scrubber could be regenerated while the others
removed all of the inlet S02 even as the buildup of sulfate became critical.
Further, it was determined that thermal regneration of the sulfite sulfur in
the solids would be essentially complete after 80 minutes at 200°C as evidenced
by the lack of acidic gases in the sweep gas used. However, if a large degree
of oxidation occurs, a longer heating time may be required since the presence
of sulfate inhibits the volatilization of the sulfite sulfur. With the presence
of oxygen in actual stack gas (about 5$ ®z) it is expected that oxidation will
be greater.
TESTS WITH COLBERT STEAM PLANT GAS
The tests were made using stack gas from Colbert units 3 and k. The S02
content of the gas varied from 2000 to 2900 ppm and the oxygen content varied
from 7 to 10-5$- Cyclic tests were made in which melamine slurry was first
used to scrub the gas and then the spent melamine was regenerated and returned
to the scrubbers. Tests were made to determine the effects of using a water-
prescrubber and of the addition of an antioxidant to the scrubber slurry. As
discussed previously, two methods of regeneration of the spent melamine were
used, thermal and chemical. Thermal regeneration may be used to drive off
absorbed water and sulfite sulfur but does not drive off any sulfate sulfur;
temperatures as low as 200°C are suitable for thermal regeneration. Chemical
regeneration, such as reaction with ammonium hydroxide, removes both sulfite
and sulfate sulfur from the melamine, but the absorbent is somewhat soluble in
the regenerating solution, which increases losses.
Equipment and Procedure
A series of four tests was made at the conditions tabulated below.
108
-------
Test
No.
1
2
3
k
Gas rate,
1/min
2
2
2
2
Prescrubber
in system
No
No
Yes
Yes
Antioxidant
in slurry3-
Ho
Yes
No
Yes
Scrubber
temp. , °C
52
52
52
52
a. 0.1';9 p-phenylenediamine added to original slurry.
As in the previous tests, four 250-milliliter gas-scrubbing bottles
numbered 1 through k were used, each with 10-millimeter open- tube gas inlet.
Each scrubber bottle was charged with a slurry of melamine that consisted of
6.0 grams of reagent-grade melamine and 100 milliliters of water saturated
with melamine. The scrubbing train consisted of three of these scrubbing
bottles in series (positions I, II, III), and the unused bottle was regenerated
while the other three were in use. The equipment arrangement was as shown in
Figure 32. Three of the four scrubber bottles (and the prescrubber, when in
use) were placed in a water bath maintained at 52°C. Stack gas was drawn
serially through the scrubbers and a condenser by pump No. 1 and was metered
at the pump exit. The condenser was used to dry the gas enough to meter it.
A second pump (pump No. 2; Figure 32) was used to pump as much gas as was
possible (about 1.5 1/min) to exhaust; this was done to keep the flow in the
incoming line as great as possible to prevent settling of fly ash in the lines
and to prevent cooling and settling of the incoming gas as much as possible.
The incoming line was a steam- traced 0.6-centimeter (l/4-in) stainless steel
tube. Separate inlet streams from points upstream of the scrubber train and
downstream of the first scrubber were passed through UV analyzers (Figure 32)
to determine the S02 content of the gas at these points; these streams were
passed through tubes filled with anhydrous calcium sulfate to remove water
and any fly ash prior to entry to the UV cells. The stream originating between
scrubbers 1 and 2 was metered as was the stream passing completely through the
scrubber train; the total of these two metered rates was the rate of gas
passing through scrubber 1.
In operation, when S02 content of the off-gases from the scrubber bottle
position I reached about ^ of the level in the inlet gas (50,, breakthrough)
poson reace aou
as measured by the UV spectrophotometers (about 1300 ppm at UV-2J , that
scrubber bottle was removed for regeneration and the other bottles were
advanced from position II to position I and from position III to position II.
The reserve scrubber bottle then was added to the system in position III. At
the next 5^ breakthrough, the same procedure was followed, with the regenerated
scrubber bottle being placed back in the system in position III. These cycles
were repeated until there was significant reduction in the effectiveness of
the recycled melamine slurry or until four scrubbing cycles with each scrubber
bottle had been completed. The melamine was not regenerated after the last
cycle for each scrubber bottle.
109
-------
H
H
O
CaS04
(Dryer)
Gas in
Prescrubber
Water bath
Exhaust
Flowmeter (100 cc/min)
Ice water
Cooled condenser
I II III
Scrubber positions
(Melamine slurry scrubbers)
Exhaust
Flowmeter
(1900 cc/min)
Figure 32. Equipment arrangement used in tests of melamine
scrubbing of Colbert Steam Plant stack gas.
-------
SPENT
SCRUBBER
Solids, weighed
Solids
OVEN
200° C
120 min
1
^ Iodine
titration
NaOH
Filtrate
Figure 33. Thermal regeneration of spent melamine slurry.
1
T
REGENERATED
SCRUBBER
SPENT
SCRUBBER
NH^OH
15 min
Ice-water washed,
ethanol dried
Solids, weighed
Iodine
titration
Filtrate
Figure J4. Chemical regeneration of spent melamine slurry.
J
REG^IKRAT!
SCRUBBER
-------
gas entry and delivery tubes attached. The solids in their container were
placed in an oven held at 200°C for a period of 2 hours during which time the
raelamine sulfite was decomposed. A flow of nitrogen was passed through the
vessel during regeneration to sweep out any released H20 and S02 and to carry
it into a sodium hydroxide solution to absorb the S02. This solution later
was analyzed for S02 content by iodine titration. The regenerated melamine
was weighed and recorabined with the filtrate from the spent scrubber slurry
to form the regenerated scrubber slurry. Any volume loss resulting from the
scrubbing or regeneration steps was made up as water saturated with melamine
at room temperature.
Chemical regeneration is depicted in Figure jif. Scrubber bottle No. it-
was subjected to this type regeneration. Spent scrubber slurry was filtered
on a 30-milliliter coarse-fritted crucible and the solids were transferred to
a 50-milliliter beaker where they were stirred in 25 milliliters of 12.5$
NH40H solution for 15 minutes. The solids then were refiltered on the same
crucible, washed with ice water five times (5 ml each wash), and then were
washed with ethanol and air dried before weighing and recombining with the
filtrate from the spent scrubber slurry. The regenerated scrubber slurry was
brought back to its initial volume by the addition of water saturated with
melamine. The filtrate from the ammonia reaction slurry was analyzed for
sulfite sulfur by iodine titration during the tests and was later analyzed for
total sulfur, total nitrogen, and ammoniacal nitrogen.
Discussion of Test Results
Data collected during the tests and indices calculated from these data
are given in Table 21. Sulfur dioxide absorption by the individual scrubbers
is shown in Figure 35, and recovery of S02 in the regeneration step is shown
in Figure 36. The theoretical amount of S02 that could have been absorbed by
the solid melamine in each scrubber would be about 1.52 grams according to
the equation:
2C3N6H6 + S02 + (l+x)H20 > (C3N6H6)2-H2S03'xH20
The melamine in solution would absorb about an additional 0.1 gram and the
water could hold about 0.1 gram more so that the total S02 holding power of
the individual scrubber would be about 1.7 grams. The amount of S02 absorbed
in each scrubber includes that collected while that scrubber was in position
I in the series plus that collected while it was in position II (breakthrough
from scrubber in position I) and position III. These amounts were calculated
from the input analysis and flow and from integration of the UV analyzer
charts (S02 analysis). The values as calculated above are shown in Figure 35-
They generally are higher for the first cycle than would be indicated by the
original capacity of the scrubbers or by the amounts of S02 liberated during
regeneration. No explanation of this inconsistency is offered. Errors in
analytical values and in measured flows would not seem to account for the
differences, particularly since the values for the second and succeeding
cycles are much more reasonable than those for the first cycle. It is
possible that there was some component of the stack gas which, when wetted
112
-------
I-1
TABLE 21. REMOVAL OF SULFUR OXIDES FROM WASTE GASES BY MELAMINE SCRUBBING
(DATA FROM LABORATORY-SCALE TESTS AT COLBERT STEAM PLANT)
Test 1
Prescrubber used? No
Scrubber bottle No. 1 2 3
Cycle la .
Input SOp, ppm 2492 2085 2439
Input Op,
-------
TABLE 21 (continued)
Test 1 Test 2 Test 3
Prescrubber used? Ho No Yes
Scrubber battle No. 1 2 5 U 1 2 3 1* 12
Cycle 3*
input SO,, ppmb 2557 2721 2623 2623 225>* 2597 2573 21*58 281*9 2791*
input os, *c 8.0 8.5 - 8.2 8.0 9-7 8.7 8.6 7.1
Recycled solids, g 7-2 6.85 6.7 6.0 5-25 5-5 5-6 5-1 6.6 6.3
Recycled slurry, g 89-5 9>*.8 87.1 86.5 118.0 116.6 112.9 109-9 128.2 12U-5
H,0 added, gd 1*0. oJ 35.7-) !»5.5J kk.si 13.2 lU.l 18.1 21.1 2.9 6.1
Time to 50* breakthrough, min 17 1*0 60 103 110 101 10? 108 75 82
Total SO, fed, ge 0.2U8 0.622 0.9H* 1-51*5 l-'Ufi 1-1»98 1.602 1.516 1.220 1.308
Removal of S0e, *e 72.2 91.1 89. U 90.8 92.2 95-2 91.0 93-2 9k. 6 95-6
S0e absorbed, gf 0.379 0.632 0.870 l.l»95 l.ltllt 1.501 1.551* 1.550 1.256 1.315
Solids before regeneration , g 11.3 15.65 lk.1 - 11.2 9-7 8.3 10.0 lU-95 15-1*
Regenerated solids, g 5-55 5-1 5.!*5 "*-T5. 7.0 6.8
Indicated S0= regenerated, g, 1.1568 1.0688 1.1168 l.l2l*n 0.5>*68 0.7168
Indicated S0a regenerated, jt1 81.8 71.1 71.8 72-5 1*3.5 5"*-5
Cycle k*
Input SOo, ppm1' 21*25 2331* 2529 2550 2633 2638
Input 0,, *c 9.0 9.0 9.0-9.5 8.2 7-5
Recycled solids, g 5-55 5.1 5-"*5 <*-75 7-0 6.8
Recycled slurry, g 120-3 112-2 U^-5 113-8 125-3 119-3
HaO added, g3 10-9 18-5 17.U 17.2 5-8 11. 5
Time to 50* breakthrough, min 9k 105 108 95 55 69
Total S0e fed, ge 1-502 1.1*00 1.1*37 1-581* 0.827 1.040
Removal of SO,, *e 92.8 90.8 90.6 95-1 92.6 91-7
SO. absorbed, g*" 1.3OS 1.56o i.k}k 1.1*16 0.81*2 1.011
Final slurry, g 130-5 126.0 129.8 131.6 121*. 1* 128.7 129.0 126.0 136. U 132.3
Final slurry pH 5-01 3-8 3-9 3-9 3-85 3-79 U.o
Weight dry solids recovered, g 7-1* 8.3 7-6 8.0 7-75 7-65 7.60 6.55 8.65 8.1*0
Filtrate, ml . - - llU 102 107 113 110 - 109
Filtrate, g 119.2° 110-3° 115-7° HO. 7 101.9 106-5 H5-6 110.8 113.8° 109. U
a Cycle consisted of bottle being operated successively in scrubber positions III , II , and I (until 50* breakthrough)
3
21*92
7.0
6.1*5
126.5
It. It
77
1.096
92.1
l.oSU
13-75
It
2687
7-5
5-1*5
120.8
10.2
87
1-555
9k. o
1-557
13-9
6.85 it .9.
0.6638 o.883n
62.5
2702
8.0
6.85
123-1*
7-5
57-5
0.887
91*. 2
0.918
132.1*
3.82
8.95
101
101.5
, then
b Inlet S02 measured with Perkin Elmer (Coleman 12U-D) UV analyzer averaged for duration of test. Gas from Colbert units 3
c Oxygen content of off -gas as measured before and after the test using Fyrite analyzer and averaged.
Water saturated with melamine at room temperature.
e Calculated by integration of UV analyzer recorder charts, assuming constant flow of 2 1/mln of gas through scrubber
f previous stage.
Includes breakthrough area while scrubber was in position II of scrubbing cycle.
; feed
66.0
27W*
8.0
1*.9
115-5
15-7
101.5
1.591
91*. 8
1.557
128.5
3.68
7.1*5
111
110.05
regenerated
and I*.
1
2877
7.0
5-55
120.5
11.0
101
1.660
95-5
1.672
12.8
5.65
1.2058
72.1
2280
9-0
5-65
122.7
8.8
125
1.628
90.6
1.61*9
137.9
lt.32
10.05
118
117-5
Test U
Yes
2
2887
7-0
5.8
125-5
5-1*
100
1.665
95-1*
1.66o
12.6
5-5
1.1U58
69.0
2221
9-5
5-5
122.5
8. It
126
1-598
88.it
1.556
155-5
l*.28
12.6
117
116.1*
includes S02 breakthrough
3
2327
9-1*
5-5
128.6
2.1*
125
1.6U8
88.0
1.525
12.9
5-55
1.1858
77.7
2203
9 .1*
5-55
120.0
11.0
1U7
1.1*90
92-5
1.551
136.6
l*.29
15.3
117
116.9
from
k
2278
9.1*
5-05
12l».2
7-1
126
1.639
88.8
1.6U3
10.8
't .70
1.119"
68.1
2030
11.5
It. 70
126.6
It .7
252
2.922k
.
-
136.U
5-95
9-5
122
120.5
8 Sulfur dioxide evolved during regeneration was absorbed in NaOH (3-8 g NaOH/100 g H20) and analyzed by iodine titration.
Sulfur dioxide in filtrate from chemical regeneration of solids with 25 ml of 12* NH4OH, filtered to recover solids
portions of ice water, does not include sulfate sulfur.
* 100(SOS absorbed - S0e regenerated)/SOe absorbed.
3 Increased from 100 to 125 ml at beginning of third cycle to maintain fluidity during scrubbing test.
? Air in-leakage detected in slip-stream of gas to UV analyzer; caused operation well beyond 50* breakthrough.
Filtered after standing overnight; pH was measured on filtrate.
m Calculated from final slurry weight less weight of wet solids from filtration.
for recycle and solids washed with five
5 -ml
-------
(.
^
bJ
•a
n
123.
Cycle
Test 1: "Dirty" gas; no antioxidant
Scrubber No. Regeneration
Thermal
Thermal
Thermal
Chemical
0.0
2 3 »i
Cycle
Test 2: "Dirty" gas; "ith antioxida^t
3-0
2.0
1.0
0.0
I
2 3
Cycle
Test 3: "Clean" gas; no antioxidant Test U:
Cycle
"Clean" gas vith antioxidant
35. Absorption of S02 from Colbert Steam Plant gas by melamine scrubbers,
115
-------
to
O -H
M -P
O -P
O -H
O -P
c\j M
O
C/3
if .0
»-,
i!
-,
s
0.0
•est 1
1
^T "
^>x. n
^0
A
1
L 2 3
Cycle
"Dirty" gas; no antioxidant
2.0
1.0
0.0
I
123
Cycle
Test 2: "Dirty" gas with antioxidant
2.0
hO
tJ O
0 -H
!H -P
0) CO
O -H
0) -P
O
00
0.0
Scrubber No. Regeneration
O 1
A 2
n 3
*• h
I
2
Cycle
Thermal
Thermal
Thermal
Chemic al
Test 3: "Clean" gas: no antioxidant
2.0
1.0
0.0
I
123
Cycle
Test 4: "Clean" gas with antioxidant
Figure 36. Recovery of r,0% during regeneration of spent melamine scrubbers.
116
-------
with the aqueous slurry, absorbed sulfur dioxide until the liquid became
saturated with that component or its reacted salt. No speculation as to the
identity or nature of such a component is offered.
Several trends may be seen in the data of Figure 35. First, when dirty
gas was used (no prescrubber, test l), there was a considerable spread in the
absorption of S02 between scrubber bottles in the first cycle; this spread was
not noted when clean gas (prescrubber used; test 3) was scrubbed. Next, when
no antioxidant was used (test 1 and 3), the absorptive capacity of the
scrubber bottles receiving thermal regenerative treatment (bottle Nos. 1, 2,
and 3) declined throughout the test while the one receiving chemical regenera-
tion (bottle No. Ij-) retained absorptivity longer. Third, when antioxidant
(0.1$ p-phenylenediamine) was present (tests 2 and k), the absorptive capacity
of the scrubbers remained about constant after the first cycle had ended and
the regeneration method did not make any perceptible difference.
The amounts of S02 recovered during regeneration and measured by iodine
titration are shown in Figure 36. The trends, discussed in the preceding
paragraph, except for the spread in S02 absorbed during the first cycle, are
reflected in the recovery data presented in Figure 36. There was a general
falling off of S02 recovery during successive cycles when no antioxidant was
added (tests 1 and 3) and, with the exception of the first cycle, slightly
more S0a was recovered by chemical regeneration than by thermal regeneration.
When antioxidant was present (tests 2 and 10, the amounts of S02 recovered in
successive cycles were nearly constant and there was no difference between
chemical and thermal regeneration. Best results were obtained when both
the prescrubber and the antioxidant were used.
The spent scrubber solids were not regenerated after their last use in
the scrubbers. These solids were filtered from the slurry, washed with
ethanol, and dried to constant weight on a filter with dry nitrogen passing
through them. Nitrogen was used to preclude oxidation. Both dry solids and
the filtrates were analyzed to determine the sulfur and melamine content; the
solids were analyzed also for sulfite sulfur so that the proportion oxidized
could be determined. These values are given in Table 22. In all tests and in
all scrubbers, the analyses of the solids indicated a fairly high degree of
utilization of the melamine. The desired reaction product was
(C3N6H6)2-H2S03«xH20, which has an S:C3N6H6 mole ratio of 0.5- When oxidized
sulfur (sulfate) is absorbed, the reaction was considered complete when
(C3N6He)2H2S04*xH20 was formed; this compound also has an S:C3N6H6 mole ratio
of 0.5. The attained S:C3N6H6 ratios were in the range O.U8 to 0-50 except in
two of the scrubbers of test 1 where the ratios were O.lt-3 and OAU. These two
may have been low because of the high degree of oxidation and consequent low
pH in that test. As mentioned, oxidation was highest in test 1. The solids
from scrubbers 1, 2, and 3, in which thermal regeneration had been used,
showed 82 to 92% oxidation in test 1 (average of nearly 30^/cycle for three
scrubber cycles). The solids from scrubber k (chemical regeneration) showed
^6% oxidation; this value probably reflects insufficient regeneration of
sulfate during the initial cycles since the sulfate should have been regenerated
and the indicated oxidation should have resulted from the final cycle only. It
is probable that more ammonium hydroxide and/or a longer regeneration period
flight be required when oxidation is high.
117
-------
TABLE 22. CHEMICAL ANALYSES OF FINAL PRODUCTS FROM SCRUBBING STACK GAS WITH MELAMINE SLURRY
00
Test 1
Preaerubber used?
Antioxidant used?
Scrubber bottle No.
Chemical analysis of solids, %
Total S
Sulfite S
Nitrogen
S:C3ffeNs mole ratio
Composition/100 g solids
Moles melamine compounds
As C3HeNe
As (C3HSNB)2-H2S03.2H20
As (C3HeNe)s-H2S04.2HsO
S oxidized
Chemical analysis of filtrates,
-------
The high proportion of sulfur oxidized to sulfate form in test 1 was due
to the lack of prescribing and the absence of antioxidant rather than to any
unusually high oxygen content of the incoming gas. When antioxidant was added
^5S? rf/thS p^°P°rtion of sulfur oxidized to sulfate form dropped to about
1M> (5-5V cycle) when thermal regeneration was used and to 5.7 ' when chemical
regeneration was used. When oxidation was fairly low, the chemical regenera-
tion procedure used apparently was adequate.
When a prescrubber was used without antioxidant in the scrubbers (test 3),
oxidation averaged 60^ (15^/cycle) when thermal regeneration was used and
22.5$ when chemical regeneration was used. The relatively high value in the
chemically regenerated scrubber again indicated the need for more severe
conditions for chemical regeneration when oxidation is high.
Best results were obtained in test k, in which both the prescrubber and
the antioxidant were used. Oxidation averaged only 12^ (about 3 fj/cycle) when
thermal regeneration was used and % when chemical regeneration was used.
The filtrates from the slurries in the spent scrubbers (Table 22)
contained an average of 0.15^ sulfur. The nitrogen content of these solutions
averaged about O.U3$, which is the equivalent of 6.5 grams melamine per liter.
This dissolved melamine is returned to the scrubber in the process.
Solutions obtained from the chemical regeneration of solids from bottle
No. k in each cycle of each test were analyzed also (Table 22). These solu-
tions contained an average of about 0.89"^ nitrogen from melamine, which is
equivalent to about 1J.U grams melamine per liter. This dissolved melamine
would be lost from the system unless a means of recovery is devised. The
regeneration solution used in the present work was quite dilute, which would
increase melamine loss by solubility and would require special methods for
recovery of solid ammonium sulfate. However, by recycling the regeneration
solution and using anhydrous ammonia instead of ammonium hydroxide solution,
the concentration of ammonium sulfate could be increased substantially and
losses of melamine reduced. Also, in practice, chemical regeneration would
follow thermal regeneration and involve only the amount of solids required to
maintain the proportion of melamine sulfate at an acceptable level; this would
obviate an otherwise necessary oxidation step for the ammonium sulfate-sulfite
solution.
Conclusions
It was concluded from these tests that melamine slurry scrubbing is a
chemically feasible method for removal of sulfur oxides from power plant waste
gases. It was shown that the absorptive capacity of the melamine could best
be maintained by the addition of an antioxidant to the system and that the
antioxidant was not destroyed during regeneration (the antioxidant probably
stayed in the filtrate and was not subjected to regeneration procedures).
Further improvement was attained in tests in which a prescrubber also was
used. The prescrubber should remove a large part of both the fly ash entrained
in the gas stream and most of the sulfur trioxide present; this would lower the
proportions of melamine tied up as sulfate per pass and would alleviate the
problem of removing fly ash from the melamine on continued operation.
119
-------
Almost complete removal of S02 was obtained in a single scrubber until
the melamine had been about 80$ exhausted; thus, a two-stage scrubber might
prove sufficient in a larger scale unit. The existence of solids in the
scrubber solution may preclude the use of packed scrubbers, particularly where
oxidation may be fairly high. Preliminary tests had shown that, when melamine
sulfate was formed in large proportions, a sticky mass plugged the gas inlet
tube; this probably would foul packing.
Even though a prescrubber and an antioxidant are used, a gradual increase
in the proportion of melamine tied up as sulfate may be expected if thermal
regeneration is used. In that case, a side-stream of spent slurry solids
probably will have to be regenerated chemically.
Further testing, on a larger scale and over a greater number of cycles,
will be needed to test the size of the side-stream requiring chemical regenera-
tion and the makeup rates of melamine and antioxidant. None of the currently
reported work was done to determine the most effective antioxidant or the most
effective concentration.
MINIPILOT-PLANT TESTS
Previous tests of melamine slurry scrubbing for the removal of sulfur
oxides from stack gases were limited to scrubbing in laboratory-type gas
washing bottles. A minipilot-plant scrubber, which had been used successfully
for lime and limestone slurry scrubbing, was operated in limited tests using
6% melamine (C3H6N6) slurry as the absorbent. The gas used was produced by
adding sulfur dioxide to the combustion products of natural gas to obtain a
mixture containing JOOO ppm S02. Although testing of a variety of conditions
was planned, operating problems restricted the tests; only one scrubbing
stage, 100$ stoichiometry, and a water prescrubber to protect the glass column
from the heat of the inlet gas were tested. (A C3H6N6:S02 mole ratio of 2 is
considered to be 100$ stoichiometry.) Because of the operational problems,
the operating periods were short, liquid to gas ratios (L/G) were higher than
desired, and the forward flow probably was too high at times.
The scrubber arrangement used is shown in Figure 37- The scrubber
consisted of a glass tower with a full cone spray nozzle at the top, spraying
downward, opposing the upward flow of stack gas which had previously been
passed through a similarly constructed water prescrubber at a lower level in
the tower. The slurry was removed from the tower through a centered funnel
into which it was directed by a circular glass deflector and diverter screen.
The slurry was drained into a 500-milliliter surge tank from which it was
recirculated to the spray nozzle through a centrifugal pump, the rate being
controlled by the size of the spray nozzle. Fresh slurry was added to the
tank as required by a metering pump; the level in the tank was controlled by
an overflow line which vented the spent slurry to a holding tank. The forward
flow of slurry to the tank was about 60 milliliters per minute, so that the
retention time of the slurry in the scrubber loop (scrubber + surge tank) was
about 8 to 9 minutes. The recirculation rate through the spray nozzle during
the more successful test periods was about 3.8 liters per minute; this resulted
in an L/G of 33 liters per cubic meter and a liquid retention time of about 8
seconds in the surge tank.
120
-------
Gas out
Spray
(3000 ppm S0e)
r
S\orge tank
(500 ml)
Spent
slurry
3•8 1/min
Figure 37- Spray scrubber used in melamine
scrubbing tests--configuration 1.
121
-------
Operating problems were occasioned by the properties of the melamine and
its reaction products. Although the screen analysis of the feed melamine
showed > 95$ minus k8 mesh, a few large particles were present that stopped up
the 50-mesh strainer in the inlet line to the spray nozzles (see Figure 37).
When the screen was removed, the nozzle plugged (orifice dia., approx. 1 mm,
the equivalent of a l6-mesh screen). Since a larger strainer was not
immediately available, the nozzle size was increased to one with an orifice
about 1.5 millimeter in diameter. This nozzle also plugged. The slurry then
was removed from the system and passed through a M3-mesh screen (0.297-mm
opening). Oversize particles were discarded and the slurry was returned to
the system. The dry melamine feed thereafter was scalped using the ij-8-mesh
screen, and a l6-mesh strainer was installed. Little further trouble occurred
due to stoppages in the slurry recirculation system.
Concurrent with the problem of strainer and nozzle stoppages, there was a
buildup of solids on the diverter screen at the bottom of the tower. The
slurry does not pass through this screen; the incoming gas does. The screen
is present to direct the scrubbing medium into the funnel through which it
exits the tower. Solids buildup was fairly rapid but, until stoppages in the
feed system were eliminated, the cause could not be separated from the effects
of the interrupted flow. Solids built up so rapidly on a 20-mesh diverter
screen that gas flow through it became severely restricted within an hour of
operation. This screen was replaced with one of 10 mesh; this plugged within
about the same operating time. The solids buildup apparently began on the
outside of the screens, the side through which the gas entered, and then
proceeded to grow toward the slurry side. This is depicted in the photographs
in Figure 38. It is assumed that the mechanics of this deposition include a
dissolution of melamine in the incoming slurry, the reaction of dissolved
melamine with the incoming sulfur dioxide, and subsequent rapid precipitation
of melamine sulfite on the diverter screen surfaces. Petrographic analysis
indicated the solids to be, primarily, melamine sulfite.
The longest period of operation in the mode indicated by Figure 37 was
less than 2 hours. Throughout the sustained operating periods, however,
sulfur dioxide removal was very good; better than 99% removal was attained
most of the time. From these data, it appeared that melamine slurry would be
a suitable absorbent in a spray scrubber provided the problem of solids build-
up could be accommodated. The use of packed column-type scrubbers, however,
would not seem advisable because of probable plugging problems.
One method of accommodating solids buildup would be to remove the surfaces
to which the solids cling. During a period of the test, the diverter screen
was removed. However, the slurry tended to flow around the edge of the stage
bottom and to fall into the stage below to such an extent that continued
operation was not possible.
Next, the scrubber was modified as indicated in Figure 39. The water
prescrubber was removed, the diverter screen and collecting funnel were
removed, and the surge tank was placed to catch the drain from the bottom of
the scrubber. This configuration now resembles a wet-bottom scrubber with the
gas entry at the side near the bottom. The internals at the wet-dry interface
are eliminated. It is noted that, with the configuration used earlier, the
122
-------
Top view
Side view (tilted)
Figure 38. Photograph of diverter screen
after use in raelamine scrubbing test.
123
-------
Gas out
Gas in,
113 1/min
(JOOO ppm S02)
8o°c
Spray head
Glass
column "
i
302)
Stainless
steel
1 fitting 1
1
-*£j^~~£
/n\
«-10 cm-*
i
^_-
1
— -
A ^
fl
i
35
\
i
cm
£
r
t
r
Surge tank
(500 ml)
Spent
slurry
Screen
3.8 l/min
melamine
slurry
Figure 39. Spray scrubber used in melamine
scrubbing tests—configuration 2.
-------
incoming gas leaving the prescrubber first met the slurry at the diverter
screen where it made fairly good contact and then passed through the spray
where it contacted the slurry again. In effect, the minipilot-plant spray
scrubber stage was a double contact countercurrent scrubber. Omitting the
diverter screen and funnel reduced the scrubber to one having a single contact.
Omitting the prescrubber allowed the gas to contact the slurry at a higher
temperature. Both these actions should be detrimental to absorption
efficiency.
Operating time in the configuration described in Figure J9 was limited to
I hour by the work schedule. Sulfur dioxide removal during that period
dropped from an initial value near 100$ to a final value of about 90% and the
trend was still downward. Inspection of the area near the gas inlet showed
little tendency to scale.
A minor problem that had been noted in the minipilot-plant run was the
tendency of the melamine slurry to drain toward the inlet end of the gas duct
at the bottom of the scrubber when in configuration 2. The stainless steel
duct was modified as shown in Figure kO to prevent such drainage.
A second test then was made of the single-stage, single-contact scrubber
using a 6% melamine slurry containing 0.1% p-phenylenediamine as an antioxidant.
The test was operated for 6 hours. Three sets of conditions were tested as
shown in the following tabulation.
Test
time , min
0-l80
180-235
235-300
Gas
flow,
irfVmin
0.113
0.113
0.113
S02
content ,
ppm
3200
3180
3275
Forward
flow
ml/mina
6k
6k
78
Recirculation
rate , ml/min
1068
13^7
13^7
L/Gb
9-5
11.9
11.9
Melamine
stoichi-
ometry, %c
9k
95
112
S02
removal ,
$d
75
78
80
a. 6% slurry of re agent-grade melamine in water.
b. Liters slurry per cubic meter of gas.
c. Based on equation 2C3N6H6 + H2S03 + UH20 > (C3N6H6)2'H2S03tUH20.
d. As determined by UV analyses of inlet and outlet gases.
During the first 3 hours the melamine feed was maintained at about 1.88 moles
per mole of incoming S02 in the stack gas (9k% of stoichiometric amount of
melamine). The L/G in the scrubber was 9-5 liters of slurry per cubic meter
of gas. The scrubber inlet gas temperature was about 50°C, being cooled from
132°C by spraying water on the outside of the stack from the burner. At these
conditions, the slurry reached about k8°C and a pH of about k.6. Sulfur
dioxide loss averaged 25$. The L/G was increased to 11.9 liters per cubic
meter at 180 minutes. The pH and temperature remained about constant but the
S0a loss decreased slightly (about 22$). An increase in forward flow at^235
minutes increased the melamine to S02 mole ratio to 2.25 (112$ of^stoichio-
metric). This decreased the slurry temperature slightly (about 2°C) and
further decreased loss of S02 to about 20$. Inspection after the test showed
little buildup at the bottom of the scrubber. There was some buildup on^the
scrubber walls above the spray head where melamine spattered or was carried
125
-------
Gas out
Spray head
Glass
column
Gas in,
113 1/min
(3000 ppm S02)
Bo°c
Stainless
steel
fitting
ll\
I I V
-10 cm-
35 cm
Screen
Surge tank
(500 ml)
Spent
slurry
3.8 1/min
melamine slurry
Figure ho. Spray scrubber used in melamine scrubbing tests—configuration 3-
126
-------
up as a mist. The material in this "buildup was almost; entirely a weak, pasty
cementation of melamine sulfite /(C3N6HS)2'H2S03'1+H20/ that was easily removed
with a low velocity stream of water or by scraping. It should be easy to
remove such material from a mist eliminator with makeup water.
Sulfur dioxide removal was such that acceptable removal would have been
possible with two such stages in series or with a crossflow scrubber.
Conclusions
It was concluded that melamine scrubbing could be used to remove S02 from
waste gas streams but that packed scrubbers would probably cause trouble with
plugging. If spray scrubbers are used, two such scrubbers probably will be
required; a crossflow type might be suitable. The inclusion of an antioxidant
is suggested to limit oxidation as much as possible.
When using actual stack gas, the use of a prescrubber to cool the gas,
remove most of the S03 present, and remove most of the fly ash is suggested.
The cooler gas will react more completely with melamine; melamine sulfite
exhibits a vapor pressure (S02) at moderate temperatures. The removal or
limiting of S03 is important because melamine sulfate is not regenerated
thermally. Limiting fly ash will make processing the melamine easier and
probably will help to limit oxidation; the removal of fly ash from the
melamine, when it becomes necessary, will be expensive.
127
-------
SODIUM SULFITE-MELAMINE PROCESS
PRODUCTION OF MELAMINE SULFITE FROM SODIUM BISULFITE SOLUTION AND MELAMINE
A cyclic process has "been proposed (13) for the removal and recovery of
sulfur dioxide from waste stack gas in which the gas would be scrubbed with a
sodium sulfite (Na2S03) solution to produce sodium bisulfite (NaHS03) solution,
leaving a waste gas essentially free of S02 to be expelled into the atmosphere.
As a second step, the NaHS03 solution would be reacted with melamine (C3HSN6)
to regenerate the Na2S03 solution and precipitate melamine sulfite
J/(~C3H6N6)2'H2S03'xH2o7. Next, the melamine sulfite would be regenerated
thermally or chemically to recover C3H6N6 for recycle and S02 in useful form.
Laboratory- scale tests now have been made with reagent-grade chemicals to
determine the feasibility of the C3H6N6-NaHS03 reaction step as shown in the
following equation:
2NaHS03 + 2C3H6N6 - J> Na2S03 + (C3H6N6)2-H2S03 -xH20
In the present work, the variables studied were the C3H6N6:NaHS03 mole
ratio, temperature, sodium bisulfite concentration, and test time. Test
procedure was as follows: The required amounts of NaHS03 and water for each
test were combined and placed in a water bath of controlled temperature. This
solution was given time to equilibrate while being stirred. When test tempera-
ture had been reached, the C3HsN6 was added as a finely divided powder (96$>
-48 mesh) to the NaHS03 solution and stirred for the required test period.
The final slurry was filtered through glass fiber filter paper on a medium-
frit funnel. The cake was slurried twice with water of weight approximately
1-1/2 to 2 times the weight of C3H6N6 used, and was refiltered. The cake then
was washed twice with ethanol and allowed to dry at room temperature under
house vacuum. After drying, the solids were submitted for analysis to
determine the percent sodium, nitrogen, total sulfur, and S02-sulfur (sulfite
sulfur). The solids were also examined microscopically for identification of
the crystal phases present and for comparative particle size of the melamine
sulfite crystals. Collected data and some calculations from these data are
shown in Table 2J. The NaHS03 used for these studies also was analyzed; this
analysis along with the percent of the NaHS03 oxidized, as calculated from
this analysis, is given as a footnote to Table 23.
The sodium analysis of the product solids indicates the efficiency of the
washing procedure. The only test which indicates the need for more thorough
washing is test 3 in which the more concentrated NaHS03 solution (25$) was
used. This also is the only test in which the amount of sulfur associated
with the sodium is large enough to affect the total sulfur analysis signifi-
cantly. Since, as indicated later, the melamine apparently exhibited a
preferential reaction with sulfite, the sodium probably would be present as
128
-------
TABLE 23. PRODUCTION OF MELAMINE SULFITE FROM MELAMINE
AND SODIUM BISULFITE SOLUTION (LABORATORY TEST DATA)
Test No.
C3HeN6:NaHS03a mole ratio
NaHSOs concentration in solution, %
Test time, min
Temperature, ° C
Input formulation, g
NaHS03
Melamine
HaO
Cake weight , g
Chemical analysis of solids , %
Na
Total S
S02-S
N
Calculated indices
Sulfur oxidized, $ of total
S:C3HsN6 mole ratio
Residual melamine, % of input
Bisulfite reacted, $ of input
Particle size
Av . microns
Av. area compared with test 4, $
2B
0.8
15
60
52
10.41
10.9
58.97
13.45
o.o4
5-7
5-3
51.6
7.01
0.29
4,:. 03
46.40
11 x 35
6
3
0.8
25
60
52
10.41
10.09
31.22
18.6
0.58
5-7b
5.0
37-7
12' . 3b
o.4ob
20.63b
64.00
37 x 63
37
4
0.8
15
60
38
10.41
10.09
58.97
15-0
0.05
7-1
6.3
43.8
11.3
0.43
13-79
68.80
43 x 148
100
5
1.0
15
30
52
10.41
12.61
58-97
16.15
O.o6
5-1
4-7
50.7
7-8
0.26
47-92
52.00
24 x 59
22
6
0.8
15
30
52
10.41
10.09
58.97
13.4
O.o4
6.4
6.0
48.7
6-3
0.34
31-97
54.40
17 x 36
10
7
0.6
15
60
52
io.4i
7-57
58.97
9-93
0.05
6.5
6.5
47.8
0
0.36
28.57
43.22
8 x 21
3
8
0.6
15
60
38
io.4i
7-57
58.97
10-94
0.02
7-3
7-3
44.0
0
0.44
r-i.o4
52.82
40 x 108
68
9
0.8
15
60
27
io.4i
10.90
58.97
15.30
0.01
7.2
7.1
45.6
1.4
0.4l
18.0:;
65.00
31 x 85
41
10
0.8
15
60
16
10.41
10.90
58.97
15.31
0.01
7-5
7-3
45-5
2.7
0.45
14. lu
68.60
43 x c9
47
Chemical analysis of NaHS03:
£ Calculated
Na Total S SQ~-S % oxidized
26.9 31-5 29-8
5.4
Corrected value for sulfur not combined with sodium.
-------
sodium sulfate. Therefore, calculations were made to correct the total and
sulfate sulfur analysis by excluding that tied up with the sodium. The
corrected values are given in the tables.
Evaluation was made of the measured responses and calculations from those
responses. Calculations include: percent of total sulfur oxidized; S:C3HSN6
mole ratio in the product; percent residual melamine; percent bisulfite
reacted; and average particle size of the melamine sulfite crystals in the
solid product. Oxidation to the sulfate form is undesirable since melamine
sulfate does not regenerate thermally to C3H6N6, and chemical regeneration
would be required to recover that portion of the C3H6W6. The S:C3H6W6 ratio
indicates the degree of reaction. In the reaction of NaHS03 solution with
C3H6N6, it is advantageous to react as much of the sulfite as possible; use
of an excess of C3H6N6 (low S:C3H6N6) would encourage high degree of reaction.
However, in the thermal regeneration step it is desirable to have as high an
S:C3H6N6 ratio as possible in order to reduce the energy requirement. An
optimization of this ratio will be a major consideration in the design of the
system. The nitrogen analysis offers a means of measuring the C3HsNe content
of the product (including that present as melamine sulfite and sulfate). The
percent residual melamine and percent bisulfite reacted are measures of the
degree of completion of the desired reaction. Particle size of the product
affects filtration rates; therefore, larger particles are desirable. The
effect of the four variables on these responses is shown in Table 2k. The
values footnoted b are so small as to be equal to or less than the precision
of the analyses and therefore probably are not significant. Those values
footnoted c_ lie within twice the value of the accuracy of the analyses and
are therefore of questionable significance.
Effect of Sodium Bisulfite Concentration
The effect of NaHS03 concentration on the responses measured and calcu-
lated is given in the first column of Table 2k. With increase in concentra-
tion from 15 to 25$, the melamine sulfite crystal size increased by 81$, but
kj% more of the total sulfur underwent undesirable oxidation to sulfate, while
S02-sulfur decreased, nitrogen decreased, and the percent residual melamine
decreased by 51$. The S:C3H6N6 mole ratio indicated more complete reaction
(S:C3H<3NS = O.lj-0, 27$ greater) with the higher concentration (desired S:C3H6N6
mole ratio is 0.5)« Percent of bisulfite reacted increased from k6 to 6k%
(Table 2j).
Effect of Temperature
A maximum temperature of 52°C was selected for testing since aqueous
media scrubbers on coal-fired boiler stack gas will equilibrate at about that
temperature. It should be possible to cool the reaction to 38°C using river
water as the coolant, so tests were made at this temperature, also. A
comparison of the results at 52°C and 38°C revealed such great differences,
with the values moving in the desirable direction at the lower temperature,
that tests were made also at 27°C and l6°C.
130
-------
TABLE 2k. EFFECT OF VARIABLES--DIFFEKENCES IN RESPONSES DUE TO VARIABLES
Variable
Test No.
Analysis , %
Na
Total S
SOp-S
N
Melamine sulfite
comparative particle size^
Calculated indices
S oxidized, %
S:C3HsNs mole ratio
Residual melamine , ?. of input
Bisulfite reacted, «f of input
Other conditions
OsHgNs iMaHSOg mole ratio
Temperature , °c
Time , min
NaHSO^ concentration, %
NaHS03
concentration,
15 25
2B 3
+0.54(93)"
0
-0.3 (6)
-13-9 (27)
+31 (81.)
+5-29(1*3)
+0.11(27)
-21.1*0(51)
+17.6 (28)
0.8
52
60
Temperature , °c C^H«NA
52 38
2B 1*
+0.01(20)b
+1.1* (20)
+1.0 (16)
-7-8 (15)
+91* (91*)
+1*. 29(38)
+0.14(33)
-28. 21* (67)
+22.1* (33)
52 27
2B 9
-0.03(75)
+1.5 (21)
-1.8 (25)
-6.0 (12)
+37 (90)
-5.6l(8o)
+0.12(29)
-21*. 00(57)
+19.2 (£9)
52 16
2B 10
-0.03(75)
+1.8 (21.)
+2.0 (27)
-6.1 (12)
+1*1 (87)
-4.31(62)
+O.ll*(33)
-27.87(66)
+22.1* (33)
38
4
27
9
-o.oi*(8o).
+0.1
+ 0.8
+ 1.8
-59
-9-9
-0.02
+4.2U
-3.2
i A
(D°
(11)
(59)
(88)
(5)b
(2"*)
(5)
38 16
4
-0,
+ 0
+1
+1
-53
-8
0
+0,
0
10
,04(80)
•>> (5)
.0 (14)
.7 (4)
(53)
.6 (76)
.37 (3)
(0)
27 16
9 10
0
+0.3 (4)
+0.2 (rf
-0.1(0.2)b
+6 (13)
+1.3 (48)
+0.02 (5)c
-3-87(21)c
+3-2 (5)
^
.
60
15
52 38
7 8
-0.03(60)
+0.8 (11)
+0.8 (11)
-3.8 (8)
+65 (96)
0
+0.08(18)
-15.53(54)
+9.6 (18)
0.6
60
15
0.8 0.6
2B 7
+0.01b(20)
+0.8 (12)
+1.2 (18)
-3.8 (7)
-3 (50)
-7.01(100)
+0.07 (19)
-13.46 (32)
-3.18 (7)
52
60
15
:NaHSO-, mole
0.8 0.6
4 8
-0.03 (60)
+0.2C (3)
+1.0 (14)
+0.2C(0.5)
-32 (32)
-11.3 (100)
+0.0l» (2)
-0.75b (5)
-15-98 (23)
38
60
15
ratio
1.0 0.8
5 6
-0.02C(33)
+1-3 (20)
+1.3 (22)
-2.0 (4)
-12 (55)
-1.5 (19)
+0.08 (24)
-15-95 (33)
+2.4 (4)
_
52
30
15
Time , min
60 30~
2B 6
0
+0.7 (11)
+0.7 (12)
-2.9 (6)
+4 (40)
-0.71(10)
+0.05(15)
-10.0 (24)
+8 (15)
0.8
52
15
*J Numbers in parentheses represent ^ difference in the two tests.
c Insignificant.
d Significance questionable.
Average crystal size area in each test compared with average crystal size area in test 4 using lOOjt for test 1*.
-------
Effects of temperature are most easily seen in Figure 14-1. The proportion
of sodium in the cake decreased as temperature decreased, which indicated
either better washing or better exclusion of the liquid phase from the crystals.
As the temperature decreased, total and sulfite sulfur contents increased and
the nitrogen content decreased; these changes, reflected in higher S:C3H6N6
ratio, indicated a more complete reaction at the lower temperature. Percent
of bisulfite reacted at 38°C was 69% versus k6% at 52°C (Table 23). At 27°C
and l6°C, the percent reaction was 66 and 69%, respectively. More complete
reaction is indicated also in the curve showing residual melamine (calculated
as % of C3H6NG above a C3H6N6:S mole ratio of 2). The particle size of the
melamine sulfite crystals, indicated by the average area of one face of the
crystals as viewed under the microscope (random orientation), increased
drastically as the temperature decreased from 52°C to J>8°C but then decreased
as temperature was lowered to 27°C and l6°C. Petrographic examination showed
the products at 38°C and lower to be a single phase, probably
(C3H6N6)2'H2S03'UH20, while the product at 52°C was a mixture of at least
three different salts. The proportion of sulfur oxidized, when the
C3H6NS:NaHS03 input mole ratio was 0.8, tended to follow the pattern exhibited
by particle size; maximum oxidation occurred at 38°C. It is noted that, when
the CsHeNs:NaHS03 input mole ratio was 0.6 (tests 7 and 8), the oxidation
value was 0; this may indicate a preferential reaction of melamine with
sulfite.
Effect of Melamine to Sodium Bisulfite Mole Ratio
This effect is best seen in Figure k2. Sodium content of the filter cake
was little affected but tended to trend downward as the ratio was decreased
from 1.0 to 0.8 to 0.6 mole C3H6N6 per mole NaHS03. As expected, total and
sulfite sulfur in the filter cakes increased and nitrogen decreased as the
ratio decreased. The S:C3HeN6 mole ratios in the products reflected chiefly
those in the raw materials; residual melamine decreased as the proportion of
melamine in the raw materials decreased. Response of percent bisulfite
reacted was greatest in the tests at ^8°C (tests h and 8); there was a 23%
decrease in percent reacted (53 vs. 69% reacted) as a result of lowering
C3H6N6:NaHS03 mole ratio from 0.8 to 0.6. The most meaningful responses were
in particle size which tended to decrease with decrease in the input ratio,
and in the proportion of sulfur oxidized which also decreased with decrease
in the input ratio.
Effect of Time
This effect is shown in the last two columns of Table 2k. The direction
of all the responses, excepting possibly oxidation, is the reverse of that
expected. Most of the differences in response were fairly small, however.
Further testing of this variable will be required to clarify the results.
Petrographic Analyses
From petrographic analyses, at least four different phases were found in
the product solids. As many as three of these phases were found to exist
simultaneously in the product from some of these studies. In the products
with the largest melamine sulfite crystals (tests k and 8) only one phase was
132
-------
Na, %
Total
sulfur,
H,
.06
.oi+
.02
S03-S, 6
NJ
52
38 27 16
52
o-'
38 27 16
^-O
I
52
38 27 16
52
*• -,
Particle
area,
microns2
(one side)
38 27 16
12
10
Sulfur o
oxidized,
% of 6
total k
2
0
0.^5
o.4o
mole ratio 0.35
0.30
0.25
O n
52
o/
38 :^7 16
38 27 16
Residual
melamine,
% of input
37
32
27
22
17
12
Symbol
A
O
C3HQN6:NaHS03
mole ratio
0.8
0.6
Time,
mm
60
60
NaHS03
concn , %
15
15
Temperature, ° C
Temperature,
52 38 27 16
Temperature, ° C
Figure hi. Effect of reaction temperature on composition of filter cake
from reaction of melamine with sodium bisulfite solution.
-------
0
0
Na, 5
0
0
Total
sulfur,
S02-S,
N, %
.06
.Ok
.02
8
7
5
8
6
k
52
50
kQ
kk
h.9
A--..^ <\>^* TOGO
^ 6000
5000
I | | Particle
1.0 0.8 0.6 area, ^»
microns^ JQOO
(one side)
2000
o — -° 1000
s ^S^*^ 0
ft'' \ 1
1.0 0.8 0.6 12
10
Sulfur g
s® oxidized, ..
^ n rf n-F 6
O^ .x^ i°
'X/^ total 4
X'1^
III 0
l.o 0.8 0.6
n 0.1*5
- ^"^.^NV 0>lf0
mole ratio
0-30
0 O
| 1 1 0.2S
Temp., Time, NaHS03
Symbol °C min concn,
^\
\ D 52 60 15
\ o 38 60 15
« A 52 30 15
>^
1.0 0.8 0.6
1 ^-^X
1 \\
\ r\
lib U? \ n
1.0 0.8 0.6 \^ NV
\ \
r Residual 32 _ ^ \
°~~ melamine, 27 a
% of input
A y^ 22 ~
- ,-'' n^ ^ -
l.o 0.8 0.6
iNaHSOs mole ratio
mole ratio
C3HeN6:NaHS03 mole ratio
Figure k2. Effect of C3H6N6:NaHS03 mole ratio on composition of filter cake
from reaction of melamine with sodium bisulfite solution.
-------
found and this phase, later identified as melamine sulfite tetrahydrate
/(C3H6N6)2-H2S03-l|.H20/, was different from those in the products containing
three phases. The surfaces of the crystals were pitted as a result of
deterioration.
Because of the deterioration of the crystals, there was a question
concerning a possible change in crystal phase as a result of drying the
product solids. Therefore, repetitions of tests k and 7 were made excluding
the drying procedure. One repetition of test k (Ij-V-l) and two repetitions
of test 7 (7-V-l and 7-V-1A) were made. Data from these tests are shown in
Table 25- QIn test 7-V-l there was difficulty maintaining a constant tempera-
ture of 52°C. The temperature tended to rise. Temperature was stable in
tests 14--V-1 and 7-V-1A. Petrographic analysis of the product solids was made
immediately following the tests and again 16 hours later. During this 16-hour
interval the crystals remained in the mother liquor at room temperature.
Petrographic analysis revealed that immediately following the tests, the
melamine sulfite crystals in tests i(-V-l and 7-V-1A were of the same phase
while the melamine sulfite crystals in test 7-V-l were of a different phase
and much smaller than those from the other two tests. The melamine sulfite
phase existing in tests Ij-V-l and 7-V-1A was the same as that in tests h and
8 and was identified as melamine sulfite tetrahydrate /CC3H6N6)2-H2S03'to2o7.
After 16 hours, the crystals in test 7-V-l had grown to be significantly
larger than the crystals of tests U-V-1 and 7-V-1A and were now the same
phase as in those tests. None of the original phase existed in test 7-V-l
after l6 hours.
Thermal Analysis
Analysis then was made of the melamine sulfite tetrahydrate by
differential scanning calorimetry at 20°C per minute. Three peaks were shown
(Figure kj>) upon heating from 27°C to 150°C. The first peak reached a maximum
at oVC, the second at 9^°C, and the third at 119°C. In further tests at 5°C
per minute, these peaks were separated. The weight loss represented by the
first peak was 9$ of the original weight of the sample which is equivalent to
2 moles of water. The weight loss depicted by the second peak is equivalent
to 1 mole of water, and the weight loss after completion of the third peak is
equivalent to an additional mole of water plus 1 mole of H2S03. This analysis
then supports the microscopic analysis, indicating that the melamine sulfite
analyzed exists as the tetrahydrate. It also offers a possible method of
separation of a major portion of the water from the S02 on thermal decomposi-
tion. The indication is that controlled heating could be used to dehydrate
melamine sulfite tetrahydrate prior to further thermal treatment for regenera-
tion.
Conclusions
It appears that the degree of completion of the reaction (indicated by
the S:CoHfiN6 ratio, % residual melamine, and bisulfite reacted) may be
controlled Sy adjustment of the C3H6N6:NaHS03 ratio and the temperature of the
reaction. In these tests 87^ utilization of melamine (test 8) and 69^
reaction of bisulfite (tests k and 10) were the best obtained. The degree of
135
-------
TABLE 25. PRODUCTION OF MELAMINE SULFITE FROM MELAKENE AND SODIUM
BISULFITE SOLUTION--ANALYSIS OF PRODUCTS
Test No.
C3%N6:NaHS03 mole ratio
NaHSOs concentration in
solution, %
Test time, min
Temperature, °C
Input formulation, g
NaHS03
Melamine
H20
4-V-l
0.8
15
60
38
io.4l
10.09
58.97
7-V-l
0.6
15
60
52a
10.41
7-57
58.97
7-V-1A
0.6
15
60
52
10.41
7-57
58.97
Particle size (av.),
microns
Immediate 4o x 40 x 120 1 x 25 x 25 100 x 100 x 150
16 hours 40 x 40 x 120 150 x JOO x 600 100 x 100 x 150
Indices of refraction
Immediate
o< 1.478
# 1.703
y 1.721
26 hours
oL 1.478
# 1.703
o 1.721
1.470
1.720
1.764
1.478
1.703
1.721
1.478
1.703
1.721
1.478
1-703
1.721
a
Difficulty in holding temperature, fluctuation upward.
136
-------
Endothermic
Figure ifj. DSC trace of melamine sulfite tetrahydrate.
137
-------
hydration may be controlled by the temperature of the reaction, also. Control
of the type and size of the reaction product should be possible. The degree
of oxidation encountered probably will not be harmful to the C3H6N6-NaHS03
reaction, but can be a problem in filtration and thermal regeneration. The
work performed lends credence to the feasibility of the proposed process.
PRODUCTION AND CHARACTERIZATION OF MELAMINE-SULFTJR OXIDE ADDUCTS IN PURE FORM
Further studies were made to produce, in relatively pure form, the four
products of the reaction of melamine with sodium bisulfite found in the
earlier tests and to characterize them. All four were produced, but only
three of these in homogeneous products. The optical descriptions of these
four materials and their empirical formulas (calculated from chemical analyses)
are given in Table 26.
Test Plan
All test products were made in a 15$ sodium bisulfite solution with a
melamine to sodium bisulfite mole ratio of 0.6, and all slurries were stirred
continuously during the test period. Test time and temperature varied
(Table 27). Test solutions 3-1 and 3-2 contained 0.05$> p-phenylenediamine,
an antioxidant; the other test solutions contained no antioxidant. Heated
test products remained in the mother liquor while cooling slowly to ambient
temperature to allow maximum crystal growth. The solids were then filtered,
washed, dried, and submitted to the petrographic lab for analysis. Products
were analyzed chemically for percent total sulfur, S02-sulfur, and nitrogen.
Differential scanning calorimetry (Perkin-Elmer Differential Scanning Calorim-
eter, Model DSC-l) tests were made on each product, using about 10-milligram
samples. Test time and temperature, chemical analyses, calculations from
those analyses, and calculated weight loss during differential scanning
calorimetry (DSC) testing are given in Table 27 along with refractive indices.
Test Results
The monoclinic prism produced in test 2-1 was identified as melamine
sulfite tetrahydrate, (C3H6NS)2•H2S03'1)-H20, with lk% oxidation. This compound
was produced in several tests with varying degrees of oxidation and can be
identified petrographically, although the percent oxidation cannot be
determined petrographically. When this salt is produced in a sodium bisulfite
solution with melamine added to the solution, it is several hours before all
the melamine reacts. The melamine reacts more quickly when the slurry is
heated, but at temperatures above kO°C salts other than the tetrahydrate are
formed, largely the dihydrate. When the dihydrate is allowed to remain in
the mother liquor at temperatures below 25°C for several hours, the tetra-
hydrate begins to form. On exposure to air the tetrahydrate crystals
deteriorate; acetone also erodes these crystals. The weight loss on heating
in the DSC was 36.8$; calculated H2S03 + H20 content amounted to 36.7$. It
appears that all of these components are lost on heating to l87°C.
138
-------
TABLE 26. OPTICAL DESCRIPTION OF ADDUCTS
Common habit
b-l
1100
-\-y
Monoclinic prism
Test No. and crystal description
Test 2-1
These crystals precipitate as monoclinic, 2/m
prisms elongated along £, modified by the (010)
and (100) and terminated by {oil}, b = y
n oc l.Vf8, n
Probable formula:
1.703, n 7 1.721
001
Monoclinic plate
x
\
001-4
x'
010
Monoclinic blade
Test 2-2
This salt crystallizes as orthorhombic tablets ,
tabular on the x-y plane and slightly elongated
along x.
n c*. 1.^50, n £
Probable formula:
(C3HsN6)3-H2S04-ta20
1.721, n V 1-735
Test 2-5
These crystals precipitate as monoclinic, 2/m
plates elongated along £, modified by (010)
and (001). b z
n of. 1.UT3, n/
Probable formula:
Test 3-1
This salt crystallizes as monoclinic, 2/m
(010) blades elongated along £ modified by
small (001) and (100). This salt also exists
as elongated plates and rods, b = z
n o<
Probable formula:
159
-------
TABLE 27. SODIUM SULFITE-MELAMINE PROCESS--PREPARATION
OF PURE REACTION PRODUCTS FOR CHARACTERIZATION
Test No.
C3H^Ne:S mole ratio in mixture
Test time, hr
Before heating
Heating
Stored at room temperature
Temperature to which heated, °C
Crystal habit of solids'3
Refractive indices
o*.
£
rf
Chemical analysis, %
Total S
soa-s
Nitrogen
Calculated indices
S oxidized, % of total
C3H6Ne:S mole ratio
H20:S8 mole ratio
Weight loss on heating, %
DSC test weight loss, %*•
a 0.05$ p-phenylenediamine added
b MPR = monoclinic prism, OTB =
blade, MRD = monoclinic rod.
° 60J& of solids.
d 20% of solids.
e 50]6 of solids.
f 35# of solids.
2-1
0.6
0
0
21
Ambient
MPR
1.1*78
1.703
1.721
7-66
6.61
1*0.10
13-71
2.00
*-57
36.7
36.8
2-2
0.6
18
6
0
70
OTB
1.1*50
1.721
1-735
5-*
0.3
1*6.2
9U. 1*1*
3-25
1*.71
15.1
13.8
2-3
0.6
18
6
23
> 70
MBL
1.1*60
1.735
1.768
8.9
3-1
1*2.2
65.17
1.81
2.20
19.0
15.6
2-5
0.6
72
1
13*
88
MBLC MPLd
1.1*71 1.1*73
1.7*2 1.7*7
1.765 1.76*
8.6
5-35
37-79
1-93
2.26
2k .6
23.8
3-1
0.6a
0
5
16
70
MBL
1.1*78
1.7*6
1.77*
8.1*
7-1
M* .6
15- *8
2.02
2-31
29.1
29.1
3-2
0.6a
72
1-75
21
88
MBLe MRDr
1.1*72 1.
1.738 1.
1.768 1.
8.5
7.1
1*5.0
16.1*7
2.02
2.10
28.2
30.2
1*75
7*0
772
as antioxidant.
orthorhombic tablet, MPL
g Pnlniltrt-i a- 7100 ' **' CaH«N8 ' **' Hs>SO* '
wt. HPS(
= monoclinic
Vi XL^= H
plate, MBL = monoclinic
[->0:S mole
i ratio.
18 J
h Loss of HaO + H2S03. No loss of H2S04 or C3HsN6.
1 Differential scanning calorimeter tests heating 10-mg samples to 187" or 227°C at 20°C/min.
-------
The orthorhonibic tablets produced in test 2-2 were 9^1 sulfate. These
crystals are eroded by acetone but do not deteriorate on air drying. Chemical
analysis indicates this compound has 3 moles of melamine to 1 mole of sulfur
with approximately four waters of hydration /rc3HsN6)3-H2S04'te2p7. Calcula-
tions from the chemical analysis indicate that on heating, this compound could
have a 15.1% weight loss. DSC analysis showed that on heating this material
from Vf to 20T C, the weight loss was 13.8^.
Tests 2-2 and 2-3 were begun in the same slurry. Both orthorhombic
tablets and monoclinic blades formed in the slurry with the blades so small
as to pass through the filter (medium-fritted Gooch crucible) with the filtrate.
The tablets then were air dried and designated as test product number 2-2
(C3H6N6)3'H2S04-lj.H20. The filtrate containing the blade crystals was heated
until a clear solution resulted and then was allowed to cool slowly to form
large blades which could be filtered. This sample was designated 2-3. Because
these monoclinic blades were formed in the same slurry as a 9^ sulfate salt,
they may represent approximately the maximum sulfate substitution (about 65")
in the monoclinic blade salt. DSC analysis showed a loss of 15.61 versus a
theoretical value of 19^.
Twenty percent of the product from test 2-5 was monoclinic plates, thought
to be (C3H6N6)2-H2S04-2H20. This salt was not produced as a homogeneous
product in these tests. The bulk of the product from test 2-5 (60-1) consisted
of melamine sulfite dihydrate with the balance as decomposing melamine sulfite
tetrahydrate. DSC analysis showed a loss of 23.8$ versus a theoretical 2k.6^.
Test 3-1? in which the antioxidant was added, resulted in the formation,
after heating to 70° C and overnight cooling, of monoclinic blade crystals.
About 15% of the sulfur was in sulfate form. DSC testing resulted in 29.1;1
weight loss which is the theoretical value.
Test 3-2, a repeat of test 2-5 except with antioxidant added, resulted
in a mixture of monoclinic blades and monoclinic rods. About ±6% of the
sulfur was oxidized to sulfate. DSC testing resulted in 30.2$ weight loss
versus a theoretical 28.2$.
The majority of the products from this series of tests consisted of mono-
clinic blades which were first identified in test T-V-1. Chemical analyses of
these products indicate 2 moles of melamine per mole of sulfur with approxi-
mately 2 moles of water. The degree of oxidation varied. The highest degree
of oxidation appeared in the products which were allowed the longest reaction
time except when an antioxidant was added to the slurry. Products exhibiting
the least oxidation were from tests where antioxidant was added to the test
slurry. The weight loss during DSC tests equaled or nearly equaled the
calculated values.
Although the refractive indices of the monoclinic blades vary, this
variation does not form a predictable pattern for determining the degree of
oxidation of a particular sample. There is some indication that over a period
of time there is a slight variation in the indices themselves. It is possible
that this variation might be caused by small changes in the quantity of water
associated with the crystals.
-------
The DSC test curves for the three fairly pure products are shown in
Figure kk. As reported earlier, the test for the monoclinic prisms,
(C3H6N6)2'H2S03'l)-H20, indicates that a major portion (about J> moles) of the
water might be separated from the S02 on thermal decomposition. The weight
loss on heating at 20°C per minute to 90°C was 9.3% and to 103°C was 13.5$
(2 moles of water would be 9-5% a-ftcl 3 moles of water would be 13«3^) • Testing
of the off-gas over this range indicated little S02 to be present. The
remainder of the weight loss, equivalent to 1 mole of water and 1 mole of
sulfurous acid, was lost between 103° and 150°C. Separation of a major portion
of the water from the sulfurous acid should be possible in the proposed process.
The DSC curve for the monoclinic blade product, (C3HeN6)2•H2S03•2H20,
shows all the reaction to take place in one large endothermic peak. All the
weight loss (2 moles of water plus 1 mole of sulfurous acid) takes place
between 92° and ~Lk5°C. No separation of water from sulfurous acid by thermal
decomposition is indicated.
DSC analysis of the orthorhombic tablets (C3H6N6)3'H2S04'i»-H20, shows the
reaction to take place in three steps. Apparently the water was evolved with
1 mole being lost in heating to 100°C (about 3$ wt. loss), an additional 2
moles at 150°C (about 6% wt. loss), and the last mole at about l80°C (about 5%
wt. loss). No sulfur should be lost in this test since the sulfate is not
volatile at the temperatures involved.
DSC test curves for the mixed products and those of major proportions of
both sulfate and sulfite are exemplified by the curves in Figure 45- These
curves are not definitive enough to identify the specific losses at any given
temperature although the total losses are very near the theoretical values.
These products would not allow separation of water from S02.
Recommendations
In the proposed process the crystal form most likely to be produced is
the blade-type crystal (C3HSN6)2'H2S03'2H20. Products consisting of this
crystal habit are very slow filtering and drying, especially when compared
with the filtering time of the tetrahydrate products. If the dihydrate
crystals were allowed to remain in the mother liquor containing antioxidant
and at temperatures of 25°C or below for a period of time to allow the tetra-
hydrate to form, the tetrahydrate could then be easily filtered and heated to
about 103°C to drive off 3 moles of water. The remaining product could then
undergo melamine regeneration and with addition of oxygen result in 8k.5%
H2S04. The economic feasibility of such a process or of one using the
dihydrate directly would need to be studied.
PRODUCTION OF MELAMINE SULFITE FROM MELAMINE AND SODIUM BISULFITE SCRUBBER
LIQUORS
Exploratory operation of the minipilot plant was carried out in the mode
of the sodium sulfite-melamine double reactant process (13). A sodium sulfite
solution containing 0.1$ p-phenylenediamine as an antioxidant was fed and the
spent solution was reacted with melamine. The regenerated solution was
142
-------
227
2Of
187
167
g
cu
EH 127
107
87
67
- lj-20
°K
500
WO
1^60
380
360
3^0
320
Empirical formulas
Test 2-1:
Test 2-2:
Test 3-1:
Test No.
(C
(C3H6N6)2H
2-1
Crystal habit:
Monoclinic prisms
Q.h-mg sample
Sensitivity, 8
S03)0.
S03)0<
S03)0,
Test No.: 2-2
Crystal habit:
Orthorhombic
8.85-mg sample
Sensitivity, 8
Test No.: 3-1
Crystal habit:
Monoclinic blades
10.5-mg sanrple
Sensitivity, 16
tablets
Endothermic
reaction
Figure kk. Sodium sulfite-melamine process--
differential scanning calorimeter analysis of pure products.
-------
>K
227 i- 500
207
187
167
-------
recycled to the scrubber and the spent melamine was held for later recovery.
In the run ^ it was found that the use of the stoichiometric amount of melamine
to react with the S02 in the incoming gas did not result in sufficient regenera-
tion of sodium sulfite solution to maintain high S02 removal. The use of
significantly greater portions of melamine did not improve the situation.
Apparently, the acceptance of a low regeneration efficiency, a higher forward
flow in the scrubber, and a lower pH in the scrubber would be required in a
viable process. It was decided to use the spent scrubber solution collected
during the latter part of the run to further study the reaction with melamine.
Eight tests were made in which the melamine stoichiometry, as illustrated
by the following equation, was varied from Uo/o to
2NaHS03 + 2C3H6N6 - > (CaHeN6)2'H2S03':xH20 + Wa2S03
In each test 150 milliliters of the spent scrubber liquor (0.231 g/ml total
sulfur, 0.228 g/ml sulfite sulfur, 0.22U g/ml bisulfite sulfur) was placed in
an 8-ounce sample bottle and allowed to equilibrate in a water bath at 52°C.
The desired amount of melamine then was added batchwise with constant stirring.
In the first five tests the reacting slurry was left open to air (aerobic); in
the last three tests the reacting slurry was covered with a nitrogen blanket
and closed with a rubber stopper to minimize the effects of air contact
(anaerobic). Slurry samples were taken at various time intervals throughout
each test and were filtered immediately (coarse-frit funnel with glass-fiber
filter paper) using a minimum of vacuum. The filter cakes were washed with
ethanol (about 10 ml) and air dried for 5 minutes using the house vacuum.
Microscopic examination of the dried cakes and chemical analysis of the
filtrates (bisulfite sulfur) were performed by personnel of this laboratory.
The slurry remaining at the end of each test was filtered using the same
procedure and those filter cakes were submitted to the analytical laboratory
for chemical analysis. The test data are listed in Table 28.
Results of Filtrate Analyses
The proportion of NaHS03 reacted ($ of input) calculated from the
analyses of the filtrate samples taken throughout each test are listed in the
table. These results indicate an average of 35$ of the input NaHS03 reacted
at the end of all the tests. The maximum proportion reacted was h9% of input
in test 8 (100$ stoichiometry, anaerobic) after a total reaction time of 52
hours. In the aerobic tests, the extent of the reaction, based on the
bisulfite analysis of the filtrate, never exceeded ^0;'' but reached a maximum
value and then declined before rising again as shown in Figure h6 (only four
tests plotted for simplicity) . The timing of the first maximum depended on
the melamine stoichiometry; higher stoichiometry resulted in more rapid
attainment of the maximum. Linear regression analysis gives the equation
for this relationship as follows:
Time of maximum reaction, min = 59-7^ 0.1^6 (stoichiometry, $)
The correlation coefficient, 0.981, indicates a pronounced linear relationship
between the variables. This relationship is shown in Figure Iff. The reason
for the reaction to reach an indicated maximum and then to decline is not
-------
TABLE 28. PRODUCTION OF MELAMINE SULFITE FROM MELAMOTE
AND WaHS03 SCRUBBER LIQUOR—LABORATORY TESTS
Test No.
Formulation, g
NaHS03a
C3HeN6
CgHsNerNaHSOg mole ratio
Temperature , ° C
Time , min ( total )
Air contact
NaHSOs reacted, % of input
at time indicated, min
2
15
30
^5
60
75
105
135
150
180
240
420
1380 (23 hr)
1680 (28 hr)
i860 (31 hr)
3120 (52 hr)
3300 (55 far)
Final solids analysis, %
N
S
Sulf ite S
Calculated indices6
C3HeNe:S mole ratio
Residual C3HeN6, % of
solids
NaHSOs reacted, % of
input
S oxidized to sulfate, %
1
10.9
13.2
1.0
52
240
Yes
_
27.7
30.8
34.8
29.0
32.3
_
-
50.6
3.9
2.1
4.95
59-5
40.5
46
2
10.9
26.5
2.0
52
180
Yes
_
35-3
38.8
35-3
34.4
34.4
33.9
-
37-1
-
59-6
1.6
0.24
14.09
85.9
28.2
85
3
10.9
52.9
4.0
52
135
Yes
33.8
33-1
31-3
31.7
29-5
-
-
-
63.2
0.80
0.13
30.30
93-^
26.6
84
4
10.9
8.0
0.6
52
150
Yes
23.7
30.4
33-9
34.4
_
33.9
35-3
-
_
-
-
48.3
5-3
5-2
3-47
42.4
34.6
1.9
5
10.9
5-3
0.4
52
195
Yes
17.0
25.9
29.0
30.4
34.8
34.8
34.4
-
-
33-0
_
_
-
-
-
41.3
7-8
7-5
2.01
0.8
39.7
3.9
6
10.9
13.2
1.0
52
75
No
_
36.6
37-1
37.1
38.4
35-3
_
_
_
-
51.3
2.9
2-3
6.76
70.3
29.7
21
7
10.9
13.2
1.0
52
60
No
32.6
37-5
36.2
35-7
33-0
_
-
47-5
2.7
2.1
6.71
70.2
29.8
22
8
10.9
13-2
1.0
52
3300
No
_
=
=
_
38.4
45.1
45.5
47.4
42.2
49.1
44.2
52.4
2.9
2.4
6.90
71.0
29.1
17
Fed as 150 milliliters of spent scrubber liquor from minipilot-plant run; contained
b 231 g/1 total sulfur, 288 g/1 sulfite sulfur, 244 g/1 bisulfite sulfur.
1 mole C3HeNs per mole of NaHS03 is 100# stoichiometric using following equation;
2NaHSQ3 + 2C3HeN6 > (C3HeN6)2-H2S03'xH20 + Na2S03
As determined from bisulfite analyses as determined by potassium iodide iodate
titration.
Residual solids after filtering, washing with ethanol, and drying in air.
_ Based on final solids analysis.
Based on residual melamine proportion and stoichiometry.
-------
30
V
-P
2 20
10
0
Symbol
O
A
0
X
Test
5
1
2
3
C3HSNS
stoich. , %
ko
100
200
Uoo
I
I
0
60
120
Time, min
180
Figure k6. Production of melamine sulfite
from melamine and NaHS03 scrubber liquor--
proportion of NaHSQs reacted versus time.
-------
0
100 200 300
Input melamine stoichiometry,
Figure hj. Production of melamine sulfite from melamine
and WaHS03 scrubber solution—effect of stoichiometry
on time required to reach maximum proportion of NaHS03 reacted.
148
-------
understood but may be associated with oxidation to sulfate. The aerobic tests
were not continued as long as test 8 in which the greatest degree of reaction
was attained but were continued for about as long as could reasonably be
allowed in a continuous scrubbing process.
The rates of reaction during the first 2 minutes of the aerobic tests
(tests 3, 4, and 5) were 16.9, 11.8, and 8.5£ when the melamine s to ichiome tries
were 1+00, 60, and 40$, respectively. The pH rise for the same periods (from
the initial 4.8) were 0-9, 0.8, and 0.7, respectively. These data indicate
the effect of melamine stoichiometry on rate; increased amounts of melamine
increased the rate of bisulfate reaction but not the extent.
The three anaerobic tests (6, 7, and 8) were made at 100-i melamine
stoichiometry with the time of the tests being the only intentional variable.
Tests 6 and 7 were continued for 60 and 75 minutes , respectively. Results
were essentially parallel; excepting the analyses at 60 minutes, the values
for proportion of bisulfite reacted varied less than l.Up. Test 8 was a longer
term test. It was analyzed first at 240 minutes and was maintained for 3500
minutes ( 55 hr) . The degree of reaction of the bisulfite became fairly
constant at about 45$> after 7 hours.
Results of the Final Filter Cake Analyses
The chemical analyses of the final filter cakes and the indices calculated
from those analyses are listed in Table 28. It is noted that these results
show some variations from the NaHS03 proportions reacted calculated from the
filtrate analyses. These variations probably result from the calculations
having been based on the bisulfite sulfur in one case but based on the
residual melamine proportion in the other case. Except in test 8 (55 hr) the
values varied less than 10 percentage points.
The proportion of sodium bisulfite equivalent reacted in the final filter
cake (calculated from the total sulfur analysis) averaged 34$ for aerobic
tests and averaged 30$ for the anaerobic tests. The range of values for the
aerobic tests was from 27$ in test 3 to 41$ in test 1. The range in the
anaerobic tests was less than 1$ (29 -5-0'^) •
The effect of stoichiometry on the CgHeNgtS mole ratio in the solids can
be illustrated by Figure 48. This plot shows that the C3H6N6:S mole ratio in
the cake increased linearly with an increase in the input melamine stoichiom-
etry. The equation of the line (using the results from all eight tests) is:
C3H6N6:S mole ratio in filter cake = 0.080 ( stoichiometry) - 1.46
The correlation coefficient, 0.998, indicates the high probability of a linear
relationship between the variables and that the low stoichiometry (4056) did
not adversely affect the degree of reaction.
It was noted also that increasing the stoichiometry generally resulted
in increased oxidation of sulfur. The proportion of sulfur oxidized reached
a maximum of 85$ at 200$ stoichiometry and held that level at 400$ stoichiom-
etry. This result can be illustrated by a graph of sulfur oxidation versus
149
-------
0
100 200 300
Input melamine stoichiometry,
Figure k&. Production of melamine sulfite from melamine
and NaHS03 scrubber liquor—effect of stoichiometry
on C3H6N6:S mole ratio in the final filter cake.
150
-------
stoichiometry as in Figure !#. These results indicate a possible preferential
reaction with sulfite sulfur (also noted in the previous work with reagent-
grade materials). The effect of keeping air away from the slurry at the 100$
stoichiometric level was a 57$ decrease in sulfur oxidized.
Results of Microscopic Examination
Microscopic examination of the solid samples taken throughout each test
revealed that at least three phases were present in addition to melamine. Two
of these phases were identified as (C3H6N6)2-H2S03-ta20 and (C3HeN6)-H2S03-2H20.
The third appears to be the (C3H6N6)2-H2S03-to20 salt having a different
crystalline habit. All probably had some degree of sulfate substitution. The
proportion of each phase varied greatly between some tests but was similar for
other tests. However, no correlation between the phase or phases precipitated
and any of the test conditions was noted.
Conclusions
Under the conditions used in these tests, it was determined that an
average of about 33% of the input NaHS03 reacted with the melamine added. It
was found that increasing the melamine stoichiometry resulted in (l) an
increase in the C3H6N6:S mole ratio in the final filter cake, (2) an increase
in sulfur oxidized in the final filter cake, and (3) an increase in the rate
of reaction. Also, keeping the reaction slurry from contact with air signif-
icantly reduced the proportion of sulfur oxidized at 100?; stoichiometry. The
test results were reproducible within reasonable accuracy.
The results most adaptable to the proposed double reactant process were
obtained in test 5 in which kofy stoichiometric melamine was added. A product
containing less than 1% free melamine was obtained in which oxidation was less
than !<-$. The regeneration of such material would be much less expensive than
products containing more melamine and higher degrees of oxidation. Operating
in this manner would require higher forward flow rates through the scrubber
circuit, however, and the scrubber efficiency would have to be greater than
if a scrubber feed solution containing little or no bisulfite could be
produced.
151
-------
G
•H
0)
H^
2 <*H
w H
W CO
O
a
A-H
a -P
O a5
•H -P
-P -H
d P<
^ -H
•H O
X 0)
O !H
P)
100
80
60
20
0
Symbol Tests
O
0
1-5
6-8
j
0
100 200 300
Input melamine stoichiometry,
koo
Figure h9- Production of melamine sulfite from melamine
and NaHS03 scrubber liquor—effect of stoichiometry
on sulfur oxidation in the final filter cake.
152
-------
REFERENCES
1. Linke, W. F., and A. Seidell. Solubilities of Inorganic and Metal Organic
Compounds, Vol. II, pp. 29^-295, American Chemical Society, Washington,
DC, 1965-
2. Davies, 0. L. Design and Analysis of Industrial Experiments, published
for Imperial Chemical Industries "by Oliver and Boyd, Hafner Publishing
Company, New York, 1956.
3. Lele, V. N., and J. M. Joshi. "Reactions in the Reduction Process for
Potassium Carbonate." Chem. Age India 19, Wo. 1, 1J-1^, January 1968.
k. Hori, S. (Tokyo, Japan). Apparatus for Causing Intimate Contact Between
Gas and Liquid. U.S. Patent 3,802,67^, April 9, 197^.
5- Bard, M. L. (Bard and Bard Company, Joplin, Missouri). Turbine Mixer.
U.S. Patent 3,7^,765, July 10, 1973-
6. Slack, A. V., and G. A. Hollinden. Sulfur Dioxide Removal from Waste
Gases. Second Edition, pp. 275-276, Noyes Data Corporation,
Park Ridge, New Jersey, 1975*
7. Dakan, R. B., R. A. Edwards, and R. E. Rush. Froc. , EPA Symposium on
Flue Gas Desulfurization. Interim Report on Chiyoda Thoroughbred
101 Coal Application Plant at Gulf Power's Sholz Plant, New Orleans,
Louisiana, March 197°"•
8. Fraser, J. H., and H. E. Davis. "Laboratory Investigations of
Concentrating Industrial Wastes by Freeze Crystallization." Paper
presented at the 79th National Meeting of the American Institute of
Chemical Engineers, Houston, Texas, March 16-20, 1975-
9. Barduhn, A. J. "Desalination by Crystallization Processes." Chem. Eng.
Prog. 63, No. 1, 98-103, January 1967.
10. Barduhn, A. J. "The Status of Freeze-Desalination." Chem. Eng. Prog.
71, No. 11, 80-87, November 1975-
11. Ross, W. H., and R. M. Jones. "The Solubility and Freezing-Point Curves
of Hydrated and Anhydrous Orthophosphoric Acid." J. Am. Chem. Soc. hj,
2165-2170, August 1925.
155
-------
12. McCullough, J. F., and J. J. Kohler. Technical Suggestion 1676.
Tennessee Valley Authority, Muscle Shoals, Alabama, October 7? 197^-
1J. Potts, J. M. Technical Suggestion 1760. Tennessee Valley Authority,
Muscle Shoals, Alabama, December 22, 1976.
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing/
1. REPORT NO.
EPA-600/7-78-216
3. RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE
Advanced Concepts: SO2 Removal Process
Improvements
. REPORT DATE
November 1978
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
John M. Potts and John E. Jordan
8. PERFORMING ORGANIZATION REPORT NO.
TVA Y-139
. PERFORMING ORGANIZATION NAME AND ADDRESS
Tennessee Valley Authority
Division of Chemical Development
Muscle Shoals, Alabama 35660
10. PROGRAM ELEMENT NO.
EHE624
11. CONTRACT/GRANT NO.
EPA Interagency Agreement
IAG-D6-E721-BM
2. SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development
Industrial Environmental Research Laboratory
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND PERIOD COVERED
Final; Through 12/77
14. SPONSORING AGENCY CODE
EPA/600/13
5. SUPPLEMENTARY NOTES IERL-RTP project officer is Robert H. Borgwardt, MD-65, 919/
>41-2336.
6. ABSTRACT
The report gives results of a study of a potassium scrubbing system that
recovers useful forms of sulfur from pollutants while utilizing a low-energy process
o regenerate the absorbing medium. The report also describes two versions of a
new, regenerable process for SO2 scrubbing that show promise for further develop-
ment; they are based on absorption in a melamine slurry, followed by low tempera-
ure thermal or chemical regeneration. No solids requiring disposal are generated
jy the potassium scrubbing system. Laboratory tests of the scrubbing, thermal
decomposition, and reduction steps show that the process (based on the precipitation
f potassium pyrosulfite) is feasible. The absorbing solution is regenerated by ther-
mal stripping of one-third of the pyrosulfite and chemical reduction of the remainder
o form hydrogen sulfide. Elemental sulfur is obtained via the Claus reaction. The
report includes studies of the oxidation of various sulfite compounds produced in
everal different SO2 scrubbing processes. Using laboratory, bench-scale equipment
f new design, the sulfites of ammonia, potassium, zinc, and calcium were success-
ully oxidized with high efficiency. Such oxidation will be beneficial to the value of
he products and to improve physical properties or reduce chemical oxygen demand
)f the waste.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
c. COSATI Field/Group
Air Pollution
?lue Gases
Desulfurization
Scrubbers
Potassium
Sulfur Oxides
Fertilizers
Sulfur
Regeneration
Absorbers
Hydrogen Sulfide
Melamines
Air Pollution Control
Stationary Sources
Phosphate Fertilizer
Potassium Pyrosulfite
Claus Reaction
13B 02A
21B
07A,07D
131 11G
07B
07C
8. DISTRIBUTION STATEMENT
Unlimited
19. SECURITY CLASS (This Report)
Unclassified
21. NO. OF PAGES
166
20. SECURITY CLASS (Thispage)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
155
------- |