EPA-R2-73-214
April 1973 Environmental Protection Technology Series
Evaluation of Problems
Related to Scaling
in Limestone Wet Scrubbing
Office of Research and Monitoring
U.S. Environmental Protection Agency
Washington, D.C. 20460
-------
EPA-R2-73-214
Evaluation of Problems
Related to Scaling
in Limestone Wet Scrubbing
by
Dr. Joan B. Berkowitz
Arthur D. Little, Inc.
20 Acorn Park
Cambridge, Massachusetts 02140
Contract No. 68-02-0215
Program Element No. 1A2013
EPA Project Officer: R. H. Borgwardt
Control Systems Laboratory
National Environmental Research Center
Research Triangle Park, North Carolina 27711
Prepared for
OFFICE OF RESEARCH AND MONITORING
U.S. ENVIRONMENTAL PROTECTION AGENCY
WASHINGTON, D.C. 20460
April 1973
-------
This report has been reviewed by the Environmental Protection Agency and
approved for publication. Approval does not signify that the contents
necessarily reflect the views and policies of the Agency, nor does
mention of trade names or commercial products constitute endorsement
or recommendation for use.
11
-------
BIBLIOGRAPHIC DATA '• ReP"« No.
SHEET EPA-R2-73-214
Evaluation of Problems Related to Scalir
Scrubbing
7. Author(s)
Dr. Joan B. Berkowitz
9. Performing Organization Name and Address
Arthur D, Little, Inc.
20 Acorn Park
Cambridge, Massachusetts 021bO
12. Sponsoring Organization Name and Address
EPA, Office of Research and Monitoring
NERC/RTP, Control Systems Laboratory
2- 3. Recipient's Accession No.
5. Report Date
ig in Limestone Wet April 1973
6.
8. Performing Organization Rept.
No.
10. Project/Task/Work Unit No.
11. Contract /Grant No.
68-02-0215
13. Type of Report & Period
Covered
Final
Research Triangle Park, North Carolina 277! 1 '«•
15. Supplementary Notes
16. Abstracts ^ repor(. deflnes the relatior
chemistry, based on thermodynamic analyse
includes a review of scaling problems in
that mechanical scale deposition (due to
gas/liquid distribution) can be controllc
controlling chemical scale deposition arc
dence time In a holding tank, temperature
that SO?, removal efficiency is a linear
CaCOB and 0.6% limestone slurries, and 1'
non-scaling, but removal efficiencies arc
50$. In CaC03 scrubbers, S03 scaling can
al efficiencies, if the pH of the scrubl
of 5«7-5.8 where CaS03 is fairly soluble
17. Key Words and Document Analysis. 17o. Descriptors
Air Pollution Anhydrite
•"Limestone
*Scrubbers
*Scaie (Corrosion)
Slurries
Sulfites
Sul fates
Carbonates
PH
Sulfur Dioxide
17bt Identifiers/Open-tnded Terns
Air Pollution Control
*Wet Scrubbing Process
Blsulfites
17e. COSAT1 Field/Group 07A, 07B, 07D, 1 3B
IB. Availability Statement
Unlimited
iship between scale formation and scrubber
is and bench-scale scrubber experiments, and
limestone wet scrubbing systems. It is assumed
wet/dry interfaces, stagnant areas, or poor
>d by good engineering design. Approaches to
;: pH control, high L/G, seeding, high resi-
5 control, and oxidation control. It is shown
function of pH for 0.5$ CaS03 slurries, 0.5$
t CaC03 slurries. The CaC03 scrubbing system is
> limited by the S03/HS03 equilibrium to about
be minimized while maintaining 80-90$ reraov-
>er effluent is controlled in the neighborhood
. A 100-125 F temperature Increase seems to
increase the rate of scale deposition and,
in the presence of 02, can lead to catastro-
phic scaling via a cementing reaction
initiated by the precipitation of anhydrite.
19.. Security Class (This 21. No. of Pages
Report)
LJ^CLASSIFIED
20. Secuiity Class (This 22. Price
Page
UNCLASSIFIED \
FORM NTIS-35 IR6V. 3-72)
USCOMM-DC M9S2-P72
-------
ABSTRACT
Scaling problems in limestone wet scrubbing systems have been reviewed.
Thermodynamic analyses and bench scale scrubber experiments have been
carried out to define the relationship between scale formation and
scrubber chemistry. It has been assumed that mechanical scale deposi-
tion, due to wet-dry interfaces, stagnant areas, or poor gas-liquid dis-
tribution can be controlled through good engineering design. The ap-
proaches to controlling chemical scale deposition are: pH control,
high L/6, seeding, high residence time in a hold tank, temperature con-
trol, and control of oxidation. It has been shown that sulfur dioxide
removal efficiency is a linear function of pH for 0.5% CaSO, slurries,
0.5% CaCO- and 0.6% limestone slurries, and 1% CaCO_ slurries. The
calcium sulfite scrubbing system is non-scaling, but removal efficiencies
are limited by the sulfite-bisulfite equilibrium to about 50%. In cal-
cium carbonate scrubbers, sulfite scaling can be minimized while main-
taining 80-90% removal efficiencies, if the pH of the scrubber effluent
is controlled in the neighborhood of 5.7-5.8, where calcium sulfite is
fairly soluble. An increase in temperature from 100 to 125°F seems to
increase the rate of scale deposition, and in the presence of oxygen
can lead to catastrophic scaling via a cementing reaction initiated by
precipitation of anhydrite.
Arthur D Little Inc
-------
TABLE OF CONTENTS
Page No.
SUMMARY 1
CONCLUSIONS 3
RECOMMENDATIONS 5
CHAPTER I - INTRODUCTION 6
PURPOSE AND SCOPE 6
BACKGROUND 7
THE HOWDEN-ICI PROCESS 8
MODERN WET SCRUBBING SYSTEMS 12
CHAPTER II - THE CHEMISTRY OF LIMESTONE WET SCRUBBING 22
INTRODUCTION 22
PHASE RULE CONSIDERATIONS 22
CHEMICAL EQUILIBRIA IN SCRUBBER SLURRIES 25
KINETICS IN WET LIMESTONE SCRUBBING 36
CHAPTER III - BENCH-SCALE SCRUBBER EXPERIMENTS 41
INTRODUCTION 41
EXPERIMENTAL APPARATUS 41
DESIGN PARAMETERS 43
ABSORPTION COLUMN 43
GAS HANDLING SYSTEM 43
LIQUID HANDLING SYSTEM 45
ANALYTICAL PROCEDURES 45
PRELIMINARY SHAKEDOWN EXPERIMENTS 47
COMPARISON WITH OTHER SCRUBBERS 53
TEST PROGRAM 53
RESULTS 59
0.57. CALCIUM CARBONATE SLURRIES 59
SCRUBBING EFFICIENCY AND SCALING IN THE 59
ABSENCE OF OXYGEN
ii
Arthur D Little Inc
-------
TABLE OF CONTENTS, continued
Page No.
1% CALCIUM CARBONATE SLURRIES 59
SCRUBBING EFFICIENCY AND SCALING IN THE 59
ABSENCE OF OXYGEN
0.6% LIMESTONE SLURRIES 70
SCRUBBING EFFICIENCY IN THE ABSENCE OF OXYGEN 70
SCRUBBING EFFICIENCY AND SCALING IN THE 85
PRESENCE OF 4% OXYGEN
0.5% CaS03 SLURRIES 88
SCRUBBING EFFICIENCY IN THE ABSENCE OF OXYGEN 88
SCRUBBING EFFICIENCY IN THE PRESENCE OF 88
4% OXYGEN
SCRUBBING EFFICIENCY IN THE PRESENCE OF 4% 88
OXYGEN AND 13% CARBON DIOXIDE SCALING
EXPERIENCE
SLURRY ANALYSIS AND MASS BALANCE 101
DISCUSSION 104
EFFECT OF OPERATING VARIABLES ON SO, REMOVAL EFFICIENCY 104
(NO OXYGEN) Z
PH 104
L/G 106
TEMPERATURE 106
KINETIC OBSERVATIONS 106
pH LESS THAN 4.3 (0.5% CALCIUM SULFITE SLURRY) 106
pH 4.3 to 6.4 (0.5% CALCIUM CARBONATE SLURRY) 107
pH GREATER THAN 6.4 (1% CALCIUM CARBONATE SLURRY) 107
SCALING PHENOMENA 108
TYPES OF SCALE 108
EFFECT OF MATERIALS OF CONSTRUCTION 108
CHAPTER IV - POTENTIAL SCALE CONTROL METHODS 111
SUMMARY 111
SCALE CONTROL TECHNIQUES USED IN INDUSTRIAL PROCESSES 114
SCALE CONTROL BY SEEDING 115
iii
Arthur D Little Inc
-------
TABLE OF CONTENTS, continued
Page No.
CHAPTER IV. continued
SODIUM SULFATE MANUFACTURE 115
EARLY SULFUR DIOXIDE WET SCRUBBING 116
PAPER INDUSTRY 117
REFINING OF SUGAR . 119
SALT MANUFACTURE 120
DESALINATION 120
SUMMARY 127
SCALE CONTROL BY THRESHHOLD TREATMENT 128
CALCIUM CARBONATE DEPOSITION 128
CALCIUM SULFATE DEPOSITION 130
MAGNESIUM HYDROXIDE DEPOSITION 130
OTHER INORGANIC AGENTS 130
USE IN SEAWATER EVAPORATORS 131
STUDIES OF POLYORGANIC COMPOUNDS 132
SUMMARY 127
APPENDIX METHOD OF CALCULATING IONIC EQUILIBRIA 138
iv
Arthur D Little Inc
-------
LIST OF FIGURES
Page No.
1. TOTAL DISSOLVED CHEMICAL SPECIES IN A SULFITE, CARBONATE 31
SLURRY SYSTEM
2. TOTAL DISSOLVED CHEMICAL SPECIES IN A SULFITE, CARBONATE, 32
SULFATE SLURRY SYSTEM
3. CONCENTRATIONS OF ALL IONIC SPECIES IN A SULFITE, CARBONATE 33
SLURRY SYSTEM
4. CONCENTRATIONS OF ALL IONIC SPECIES IN A SULFITE, SULFATE, 34
CARBONATE SLURRY SYSTEM
5. SOLUBILITIES OF THE SOLID PHASES IN THE CALCIUM SULFATE 39
WATER SYSTEM
6. LINE DIAGRAM OF EXPERIMENTAL SCRUBBING APPARATUS 42
7. SULFUR DIOXIDE REMOVAL WITH NaOH AS A FUNCTION OF GAS RATE 49
8. SULFUR DIOXIDE REMOVAL AS A FUNCTION OF LIQUID TO GAS 52
RATIO FOR DIFFERENT SCRUBBING AGENTS (100°F)
9. CHANGE IN SULFUR DIOXIDE REMOVAL EFFICIENCY WITH TIME 54
AFTER STARTUP
10. COMPARISON OF EFFECT OF L/G ON SCRUBBER EFFICIENCY - 55
MEDIUM SURFACE (VELOCITY) SCRUBBERS
11. COMPARISON OF EFFECT OF L/G ON SCRUBBER EFFICIENCY - 56
HIGH VELOCITY (SURFACE) SCRUBBERS
12. MULTI-GRID SCRUBBER - EFFECT OF GAS VELOCITY ON EFFICIENCY 57
AT CONSTANT L/G
13. SULFUR DIOXIDE REMOVAL EFFICIENCY AS A FUNCTION OF pH - 63
0.57. CaC03 (125°F, NO OXYGEN)
14. SULFUR DIOXIDE REMOVAL EFFICIENCY AS A FUNCTION OF MAKE 64
UP RATE - 0.5% CaC03 (125°F, NO OXYGEN)
15. SULFUR DIOXIDE REMOVAL EFFICIENCY AS A FUNCTION OF pH - 65
0.5% CaC03 (100°F, NO OXYGEN)
Arthur D Little, Inc
-------
LIST OF FIGURES, continued
Page No.
16. SULFUR DIOXIDE REMOVAL EFFICIENCY AS A FUNCTION OF 66
MAKE UP RATE - 0.5% CaC03 (100°F, NO OXYGEN)
17. PRESSURE DROP AS A FUNCTION OF OPERATING TIME - 0.57% 67
CaC03 (125°F, NO OXYGEN)
18. PRESSURE DROP AS A FUNCTION OF OPERATING TIME - 0.5% 67
CaC03 (100°F, NO OXYGEN)
19. SULFUR DIOXIDE REMOVAL EFFICIENCY AS A FUNCTION OF pH - 71
1% CaC03 <125°F, NO OXYGEN)
20. SULFUR DIOXIDE REMOVAL EFFICIENCY AS A FUNCTION OF MAKE 72
UP RATE - 1% CaC03 (125°F, NO OXYGEN)
21. SULFUR DIOXIDE REMOVAL EFFICIENCY AS A FUNCTION OF pH - 73
1% CaC03 (100°F, NO OXYGEN)
22. SULFUR DIOXIDE REMOVAL EFFICIENCY AS A FUNCTION OF MAKE 74
UP RATE - 1.0% CaC03 (100°F, NO OXYGEN)
23. PRESSURE DROP AS A FUNCTION OF OPERATING TIME - 1.0% 75
CaC03 (100°F and 125°F, respectively, NO OXYGEN)
24. SULFUR DIOXIDE REMOVAL EFFICIENCY AS A FUNCTION OF pH - 82
0.6% SHAWNEE LIMESTONE (125°F, L/G=100, NO OXYGEN)
25. SULFUR DIOXIDE REMOVAL EFFICIENCY AS A FUNCTION OF MAKE 83
UP RATE (STOICHIOMETRY) - 0.6% SHAWNEE LIMESTONE (125°F,
L/G=100, NO OXYGEN)
26. PRESSURE DROP AS A FUNCTION OF OPERATING TIME, 0.6% 84
SHAWNEE LIMESTONE (125°F, L/G=100, NO OXYGEN)
27. SULFUR DIOXIDE REMOVAL EFFICIENCY AS A FUNCTION OF PRESSURE 86
DROP - 0.6% SHAWNEE LIMESTONE (125°F, L/G=100, MAKE UP
RATE =2.9 gph)
28. SULFUR DIOXIDE REMOVAL FOR SHAWNEE LIMESTONE AS A FUNCTION 87
OF MAKE UP RATE, 4% OXYGEN
29. SULFUR DIOXIDE REMOVAL EFFICIENCY AS A FUNCTION OF pH - 93
0.5%CaS03 (125°F, NO OXYGEN)
vi
Arthur D Little, Inc
-------
LIST OF FIGURES, continued
Page No.
30. SULFUR DIOXIDE REMOVAL EFFICIENCY AS A FUNCTION OF MAKE 94
UP RATE - 0.5% CaS03 (125°F, NO OXYGEN)
31. SULFUR DIOXIDE REMOVAL EFFICIENCY AS A FUNCTION OF pH - 95
0.57. CaS03 (125°F, 4% OXYGEN)
32. SULFUR DIOXIDE REMOVAL EFFICIENCY AS A FUNCTION OF MAKE 96
UP RATE - 0.5%CaS03 (125°F, 4% OXYGEN)
33. SULFUR DIOXIDE REMOVAL EFFICIENCY AS A FUNCTION OF pH .- 97
0.5%CaS03 (100°F, 4% OXYGEN)
34. SULFUR DIOXIDE REMOVAL EFFICIENCY AS A FUNCTION OF MAKE 98
UP RATE - Q.5% CaSO_ (100°F, 4% OXYGEN)
35. SULFUR DIOXIDE REMOVAL EFFICIENCY AS A FUNCTION OF pH - 99
0.5% CaS03 (100 F, 4% OXYGEN, 13% CARBON DIOXIDE)
36. SULFUR DIOXIDE REMOVAL EFFICIENCY AS A FUNCTION OF MAKE 100
UP RATE - 0.5% CaS03 (100°F, 4% OXYGEN, 13% CARBON DIOXIDE)
37. PRESSURE DROP AS A FUNCTION OF OPERATING TIME-0.5% CaSO 102
38. EFFECT OF pH ON S0_ REMOVAL FOR DIFFERENT SLURRIES 105
(NO OXYGEN)
39. APPARATUS USED TO STUDY SCALE CONTROL BY USE OF SEEDS 118
40. A VAPOR COMPRESSION SEAWATER DISTILLATION UNIT FITTED 121
WITH A CONTACT STABILIZATION UNIT
41. A LABORATORY FORCED CIRCULATION EVAPORATOR USED TO TEST 123
CALCIUM SULFATE SEEDING AS A METHOD OF SCALE CONTROL
42. SLURRY RECYCLE PROCESS USED IN A PILOT FLASH DIS- 125
TILLATION PLANT
43. EFFECT OF SODIUM HEXAMETAPHOSPHATE ON THE PRECIPITATION 129
OF CALCIUM CARBONATE
44. EFFECT OF POLYMER CONCENTRATION ON THE INHIBITION OF 132
GYPSUM CRYSTALLIZATION
45. RECIPROCAL OF SUPERSATURATION AS A FUNCTION OF TIME FOR 134
GYPSUM SOLUTIONS CRYSTALLIZING AT 100°C IN THE PRESENCE
OF POLYMER
46. ANTISCALING ACTION AS A FUNCTION OF CONCENTRATION OF A 136
POLYETHACRYLITE (DOREX 40)
vii
Arthur D Little Inc
-------
LIST OF TABLES
Page No.
I. EFFECT OF SULFITE-SULFATE MAKE ON SCALING BEHAVIOR 11
II. SCALING EXPERIENCE IN WET SCRUBBING FACILITIES 16
III. PHASE RULE CONSIDERATIONS FOR TYPICAL WET SCRUBBER 24
SYSTEMS
IV. EQUILIBRIUM CONCENTRATIONS AT 50°C, pH=6 (SOLID 27
SULFITE, CARBONATE)
V. EQUILIBRIUM CONCENTRATIONS AT 50°C, pH=6 (SOLID 28
SULFITE, CARBONATE, AND SULFATE)
VI. EQUILIBRIUM CONCENTRATIONS AT 50°C, pH=5.5 (SOLID 29
SULFITE AND SULFATE: DISSOLVED CARBONATE, p =1 atra)
CO 2
VII. DESIGN PARAMETERS FOR SEVERAL SCRUBBING UNITS 44
VIII. SUMMARY RESULTS - SCRUBBING LIQUID, WATER 48
IX. SUMMARY RESULTS - SCRUBBING LIQUID, SODIUM HYDROXIDE 48
X. SUMMARY RESULTS - SCRUBBING LIQUID, 5% CaSO- SLURRY 50
XI. SUMMARY RESULTS - SCRUBBING LIQUID, 5% CaC03 SLURRY 50
XII. SUMMARY RESULTS - SCRUBBING LIQUID, 2% CaCO- SLURRY 51
XIII. SUMMARY RESULTS - SCRUBBING LIQUID, 1% CaCO,, SLURRY 51
XIV. EXPERIMENTAL TEST PROGRAM 58
XV. SUMMARY RESULTS - SCRUBBING LIQUID, 0.5% CaCO-J 60
LIQUID TEMPERATURE 125°F; NO OXYGEN
XVI. SUMMARY RESULTS - SCRUBBING LIQUID, 0.5% CaCO ; 61
LIQUID TEMPERATURE 100°F; NO OXYGEN
XVII. SUMMARY RESULTS - SCRUBBING LIQUID, 0.5% CaCO_; 62
LIQUID TEMPERATURE 100°F
viii
Arthur D Little; Inc
-------
LIST OF TABLES, continued
Page No,
XVIII. SUMMARY RESULTS - SCRUBBING LIQUIDS 1% CaCO,; 68
LIQUID TEMPERATURE 125°F, NO OXYGEN
XIX. SUMMARY RESULTS - SCRUBBING LIQUID, 1% CaCO,; 69
LIQUID TEMPERATURE 100°F, NO OXYGEN
XX. SUMMARY RESULTS - SCRUBBING LIQUID, 0.6% SHAWNEE 76
(FREDONIA WHITE) LIMESTONE: LIQUID TEMPERATURE
125°F, NO OXYGEN
XXI. SUMMARY RESULTS - SCRUBBING LIQUID, 0.5% SHAWNEE 77
LIMESTONE: LIQUID TEMPERATURE 100°F, 4% OXYGEN
XXII. SUMMARY RESULTS - SCRUBBING LIQUID, 0.5% SHAWNEE 78
LIMESTONE: LIQUID TEMPERATURE 125°F, NO OXYGEN
XXIII. SUMMARY RESULTS - SCRUBBING LIQUID, 0.6% SHAWNEE 79
LIMESTONE: LIQUID TEMPERATURE 125°F, 4% OXYGEN
XXIV. SUMMARY RESULTS - SCRUBBING LIQUID, 0.6% SHAWNEE 80
LIMESTONE; LIQUID TEMPERATURE, 125°F, NO OXYGEN
XXV. SUMMARY RESULTS - SCRUBBING LIQUID, 0.6% SHAWNEE 81
LIMESTONE; LIQUID TEMPERATURE 125°F; 4% OXYGEN
XXVI. SUMMARY RESULTS - SCRUBBING LIQUID, 0.5% CaSO , 89
125°F, NO OXYGEN) J
XXVII. SUMMARY RESULTS - SCRUBBING LIQUID, 0.5% CaSO,, 125°F 90
4% OXYGEN J
XXVIII.
XXIX.
SUMMARY RESULTS - SCRUBBING LIQUID, 0.5% CaSO ,
100°F, 4% OXYGEN
SUMMARY RESULTS - SCRUBBING LIQUID, 0.5% CaSO.,
100°F, 4% OXYGEN, 13% C0
91
92
ix
Arthur D Little Inc
-------
SUMMARY
Long-term reliable operation of limestone wet scrubbers for removal
of sulfur dioxide from flue gas has been limited in many cases by the
deposition of calcium sulfite and sulfate scale on process equipment
surfaces. The present program was undertaken to identify the fundamental
factors responsible for such scaling and plugging phenomena, and to
assess the possibility of controlling these factors to minimize the
problem.
The literature relating to the occurrence and control of scaling
in industrial processes was surveyed, and several visits were made to
lime and limestone wet scrubbing installations to discuss and observe
the kinds of scaling problems that are encountered in the field. In
wet scrubbers, several different types of scale may be identified as
follows: (1) gas inlet deposits; (2) reheater or demister deposits;
(3) soft, easily removed sulfite scale; (4) hard, thin sulfate deposits
on walls; and (5) catastrophically formed scale deposits which lead to
rapid and complete plugging of the process equipment. In many cases,
the scale appears to be mechanically deposited at wet/dry interfaces,
at sharp bends in the slurry flow circuit, or in regions of poor gas/
liquid distribution. Even in the absence of mechanical deposition,
however, serious scaling problems have been encountered due to chemical
precipitation of the reaction products, calcium sulfite and calcium
sulfate, on process equipment surfaces. This program has been aimed
primarily at identifying factors responsible for chemical scale deposition
and developing potential control methods.
The chemical equilibria existing in scrubber slurries have been
calculated for a range of typical operating conditions using computer
iteration techniques. Equilibrium concentrations of the various chemical
species are plotted as a function of pH and provide a useful framework
for discussion of scaling and scrubber operating phenomena. Examination
of the effect of pH and different chemical conditions on the concentration
of chemical species active in scaling reactions yielded insights into
scaling mechanisms. For example, the calculations identified a critical
pH region between 5.7 and 5.9 where carbonate solid is releasing C02 at
one atmospheric pressure; below this region no solid carbonate exists
at equilibrium and the system is net dissolving with respect to calcium
sulfite.
A bench-scale scrubber was designed and constructed to simulate
the operation of larger sized units. Experiments were carried out with
and without oxygen in the gas stream and with slurries of calcium carbonate,
calcium sulfite and Shawnee (Fredonia White) limestone. The experiments were
designed to explore the effects of principal operating parameters "on
both S02 absorption and scale deposition. The two phenomena are related
in that the main products of the absorption reaction, calcium sulfite
and calcium sulfate, are also the most troublesome scale forming solids.
Furthermore, any scale control methods developed should not seriously
Arthur D Little, Inc
-------
compromise scrubbing efficiency. The principal experimental findings
with respect to S02 absorption may be summarized as follows. In all
cases, sulfur dioxide removal efficiency increased linearly^with pH
over the range_4 to 7. Greater than 90% SO? removal was obtained with
Tjoth limestone and calcium carbonate when the stoichiometry was > 1*.
Removal efficiency fell off rapidly at stoichiometries < 0.7, coincident
with a fall in pH below 5.7. The maximum removal efficiency with a
calcium sulfite slurry alone is limited by the sulfite-bisulfite equili-
brium under typical operating conditions to about 40%.
Several different types of scaling phenomena, similar to those
encountered in the field, were duplicated in the bench-scale scrubber
experiments. In the absence of oxygen in the flue gas, only calcium
sulfite scale was formed. It built up gradually, and was manifested by
an increasing pressure drop across the scrubber, and by negative and
equal calcium and sulfite mass balance results. Calcium sulfite deposits
formed at low L/G could be partially removed at higher L/G levels. The
rate of sulfite deposition was found to be considerably more rapid at
125°F than at 100°F. Sulfite scale also formed a thin, brittle and
adherent film on the wetted glass wall surfaces of the scrubber.
In the presence of oxygen, when sulfate species were allowed to
accumulate, a catastrophic scaling reaction was observed, which caused
sudden plugging of the flow lines. The precipitated material contained
calcium sulfate, sulfite and carbonate, suggesting a cementing action
initiated by the presence of sulfate.
On the combined basis of the literature review, thermodynamic
analyses, and supplemental experimental work, we have identified several
factors which seem to have a significant influence on chemical scale
formation. These are pH, which should be as low as possible in the
scrubber compatible with adequate removal efficiency (probably in the
range of 5.7 - 5.9); L/G, which should be as high as possible; tempera-
ture, which should probably be less than 100°F, the transition point
between gypsum and anhydrite; and oxidation, which appears to be impli-
cated in catastrophic scale formation. The exact nature of the relational
ship between oxidation and scale formation has not been determined and •»
requires further study.
Avenues of scale control that might be explored in addition, include
the use of polyphosphates and polyacrylates, which have been found to
be very effective in the ppm range for control of scale in many industrial
processes. There is also evidence from many sources that addition of
calcium sulfite and/or gypsum seed crystals to the scrubber slurry can
reduce the rate of chemical scale deposition. The most effective seed
concentrations, particle size, and crystal structure for given scrubber
operating conditions have yet to be determined.
*Stoichiometry. Defined as the mole ratio of calcium carbonate or
limestone fed to the system to the S0_ fed to the system
Arthur D Little Inc
-------
CONCLUSIONS
Several different types of scale that have been encountered In
limestone wet scrubbing installations in the field are: (1) gas
inlet deposits; (2) reheater or demister deposits; (3) soft,
easily removed sulfite scale; (4) hard, thin sulfate deposits on
walls; and (5) catastrophically formed deposits which rapidly
plug the process equipment. All but type (4) have been dupli-
cated in bench-scale experiments performed under the present
program.
Gas inlet deposits occur due to the hot gas/liquid interface in
the gas inlet section. To alleviate this problem, it has been
recommended that the gas duct be bent downward, and that wetted
wall entrances be provided for venturi, impingement plate and
flooded-bed scrubbers. For the TCA scrubber, a large rectangular
inlet nozzle, which creates a sharp interface at the point where
hot gases contact scrubbing liquid, has been suggested. Install-
ation of half-track soot blowers has also been Affective in
eliminating massive inlet deposits in several field scrubber
units.
Reheater or demister deposits at the gas outlet from the scrubber
were never sufficiently serious in our bench-scale unit to force
a shut-down of operation. In large scale TCA and hydrofilter
units the problems can be quite severe and no totally satisfactory
solution has emerged. When the problem is in the reheater above
the demister, operating experience has been better with a widely
spaced steel tube-steel fin reheater coil than with a closely
spaced copper coil-copper fin unit. For the demister, fiber-
glass appears to be a better material than steel, but periodic
washing and mechanical removal of the solids deposited is still
necessary.
Sulfite scale, which has been formed exclusively when our bench-
scale scrubber is operated in the absence of oxygen, builds, up
gradually and is manifested by a slow Increase in pressure drop
across the scrubber with time on line. The deposition is
partially reversible at high L/G levels.
The extent of sulfite scale deposition in the scrubber bed can be
estimated from mass balance results on calcium and sulfite,
which have been shown to be negative and equal when scaling
occurs.
Since the solubility of calcium sulfite increases strongly below
pH 6, the sulfite scaling problem can be minimized.by operating
the scrubber at as low a pH as possible consistent with adequate
S02 removal. In our bench-scale experiments, removal efficiency
Arthur DLittklnc
-------
was greater than 80% down to pH 5.7 at the scrubber outlet,
but dropped rapidly at lower pH levels. For other scrubbers
the operating parameters should be established empirically to
correspond to the optimum pH.
jCatastrophic scaling was observed in the laboratory only in
'the presence of oxygen, and therefore is believed to be
associated with the conversion of sulfite to sulfate by oxygen
in the flue gas. However, the deposit that actually plugged
the lines was shown to contain relatively little calcium sulfate
Catastrophic scaling may be thus initiated by deposition of
anhydrite, the stable form of calcium sulfate above 100°F, or
the metastable hemihydrate (plaster of paris). Such deposits
can "set" and harden under water forming a cement-like material
which could result in sudden catastrophic plugging of the flow
^ --.^ system.
An increase in scrubber liquor temperature from 100°F to 125°F
results in both an increase in the rate of deposition of
calcium sulfite scale, and a decrease in the number of hours
on line prior to catastrophic scaling when the flue gas contains
oxygen.
Arthur D Little. Inc
-------
RECOMMENDATIONS
A number of additives, primarily polyphosphates and polyacrylates, \ , C/
have been developed and found to be effective, in the ppm range,
for control of scale in many industrial processes. The feasi-
bility of using such additives to control scale in limestone
wet scrubbers should be explored.
Although oxidation of sulfite to sulfate is implicated in cata-
strophic scaling, there are serious gaps in the literature on
the mechanism and control of oxidation in the optimum pH
and temperature range for limestone wet scrubber operation.
An investigation should be carried out to evaluate means for
inhibiting and promoting oxidation and to determine the effect
of such treatment on scale formation.
Seeding is a well recognized scale control technique but a
systematic investigation should be carried out on optimum seed
concentrations and crystal structures for maximum effectiveness.
A dynamic mathematical model of the anticipated limestone
scrubbing system should be developed to account for all steps
between sulfur dioxide absorption and sulfite or sulfate
precipitation. Operating conditions which minimize potential
for precipitation in the scrubber should be identified by
means of the model and tested experimentally.
Arthur D Little Inc
-------
CHAPTER I
INTRODUCTION
Purpose and Scope
The objectives of the program were to develop a basic understanding
of the factors responsible for scaling in limestone wet scrubbing pro-
cesses and from such understanding to devise means for alleviating the
problems. To meet the objective, the following four tasks were under-
taken.
Task I - Information Survey - A literature survey was prepared
covering both scaling experience in alkaline wet scrubbers, and scale
deposition and prevention in other industrial processes.
Task II - Thermodynamics - Equilibrium data and existing programs
for calculating the distribution of components among ionic and neutral
species in solution, as well as solid phases were critically reviewed.
Several simple computer programs were written, based on.the. phase rule,
for calculating activities and compositions of all species in solution
in equilibrium with given typical scale forming solid phases, as a
function of pH. These programs were designed to provide equilibrium
data in a form most readily comparable with the experimental data obtained
under Task III.
Task III - Bench-Scale Scrubber Experiments - A bench-scale scrubber
was designed and built to provide a reasonable simulation of larger
units in terms of L/G, gas and liquid temperature, and gas composition.
A packed-bed scrubber was selected, as representative of one of the
most scale-prone systems. Experiments were carried out with a view
towards isolating the basic reactions of importance to the over-all
scaling problem. The effects of operating parameters - L/G, stoichio-
raetry, and temperature - on scale formation were investigated for
controlled gas and scrubber feed compositions. A synthesized flue gas
containing N2, S02, 02, and C02, in any desired proportions was used to
separate the processes of S02 absorption and oxidation. Scrubber feed
compositions investigated included 0.2N NaOH, Na2S03-NaHS01, and water in
shake-down runs; 1, and 0.5 wt % CaC03 slurry; 0.5 wt % CaS03 slurry;
and 0.6% Shawnee limestone slurry.
Task IV - Structural Materials Evaluation - Coupons of Monel 400,
Teflon, OFHC copper, 316 stainless, 1008 low carbon steel, and boro-
silicate glass were immersed at six different locations within the
scrubber during operation, and limited data were obtained with respect
to scale deposition and corrosion.
Arthur D Little Inc
-------
Background
Virtually every lime or limestone wet scrubbing installation, -
laboratory, pilot plant, or full scale unit - has been plagued at some
point with scaling problems. In this context "scaling" may be defined
as any kind of solid phase precipitation or plugging which interferes
with normal continuous process operation. While there is little basic
understanding of the mechanism or mechanisms of scale deposition,
several qualitatively distinguishable types of scaling problems have
been encountered in practice. In some cases, physical factors are
clearly responsible - solid's build-up at wet-dry interfaces, solid's
settling in stagnant regions, splash, filtration effects through packing,
etc. In other cases, nucleation and growth of crystalline solids onto
process equipment surfaces is the inevitable result of circulating the
metestable supersaturated solutions of calcium sulfite and calcium
sulfate formed as a consequence of the basic scrubber chemistry. It is
not always easy to distinguish chemical from physical scaling, and the
latter is, in any case, probably aggravated by high degrees of super-
saturation in the re-circulating scrubber liquor.
The types of scaling problems that have been reported may be
summarized as follows. Gas inlet and demister or reheater deposits are
quite universal, and have been encountered in virtually every wet scrubber
system, at least during start-up. A soft sulfite scale, which is rela-
tively easy to clean, has been reported by Kansas Power and Light in
Lawrence and by Research-Cottrell. A hard sulfate deposit, about 1/16"
thick, which forms on the scrubber walls but does, not appear to grow has
been observed by TVA at Colbert, and in the Bischoff and Bahco scrubbers.
A catastrophic scale, which forces immediate shut-down of the apparatus,
was encountered in the early ICI work and has been seen more recently
by Chemico. The nature of the catastrophic scale is unclear. Our own
work suggests that the oxidation of sulfite to sulfate is implicated,
but the scale itself may be either calcium sulfite, calcium sulfate, or
more probably a complex mixture.
The literature offers a number of control measures which can be
taken to minimize scale, but these are more in the nature of general
principles than practical implementation plans. The basic principles
are as follows:
1. Intermittently wet and dried areas must be strictly avoided.
2. All surfaces exposed to slurry must be kept well irrigated.
3. L/G should be maintained at as high a level as possible.
4. Close regulation of pH in the scrubber is highly desirable.
5. Circulation of calcium sulfate and calcium sulfite seed
crystals may be used to dissipate supersaturation in the
scrubber.
Arthur D Little. Inc
-------
6. Provision of a delay tank should further aid in desuper-
saturation of the circulating scrubber liquor.
Quantitative details on tolerable pH levels, seed crystal structures
and concentrations, delay times, and optimum L/G levels for various
practical scrubber systems are largely lacking, as are diagnostic
techniques and remedial methods for obtaining acceptable gas-slurry
distributions.
The Howden ICI Process
Two of the earliest wet scrubbing systems for control of sulfur
dioxide emissions from power plants were set up in Fulham and Billingham
around 1935. A lime or chalk slurry was used as the scrubbing medium.
At Billingham, flue gas entered the base of the scrubber and ran first
through primary scrubbing elements (deep vertical plates) , through a
grid packing, and finally to the chimney via a demister. The recircu-
lating slurry flowed countercurrently through a nozzle plate at the
top of the scrubber, onto primary liquor distributors, down through the
grid packing and primary elements, into a hold tank, and finally back
to the scrubber. Additional alkali, as needed to maintain the pH of
the liquor leaving the scrubber at 6.2, was introduced to the scrubbing
liquor as it entered the hold tank.(l)
Many of the features of the scrubber described evolved over a period
of time in response to very severe scaling problems. In the first long-
term closed-loop tests conducted at Fulham, increased resistance to gas
flow was noted after 25 hours of operation, and the blockage became
severe enough to force a shut-down after 72 hours. A scale 2-3 inches
thick was found on portions of the scrubber surfaces. Originally, it
was thought that the scale formed by mechanical entrapment of suspended
particulates, which cemented together upon drying in relatively stagnant
areas. However, when hard deposits of pure gypsum were found in an
area of the scrubber that had only been exposed to clarified liquor, it
became clear that a crystallization phenomenon or chemical scaling was
involved. In order to find ways to cope with the problem, Lessing'^)
undertook a rather intensive study of crystallization from supersaturated
solutions of calcium sulfate. The first essential conclusion was that
a 5% addition of CaS04'2H20(s) crystallites to the scrubber liquor
could reduce the concentration of a supersaturated solution to normal
solubility in 30 seconds under conditions where the same solution would
remain in a metastable state for more than 6 minutes. The rate equation
derived to fit the experimental data is:
dc , / . 2 .. .
- c) (1)
where c = concentration of CaSO^ in solution at time t
c- = initial concentration
c~ = final concentration
p = amount of CaSO^-Zl^O in suspension throughout the test
8
Arthur DLittldnc
-------
Equation (1) may be integrated to give the following expression:
c - c
The value of the rate constant k depends on the shape of the suspended
crystallites. Small amounts of gypsum were ineffective in promoting
desupersaturation. Amounts larger than 5% offered no particular advantage
from the point of view of supersaturation and added to the severity of
the pumping problems.
Lessing's studies of supersaturated solutions were limited to
calcium sulfate. The primary over-all scrubbing reaction, with slaked
lime or limestone, involves, of course, the formation of calcium sulfite:
Ca(OH)2 + S02 -»• CaS03'l/2H20 -1-
CaC03 + S02 + 1/2H20 + CaSC
The presence of sulfate is believed to be due to oxidation of sulfite
or bisulfite species in solution. It is not known to this day whether
chemical scaling due to nucleation and growth of calcium sulfite from
supersaturated solutions ever occurs. Lessing did find that the presence
of calcium sulfite substantially reduces the effectiveness of gypsum as
a desupersaturating agent. The rate of desupersaturation by a 5%
CaS04«2H20 slurry was halved by the presence of 4% CaS03. Lessing there-
fore recommended the installation of an oxidizing tower after the
scrubber and prior to the hold tank. At Billingham, provision was made
for installation of such a tower, but it was never in fact installed.
Subsequent efforts have also failed to clarify the relationship, if any,
between oxidiation and scaling in wet scrubber systems.
The key to eliminating scaling in the ICI wet scrubbing system was
believed to lie in the understanding and control of supersaturation
behavior in calcium sulfite and calcium sulfate solutions. It must be
reiterated, however, that most of the fundamental work was done with the
sulfate, and an analogous behavior was then assumed for the sulfite.
Three steps were recommended as follows, to minimize problems due to
supersaturation :
1. Limit the amount of calcium sulfite and sulfate generated
in a single pass through the scrubber (the "make per
pass") .
2. Provide a delay tank or hold tank in the recirculation
loop.
3. Circulate crystallites of calcium sulfate and sulfite,
3-5% of each in the slurry.
Arthur DLittklnc
-------
The reason for limiting the make per pass is to assure that the
scrubbing liquor never becomes anything more than slightly supersaturated.
In accord with equation (1), the lower the degree of supersaturation
(c - C2>, the slower the rate of precipitation (f*£). In order to establish
quantitatively the make per pass which corresponded to an acceptable
level of supersaturation, a series of experiments were carried out in
which sulfite and sulfate make were measured as a function of L/G and
absolute liquor and gas flow rates. In one experiment, for example,
conditions were changed in steps during a single run as indicated in
Table I to increase the make per pass from 4-5 millieq/1 to 7-8 millieq/1.
At the low make level, there was no evidence of scaling, either visually
or by pressure drop measurements. An increase in the make to the 5.2 -
5.7 millieq/1 range produced definite evidence of some scale development.
At the high make level, the increase in pressure drop due to extensive
scaling eventually forced a termination of the experiment.
It should be noted that the ICI system was operated at exceedingly
high values of L/G compared to those used in modern pilot plants. Even
under the final operating conditions shown in Table I, where severe
scaling was observed, L/G was 88. In current pilot plant practice
values of L/G are typically in the 20 to 80 range, and it is seldom
practical to go higher.
The delay tank is introduced into the scrubber loop to provide
sufficient time to desupersaturate the scrubbing liquors before they
are returned to the scrubber proper. Furthermore, the desupersaturation
and resultant precipitation is forced to occur in a convenient place.
The four-minute delay time built into the ICI system was found empirically
to be adequate for the particular scrubber design and operating conditions
specified. It is not necessarily a universally applicable figure.
Since supersaturated solutions are only metastable, at best, some
desupersaturation is to be expected in the scrubber tower. When
crystallites of CaS03-l/2H20 and CaSO^-2H20 are circulated with the
liquor, they provide preferential sites for precipitation, and hence
help to prevent nucleation on solid surfaces within the scrubber. At
the ICI plant in Billingham, 3-5% crystallite concentrations were main-
tained for both the sulfite and sulfate. In the previous work at Fulham,
cited above, it was shown that sulfite crystals inhibit the desupersatura-
tion efficiency of sulfate crystals. It is possible, however, that in
conjunction with the four-minute delay time provided, the crystallite
concentrations may be optimum. It is interesting to note that in one
of the early pilot plant experiments at Billingham, the total suspended
solids varied from 4.2% at the beginning of the run to only 9.4% at the
end of 129 hours. Scaling of the grids, primarily with hard crystalline
aggregates of CaS04«2H20, was very severe. Since the degree of oxidation
was only 25%, the suspended gypsum in the circulating liquor stream was
no more than 2%, which is woefully insufficient to discharge supersatura-
lLon in the 3-4 minute delay time permitted.
10
Arthur D Little Inc
-------
TABLE I
Gas Flow Rate (N M /hr)
3
Liquor Flow Rate (M /hr)
L/G (gpm/1000 ft3)
Alkali
Make of S per mass
(millieq/1)
Delay Time (min)
Sulfur in coal (%)
pH exit liquor
Alkali consumption (%)
% Suspended Solids
% Suspended CaSO,'2H_0
% Suspended CaCO_
% Suspended CaSO-
% CO- in flue gas
Oxidation (%)
Scaling
.FATE MAKE ON
First
351 Hours
5740
100
130
Lime (106
hours)
Limestone
4-5
3-4
1.6-2
6
100 Lime
70-80
17-20
4-5
0.5 (Lime)
0.8 (Chalk)
5
11-13
50
None
SCALING BEHAVIOR
Next 100
Hours
4100
80
146
Limestone
-
4
1.6-2
6.1
90
15
5
0.6
5
11-13
55
Next 200
Hours
4100
70
128
Limestone
5.2-5.7
4
2
6.1
90
15
5
0.6
5
11-13
60
Scale on
wood
Final 150
Hours
5950
70
88
Limestone
7-8
4
2
6.1
120
15
5
1.0
5
11-13
40
40
11
Arthur D Little Inc
-------
In the ICI system, limiting the make per pass, introduction of a
delay tank, and circulation of calcium sulfate and sulfite crystallites
helped to control supersaturation, and hence to alleviate scaling, if
not to completely eliminate it under all conditions. Additional features
incorporated into the scrubber design to cope with the scaling problems
were:
A. Addition of fresh alkali into the circulating liquor stream
just prior to the delay tank.
5. Insertion of deep, vertical plates in the lower section of
the scrubbing tower.
The solubility of calcium sulfite decreases dramatically with
increased pH. Hence, the addition of alkali to the scrubber effluent
liquor just before it enters the delay tank helps to dissipate any
sulfite supersaturation.
The vertical plates with pear-shaped tops in the bottom portion
of the ICI scrubber.helped to assure uniform gas distribution across
the grid packing. Furthermore, scale formation on the primary elements
is inhibited by the high liquor velocity and fairly high film thickness
associated with them. It was found experimentally that about 50% of
the sulfur was absorbed into the liquor flowing down the primary elements.
This reduced the burden on the grid packing, and hence contributed sub-
stantially to safe and efficient scrubber operation.
Inspite of the significant progress made by ICI in the 1930's
towards prevention of scaling by proper adjustment of design parameters,
occasional scaling problems were still encountered when ICI resumed
scrubbing operations in the 1950's. Most of the problems were mechanical
rather than chemical, however. That is, the scaling was usually due to
incomplete irrigation of the scrubbing tower grids, which in turn
resulted from accidental blockage of flow elsewhere in the system.
Modern Wet Scrubbing Systems
A very brief summary of operating conditions and scaling experience
in a number of recently constructed wet scrubbing installations is
presented in Table II. In the published literature, at least, there is
a dearth of systematic studies aimed at defining a range of operating
parameters for,scale-free operation of the various scrubber types.
Scaling has been a serious problem in most wet scrubbers, and remedial
measures have been taken, with varying degrees of success, to alleviate
the problems. What appears to be lacking, however, is a clear delineation
of the factors responsible for scaling, and the degree to which such
factors must be controlled for long-term scale-free operation. It
should be recognized that, it is almost as Important to know that a given
set of conditions will always produce scale, as it is to know safe
operating ranges for the significant parameters.
12
Arthur D Little; Inc
-------
In the Howden-ICI process, scaling was alleviated to a considerable
extent by (1) limiting the sulfite-sulfate make per pass to 4-5 millieq/1,
(2) introducing a 3-4 minute delay time into the recirculation loop, and
(3) circulating a slurry containing 3-^5% of both gypsum and calcium
sulfite crystallites. These values of the parameters were developed
empirically for the Billingham scrubber. While the same type of control
measures are probably applicable to other systems, the numerical values
of the parameters are probably not optimum for different scrubber designs.
The make per mass (Ms) of total sulfur in millieq/1 is related to
L/G in gal/1000 ft3, the amount of sulfur in the flue gas in ppm (Sppm),
and the efficiency of the scrubber for S0£ removal (e) as follows:
M (millieq/1) = 0.385e (Sppm)
In the ICI work, the measured safe make per pass was 4-5 millieq/1.
For a modern wet scrubber, operating on 2000 ppm of S02 in the flue gas,
with typically an L/G = 50 and an efficiency of 90% (E = 0.9), one calcu-
lates Ms = 11 millieq/1, almost three times higher than the value pre-
scribed by ICI. Delay times in the ICI scrubber, however, were 3-4
minutes. In modern practice, 10-minute delay times are more typical.
It is not clear at the present time whether a trade-off is possible be-
tween delay time and make per pass. That is, it is not known whether
the tolerable make per pass, or maximum acceptable supersaturation in
the scrubber, is a function of the time allowed for desupersaturation
in the delay tank outside of the scrubber loop.
The use of circulating sulfate and sulfite seed crystals to reduce
supersaturation in the wet scrubber systems, and hence to reduce scaling,
is a poorly understood phenomenon at best. For currently operating
systems, very little analytical data has been recorded on percentage
total solids and composition of the circulating slurry. Research-Cottrell^
did find that an increase in slurry concentration from 2% to the 4-8%
range did help to reduce scaling. The results were in part confirmed
by TVA in their 4-8 cfm scrubber&> , when circulation of 1% solids led
to rapid scaling, while 4% solids helped to alleviate scaling, However,
delay times were also different in the two cases. MitsubishiUO) reports
the use of seed crystals as one scale control measure, but does not
present quantitative data on total solids or percentage of sulfite and
sulfate in the scrubber liquor. Ontario-Hydro(5) reports 12% total
solids in the scrubber liquor to their spray tower - 3.6% CaS04'2H20;
6.7% CaS03«l/2H20; and 1.4% CaC03. Most of the scaling problems appear
to have occurred at wet-dry interfaces and on poorly irrigated surfaces.
On the other hand, the very high percentage of gypsum in demister deposits
suggests the possibility of true chemical scaling from a supersaturated
sulfate solution.
Much of the scaling experience summarized in Table II appears to be
physical or mechanical, rather than chemical in nature. It is generally
recognized that areas of stagnant flow, and intermittently wet and dried
13
Arthur D Little. Inc
-------
areas, are to be avoided. Furthermore, experience has amply demonstrated
that all surfaces exposed to the scrubber slurry should be well-irrigated.
Gas inlet deposits have been extremely common, particularly in the
initial phases of scrubber operation. In most cases, these have been
brought under control without undue difficulty. Combustion Engineering
successfully eliminated problems of massive inlet deposits on Meramac //2
and on K, P, and L //4 by installation of half-track sootblowers on the
scrubber inlet line. Prior to installation of sootblowers, the gas inlet
line at Meramac became 40-50% plugged within 24 hours on line. Bechtel
recommends that the gas duct be bent downward to prevent solids build-
up, and further that a wet collar be provided in the duct ahead of the
scrubber inlet. Wetted wall entrances, formed by use of an overflow
weir that wets the inlet wall, are suggested for the venturi, impinge-
ment-plate and flooded-bed scrubbers. For the TCA scrubber a large
rectangular inlet nozzle, which creates a sharp interface at the point
where hot gases contact scrubbing liquid, has been proposed.
In the scrubber proper, both Research-Cottrell and Combustion
Engineering have found that scale build-up can be reduced by increasing
L/G. This could be due to a reduced make per pass, or simply to a
mechanical flushing of surfaces where deposits might otherwise form.
In the TVA packed cross-flow scrubber , scaling was almost certainly
due to nucleation and growth from a supersaturated solution, since the
deposit formed was almost pure gypsum. In contrast, in the TVA open
spray tower'"), mechanical deposition of a mixture of limestone, calcium
sulfite, and calcium sulfate occurred in a poorly wetted region. Further-
more, the chemically deposited gypsum was hard and highly adherent, while
the mechanically deposited mixture was quite soft.
In the Bischoff process , the formation of a pure calcium sulfite
scale suggests chemical deposition. However, it is not necessarily
deposition from a supersaturated solution, since precipitation of calcium
sulfite is highly pH dependent. A local increase in pH could throw HS03~
out of solution as the solid calcium sulfite. In any case, improved
flow conditions and addition of spray nozzles for intensive washing of
the scrubber walls greatly alleviated the deposition problem. In general,
sulfite scale is probably easier to deal with than sulfate scale, in
the sense that the sulfite, even after deposition, seems to be amenable
to at least partial removal by thorough flushing. Calcium sulfate tends
to form a hard, tenacious scale that often has to be literally hacked
out.
Scaling of demisters and reheaters on the gas outlet line has fre-
quently been very severe. In TVA's pilot plant at Colbert, scale-free
scrubber operation with 90% S02 removal was achieved in a simple multi-
screen unit, run with a relatively high gas velocity, 16 ft/sec., and
L/G = 30. However, the mist eliminators scaled severely under these
conditions, inspite of the fact that they were intermittently washed
14
Arthur D Little, Inc
-------
with settling tank supernatant. At K, P, and L, Lawrence, a closely
spaced copper coil-copper fin reheater coil above the demister plugged
regularly. Substitution of a steel tube-steel fin coil with bigger
spacing alleviated the problem to a considerable extent.' For the
demisters, fiberglass seemed to be less prone to scaling and easier to
clean by periodic washing than steel. Nonetheless, there remains some
sulfate scale build-up which in practice is mechanically removed by
rapping with a stick.
15
Arthur DLittklnc
-------
TABLE II
SCALING EXPERIENCE IN WET SCRUBBING FACILITIES
Company
Plant
Capacity
(cfm) Scrubber Type
L/G pH Solids
Delay Stoichio- Eff.
Time metry % %
Scaling Experience
Research-
Co tt relit ;
Tidd 1000
Venturi & Packed 30- 4.6-
Tower Packing is low 45 6.0
AP, high surface
area, neoprene coat-
ed asbestos
Limestone slurry
10-
30
Research-
May, 1971
3 Bischoff(11)
-t
D
"
Tidd 1000
Steag 82,400
same; limestone 45
slurry
Two-stage with
central spiral jet
nozzles; stages
separated by a slop-
ing plate; gas enters
at the top
4-8
10
100-120
150
80
85
• Measure scale deposition by
weighing the packing.
• High L/G gives less scaling
than low L/G.
• Hold tank residence time has
no effect between 10 and 30
minutes.
• Little scale at the top of the
packed bed. Consistent pre-
cipitation in the bottom four
elements.
• Most scale was on the packing
periphery; it was like mud;
not hard scale.
• No scale on well-irrigated
surfaces.
• 10.5 Ibs of solid build-up
in 40 hours.
•12.5 Ibs of additional solid
build-up in the next 40 hours.
• 2% slurry concentration
gave much greater scaling
than 4-8% slurry concentration.
• Incrustations and asymmetric
deposits in the conical
section and upper third of
the cylindrical scrubber
shell; analyzed almost pure
CaSO.,
-------
TABLE II (Continued)
Company
Plant
Capacity
(cfm)
Scrubber Type
L/G pH Solids
Delay Stoichio- Eff.
Time metry % %
Scaling Experience
Combustion-
Engineering
Union 140 Mw
Elec.-
Merri-
mac
KPL 125 Mw
Lawrence
Furnace injection;
water sprays and
marble bed
• Improved by better flow and
additional spray nozzles
for intensive shell washing.
• Rate of growth of deposits
on the shell surface depended
on thickness and velocity
of film running down the
surface.
• Sootblowers on the scrubber
unit eliminated otherwise
troublesome inlet deposits.
• Problems of gas distribution;
ladder vanes scaled.
• Fiberglass is a good material
to avoid corrosion.
• Wash nozzles and quenching
nozzles at scrubber inlets
reduced deposition.
• Use high L/G to prevent
mechanical deposits.
• Sulfite scale in drain line
to clarifier; some CaO may
fall into scrubber bottom,
dissolve, raise pH and force
as sulfite.
Combustion-
Engineering
W
«->•
{L
8
Windsor,a)12,000 a) Marble bed
Conn. b) 1200 b) Marble bed and rod
type contactors
c) Bench-scale pack-
ed bed
d) CFSTR-continuous
flow stirred tank
reaction
-------
TABLE II (Continued)
Company
Plant
Capacity
(cfm)
Scrubber Type
L/G pH Solids
Delay Stoichio- Eff.
Time metry % %
Scaling Experience
Detroit-
Edison^12)
River 2500
Rouge
Bechtel
(13)
Mohave/ 4000
Navajo
00
Ontario-
Hydro (5)
Toronto 4000
I"
D
cr
1) Series of two
Venturis
2) Venturi followed
by a sieve tray
3) Venturi followed
by TCA
Limestone Slurry
1) Single stage
Venturi
2) TCA
3) Impingement tray
4) Low AP egg-crate
packing (Heil)
Lime or limestone
slurry; soda ash;
regeneration
Spray tower
Limestone slurry
30- 6.0
40
100-120
60-
90
1) Series of Venturis has been
run for 120 hours without
plugging or scaling.
2) The sieve tray absorber did
scale due to precipitation
of dissolved solids.
3.3- 7,9
100
10
12- 5.6- 12;
75 6.0 3.6%
gypsum
6.7%
CaSO-:
1.'
CaCO,
2-17 120-130
70-
80
>3>
• Soft, easily removed deposits
in the demister and the flue
gas inlet zone.
• Demister deposits were 74%
gypsum, 3% CaC03, and 23% ash.
• Gas inlet deposit was 51% ash,
30% gypsum, 15% CaC03 and 4%
CaS03.
• Some deposits occurred on
surfaces not adequately
irrigated with slurry
solution.
-------
TABLE II (Continued)
Company
Capacity
Plant (cfm) Scrubber Type
% Delay Stoichio- Eff.
L/G pH Solids Time metry % %
Scaling Experience
TVA
(9)
Muscle 4-8
Shoals
Union
Carbide
Chemico
(14)
Marrietta 1500
TVA
<6>
Colbert 2000
I
D
Two-stage, counter
current spray
scrubber
40- 6.1- 1-5
100 5.2
3-7 100
Two-stage Venturi
(Chemico)
Carbide sludge
Limestone slurry
1) Venturi followed
by Tellerette
packed cross flow
scrubber (gas j^
to liquid)
2) Three-stage
mobile bed
(TCA)
3) Venturi followed
by open spray
tower
20-
60
40
4.7- 10-
6.8 15
25 90-150
130
6 130-150
48 6.2
97 • Tests with Ca(OH)2 - 1% solids
and one-minute delay time gave
rapid scaling;
4% solids, 10-minute delay
time eliminated scaling;
3% solids, 10-minute delay
time was intermediate
70- o Insignificant scale
90 deposition in a three-day
open loop test.
70
34- • Packed cross-flow scrubber
87 operation was stopped after
125 hours due to scaling;
pressure drop increased
continuously during the
90- run; scale was very hard
92 and tenacious. Scale was
77 70-95% CaSC-4-2H20; 10-20%
CaS03'l/2H20 with minor
amounts of CaC03 and
CaxMgl_x(C03).
• Spray tower was run for
354 hours with insignificant
scaling; soft deposits of
limestone and reaction
products, several inches
thick, formed in areas of the
tower above the spray bank and
not wetted by the spray.
-------
TABLE II (Continued)
Company
Plant
Capacity
(cfn)
Scrubber Type
L/G pH Solids
Delay Stoichio-
Time me try %
Eff.
Scaling Experience
Zurn
(7)
Key
West
Paducah
1500
to
o
"Dustraxtor" -
mobile unit;
turbulent contact
scrubber; very high
effective L/G (100-
500 gal/MCFM)
100- 4.6- 1-5
500 6.2
100
ICI
(1)
Chemico
Nov. 1971
Billing- 3500
ham
1500
Vertical plate
primary elements
and grid packed
tower in series
Two Venturi
scrubbers in
series
130- 6
140
15
3.4 80-120
• The TCA was run for 172 hours,
closed loop. The only solids
build-up was a small accumula-
tion below the bottom retaining
grid of the first stage.
50- • Scale build-up on blower blades
90 down-stream of the scrubber. ,
• Scale at the top of the
scrubbing tube in an area
of high turbulence,
directly in the gas path;
Analysis: 30.5% gypsum;
68% CaC03; 1.5% MgC03.
• Scale also occurred on the
scrubber housing at the
water line.
85- • See text.
99
D
• More scaling and build-up
in closed loop than in open-
loop operations.
• It is advisable to keep
the scrubber slurry outside
the path of gas flow.
-------
TABLE II (Continued)
Company
Bahco(8)
Capacity
Plant (cfm)
Seders- 17,700
jukhuset
%
Scrubber Type L/G pH Solids
Two-stage Venturi; 5.5-
lime slurry; high 6.0
turbulence ; rigorous
cascade of droplets
on gas-liquid
contact
Delay Stoichio- Eff.
Time metry % % Scaling Experience
96- • 1" of scale on the scrubber
98 walls in 7 months. Manual
cleaning took 3 men 32 hours
Mitsu-
Hiro- 1765 Plastic Grid Packed 50
shima Tower
Lime Slurry
<9
NOTES: (1) The pH values generally refer to the scrubber outlet.
(2) The values of L/G are given in units of gal/1000 ft3, as reported by the
individual companies. No attempt has been made to convert the gaseous
flow rates to a common basis.
90- • Free from scale at liquid
98 flow of 70m3
m -hr ,
and gas flow of 1.2 x 10 m /
• Soft deposit is usually CaCC>3
or CaS03- It is controllable
via pH and by avoiding stagnant
positions in the scrubber.
• Hard scale is due to growth
of gypsum. Desupersaturation
is the remedy-addition of seed
crystals and use of delay tank.
• Must avoid stagnant flow or
positions in the scrubber
dried intermittently.
Gypsum often grows when liquid
concentration increases in
these places.
-------
CHAPTER 2
THE CHEMISTRY OF LIMESTONE WET SCRUBBING
INTRODUCTION
The overall chemical reaction taking place in a limestone wet
scrubber is quite simple:
CaC03(solid) + 1/2 H20 -t-CaSO^l,^ H20(solid)
However, the details of the equilibria established between the many gase-
ous, dissolved and solid species in the system and of the kinetic inter-
actions between these species are complex. As many as six separate
phases, and over twenty dissolved species may exist in scrubber slurries.
Oxidation of sulfite species to sulfate species occurs simultaneously
with sulfur dioxide removal from the flue gas. Additional chemical
entities are continuously introduced to the system in the fly ash and
in the various mineral forms of the limestone feed. The ultimate con-
centration level of these entities in the slurry is governed by their
chemistry and the operating characteristics of the scrubber. Solid
phases may exist in one of several crystal forms depending on temperature
and on kinetic factors. The solubility of these solids depends on their
crystal form and may be altered dramatically by supersaturation.
In order to provide a framework for the understanding of the chemis-
try of scrubbers, we first considered the phase rule restraints on the
system, and proceeded to calculate the equilibrium concentrations of
various species under likely scrubber operating conditions. We then
considered the various thermodynamic and kinetic implications of these
results for scaling and scrubbing reactions.
PHASE RULE CONSIDERATIONS
Determination of the equilibria existing in a scrubber slurry is
much simplified by considering the restraints placed on the system by
the phase rule. We will find, under the most likely operating condi-
tions when steady state has been reached, that the equilibria can be
fully specified by only two operating variables, such as pH and tem-
perature. (In the absence of certain solid species more variables may
be required.)
The phase rule defines the relationship between the following
parameters:
22
Arthur D Little, Inc
-------
F - the number of degrees of freedom of the system or the variables
needed to completely specify it,
P - the number of phases present (gaseous, liquid and solid ), and
C - the number of components which are required theoretically to
synthesize the system.
Then,
F = C - P + 2
Table III summarizes the application of the phase rule for five
cases of interest in wet scrubbing. Consider the simple scrubbing
system (A) consisting of a calcium sulfite slurry interacting with gaseous
sulfur dioxide. The number phases, P, is three - gas, liquid and solid
calcium sulfite, and the number of components is also three (e.g.,
sulfur dioxide, water and lime). Therefore, the number of degrees of
freedom is two. Temperature is conveniently taken as one of these
variables. At a given temperature, the equilibria in the system are
fully defined by the specification of one additional thermodynamic
variable - say, pH or the equilibrium partial pressure of S09 above the
slurry. If we now add one further component, C02, and a further phase,
solid calcium carbonate, the number of degrees of freedom remains the
same, and the system (C) is fully specified by temperature and pH.
If we now assume, for example, that some oxidation of sulfite to
sulfate has occurred, but that no solid sulfate phase has yet appeared,
we have added a new component (sulfate) and so must specify a further
degree of freedom (e.g., sulfate concentration) to fully define the
equilibria in the system (D). Once solid calcium sulfate precipitates,
a new phase has been added; as one degree of freeedom is removed, pH
and temperature once again are sufficient to define the system (E).
A further complication is the addition of other chemical species
from the fly ash and the limestone. Of particular significance are (1)
magnesium compounds which will introduce a further component, magnesium
carbonate, and (2) sodium and chloride which will then yield further
degrees of freedom with associated variables (their concentrations)
which will require specification. In this report, we do not deal with
these additional materials in detail.
Several conclusions and observations may be made based on this
analysis which are relevant to scaling phenomena:
*
All solids are unique phases.
23
Arthur D Little Inc
-------
o
TABLE III
PHASE RULE CONSIDERATIONS FOR TYPICAL WET SCRUBBER SYSTEMS
System Solid Phases Components
Phases
Elected Variables
(C - P + 2)
CaSO-'1/2 H20
S02; Ca(OH)2; H
gas, solution, T; pH.
one solid
CaSO.j-1/2
S02;
H20
,; Ca(OH)2;
gas, solution,
one solid
T; pH; carbonate in solution
CaSO.-1/2
CaCO^
H90
; Ca(OH>2;
gas, solution, T; pH
two solids
CaSO -1/2 H.O; S02; C02; Ca(OH)?;
CaCO^ H 0; "-""
gas, solution,
two solids
T; pH; sulfate in solution
CaS03«l/2 H20; S02; C02; Ca(OH)2;
CaCO.; H-0; CaSO,
gas, solution, T; pH
three solids
>
c
-i
cr
Notes:
1 - Other chemical components may be selected to totally specify the system, but the total
number will remain the same.
-------
1. Scales containing any proportion of calcium sulfite, calcium
carbonate and calcium sulfate can be formed from typical
scrubber slurries, and be at equilibrium over a range of pH
and temperature. Limitations of these ranges may be set by
kinetic factors and the boundary conditions imposed on other
derivative variables of the system. For example, as the pH
falls, the partial pressure of carbon dioxide rises and clearly
can reach a limit of no more than 1 atmosphere in a practical
system.
2. The amount of solid in the scrubber liquor can in no way effect
the system thermodynamically; only its existence is required
to establish equilibrium. Any observed change in scrubber
operating parameters arising from simply increasing the slurry
concentration must be due to kinetic factors, such as rate of
dissolution of calcium carbonate, rate of precipitation of
calcium sulfite or sulfate, or rate of solution of S0?.
3. A useful presentation of equilibria data can be made by dis-
playing the concentration of equilibrium species as a function
of pH at constant temperature.
CHEMICAL EQUILIBRIA IN SCRUBBER SLURRIES
With an understanding of the restraints placed on the system by
the phase rule, it becomes a simpler matter to calculate the equilibrium
concentrations of species in scrubber slurries. The five situations
listed in Table III are considered in turn, and yield equilibrium data
for all likely modes of scrubber operation.
The calculation of these equilibria, while simple in concept, is
tedious to execute. Calculations must be iterated to converge the
values of the activity coefficients of the dissolved species. We,
therefore, have prepared a set of computer programs for use on the
IBM 1100 machine to carry out these calculations.
Our approach differs from that used previously by the Radian Cor-
poration^-"' who developed a sophisticated program which would calcu-
late the equilibrium state achieved between given masses of input
materials. This type of calculation was required in their work on the
simulation of scrubber operation. Our objectives were different:
namely, to generate equilibrium diagrams such that if the identify of
the solids in a given slurry were known, and if certain other variables
were measured, primarily temperature and pH, then all values of concen-
trations of ionic species and gases could be readily obtained.
25
Arthur D Little. Inc
-------
The method of computation is described in detail in Appendix I. In
summary, we set up the simultaneous mass law equations for the many
equilibria in the scrubber, assumed starting values for. the activity
coefficients, and then solved the electrical neutrality equation for the
calcium concentration. Next, we calculated improved activity coeffi-
cients by modified Debye-Huckel theory and repeated the process until
satisfactory convergence was achieved. Convergency was tested by com-
paring sequential values of all significant ionic concentrations.
i
Sample data are presented in Tables IV, V and VI for the following
conditions at 50°C (122°F):
Other
Dissolved
pH Solids Species
6 calcium sulfite, calcium carbonate
6 calcium sulfite, calcium carbonate,
calcium sulfate
5.5 calcium sulfite, calcium sulfate carbonate
These data illustrate most normal operating conditions of the
scrubber and show the flexibility of the calculations.
Several observations can be made from this data. For example, the
presence of sulfate has a significant effect on the equilibria at a
given pH and temperature. Comparing Table V with Table IV, we see that
the presence of solid sulfate suppresses the dissolved sulfite concen-
tration which means that the solution has less capacity for absorbing
sulfur dioxide. Where the rate of reaction is liquid-phase controlled
(and has an L/G effect), we might expect that the pH of the solution
phase would fall more rapidly on absorption of a given amount of S0_ in
the presence of calcium sulfate and thereby reduce efficiency. On the
other hand, the partial pressure of S0_ above the solution is less in
the presence of sulfate, and so under conditions where the surface
reaction of S02 is rate-limiting (no L/G effect), then we would expect
the scrubbing efficiency to rise in the presence of sulfate.
Table VI illustrates a boundary condition of the system: as the pH
falls, there will come a time when the pCO. exceeds one atmosphere, and
calcium carbonate solid must completely dissolve. Under such conditions
and at values of pH even lower than this, we have assumed that the most
likely "quasi" equilibrium state will be the point at which pC02 = 1.
26
Arthur D Little. Inc
-------
SPECIES
CONCENTRATION
ACTIVITY
ACTIVITY
COEFFICIENT
S03--
CO3--
HSC3-
HC03-
H2SO3
H2C03
CASO3
CACO3
CAHCO3+
CACH+
H+
OH-
CA++
TOTAL CA++ =
TOTAL SO4- =
pSO2 = .
.4673-04
.1007-05
.7372-03
.1009-01
.8007-07
.1710-01
.2830-03
.3802-05
.6703-03
.5161-08
.1122-05
.6190-07
.5125-02
.6082-02
.0000
.1490-06
.2757-04
.5941-06
.6460-03
.8841-02
.8029-07
.1715-01
.2838-03
.3812-05
.5865-03
.451 7-08
.9975-06
.5417-07
.3047-02
TOTAL S03-- =
TOTAL CO3— =
pC02 =
.5899-00
.5899-00
.8764-00
.8764-00
.1003+01
.1003+01
.1003+01
.1003+01
.8751-00
.8751-00
.8893-00
.8751-00
.5946-00
.1067-02
.2787-01
.9167-00
Units: moles/litre
atmospheres
TABLE IV. EQUILIBRIUM CONCENTRATIONS AT 50°C, pH = 6 (SOLID SULFITE, CARBONATE)
O
-------
SPECIES
CONCENTRATION
N3
00
S03--
C03--
HSO3-
HC03-
H2SO3
H2C03
CASO3
CAC03
CAHCO3+
CACH+
H+
OH+
CA++
S04-
HS04-
CAS04
TOTAL CA++ =
TOTAL SO4— =
pSO2
.3317-04
.7148-06
.4182-03
.5723-02
.4190-07
.8951-02
.2814-03
.3779-05
.7215-03
.1055-07
.1178-05
.6663-07
.1291-01
.1017-01
.1042-05
.0319-02
.2014-01
.1639-01
.7845-07
ACTIVITY
.1451-04
.3127-06
.3401-03
.4654-02
.4227-07
.9029-02
.2838-03
.3812-05
.5865-03
.8580-08
.9975-06
.5417-07
.5789-02
.41 24-02
.8410-06
.6273-02
TOTAL SOS- =
TOATL C03- =
pCO2
ACTIVITY
COEFFICIENT
.4375-00
.4375-00
.81 33-00
.81 33-00
.1009+01
.1009+01
.1009+01
.1009+01
.8129-00
.8129-00
.8470-00
.8129-00
.4484-00
.4056-00
.81 29-00
.1009+01
.7328-03
.1540-01
.4825-00
atmospheres
Units: moles/litre
>
-t
^^
•3-
TABLE V. EQUILIBRIUM CONCENTRATIONS AT 50°C, pH
AND SULFATE)
6 (SOLID SULFITE, CARBONATE
-------
SPECIES
CONCENTRATION
ACTIVITY
ACTIVITY
COEFFICIENT
to
VO
S03—
CO3--
HS03-
HC03-
H2S03
H2C03
CAS03
CACO3
CAHC03+
CAOH+
H+
CH-
CA++
S04—
HSO4-
CAS04
TOTAL CA++ =
TOTAL S04- =
pSO2
.3353-04
.1471-06
.1344-02
.3743-02
.4267-06
.1855-01
.2814-03
.7695-06
.4638-03
.3274-08
.3722-05
.2103-07
.1261-01
.1026-01
.3348-05
.6220-02
.1958-01
.1649-01
.7996-06
.1477-04
.6478-07
.1095-02
.3049-02
.4303-06
.1871-01
.2838-03
.7761-06
.3776-03
.2666-09
.3155-05
.1713-07
.5688-02
.4197-02
.2726-05
.6273-02
TOTAL SOS- =
TOTAL COS- =
pC02
Units: moles/litre
.4404-00
.4404-00
.8146-00
.8146-00
.1009+01
.1009+01
.1009+01
.1009+01
.8142-00
.8142-00
.8477-00
.8142-00
.4511-00
.4089-00
.8142-00
.1009+01
.1659-02
.2276-01
.1000+1
atmospheres
TABLE VI .EQUILIBRIUM CONCENTRATIONS AT 50 C, pH
•^ DISSOLVED CARBONATE. pCO2 = 1)
5.5 (SOLID SULFITE AND SULFATE;
-------
Table VIgives data for the situation at 50°C and pH = 5.5 where pC02 is
held at a value of 1 and solid calcium carbonate is no longer present.
Clearly, if this solution were allowed to equilibrate with flue gas or
with air, the pCO- and the concentration of carbonate species in solu-
tion would fall still further; but under dynamic scrubber conditions, it
is unlikely that there will be sufficient time for the pCO? to fall back
to ambient levels.
In Figures 1 and 2 we have plotted the total solution concentrations
of the major chemical components of the system - calcium, sulfite, car-
bonate and sulfate - as a function of pH. These plots yield several
insights into scrubber operation and possible scaling mechanisms.
First, it is instructive to follow the changes in system equili-
brium as sulfur dioxide is added. Referring to Figure 1'^ we can assume
that the slurry starts at a pH of around 8. As sulfur dioxide is added
to the slurry, the pH falls; the total sulfite and total carbonate con-
centrations rise as the proportion of bisulfite and bicarbonate in
solution increases. Solid calcium carbonate dissolves, and solid calcium
sulfite is precipitated. At a pH of 'about 5.95, the pC02 reaches 1
atmosphere, and the pH will then remain relatively constant as C0« is
evolved, while more SO. is absorbed until all the solid calcium carbonate
is consumed. At this point, the pH will once again continue to fall.
The behavior of ionic concentrations beyond this point is complex. The
calcium ion concentration goes through a minimum as the level of car-
bonate species fall, and rises again as more S0? is absorbed; the con-
centration of sulfite species rises to dominate the dissolved species.
The addition of sulfate to the system alters the behavior to some
extent as shown in Figure 2. The pH at which solid carbonate is lost
from system drops to about 5.8. The carbonate and sulfite concentra-
tions remain about the same, but the calcium concentration is increased
by about a factor of five in the pH range of normal scrubber operation.
The total dissolved sulfate concentration is relatively invariant with
pH, but passes through a slight maximum at a pH of about 5.7. Thus,
as the pH drops toward this value, sulfate will be dissolving (so long
as further oxidation is not occurring); but below this value of pH and
as the pH continues to fall, sulfate is expected to be precipitating
solely due to the pH change.
In the sulfate system, the calcium concentration varies much less
with pH than it does in the system not containing sulfate.
In Figures 3 and 4, we show more detail of the individual ionic
concentrations as a function of pH in the absence and in the presence
of solid calcium sulfate. We believe that these data can be used as a
basic display for discussion of scaling and scrubber performance phenome-
na. In most cases, the scrubber will be operating in regions where
30
ArthurDLittlelnc
-------
10"
10J
100
No solid CaC0
Solid CaC03 present
_
o
E
10
Total S03=
Total C03°
Total Ca++
1.0
0.1
50"
Solid calcium sulfite
-Solid calcium carbonate
so long as pC02 > 1
8
pH
FIGURE 1. TOTAL DISSOLVED CHEMICAL SPECIES IN A SULFITE. CARBONATE SLURRY
SYSTEM
31
Arthur D Little. Inc.
-------
Solid CaS03
Solid CaS04 Present 50°C
PC02 = 1 specified
50 C
Solid calcium sulfite
Solid calcium sulfate
Solid calcium carbonate so
long as pCO, > 1
6
8
PH
FIGURE 2. TOTAL DISSOLVED CHEMICAL SPECIES IN A SULFITE, CARBONATE, SULFATE
SLURRY SYSTEM
32
Arthur D Little. Inc
-------
50
Solid calcium sulfite
Solid calcium carbonate
when pCCL > 1
Solid CaSO- and Solid CaCO, -»>LND
FIGURE 3. CARBONATE OF ALL IONIC SPECIES IN A SULFITE. CARBONATE SLURRY SYSTEM
33
Arthur D Little. Inc
-------
Solid calcium sulfite
Solid calcium sulfate
Solid calcium carbonate
FIGURE A . CONCENTRATIONS OF ALL IONIC SPECIES IN A SULFITE. SULFATE,
CARBONATE SLURRY SYSTEM
34
Arthur Dl.iule.lnc
-------
these equilibria apply - particularly those of Figure 4 where solid
sulfate is present in the slurry and/or on scaled surfaces. As the
slurry circulates in the scrubber, it experiences pH changes on the
average, and in localized areas. For example, the slurry emerging from
the gas contacter is slightly more acid than that entering at the top,
representing a shift to the left on Figures 3 and 4. The associated
changes in concentration of the ionic species can readily be appreciated.
While it is more productive to discuss specific phenomena with
respect to these data, we can make some general observations about
scaling and about scrubber operation:
1. The evolution of CO- and loss of solid calcium carbonate at
pH of about 5.9 effectively "buffers" the system at this pH.
Many scrubbers, when operating at relatively low stoichiome-
tries, reach this pH, particularly in the effluent from the
gas-contacting region.
2. If the pH falls below this value and equilibrium is approached,
all solid calcium carbonate will be lost, and solid calcium
sulfite will begin to be dissolved. This suggests a method of
scale control. Under normal circumstances, it will be desir-
able to operate at relatively high stoichiometries and pH where
scrubber efficiency is high; but in this region, scaling can
be expected due to the simultaneous precipitation of calcium
sulfite and calcium sulfate. Calcium carbonate may also appear
in the scale due to entrainment. If the stoichiometry is
lowered or the L/G reduced, the pH may be permitted to fall
below 5.8 where calcium sulfite (and, of course, any calcium
carbonate) is dissolving. Referring to Figure 2, we see that
solid calcium sulfate will be relatively nonprecipitating
between pH = 5.9 and pH = 5.0, but will be precipitating if
the pH falls below this. Thus, a potentially promising scale
control method may be to let the pH take occasional excursions
to a value approaching but not less than 5. Such a method is
rather analogous to the acid treatments used for alkaline scale
control in evaporators.
3. If, under some circumstances, a significant reaction step in
the kinetic sequence occurs between dissolved species, we can
choose appropriate candidates from Figure.4 and see how their
concentration varies with pH. The reaction of bicarbonate ion
with S02 or H SO :
HCO ~ -I- S02 -> HSO " + CO
35
Arthur D Little. Inc.
-------
has been proposed as a rate-limiting step in the kinetic se-
quence. Figure 4 shows that the concentration of bicarbonate
rises rapidly as the pH increases to about 5.9; at still
higher pH, it falls again. This is compatible with the ob-
servation that scrubber efficiency rises as the pH rises, at
least up to pH = 5.9. (Of course, other mechanisms may also
explain this. For example, if interfacial diffusion of SO-
into the liquid phase were rate-limiting, which it clearly is
at very low pH values, then we would again expect scrubber
efficiency to increase with pH: the partial pressure of SO
above the scrubber liquid falls as pH rises, and the driving
force for SO- absorption increases.)
4. Noted on Figures 3 and 4 are the pH values at which the pSO
above the slurry reaches typical scrubber values and at which
no further absorption would take place. In Figure 4, pSO
-4 -3
reaches 1 x 10 and 2 x 10 atmospheres at pH = 4.4 and 3.7,
respectively.
KINETICS IN WET LIMESTONE SCRUBBING
There is still considerable controversy concerning the principal
chemical reactions that control absorption efficiency and scaling in a
limestone wet scrubber. The following highly simplified set of reac-
tions may serve as a convenient starting point for discussion.
S02(g) «- S02(aq)
(1)
S02(aq) + H20 J H
HSO
(2)
(3)
>SO_ absorption
HS03 H + SO =
CaC03 (g) •<- CaC03(aq)
ll+ + CaC03(aq) *- Ca++ + HCO "
+ S03~ <- CaS03(aq)
di
(6) VCaC03 dissolution
(7)
36
Arthur D Little. Inc
-------
CaS03(aq) + 1/2 H20
1/2
HC03 + C03 + H
H+ + HC03" ? H2C03(aq)
H2C03(aq) * C02(g) + H20
HSO ~ + 00 •*• HSO "
324
HSO ~ 2 H+ + SO. =
4 4
-H- = ->•
Ca + SO, t- CaSO.(aq)
k
CaSO (aq) + x H.O ->
C02 release
Sulfite oxidation
The first ten equations describe the overall scrubbing reaction:
CaC03(s) + S02(g) + 1/2 H20
1/2
(16)
The last four equations describe the oxidation reaction, which ultimately
can lead to precipitation of solid calcium sulfate.
The starred reactions are likely candidates for kinetic control.
Equation (1) represents gas-liquid transfer for S02. It would include
diffusion of SC»2(g) to the gas-liquid interface, as well as transfer of
SO. into the liquid phase and diffusion away from the interface. If the
net transfer of S02 into the scrubber liquor is controlled to some extent
by the solubility of the gas, which increases with decreasing temperature,
removal of S0_ from the flue gas should become more efficient as the
temperature is reduced. Such an effect has been reported by TVA for
gas inlet temperatures of 150-300°F (9). Once S0_ has dissolved in
the liquid phase, the equilibria (2) - (4) should Be established very
rapidly. In addition, the partial pressure of SO- in the liquid phase will
reduce the driving force for SO absorption. This partial pressure is
increased as the pH falls; this is compatible with an observed reduction
in scrubber efficiency, at low pH.
Equation (5) represents dissolution of CaCO-(s) to form the neutral
molecular species CaCO ° or CaCO (aq). Reaction of dissolved CaCO °
+
with H , perhaps in the boundary layer . surrounding each CaCO.(s) par-
ticle, via Equation (6) would contribute to the carbonate dissolution
process.
37
Arthur D Little. Inc
-------
Equation (7) corresponds to an equilibrium among Ca , SO = , and
neutral CaS03°. Equation (8) represents the precipitation kinetics of
CaS03 • 1/2 H20(s). The higher the concentration of dissolved CaSO °,
the more rapid may be the precipitation of solid. /The concentration of
CaS03° depends in turn on the activity product [Ca ][SO ~] for calcium
and sulfite ions.
Equations (9) and (10) represent the various carbonate equilibria.
The forward reaction in Equation (11) represents the release of C0_(g)
from the scrubber liquor; the backward reaction corresponds to transfer
of C02 from the flue gas into the solution.
In summary, the general kinetic steps involved in the scrubbing of
S02(s) from flue gas with CaC03 slurries are: dissolution of the gas in
the liquid; dissolution of solid calcium carbonate; precipitation of
CaC03 • 1/2 H20(s); and evolution of C02(g) from the slurry into the
vapor phase. Each of these general kinetic steps is mechanistically
complicated. Furthermore, the individual steps are interrelated.
Danckwerts(l6) provided some data on gas-liquid transfer of sulfur
dioxide and carbon dioxide, but most of the work has been confined to
the alkaline range. Radian is exploring the precipitation kinetics 6f
calcium sulfite using a CSTR^l?). The dissolution kinetics of solid
calcium carbonate has been studied to some extent by EPA. ' From
the point of view of scaling, the precipitation kinetics are ultimately
critical, although they depend upon the multitude of other kinetic and
equilibrium steps occurring. While precipitation per se is not scaling,
it can lead to scaling if it occurs on process equipment surfaces.
Precipitation is a necessary component of the net wet scrubbing reaction,
Equation (16). If it is not to lead to scaling, precipitation must be
made to occur within the liquid phase, on the surface of seed crystals,
or on surfaces outside of the scrubber loop. If the pH is carefully
controlled, it should be possible to favor HSO, over SO and hence to
minimize the rate of calcium sulfite precipitation without too great
an adverse effect on S02 removal efficiency.
On the other hand, oxidation is always a factor in limestone wet
scrubbing, and as indicated in Equations (12)-(15), oxidation is
believed to proceed through a HSO intermediate. Once HSO ~ has been
oxidized to HSO ~ (Equation (12)), it will rapidly equilibrate with H+
= | | =
and SO. . Interaction between Ca and SO, can then lead to precipi-
tation and/or scale of calcium sulfate. According to the phase diagram
shown in Figure 5 for equilibrium sulfate precipitation, gypsum is
the stable form below 100°F, while anhydrite is favored above 100°F.
38
Arthur D Little. Inc
-------
32 50 1000 150 200 250 300
TEMPERATURE, °F
330
400
FIGURE 5. SOLUBILITIES OF THE SOLID PHASES IN THE CALCIUM SUL-
FATE WATER SYSTEM
39
Arthur I) Little. Inc.
-------
In many practical situations, it is the metastable hemihydrate which
actually precipitates at the higher temperatures. In any case, precipi-
tation of anhydrite or hemihydrate (plaster of paris) on process equip-
ment surfaces can initiate a cementing reaction by conversion of the
deposit to gypsum. The setting of plaster of paris (conversion of
hemihydrate to gypsum), a reaction with occurs under water, could lead
to catastrophic blockage if it occurs in the scrubber flow lines.
40
Arthur D Little Inc
-------
CHAPTER III
BENCH-SCALE SCRUBBER EXPERIMENTS
INTRODUCTION
The aim of the laboratory study is to define the mechanisms of
scale formation in limestone based, wet scrubber systems. As indicated
in the preceding chapter, the anticipated end products of the scrubbing
reaction are known to be potential scale forming compounds. Therefore,
the complex chemical kinetics of all vapor-liquid-solid reactions
occuring in the scrubber have direct bearing on the scaling problem,
and identification of the elementary rate controlling reactions may
hold the key for any approach to circumvent scaling conditions by process
modifications. Scaling reactions and S02 removal reactions must be
considered in parallel, not only because some of the critical reaction
steps are likely to be common to both pheonomena, but also because
acceptable solutions to the scaling problem may affect removal efficiency.
The bench-scale scrubber was designed so that individual potential
scaling reactions might be isolated and studied in terms of scrubber
operating parameters. In particular, the proportions of S02, Q£ and
in the gas phase, as well as the amounts of CaC03, CaS03, and CaSO^ in
the scrubber liquor can be varied at will from zero to any desired level.
The design parameters were selected to permit as close a simulation
as possible to other small-scale scrubbers described in the literature.
A packed-bed absorption column was selected as representative of the
worst scaling situation, but sufficient flexibility was introduced into
the apparatus design so other types of absorber might readily be sub-
stituted as appropriate.
EXPERIMENTAL APPARATUS
The schematic diagram shown in Figure 6 identifies the major
components of the bench-scale apparatus. Gas with a constant inlet S02
concentration of 2,000 ppm passes once through the scrubber. The exit
sulfur dioxide concentration is monitored continuously with a MSA
infrared anaiyj,LJL. Liquid inventory in the scrubber circuit is about
SfcUiluiij MlLli lllfist of the liquid being contained in the hold tank
immediately beneath the scrubbing column. At high liquid recycle rates
examined, the liquid phase in the equipment approximates a well mixed
reactor. The make up and purge streams are pumped at the same rate and
simulate solids removal and addition of fresh absorbent. Analysis of
the purge stream gives the stationary composition in the apparatus under
well mixed conditions.
41
Arthur D Little, Inc
-------
GAS
IN
t
GAS OUT
PACKED
COLUMN
HOLD
TANK
MAKE UP
PURGE
MAKE UP
TANK
-s
RECYCLE
PUMP
Figure 6. Line Diagram of Experimental Scrubbing Apparatus
42
Arthur D Little. Inc
-------
Design Parameters
The design parameters were evolved from a consideration of several
other scrubbers with varying characteristics described in the literature
(Table VII). The ADL unit is most closely modeled on the Cottrell packed
scrubber but has a lower linear gas velocity in the absorber and a lower
liquid loading (gpm/ft2). However, the comparable values for the TVA
spray tower suggest that these lower f-igures are acceptable. The ADL
unit provides for variation in the major experimental parameters over
the following ranges: gas flow rate, 1 to 10 scfm; liquid recycle flow
rate, 0.2 to 2 gpm; L/G ratio, 20 to 200 gpm/1000 ft3; scrubber liquid
temperatures, 100° to 150°F; gas temperatures, 250° to 300°F.
Absorption Column
The absorption column and much of the associated liquid recycle
system is constructed of industrial glass from commercially available
sections which are bolted together. This has the advantage of flexi-
bility if modifications in apparatus design are required. The trans-
parency of glass allows direct visual observation of any scale forma-
tion on the apparatus and on specimens of various test materials placed
inside the apparatus. The smooth glass surface probably helps to
minimize scale formation on the apparatus and facilitates removal of
any deposits formed.
The column is operated under countercurrent flow. Hot gas enters
at the base and liquid is pumped through a liquid distributor at the
head of the column. All of the experiments reported to date have used
a 4-inch diameter column and a 1 foot packed length of 1/2" Intalox
saddles. Alternate packing materials, including 1/4" saddles, 1/4"
and 1/2" rushing rings, and 1/4" glass spheres are available, but have
not been employed to date. The absorption column is built into the
system in such a way that the packing size can be readily changed, extra
sections can be used to increase the length, the diameter can be changed
or the packing can be replaced with a sieve tray or other configuration,
as required.
Gas Handling System
The gas handling system allows various components of a typical flue
gas, i.e., nitrogen, oxygen, carbon dioxide, and sulfur dioxide to be
mixed together as desired. Two liquid nitrogen cylinders in series
allow flow rates of up to 10 scfm to be achieved. The mixed gases are
passed through an electrically heated packed tube and gas temperatures
of 300°F at 10 scfm can be attained readily. Gas temperature at the
scrubber inlet is maintained with a temperature controller attached to
the furnace and operated by a thermocouple in the exit gas stream. Gas
flow rates are measured by rotometers, except for the S02 which is
metered with a mass flow controller.
43
Arthur D Little, Inc
-------
-»
D
CT
Table VII. Design Parameters for Several Scrubbing Units
Parameter
Gas flow scfm
Recycle liquid flow gpm
gpm/ft2
L/G gpm/000 ft3
2
Cross section area ft
3
Volume ft
Linear gas velocity ft/sec
Gas residence time (sees)
ADL
packed
bed
(Bench
unit)
11.2
0.45
5.2
40
0.087
0.087
2.1
0.4
Cottrell
packed
bed
(pilot
unit)
1000
40
28.4
40
1.41
7.05
10.7
0.42
/
TVA /
TCA (
(pilot*
unit)\
2000
96
34
48
2.8
42
12.5
1.3
TVA \
spray \
tower j
(Bench /
unit)/
~"^^*^^^'
8
0.64
7.3
80
0.087
0.19
1.5
1.4
EPA
(pilot
unit)
200
8.4
19
42
0.44
3.52
6.6
1.2
-------
Liquid Handling System
Liquid or slurry from the absorber falls into a stirred, 5 gallon
spherical glass vessel which serves as the delay vessel or hold tank.
The residence time in the hold tank can be varied between 1 to 10 minutes
by adjustment of the liquid recycle rate and the liquid volume (controlled
by the height of the overflow to the mixing tank). After leaving the
hold tank, a proportion of the liquid flow is purged through a sampling
system for chemical analysis.
The remainder of the liquid stream enters the 1 gallon stirred
mixing tank, where fresh slurry feed is added before recycling the
liquid stream to the absorber. Residence in the mixing tank can be
varied from about 30 seconds to 3 minutes. From the mixing tank, the
slurry is recycled through the pump and a flow meter back to the absorber,
or through a bypass loop directly into the hold tank if it is required
to take the absorption column out of service without interrupting the
recycle operation.
The flexible impellor recycle pump will pump up to about 5 gallons
a minute against the head (about 8 ft) in the apparatus. Lower liquid
recycle rates are achieved by reducing the pump speed and throttling
the outlet from the pump. Before installation, this pump was tested with
a 10% slurry of calcium carbonate which it was able to pump for several
hours without any noticeable solids build up or variation in flow rate.
The experimental unit does not have any solids separation facility.
This effect can be simulated by variation of the purge and fresh feed
makeup flow rates and by control of the chemical composition of the fresh
feed slurry.
ANALYTICAL PROCEDURES
An on-line MSA infra-red analyzer was used to monitor S02(g) in
the inlet and outlet gas streams.
The inlet slurry (makeup line) was analyzed for pH and total
calcium. The outlet purge stream from the hold tank was analyzed as
required for pH, (on-line) total sulfite, sulfite in solution, total
calcium, calcium in solution, total sulfate and sulfate in solution.
For each sample, an aliquot of the slurry was analyzed for total calcium
and total sulfite by the methods described below. The slurry was
allowed to settle and a second sample taken of the clear solution for
determination of calcium and sulfite in solution. A sulfite balance was
obtained by comparing the S02 removal at a given makeup rate with the
total sulfite in the slurry of the purge sample. Similarly the calcium
balance was obtained from the difference between the total calcium
concentration in the slurry makeup and the total calcium in the purge
sample.
45
Arthur D Little Inc
-------
To determine total sulfite (sulfite and bisulfite) in solution or
slurry (total oxidizable sulfur) a standard iodometric technique is used.
This involves addition of the sample to excess iodine in acetic acid
followed by back titration with standard thiosulfate solution. The
variation recommended in the Radian report (Technical Note 200-004-04,
David W. DeBerry, May 22, 1970) uses buffered iodine solution (pH 6 to
6.2) with back titration with arsenite solution to avoid interference
from nitrite ions. Since nitrite ions were not present in.any of our
solutions, the standard technique proved more convenient. This procedure
effectively measures the concentration of bisulfite ion in solution be-
cause the amount of sulfite ion in solution is very small.
Calcium in solution is measured by standard atomic absorption
techniques.
Calcium in the slurry is measured by the following volumetric
titration against EDTA solution. A convenient aliquot (equivalent to
about 40 mgs Ca) is diluted with 50 mis of water and dissolved by the
addition of 10% HC1. The solution is then made alkaline by the addition
of 30 mis, 0.5 N, NaOH (or KOH).m A mixed indicator is then added and the
solution titrated against 0.1 molar EDTA, sodium salt, which, at the end
point,changes the solution from pink to purple. The mixed indicator used
is compounded from 20 grams of NaCl, 0.1 grains of calcein and 0.04
grams of murexide; about 0.2 mg is used for each titration.
Total sulfate in solution or slurry was determined by a gravimetric pro-
cedure involving acidification with HC1, precipitation with BaCl2,
filtration, and finally ignition of 83804.
Solids Collection
A sampling check made early in the program showed that the concen-
tration of solids in the purge stream was the same as the solids con-
centration in the hold tank. Subsequently all the analytical samples
were taken from the purge stream. Solids are collected, if desired,
by vacuum filtration of known volume of slurry through a previously
tared fritted glass gooch crucible with a medium porosity filter. The
solid is then dried overnight at 105°C and weighed. In most of the
experimental work, it was found simpler and more accurate to obtain a
mass balance by measuring slurry and supernatant liquor compositions,
as described above, without filtration of the solids.
Carbonate and Bicarbonate.
No carbonate analyses were made under the present program, since the
concentrations can be calculated quite accurately by combining equilibrium
tliermodynamic analysis with the measured chemical properties of the system.
46
Arthur D Little, Inc
-------
PRELIMINARY SHAKEDOWN EXPERIMENTS
The first experiments were carried out to study the absorption of
sulfur dioxide into water and into 0.2 N sodium hydroxide to determine
the lower and upper limits of scrubber efficiency. Results are summari-
zed in Tables VIII and IX. These experiments with sodium hydroxide and
water are meant to indicate the areas where gas and liquid mass transfer
resistance respectively should be the rate-controlling step. Absorption
efficiencies of carbonate or sulfite slurries should fall between these
two extremes.
Water (Table VIII) shows a low_S02 removal efficiency, reaching only
42.5% at low gas rates and very high liquid rates. The pH measured
indicates that the solution is close to the thermodynamic equilibrium.
Results with sodium hydroxide (Table IX) contrast markedly with those
for water. Scrubbing efficiencies approached 100% except at very high
gas flow rates. Changing the liquid rate to the column has little effect
on the removal efficiency. As shown in Figure 7, S02 removal is pro-
portional to the gas flow rate indicating that the gas phase resistance
is rate controlling for NaOH solution at pH 12. ,
Data obtained for slurries of 5% calcium sulfite, 5% calcium car-
bonate, 2% calcium carbonate and 1% calcium carbonate are summarized in
Tables X-XIII. Each slurry composition was studied as a function of liquid
recycle rate and liquid make up rate. Most of the data were obtained at
a liquid temperature of 100°F, although some preliminary data was also
obtained at 1176F.
The effect of changes of L/G [gpm per 1,000 ft3] on the conversion
efficiency for S02 with the various scrubbing liquors studied is shown
in Figure 8. All the results refer to the same gas rate, 5 SCFM. It
can be seen that the 1, 2 and 5% slurries with excess carbonate can
approach the removal efficiency of NaOH - with a negligible L/G effect.
In contrast the results for 0.5% calcium carbonate show a marked
dependence on L/G ratio. The results fall between those for calcium
carbonate and calcium sulfite slurries showing that L/G has a more
pronounced effect at lower stoichiometry. Under the same conditions,
the 5% calcium sulfite slurry has a conversion efficiency of about 30%,
which is not much higher than water. These trends reflect a basic
chemical difference between the two groups. Reaction with sodium
hydroxide and calcium carbonate is essentially irreversible and absorption
at the interface (or the gas phase resistance) becomes rate controlling.
Reaction with calcium sulfite and water leads to a true equilibrium (as
opposed to steady state) concentration of bisulfite ion in solution.
The pH at steady state, for example, under the reaction conditions
examined, indicates that the liquid phase for water and calcium sulfite
slurries is very close to the thermodynamic equilibrium concentrations.
47
Arthur D Little Inc
-------
Table VIII. Summary Results—Scrubbing Liquid, Water
Inlet SO- concn 2000 ppm
Initial pH 7.5
gas temp 275° F
liq. temp 100° F
£H
SO- out
PPm
Feed rate
EPh
Recycle
L/G
_SOg removed ?
% mols/hr ft
5
5
5
5
5
5
5
3
3
3
2
2
2
2
55
48
45
45
40
2.43
2.42
2.45
2.52
50
45
2.4
2.43
2.45
1750
1550
1600
1600
1600
1500
1850
1850
1500
1150
1250
1250
1200
1150
1.6
7.0
7.0
8.2
8.2
10.0
1.6
1.6
1.6
6.6
6.6
6.6
6.6
8.4
0.9
0.9
1.36
1.36
0.4
0.4
1.36
1.36
36
36
36
0.9
1.88
1.88
180
180
372
372
80
80
12.5
22.5
20
20
20
25
1.0
1.8
1.6
1.6
1.6
2.0
372
450
7.5
7.5
0.6
0.36
450
450
680
450
940
940
20
42.5
37.5
37.5
40
42.5
0.96
2.05
1.20
1.20
1.28
1.36
Table IX. Summary Results—Scrubbing Liquid: Sodium Hydroxide
Initial pH 12.8
Inlet SO, 2000 ppm
NaOH cone 0.2N
Gas temp 275° F
Liq. temp 100° F
N, flow
scfm
5
7.8
7.6
7.4
(10.0
10.0
S02 out
PPm
12.8
12.65
12.53
12.4
8.9
5.9
< 50
< 50
50
75
200
400
Make up rate
gph
0.87
0.87
0.87
0.87
0.87
-2.86
Recycle
gpm
0.9
0.9
0.4
0.26
0.26
0.26
SO. removed _
-/ : 2
L/G % mols/hr ft
180 >97.5
115 >97.5
52 97.5
96
90
80
35
26
26
> 7.9
12.0
11.5
14.6)
13.0
48
Arthur 1)1 ittHnc
-------
t-t
PC
cd
o
I
CM
O
co
o
0)
16.0
14.0
12.0
10.0
8.0
10
Gas Flow Rate SCFM
Figure 7. Sulfur Dioxide Removal With NaOH as a Function
of Gas Rate.
49
Arthur D Little Inc
-------
Table X. Summary Rc.suJ.LH—Scrubbing Liquid, 5Z C;iSO- Slurry
N. 5 scfm
Inlet SO- 2000 ppm
Initial pH 8.1
Liq. Temp 100° F
Gas Temp 255° F
£H
5.2
4.35
3.9
3.8
3.83
3.9
Temp
4.0
3.95
4.03
4.03
S02 out
ppm
250
550
1375
1600
1450
1330
. inc to
1500
1500
1200
1150
Make up
gph
1.6
1.6
1.6
1.6
1.6
2.8
115° F
1.6
1.6
1.6
2.8
Recycle S0? removed ?
gpm L/G % mols/hr ft
0.9
0.9
0.9
0.4
0.26
0.4
180
180
180
80
51
80
87.5
72.5
31
20
27.5
33.5
7.0*
5.8*
2.5
1.6
2.2
2.7
0.4
0.9
0.26
0.26
80
180
51
51
25
25
40
42.5
2.0
2.0*
3.2
3.4
*Not steady state
Table XI. Summary Results—Scrubbing Liquid 5% CaCO- Slurry
N~ rate 5 scfm
Initial pH 8.5
Inlet S00 2000 ppm
Gas temp 275° F
Liq. temp 100° F
PJi
6.6
6.5
6.4
6.7
6.72
6.77
6.78
S02 out
Make up
150
180
200
150
130
200
200
mls/min gph
190
190
190
190
116
450
62
3.0
3.0
3.0
3.0
1.8
7.1
1.0
Recycle
gpm
1.0
0.5
0.25
1.0
1.0
1.0
1.0
L/G
200
100
50
200
200
200
200
S00 removed „
7, "'mols/hr ft
92.5 7.4
91 7.25
90 7.2
92.5 7.4
93.5. 7.5
90 7.2
90 7.2
50
Arthur I)Little. Inc
-------
Table XII. Summary Results—Scrubbing Liquid 2% CaCO- Slurry
N- rate 5 scfm
S0« 2000 ppm
Initial pH 8.6
Gas temp 275° F
Liq. temp 100° F
£H
6.65
6.6
6.53
6.7
6.72
6.72
6.73
Temp
6.72
6.62
SO. out
ppm
180
180
230
100
130
100
130
increases
140
160
Make
mls/min
192
192
192
192
114
464
62
117°F
192
192
up
eph
3.0
3.0
3.0
3.0
1.8
7.3
1.0
3.0
3.0
Recycle
gpm
1.0
0.5
0.25
1.0
1.0
1.0
1.0
1.0
0.5
L/G
200
100
50
100
100
100
100
100
150
SO
1
91
91
88.5
95
93.5
95
93.5
93
92
0 removed „
"moles/hr ft
7.25
7.25
7.1
7.6
7.5
7.6
7.5
7.5
7.4
Table XIII. Summary Results—Scrubbing Liquid 1% CaCO Slurry
N- rate 5 scfm
SO, 2000 ppm
Initial pH 8.2
Gas temp 275° F
Liq. temp 100° F
Experiment
1
2
3
4
5
6
7
£H
6.58
6.5
6.45
6.6
6.62
6.63
6.7
SO out
ppm
190
240
310
185
185
185
205
Make
ml/min
190
192
192
192
100
476
50
up
gph
3.0
3.0
3.0
3.0
1.6
7.5
0.8
Recycle
gpm
1.0
0.5
0.25
1.0
1.0
1.0
1.0
L/G
200
100
50
200
200
200
200
%
90
88
84
91
91
91
90
S00 removed
mols/hr ft
.5 7.25
7.05
.5 6.75
7.3
7.3
7.3
7.2
2
51
Arthur I) Little. Inc
-------
100
80
60
40
20
u
a
-------
Therefore, the rate of removal of sulfur dioxide by calcium sulfite or
water is limited by diffusion in the liquid phase (liquid phase resistance)
with the gas liquid interface being essentially .at equilibrium. Removal
rates for sulfite slurries are higher than for water because the solution
has a higher absorption capacity.
This is confirmed in Figure 9 which shows the change in the sulfur
dioxide removal efficiency as a function of time, from the initial
introduction of sulfur dioxide into the scrubber until a steady state
conversion is achieved. Extrapolation of the results back to time zero
shows that initially, the S(>2 removal efficiencies of water, calcium
sulfite and calcium carbonate for SC>2 approach a common value corresponding
to the gas phase resistance. For water and sulfite slurries the driving
force diminishes with time as the liquid phase approaches equilibrium.
COMPARISON WITH OTHER SCRUBBERS
A comparison of our data with that reported in the literature for
S02 absorption in limestone slurries is given in Figures 10 and 11. The
agreement is generally good, particularly for the low to medium velocity
scrubbers.
It may be noted that the effect of L/G on SC>2 removal becomes less
important as L/G increases for the low velocity scrubbers. Our scrubber
simulates the performance of the two-stage TCA (at 12.6 fps) perfectly
at all L/G values up to at least 80. At L/G = about 90-100 it also gives
a good simulation of the Venturi-rod spray tower and the medium velocity
(Ontario Hydro) spray tower. The data on high velocity (multi-grid)
and large surface (3-stage TCA) types show that our data has only limited
applicability to those cases. Thus our scrubber is limited by gas
velocity (or mass transfer surface area) at high L/G values, and flow
rates of 5 scfm. Between L/G = 32 to 48 our scrubber does simulate the
operation of a multi-grid scrubber with a gas velocity of 13.5 fps (see
Figure 12).
f
TEST PROGRAM
After completion of the initial shake-down runs described above,
a systematic series of experiments was carried out to isolate some of
the reactions that have been postulated as contributing to scrubber
efficiency, and to determine the effect of these reactions on scale
formation. The test program was designed to evaluate the separate and
interacting effects of L/G, stoichiometry (or pH), temperature, and
degree of oxidation on S02 absorption and scaling for CaC03, limestone,
and CaS03 slurries. The levels of the experimental parameters investigated
are summarized in Table XIV.
53
Arthur DLittklnc
-------
100
•rl
U
•H
w
o
a
CM
O
CO
20 -
0.2N NaOH
1%
100 120
Time (mins.)
Figure 9. Change in Sulfur Dioxide Removal Efficiency With
Time After Startup
140
54
Arthur D Little. Inc
-------
100
90
ADL
Packed Column
80
TVA Colbert
2-stage TCA (P-44)
12.6fps
70
.1 60
o
£
UJ
o
01
oc 50
CM
O
CO
TVA Colbert
Venturi Rod Spray Tower
4.7 fps
40
30
Ontario Hydro
Spray Tower
8.6 fps
20
20 40 60 80
L/G GPM/1000ft3at300°Fand 1.0 ATM
100
FIGURE 1Q COMPARISON OF EFFECT OF L/G ON SCRUBBER EFFICIENCY - MEDIUM SURFACE
(VELOCITY) SCRUBBERS
55
Arthur D Little. Inc
-------
100
go
80
70
.2 60
o
CO
o
01
CC
50
40
30
18fps.
13.7fps-
3-stage TCA
4- -+—12.6fps
' 10.8 f ps-Multi-Grid
TVA Colbert (P-44)
20 40 60 80
L/G GPM/1000 ft3 at 300°F and 1.0 ATM
100
FIGURE li COMPARISON OF EFFECT OF L/G ON SCRUBBER EFFICIENCY - HIGH
VELOCITY (SURFACE) SCRUBBERS
56
Arthur I) Little, Inc
-------
100
90
80
70
L/G = 48
tt
60
,
u
c
o>
'5
Ill
g 50
u
a:
Q" 40
30
20
10
L/G = 32
ADL Simulation
6 8 10 12 14
Scrubber Gas Velocity, ft./sec.
16 18 20
FIGURE 12 MULTI-GRID SCRUBBER - EFFECT OF GAS VELOCITY ON EFFICIENCY AT CONSTANT L/G
57
Arthur D Little. Inc
-------
TABLE XIV
EXPERIMENTAL TEST PROGRAM
Scrubbing Slurry L/G Stoichiometry Slurry Temp.
caj;o3
0.5% 50, 100 0.3 - 1.95 100, 125°F
1 % 50, 100 0.9 - 2.8 100, 125°F
Limestone
0.6% 50 - 330 0.4 - 1.7 100, 125°F
0.5% 50 - 200 - 100, 125°F
Notes:
1. The S02 inlet concentration was 2,000 ppm in most experiments,
Any deviations are noted in the text.
2. The Np flow rate has 5 scfm.
3. The gas temperature was 275°F.
58
Arthur D Little; Inc
-------
RESULTS
0.5% Calcium Carbonate Slurries
Scrubbing Efficiency and Scaling in the Absence of Oxygen
Experimental runs with 0.5% calcium carbonate slurries are
summarized in Tables XV-XVII. Results are plotted in Figures 13-18-;'
In general, sulfur dioxide removal efficiency is higher at L/G = 100
than at L/G = 50 for a given pH and stoichiometry; and the effect is
more pronounced at 125°F than at 100°F. At a given effluent hold tank
pH, the removal efficiency is somewhat higher at 100°F than at 125°F,
particularly at L/G =50. A substantial pressure drop, indicative of
significant sulfite scaling, was observed at 125°F, but not at 100°F.
In Figure 13, sulfur dioxide removal for 0.5% calcium carbonate
at 125°F is seen to vary linearly with pH and to increase at the higher
L/G ratio. In Figure 14, the results are plotted as a function of
stoichiometry (proportional to make-up rate in our system). Removal
efficiencies are seen to drop quite sharply below a stoichiometry of
0.8-0.9, corresponding to complete utilization of limestone consistent
with a drop in pH below 5.8. The corresponding plots of removal
efficiency as a function of pH and stoichiometry for runs made at 100°F
are shown in Figures 15 and 16, respectively. Comparing Figures 13 and
14, we may note that at 100°F there is some increased S02 removal, less
scatter in the data points, and a smaller L/G effect. The drop in
removal efficiency below a stoichiometry of 0.9 is seen in both Figures
16 and 15 for the two temperatures investigated.
Under the conditions of run XV (0.5% CaC03, 125°F), increases in
column pressure drop were noted after about 12 hours' operating time.
Figure 17 indicates the pressure drops observed throughout the run. By
the end of 23 hours, scaling in the lines leading to the pump and
rotoraeter had become severe enough to shut down the system. In contrast,
Figure 18 shows that there was no substantial pressure drop increase
during 17 hours on line in run XVI, probably as a result of the lower
liquid temperature (100°F-125°F for run XV).
1% Calcium Carbonate Slurries
Scrubbing Efficiency and Scaling in the Absence of Oxygen
Experimental runs with 1.0% calcium carbonate slurries are summarized
in Tables XVIII and XIX. Results are plotted in Figures 19-23.
At a given makeup rate, or stoichiometry, sulfur dioxide removal
efficiency is higher at L/G = 100 than at L/G =50. In contrast to
results with the 0.5% CaC03'slurry, pH and removal efficiency change
very little with stoichiometry, and no marked temperature effects were
observed. The S02 removal rate again varied linearly with pH, but fell
much more rapidly in the pH range of 6.8 to 6.5 than with the lower
slurry concentration.
59
Arthur D Little, Inc
-------
—i
o
c
r+
Table XV. Summary Results—Scrubbing Liquid, 0.57. CaCO-
Inltal pH 8.2
Gas Temp. 275
Inlet SO, 2000 ppm
N flow 5 scfm
No oxygen
SO. Out
Sfl cop i c
22-1
22
22
22
22
22
22
22
22
22
- 2
- 3
- 4
- 5
- 6
- 7
- 8
- 9
-10
Sample
")">
£. f.
7?
f £,
22
£ £•
22
22
22
22
22
22
22
_ i
. 2
. 3
- 4
- 5
- 6
_ 7
- 8
_ 9
-10
4
5
5
6
6
5
6
4
6
4
pH
J:'
.63
.90
.28
.33
.25
.60
.35
.53
.05
.90
Cf\
so2
Removed
m. moles/1
67.5
48.0
52.2
30.8
38.9
56.0
34.7
68.3
41.5
54.3
Z
ppm
1040
540
700
350
315
325
100
990
240
675
Sulfite
Total
m. moles/!
23.8
22.7
28.5
23.5
28.3
36.9
31.4
45.7
32.5
42.8
Liquid Temp. 125 F
Make Up Rate Recycle
gph
1.3
2.7
2.3
4.9
3.9
2.7
5.0
1.3
3.9
2.2
by Titration
Solution
m. moles /I
20.1
5.5
12.8
.8
1.6
8.6
.2
?6.5
1.6
15.3
Sulfite
Balance
m. moles/1
-43.7
-25.3
-23.7
- 7.3
-10.6
-19.1
- 3.3
-22.6
- 9.0
-11.5
gpm
0.25
0.25
0.25
0.25
0.25
0.50
0.50
0.50
0.50
0.50
L/C
50
50
50
50
50
100
100
100
100
100
Ca Inlet
Total
m. moles/1
52.5
50.7
51.0
50.0
51.0
51.0
49.7
51.0
50.0
39.5
SO. Removal
2 7.
48
73
65
83
84
84
95
50
88
66
Ca
Total
m. moles/1
18.5
28.0
28.5
38.8
39.5
43.5
47.5
41.0
41.0
39.5
Stoichlometry
Exit
Solution
m. moles/1
15.5
7.3
10.5
7.3
6.3
6.8
5.0
?2.5
5.3
11.5
0.4
0.7
0.6
1.3
1.0
0.7
1.3
0.4
1.0
0.6
Calcium
Balance
m. moles/1
-34.0
-22.7
-22.5
-11.2
-11.5
- 7.5
- 2.2
-10.0
- 9.0
-11.0
_ _ . * * M
Average -17.6
-------
Table XVI•Summary Results—Scrubbing Liquid, 0.5% CaCO,
Samp 1 e
23 - 1
23-2
23-3
23 - 4
23-5
2 23-6
M 23-7
23-8
Sample
23 - 1
23 - ?
?3 - 3
23 - 4
23-5
7.3-6
23-7
23-8
Initial pH
Inlet SO.
N flow *
No oxygen
SOg Out
en Ppn
JP
5.52 450
4.48 790
6.13 215
6.03 240
5.50 350
5.93 190
4.48 750
6.13 125
Sulfite
Balance
o. moles/1
-30.3
-37.3
- 5.9
-10.0
-19.0
-12.7
-25.5
- 2.9
Average- 18.0
7.8
2000 ppm
5 scfm
Make Up Rate
gph
2.7
1.6
4.9
3.7
2.6
3.7
1.6
4.9
Ca Inlet
Total
m. moles/1
50.5
51.5
49.7
50.0
45.0
50.0
50.0
52.5
Gas Temp 275° F
Liquid Temp. 100 F
Recycle
gpm L/C
0.25 50
0.25 50
0.25 50
0.25 50
0.50 100
0.50 100
0.50 100
0.50 100
Ca Exit
Total Solution
m. moles/1 m. moles/1
21.0 8.3
27.5 18.7
40.0 6.0
42.0 5.8
38.8 9.8
37.5 6.5
38.0 22.6
43.3 8.5
4
SO. Removal
2
78
61
90
88
83
91
63
94
Calcium
Balance
m.molea/1
Stolch.
0.7
0.4
1.3
1.0
0.7
1.0
0.4
1.3
SO
Removed
m.moles/1
52.8
71.0
33.2
43.0
58.0
44.6
69.0
35.3
Sulfite by Titration
Total Solution
m.moles/1 m.moles/1
22.5
33.7
27.3
33.0
39.0
31.9
43.5
32.4
11.6
25.3
1.1
0
11.6
2.9
30.1
2.0
Average -13.9
>
IT
(L
n
-------
Table XVII. Summary Results — Scrubbing Liquid. 0.5% CaOOj Slurry
Expt.
8-1
8-2
8-3
8-4
8-5
S 8-6
8-7
ft-S
^— o
pH
r —
4.5
5.0
5.1
5.0
5.05
5.4
6.4
6.42
SC>2 out
850
785
500
710
850
400
165
165
Make up
-2.7
-2.75
-2.7
-2.9
2.9
-2.9
7.2
-5.2
Stolch.
.73
.75
.73
.79
.79
.79
1.95
1.4
Recycle
gpm
1.0
1.0
1.0
0.5
0.25
1.0
1.0
1.0
L/C
200
200
200
100
50
200
200
200
SOj removed
Z
57
61
75
64
57
80
92
92
S02
Removed
m. moles/1
38.9
39.9
50.9
40.3
50.5
32.2
Sulfite by
total
m. moles/1
-
25.1
28.2
37.4
22.5
26.0
Titration
solution
m. mo lea/1
39.8
14.8
15.7
10.9
.2
0
Sulfite
balance
m. moles /I
-
-14.8
-22.7
- 2.9
-28.0
- 6.2
-14.9
Ca
inlet
total
m. moles/1
28
51
52
52
53
52
>
c
—I
cr
?D"
Expt.
8-1
8-2
8-3
8-4
8-6
8-8
Ca Exit
Total Solution
m. moles/1
30
39
39
46
46
50
33.2
30.8
25.1
17.5
6.6
7.4
Calcium
balance
m. moles/1
2
-12
-13
- &
- 7
- 2
Average - 6.3
-------
100
90
80
70
60
50
6-S
8
-------
100
90
80
70
60
50
6.35
s-s
L/G O 100
50
pH = 4.6
Make Up Rate (gph)
345
Stoichiometry
I i
0.4 0.6 0.8 1.0 1.2 1.4 1.6
Figure 14 . Sulfur Dioxide Removal Efficiency as a Function of Make
Up Rate—0.5% CaC03 (125° F. , No Oxygen).
64
Arthur D Little. Inc
-------
100
90
80
70
60
50
pH
4.5
5.0
5.5
6.0
6.5
Figure 15, Sulfur Dioxide Removal Efficiency as a Function of
pH—0.5% CaCO (100° F., No Oxygen).
"65
Arthur D Little. Inc.
-------
100
90
80
70
60
50
L/G O 100
A 50
Make Up Rate (gph)
4 5
Stoichiometry
0.4
0.6
0.8 1.0
1.2
1.4
1.6
Figure 16. Sulfur Dioxide Removal Efficiency as a Function of Make
Up Rate—0.5% CaC03 (100° F., No Oxygen).
66
Arthur D Little Inc.
-------
1.5
1.0
0.5
Time (Hours)
16
24
rigure 17. Pressure Drop as a Function of Operating
Time—0.5% CaCO (125° F., No Oxygen).
1.5
1.0
0.5
-------
Table XVIII. Summary Results—Scrubbing Liquid. 1.07. CaCO..
oo
Sample
25
25
25
25
25
25
0
1
2
3
4
5
_£H_
6.63
6.65
6.75
6.78
6.63
Sulflte
Balance
m. moles/1
-36.4
- 4.5
- 6.3
-36.7
+11.8
-38.2
irage-18.4
Initial pH 7.9
Inlet SO. 2000 ppm
N? flew 5 . scfm
No oxygen
S02 Out
ppm
175
175
40
0
240
Ca Inlet
Total
m. moles/1
98.7
99.0
98.5
97.3
97.9
98.5
Gas Temp. 275 J?
Liquid Temp.
Make Up Rate
gph
1.8
4.7
3.9
1.8
4.9
1.8
Ca Exit
125" F
Recyc le
gpm
0.25
0.25
0.50
0.50
0.25
Total Solution
m. moles/1 m.
61.4
93.5
87.3
108.0
99.5
76.7
moles/ 1
8.4
6.0
5.0
4.4
4.6
4.8
L/C
50
50-
100
100
50
Calcium
Balance
m.moles/l
-37.3
- 5.5
-11.2
+10.7
+ 1.6
-21.8
Average -10.6
SO Remova1
2 7.
91
91
98
100
88
Stolen.
2.5
2.1
1.0
2.6
1.0
so2
Removed
m.moles/l
90.6
35.3
42.4
110.0
37.5
96.2
Sulfite by Tltratlon
Total
m.moles/l
54.2
30.8
36.1
73.3
49.3
58.0
Solution
m.moles/1
3.4
1.4
1.5
1.6
0.7
0.0
C
n>
3"
-------
Table XIX. Sumaary Results -- Scrubbing Liquid. 1.0%CaCO.
Sample
24-1
24-2
24-3
24-4
24 - 5
24-6
24-7
Initial pH 8.4
Inlet SO. 2000 ppm
H, flow 5 Beta
No Oxygen
-£»
6.60
6.53
6.63
6.68
6.70
6.78
6.55
S02 Out
ppm
175
215
125
ISO
60
40
40
Sulflte
Balance
m. moles/ 1
-25.8
-43.0
0.0
-12.9
-12.7
+ 4.5
-41.2
Make Up Rate
8Ph
2.6
1.7
5.2
3.5
2.6
3.5
1.7
Ca Inlet
Total
m.molcs/1
98.3
98.3
103.0
99.8
102.0
98.7
97.5
Gas Temp. 275° F
Liquid Temp. 100 F
L/C
50
50
50
50
100
100
100
Ca Exit
Total
m.molea/1
65.5
69.1
83.9
92.4
95.9
91.5
99.8
solution
m.moles/1
6.8
6.0
5.2
5.8
6.0
5.2
6.4
Average -18.7
SO. Removal
2 I
91
89
94
93
97
98
98
Calcium
Balance
m.molea/1
-32.8
-29.2
-19.1
- 7.4
- 6.1
- 7.2
+ 2.3
Average-14. 2
Stolch.
1.4
0.9
2.8
1.9
1.4
1.9
0.9
Removed
m. moles/1
61.7
92.3
33.1
48.5
66.5
50.9
107.0
Sulflte by Tltratlon
Total Solution
m.moles/1 m.molea/1
35.9
49.3
33.1
35.6
53.8
55.4
65.8
0
0.4
0.4
1.6
1.4
1.6
2.1
C
-------
In Figure 19, sulfur dioxide removal efficiency for 1.0% calcium
carbonate at 125°F is seen to vary linearly with pH, but much more
rapidly than for 0.5% calcium carbonate (Figure 13). Furthermore, the
L/G effect seen with the 0.5% slurry is not seen with the 1.0% slurry.
In Figure 20, results for the 1% slurry are plotted as a function of
stoichiometry. Removal rate is higher at L/G = 100 than at L/G = 50.
However, in contrast to the 0.5% slurry (Figure 14), there is no fall-
off in efficiency at low makeup rate. However, the makeup rate could
not be reduced below a stoichiometry of 0.8 because of pump limitations.
Results for the 1.0% slurry at 100°F are shown in Figures 21 and 22,
and differ very little for the corresponding plots at 1258F (Figures
19 and 20). Thus, the temperature effects noted for the 0.5% slurry are
not seen with the 1% slurry.
The efficiencies (88-100%) are somewhat higher than in the initial
shake-down runs (Figure 8), probably due to higher pressure drop through
the column. (See Figure 23.) Although the pressure drop increased con-
sistently throughout runs XVIII and XIX, column operation was still
satisfactory. After a total of 21 hours of operating time, scaling on
the packing had become very severe causing abnormally high pressure drops
at high liquid recycle rates.
0.6% Limestone Slurries
Scrubbing Efficiency and Scaling in the Absence of Oxygen
Experimental runs with 0.6% limestone slurries in the absence of
oxygen in the flue gas are summarized in Tables XX-XXV. Results are
plotted in Figures 24-28.
Figure 24 shows that at 125°F SC>2 removal efficiency varies linearly
with the pH of the scrubbing solution over the range 4.5 to 6.0, in a
similar manner to that observed with 0.5% CaC03 (cf., Figure 13).
Measurements of removal efficiency as a function of stoichiometry
over the range 0.3 - 1.4 are plotted in Figure 25. Results are similar
to those plotted in Figure 14 for 0.5% CaC03. With limestone, more than
90% of the S02 was removed down to a stoichiometry of 0.7 and a pH of
5.8. The efficiency then fell off rapidly down to 50% at a stoichiometry
of 0.3.
For limestone, run at 125°F in the absence of oxygen, sulfite scale
build-up is reflected in the pressure drop vs. time curve shown in Figure
26. During the first 8 to 10 hours on-line, the pressure drop across
the column remained low; then there was a gradual increase in pressure
drop due to the gradual build-up of sulfite scale on the column packing.
The first minor shut-down came after about 26 hours on-line when it
became difficult to maintain the recycle rate, due to scale formation
in the recycle rotometer and neighboring lines. However, this did not
.cause complete blocking and was only a problem at constrictions in the
recycle line. There was a steady build-up of sulfite scale on the
column, however; and at the end of 52 hours, the pressure drop had
70
Arthur D Little, Inc
-------
100
L/G O
100
50
90
80
70
o
a
0)
§
&
CM
O
CO
PH
5.5
6.0
6.5
7.0
Figure 19. Sulfur Dioxide Removal Efficiency as a Function of
pH—1.0% CaC03 (125° F., No Oxygen).
71
Arthur 01.ink-Inc
-------
100
90
80
70
6.75
B>!
>s^
^
O
d
cu
•H
O
•rH
y-i
W
pH = 6.63
O
CO
6.78
6.65
6.63
L/G O 100
50
Make Up Rate (gph)
Stoichiometry
0.8
1.2
1.6
2.0
2.4
2.8
3.2
Figure 20. Sulfur Dioxide Removal Efficiency as a Function of
Make Up Rate—1.0% CaCO (125° F., No Oxygen)
72
Arthur O I.ittk'hv
-------
100
90
80
70
-------
100
90
80
70
L/G O 100
50
QCM Make Up Rate (gph)
cn
4 5
Stoichiometry
i •_
0.8 1.2 1.6 2.0 2.4
2.8 3.2
Figure 22. Sulfur Dioxide Removal Efficiency as a Function of
Make Up Rate—1.0% CaCO (100° F., No Oxygen)
74
Arthur D Little. Inc
-------
2.0
1.5
1.0
0.5
00
L/G O 100
A 50
Time (Hours)
16
24
Figure 23. Pressure Drop as a Function of Operating
Time—1.0% CaCO. (100° F.
and 125° F., respectively, No Oxygen).
75
Arthur D Little. Inc
-------
c:
3"
Table XX. Summary Results — Scrubbing Liquid 0.67. Shawnee Limestone
Run 15
Initial pH 7.8 Gas temperature 275 F
Inlet S02 2300 ppm Liquid Temperature 125° F
N? flow 5 scfm
No oxygen
S0_ out Make up rate Recycle
PH PPm gph gpm L/G SO Removal %, Stoichiometry
6.08 63 5.1 0.5 100 97 12
5.90 125 3.9 0.5 100 93 l'o
5.80 235 2.8 0.5 100 90 \i
4.90 750 1.7 0.5 100 67 *4
4.75 995 1.7 0.5 100 57 *4
5.00 710 1.7 0.5 100 69 4
4.43 1025 1.4 0.5 100 55 *3
4.40 1100 1.3 0.5 100 52 .3
4.38 1140 1.4 0.5 100 50 ,*3
-------
IT
CT
^*
^»
nT
Table XXI. Summary Results—Scrubbing Liquid, 0.5'/. Shawnee Limestone
Inlet SO. 2000 ptn
Initial pH 6.8
N rate 5 scfra
Llq. Temp 100 F
Gas Temp 255 F
02 rate 12 scfh (41)
Sulfite by Titratlon
Ex|
12
12
12
12
12
3t.
- 1
- 2
- 3
- 4
- 5
JJH
5.43
4.8
4.55
5.4
5.8
SO,, out
PPm
350
950
1025
330
150
Make up
gph
5.3
2.9
2.9
5.4
5.4
Stolch.
1.4
0.8
0.8
1.46
1.46
Recycle
KPm
0.5
0.5
0.25
0.25
0.5
iZS
100
100
50
50
100
SO- removed
2 %
82.5
52.5
49
83.5
92.5
S0_ removed
m. moles/1
33.5
38.2
36.0
32.8
36.6
Total
m. moles/1
12.4
22.2
21.0
14.2
23.5
Solution
m. moles /I
6.6
22.8
22.0
5.0
0
Sulfate
Soln
11.7
17.1
0.7
Solid
1.2
0
0
Expt.
Ca Exit Ca Ic ium
Total Solution balance
m.moles/1 m. mo lea/I
12 - 1
12-2
12-3
12 - 4
17-5
- 8.2
+ 2.1
-12.1
Average -6.1
43.5
43.0
44,
44.
44.0
18.0
20.5
21.0
24.5
42.5
8.8
20.3
21.3
11.6
9.2
Average -18.6
-------
Table XXII. Summary Results—Scrubbing Liquid, 0.57. Shawnee Limestone
Inlet SO 2000 pm
Initial pH 6.S
N_ rate 5 scfm
Ixpt,
13-1
13-2
13-3
13-4
Expt.
13 -
13 -
13 -
S02 out
pH ppm
5.5 260
5.1 600
5.1 750
5.3 520
Sulfur
balance
m. moles /I
1 - 2.7
2 -13.9
3 -17.1
Average -11.2
Make up
gph
5.4
2.9
2.9
5.4
Ca Inlet
total
m. moles/1
44.5
46.5
47.0
Stolen.
Llq. Temp 1'j F
Gas Temp 255 F
Recycle
L/C
1.46
0.8
0.8
1.46
Ca Exit
Total Solution
m. moles/1
30.5 16.8
25.5 10.6
27.5 22.6
0.5 100
0.5 100
0.25 50
0.25 50
Calcium
balance
m. moles/1
-14.0
-21.0
-19.5
Average -18.1
SO removed
87
70
62.5
74
SO
removed
m.moles/1
34.5
53.0
46.7
Sulflte by Iteration
Total Solution
m.moles/1 m.moles/1
Sulfotc
Soln Solid
17.5
70.2
9.5
3.2
11.6
9.4
11.9
18.0
18.9
2.4
0.9
1.2
*
-------
Table XXIII. Summary Results—Scrubbing Liquid, 0.6% Shawnee Limestone
vo
Expc.
14 - 1
14-2
14-3
14-4
14-5
14-6
14-7
14-8
14-9
14 -10
5.93
5.45
5.3
5.6
5.9
5.9
5.6
5.4
5.7
6.1
S02 out
ppm
Inlet SO 2000 pm
Initial PH 8.1
N. rate 5 scfm
Make up
gph Stolch.
390 5.3 1.7
600 2.9 0.95
7?5 7.9 0.95
520 5.3 I.-'
280 5.3 1.7
310 5.3 1.7
520 2.9 0.95
745 2.9 0.95
540 5.2 1.7
260 5.2 1.7
Ca Exit Calcium
Total
Solution Balance
m. moles/1
28.0
24.5
31.5
31.0
39.5
42.5
34.5
34.5
54.5
9.8
11.9
22.7
16.6
11.4
10.0
14.4
13.4
8.4
m. moles/ 1
-22.0
-24.0
-18.0
-24.0
-20.5
-13.5
-21.5
-18.0
- 4.0
Average -18.4
Llq temp 125 F
Gas temp 255° F
0. rate 12 scfh
L/C
(47.)
0.5
0.5
0.75
0.25
0.5
0.5
0.5
0.25
0.25
0.5
100
ioo-
50
50
100
100
100
50
50
100
SO. removed
81
70
63.
74
86
84.
74
63
73
87
SO,
m?58Yil/i
32.1
51.3
46.0
39.1
34.0
53.5
45.7
29.8
35.8
Sulflte by Tttratlon
Total Solution
m.moles/1 m.moles/1
12.0
21.0
17.7
13.5
14.2
26.5
71.5
15.0
22.8
0
7.6
9.4
0
0
1.6
5.8
0
0
Ca Inlet
Total
m. moles/1
50.0
48.5
49.5
55.0
60.0
56.0
56.0
52.5
58.5
-------
oo
o
15
15
15
15
15
15
15
15
15
15
Expt.
- 1
- 2
- 3
- 4
- 5
- 6
- 7
- 8
- 9
-10
N,
i
5
5
5
5
5
5
5
7
3
5
,flow
icfm
PH
Table XXIV. Summary Results—Scrubbing Liquid. 0.67. Shavnee Limestone
Initial pH 8.5
Inlet S02
No Oxygen
it
ppm
Inlet S02 2300 ppm
Gas Temperature 275 F
Liquid temperature 125 F
SO out
5.75
5.50
5.20
5.60
5.35
5.80
6.03
4.73
6.30
5.55
Ca Inlet
total
m. moles/1
55.5
55.0
51.0
55.0
52.5
61.5
57.0
56.0
54.0
50.5
540
690
640
705
795
500
280
690
0
260
Ca
Total
4.8
2.9
3.9
3.9
2.8
4.7
4.8
2.8
2.8
2.7
Exit
Solution
m. moles/1
25.5
15.4
30.2
19.7
25.5
29.0
37.0
51.0
55.0
50.5
7.4
9.2
16.0
9.8
15.6
8.8
6.2
35.0
7.0
6.4
0.5
0.5
0.5
0.25
0.25
0.25
0.5
1.0
1.0
1.0
100
100
100
50
50
50
100
144
330
700
77
70
73
69
66
79
88
70
100
89
CaIclum
Balance
m.moles/l
-30.0
-38.6
-20.8
-35.3
-27.0
-31.5
-20.0
- 5.0
1.0
- 5.0
Average -21.2
SO. removed
m.moles/l
1.1
.7
1.0
1.0
.7
1.1
1.1
.5
1.1
.6
33.6
51.0
39.6
36.9
48.7
34.8
38.3
74.3
45.2
68.7
Total
m. moles/1
16.0
15.2
21.2
16.0
18.2
16.7
20.5
57.2
31.7
43.7
Solution
m. moles/1
3.8
7.6
8.0
5.2
7.4
0
.4
20.8
0
3.2
bull ice
Balance
m. moles/1
-17.6
-35.8
-18.4
-20.9
-30.5
-18.1
-17.8
-17.1
-13.5
-25.0
Average -21.5
c
tr
n>
-------
D
{1
R
Table xxy. Summary Results --Scrubbing Liquid. 0.6% Shawnee Limestone
Run 14
Initial pH 8.2
Inlet SO 2300 ppm
N flow 5 scfm
275° F
Liquid Temperature 125° F
Gas temperature
3*
5.88
5.65
5.33
S02 out
ppm
170
300
450
0 flow
Make up rate
gph
5.2
3.9
2.8
12 scfh
Recycle
?pm
0.5
0.5
0.5
L/G
100
100
100
S0n Removal %
93
87
81
Stoichiometry
1.2
1.0
.7
-------
100
50
4.0
4.5
5.0
5.5
6.0
Figure 24. Sulfur Dioxide Removal Efficiency as a Function of
pH—0.6% Shawnee Limestone (125° F., L/G = 100, No
Oxygen)
82
Arthur D Little Inc
-------
100
90
80
70
60
50
fr-S
pH =
Make Up (gph)
Stoichiometry
0.4
0.6
0.8
1.0
1.2
1.4
Figure 25. Sulfur Dioxide Removal Efficiency as a Function
of Make Up Rate (Stoichiometry)—0.6% Shawnee
Limestone (125° F., L/G = 100, No Oxygen)
83
Arthur D Little. Inc.
-------
3.0
2.0
1.0
Time (Hours)
16
24
32
40
48
Figure 26. Pressure Drop as a Function of Operating Tine
0.6% Shawnee Limestone (125° F., L/G = 100,
No Oxygen)
84
AnlunD Little. IIK
-------
risen to 2.5 in. from an initial value of less than 0.25 in. The
distributor at the top of the column became clogged with solids and was
essentially inoperable. Some local flooding of the packed column was
also observed, and the run was terminated.
During the extended runs to characterize sulfite build-up in lime-
stone scrubbing, it was noted that the removal efficiency appeared to be
increasing as the pressure drop across the column increased. Figure 27
shows that this appears to be a linear change and is about 20% for a
pressure drop increase of 1.5 in. This increase in efficiency is probably
due to an increased liquid residence time in the column. Although the
high pressure drop was not obtained in the earlier experiments, changes
in pressure drop could account for some of the variations in experimental
data.
/
Scrubbing Efficiency and Scaling in the Presence of 4% Oxygen
Scaling and plugging due to sulfate was observed on two occasions
during the runs with Shawnee limestone. In the first set of experiments,
carried out at 100°F (Table XXI) and then 125°F (Table XXII), the unit
became inoperable after 17 hours on-line with plugging at the bottom of
the hold tank, in the recycle pump and the recycle line to the scrubbing
column. The second series of experiments (Table XXIII) were carried out
at 125°F. In this instance, the plugging occurred after 12 hours on-line
at a point between the hold tank and the makeup tank but not significantly
in any other part of the apparatus. For both sets of experiments, there
was no noticeable plugging of the packed column or change in pressure
drop during operation. Deposits of scale were observed on the walls of
the hold tank, and some of this deposit appeared to have detached itself
from the walls and been carried down into the line to the makeup tank
assisting in the plugging of this line.
Figure 28 plots removal efficiency against pH for two con-
secutive runs which were carried out to establish at what stage
scaling or plugging would occur. It can be seen that there is little
or no effect of sulfate build-up on the removal efficiency at a given
makeup rate. (The repeat run is shown as solid circles and triangles.)
Sulfate build-up did not apparently affect the removal efficiency up to
the point where plugging occurred and the unit had to be shut down.
Scaling problems in the presence of oxygen were quite different from
those reported previously when oxygen was excluded from the system.
In the former case, plugging caused a complete shut-down after about 12
hours at 125°F. However, without oxygen, at the same temperature, the
run was continued for a period of about 52 hours with only one minor
shut-down to remove some scale in the recycle line.
85
Arthur D Little Inc
-------
IOC
90
80
70
50
Pressure Drop (Inches of Water)
0.5
1.0
1.5
2.0
2.5
Figure 27- Sulfur Dioxide Removal Efficiency As A
Function of Pressure Drop - 0.6% Shawnee
Limestone (125°F, L/G = 100, Make Up
Rate =2.9 gph)
Arthur D Little inc
-------
100
90
80
70
60
50
6-2
U
C
(U
i
L/G 0 100
A 50
make up (gph)
5 6
stoichiometry
0.4 0.6
0.8
1.0
1.2
1.4
1.6
Figure 28. Sulfur Dioxide Removal for Sh^wnee Limestone. Consecutive
Runs in Presence of 4% Oxygen. (0, A first run; t, A
second run.)
Arthur D Little Inc.
-------
0.5% CaS(h Slurries
In order to separate out the roles of calcium sulfite and calcium
carbonate in the SC>2 removal process, a series of experiments were
carried out using a 0.5% CaSC>3 slurry as the scrubbing medium. Results
are given in Tables XXV1-XXIX and in Figures 29-37.
Scrubbing Efficiency in the Absence of Oxygen
S02 removal rates for calcium sulfite slurries are considerably
lower than those obtained with carbonate slurries under the same process
operating conditions. Data from run XXVI is plotted as a function of
pH in Figure 29 for a temperature of 125°F. The same data are plotted
as a function of makeup rate in Figure 30. For calcium sulfite solutions,
the concept of stoichiometry is not applicable, because the chemical
equilibrium limits the reaction before the solid sulfite present has all
been utilized. In Figure 30, there is a small effect of L/G ratio (or
liquid rate) on the S02 removal efficiency at very high makeup rates;
but the curves converge to a common value at low makeup rates.
Scrubbing Efficiency in the Presence of 4% Oxygen
The presence of oxygen at 125°F (comparing Figures 29 and 31) does
not effect the results as a function of pH; and, although there is some
scatter in the data points, there appears to be no effect of different
L/G ratios within experimental error. A small increase in removal
efficiency with L/G is seen when the data are plotted as a function of
makeup rate in Figure 32; but the curves for L/G = 100, and L/G = 50
appear to converge at low makeup rate, similar to the results in the
absence of oxygen (c.f., Figure 30).
At 100°F, data from run XXVIII do show a small effect of L/G with higher
removal efficiencies for L/G = 50 at 100°F are only very slightly higher
than the S02 removal efficiencies obtain in run XXVI at 125°F (c.f., Figures
29 and 33). In Figure 34, the 100°F data are plotted as a function of
makeup rate. Removal efficiency is somewhat higher than at 125°F (c.f.,
Figure 32); and a small L/G effect is again noted.
Scrubbing Efficiency in the Presence of 4% Oxygen and 13% Carbon
Dioxide
In run XXIX (Figure 35) 13% carbon dioxide was added to the gas
stream, and the data as a function of pH appeared to be identical with
run XXVI where C02 was absent. In Figure 36, the data are shown as a
function of makeup rate. At L/G = 50, the removal efficiencies are
similar to those obtained in the absence of CC^. At L/G = 100, the
increase in removal efficiency in the presence of C02 appears to be some-
what greater than when CC>2 is excluded.
88
Arthur D Little, Inc
-------
Table XXVI. Summary Results—Scrubbing Liquid 0.5% CaSO?
CO
I
Sample
18 - 1
18-2
18 - 3
18-4
18-5
18-6
18 - 7
18 - 8
18-9
18 -10
18 *11
18 -12
18 -13
18 -14
18 -15
18 -16
4.15
4.2
4.2
4.1
4.0
3.95
3.93
3.9
4.35
4.25
4.18
4.3
4.4
4.45
4.0
4.1
SO Out
ppm
1125
1175
1250
1450
1475
1475
1600
1675
1325
1325
1300
1040
1075
1125
1625
1400
Ca Inlet
Total
m.moles/1
38.5
39.0
39.5
41.0
39.5
39.5
39.0
39.0
39.0
38.5
39.0
40.0
37.0
36.5
39.0
39.0
Initial pH
Inlet SO
N2 flow *• 5
No oxygen
Make up Rate
gph
5.2
5.2
5.2
2.8
2.9
2.8
1.7
1.7
3.9
3.9
3.9
6.8
6.8
6.8
1.7
7.7
4.8
2300 ppm
scfo
Recycle
8Pra
1.0
.50
.25
.25
.50
1.0
.50
.25
.25
.50
1.0
1.0
.50
.25
1.0
1.0
Ca Exit
Total
m. moles/1
37.0
40.5
36.5
37.0
40.2
44.7
36.0
35.7
33.2
36.7
44.5
41.7
33.0
33.7
45.2
44.5
Solution
m. moles/ 1
24.6
27.3
25.6
28.9
29.2
31.1
32.8
34.8
29.1
30.4
32.4
24.8
22.2
21.6
34.8
31.7
Gas tern;
Liquid
L/C
200
100
50
50
100
200
100
50
50
100
200
200
100
50
200
200
Ca Ic ium
Balance
m. moles/1
-1.5
+1.5
-3.0
-4.0
+0.7
+5.2
-3.0
-3.3
-5.8
-1.8
+5.5
+1.7
-4.0
-2.8
+6.2
+5.5
Average -0.2
emova1
51
49
46
37
36
36
30
27
43
43
44
55
53
51
29
39
Sulfite*
Input
m-molea/I
59
59.1
58.3
68.2
65.6
65.9
75.3
72.0
62.2
61.8
62.6
57.1
53.4
52.2
75.1
68.5
Sulfite by Tltration
Total
m. moles/1
44.75
41.1
39.5
44.1
49.1
54.2
39.2
44.5
35.7
45.2
53.1
50.5
41.3
41.8
51.0
58.8
Solution
m. moles/1
32.2
28.2
30.8
36.8
39.4
40.6
40.6
45.0
32.0
37.6
41.4
33.6
29.8
29.2
51.4
46.2
*Sulfite Input
SO, Removed - Sulfite in Feed.
Sulfite
Balance
m.moles/1
-14.2
-18.0
-18.8
-24.1
-16.5
-11.7
-36.1
-27.5
-26.5
-16.6
- 9.5
- 6.6
-12.1
-10.4
-14.1
- 9.7
Average -17.1
-------
Table XXVII. Summary Results — Scrubbing Liquid. 0.5% CaSO,
Initial ph 4.5
Sample
19 - 1
19-2
19 - 3
19 - 6
19-7
19 - 8
19 - 9
19 -10
Gas temperature 275 F
Liquid temperature 125 F
6
4
4
4
4
4
4
4
4
4
oH
.08
.0
.2
.05
.03
.18
.0
.43
.40
.43
N flow
0^ flow
SO Out
Pf"
1350
1465
1200
1650
1500
1300
1625
1075
1200
1300
Ca Inlet
Total
m. moles/1
38.5
37.5
38.0
.38.5
39.0
38.0
37.5
39.0
5 scfm
12
scfh
Make Up
3
2
6
1
2
3
1
6
5
5
Ca
gph
.9
.8
.9
.7
.8
.9
.7
.7
.2
.2
Rate
Recycle
_ 8Pm_
0
0
0
0
1
1
1
1
1
0
.5
.5
.5
.5
.0
.0
.0
.0
.0
.5
Exit
Total
m. moles/1
39
34
35
39
43
45
35
33
.0
.5
.0
.0
.0
.0
.0
.5
Solution
m. moles/ 1 m.
29.
34.
19.
30.
37.
26.
26.
23.
0
5
6
8
3
5
5
2
Average
Calcium
Balance
moles/ 1
+0.5
-3.0
-3.0
+0.5
+4.0
+7.0
-2.5
-5.5
-0.3
L/C
50
50
50
50
100
100
100
100
100
50
SO Removal
2 %
41
36
48
28
35
44
30
53
48
44
Gravimetric
m
ivera«e*
Total
.moles/1
26.7
22.1
19.0
27.4
36.1
37.0
25.2
23.0
27.1
Sulflte*
Sulflte
Input Total
m. moles/1 m. moles/1
60.6
64.2
52.6
61.9
76.4
54.7
57.0
56.7
Determination of
Solution
m. moles/1
15.5
18.0
19.8
16.8
17.8
18.4
24.6
19.5
18.6
42.0
40.2
42.0
46.7
8:8
43.2
34.3
S0,=
by Tltration
Solution
m. moles/1
32.6
40.0
29.4
37.2
49.2
30.4
30.2
32.4
Average
Sulflte
Balance
m. moles/1
-18.6
-24.0
-10.6
-15.2
-25.9
- 5.7
-13.8
-22.4
-17.0
*Sulflte Input
SO. Removed - Sulflte In Feed
cr
-------
Table XXVIII. Summary Results — Scrubbing Liquid, 0.5% CaSO
f
D
'
Initial pH 4.93 Gas temperature 275 FQ
Inlet SO. 2300 ppm Liquid temperature 100 F
N- flow 5 scfm
0? flow 12 scfh
Sample
20 -
20 -
20 -
20 -
20 -
20 -
20 -
20 -
1
2
3
4
5
6
7
8
Sample
20 -
20 -
20 -
20 -
20 -
20 -
70 -
20 -
1
2
3
4
5
6
7
8
2
pH
4.1
3.93
4.2
3.83
4.13
3.98
4.3
4.25
Ca Inlet
Total
m. moles/1
39.5
41.0
38.0
40.0
37.5
35.5
37.0
35.5
SO, Out
PPm
1415
1625
1225
1575
1175
1400
1075
965
Ca Exit-
Total
m. moles/1
38.0
39.0
37.0
40.0
37.0
43.5
33.5
36.0
Make Up Rate
gph
2.7
1.7
5.0
1.6
5.0
2.7
6.8
6.8
Solution
m. moles/ 1
35.5
40.0
79.3
40.3
32.0
39.5
76.0
25.5
Average
Ca Ic lum
Balance
m. moles/ 1
-1.5
-2.0
-1.0
0
-0.5
+8.0
-3.5
+0.5
0.0
Recycle
gpm L/C
0.5
0.5
0.5
1.0
1.0
1.0
0.5
1.0
50
50
50
100
100.
100
50
100
Gravimetric
Total
m. moles/1
26.0
21.5
18.7
21.5
18.2
21.6.
17.5
24.8
Average* 21.2
0..,*,^ Sulflte 1
_„ _ , --"-— Total
SO. Removal Input , . ,
2 « , i, m. moles/1
1 m. moles/1
39
29
47
32
49
39
53
58
Determination
Solution
m. moles/1
17.0
17.9
22.6
18.0
20.7
19.4
18.6
17.6
Average 18.9
70.2
78.5
58.0
83.0
58.1
66.5
57.5
53.8
of SO =
46.7
55.5
49.0
58.3
52.5
60.8
44.5
49.3
*Sulflte Input
>y Tit rat Ion
Solution
m. moles/1
43.8
55.0
40.2
57.6
42.8
56.0
36.2
39.4
I
« SO Removed
Sulflte
Balance
m.moles/1
-23.5
-23.0
- 9.0
-24.7
- 5.6
- 5.7
- 9.0
- 4.5
Average -13.1
n
-------
Table XXIX. Summary Results—Scrubbing Liquid. 0.5% CaS03
VO
to
Sample
21 - 1
21-2
21-3
21-4
21-5
21-6
21-7
21-8
21 - 1
21 - 2
21 - 3
21 - 4
21 - 5
21-6
21 - 7
21-8
Initial pH 4.
6
Inlet SO, 2300 ppm
M
o,2
CO
PH
4.28
3.98
4.1
4.0
4.2
3.93
4.3
4.05
Ca Inlet
Total
m. moles/1
39.2
40.7
35.5
38.0
37.0'
38.6
35.5
38.6
flow ' 4
flow 12
„ flow 40
2
SO. out
ppm
965
1515
1075
1300
1175
1575
1100
1475
Ca
Total
m.molea/1
37.5
38.0
44.0
40.5
35.0
42.8
31.8
36.0
scfm
scfh
scfh
Gas temperature
Liquid tempera tu;
Make Up Rate Recycle
gph
6.5
1.7
4.8
2.7
5.1
1.7
6.1
2.7
Exit
Solution
m. moles/1
27.5
38.8
33.5
38.3
29.3
41.6
25.0
33.5
gpm
0.5
0.5
0.5
0.5
0.25
0.25
0.25
0.25
Calcium
Balance
m. moles/1
-1.7
-2.7
+8.5
+2.5
-2.0
+2.2
-3.7
-2,6
Average -0.1
275°
L/C
100
100
100
100
50
50
50
50
S0_ Removal
58
34
53
44
49
31
52
36
Sulfite*
Input
m.moles/1
58.2
82.4
58.7
72.3
57.0
76.9
53.5
66.3
Sulfite
Total
m. moles/1
49.8
60.8
57.2
58.5
49.8
60.0
43.6
53.8
by Tltratlon
Solution
m. moles/1
40.0
60.0
46.8
55.8
43.2
59.4
37.6
40.3
Sulfite
Balance
m. moles/1
- 8.4
-21.6
- 1.5
-13.8
- 7.2
-16.9
- 9.9
-12.5
Average -11.5
—\
o
rt>
-------
60
50
40
30
20
B-S
S
w
1
CM
O
CO
L/G
4- 200
O 100
& 50
pH
3.5
4.0
4.5
5.0
5.5
Figure 29. Sulfur Dioxide Removal Efficiency as a Function of pH— 0.5%
CaS03 (125°F, No Oxygen)
93
Arthur D Little. Inc
-------
60
50-
40
30
20
L/G + 200
O 100
50
pH = 3.9
Make Up Rate (gph)
Figure 39. Sulfur Dioxide Removal Efficiency as a Function of Make Up Rate
0.5% CaS03 (125°F, No Oxygen)
94
Arthur D Little, Inc
-------
60
50
40
30
20
-H
U
•H
w
CO
O
a
CM
P"
L/G O 100
50
3.5
4.0
4.5
5.0
5.5
Figure 31. Sulfur Dioxide Removal Efficiency as a Function of pH— 0.5%
CaS03 (125°F, 4% Oxygen)
95
Arthur D Little. Inc
-------
60
50
40
30
20
L/G O 100
& 50
= 4.0
Make Up Rate (gph)
Figure 32. Sulfur Dioxide Removal Efficiency as a Function of Make Up Rate
0.5% CaS03 (125°F, 4% Oxygen)
96
Arthur D Little Inc
-------
60
50
40
30
20
&•«
o
c
0)
M-l
1
CM
o
l/D
L/G
O 100
50
pll
3.5
4.0
. 4.5
5.0
5.5
Figure 33. Sulfur Dioxide Removal Efficiency as a Function of pH— 0.5%
CaSO (100°F, 4% Oxygen)
97
Arthur D Little. Inc.
-------
60
50
40
30
20
o
c
W
0)
O
E
O
in
L/G
O 100
50
4.25
pH = 3.9
Make Up Rate (gph)
Figure 34. Sulfur Dioxide Removal Efficiency as a Function of Make Up Rate—
0.5% CaS03 (100°F, 4% Oxygen)
98
Arthur D Little. Inc
-------
60
50
40
30
20
o
ti
(U
w
cfl
O
(U
CM
O
C/l
L/G
O 100
£ 50
pK
3.5
4.0
4.5
5.0
5.5
Figure 35. Sulfur Dioxide Removal Efficiency as a Function of pH-
0.5% CaS03 (100°F, 4% Oxygen, 13% Carbon Dioxide)
99
ArthurDLittle.Inc
-------
60
50
40
30
20
g
o
e
0)
ft
Osl
O
to
L/GO100
pH = 3.9
Make Up Rate (gph)
Figure 36. Sulfur Dioxide Removal Efficiency as a Function of Make Up Rate —
0.5% CaS03 (100°F, 4% Oxygen, 13 % Carbon Dioxide)
100
Arthur D Little. Inc
-------
Scaling Experience
During a total time on-line of about 65 hours, there was no cumula-
tive pressure drop increase (Figure 37). No precipitation or scale
formation was expected during run XXVI at 125°F because of the
absence of oxygen in the gas stream and carbonate species in the liquor.
Runs XXVII-XXIX, on the other hand, were carried out in the presence of
oxygen; and qualitative analysis snowed the presence of sulfate. None-
theless, after 65 hours operation no evidence of precipitation or scaling
on the column packing or any other part of the apparatus was observed.
This contrasts markedly with the introduction of oxygen into a carbonate
slurry system where catastrophic scaling occurred after about 12 hours.
In general, the pH is much lower under a given set of process con-
ditions with sulfite than with carbonate slurries and can be quite
close to the equilibrium value for a sulfite slurry.
Slurry Analysis and Mass Balance
Analysis of the slurry for sulfite and for calcium at the inlet
(make up) and exit (purge) has been made for each set of steady-state
conditions. These results are presented in Tables XV-XXIX together
with the sulfur dioxide removal data.
At steady state, the calcium and sulfite balances for runs made in
the absence of oxygen should be zero. The deviations from zero are
therefore indicative of a non-steady state in the solids being circulated
in the scrubber due to solids hold-up and scale deposition. The sulfite
and calcium balances tend to show smaller negative values at high liquid
rates when solids deposition in the column is less pronounced.
Sulfite losses should be equal to calcium losses if the net reaction
is formation and hold-up of calcium sulfite. Average values for runs
made in the absence of oxygen give good agreement although individual
samples show some scatter. For runs where oxygen was present, the oxida-
tion of sulfite to sulfate must be taken into account in computing a
sulfur balance.
For runs made with the 0.5% CaS03 slurry (Tables XXVI-XXIX), the
average calcium balance is close to zero as expected because there is
no net precipitation in the scrubber. (Absorption of S02 leads to
Ca(HS03)2 formation.) The variations recorded for individual analytical
results can probably be ascribed to mechanical hold up in the apparatus
(and experimental error). The calcium analysis shows that a very good
mass balance can be obtained with the apparatus under conditions where
no precipitation or scaling is occurring.
Results for the sulfite balance show a consistent negative error
which cannot be due to precipitation in the scrubber as in runs with
carbonate slurries. Some of the discrepancy could be due to sulfate
101
Arthur D Little, Inc
-------
1.0
0.5
o
to
M
0)
4J
L/G O 100
A 50
16
24
32 40
Time (Hours)
48
56
64
Figure 37. Pressure Drop as a Function of Operating Time — 0.5% CaSO
>
-\
c
C"
o
-------
formation in runs XXVII to XXIX. However, oxygen was not present in
run XXVI and the sulfite balance shows a large negative value.
Examination of the data also shows that the values of calcium to
sulfite in solution approach a 1:1 ratio rather than the 1:2 ratio
expected for calcium bisulfite. The sulfite and calcium values are both
in excess of the predicted equilibrium values which either indicate a
supersaturated solution of calcium sulfite or incorrect assumptions in
the thermodynamic calculation.
Various checks were carried out on the analytical method for sulfite
and the sampling technique to discover the source of any errors contri-
buting to the sulfur losses. Samples have been fixed with potassium
hydroxide to prevent loss of S02 from the solutions while handling.
Samples have also been analyzed immediately on removal from the apparatus.
In each case the results were within experimental error (5%) of control
samples. Tests for thiosulfate were also negative.
The negative sulfite balance also appears related to the experimental
variables such as L/G ratio and SC^ removal efficiency. One possible
explanation of the loss of sulfite is evolution of S02 in the hold tank,
where the residence time is 4-16 minutes, before the analytical sample
is taken. This evolution of S02 may be quite significant at the low pH
obtained in CaS03 slurries but would be much smaller in CaCC>3 slurries.
For the runs summarized in Tables XV-XXV, the average sulfite
losses are about the same for all runs, ranging from 17.6 to 18.7 m.moles/1.
There is also good agreement for the calcium balance, except for a some-
what lower value in run XVIII. Therefore the losses appear to be
independent of temperature and calcium carbonate concentration. The
sulfite loss is consistently greater than the calcium loss by a small
amount and may be due to loss of S02 from the hold tank before sampling.
The equilibrium concentration of sulfite ion in solution varies
with pH, falling rapidly above pH 6. Calcium in solution also falls
with increasing pH and reaches the calculated thermodynamic equilibrium
value at about pH 6. Below pH 5.5 the concentration of calcium is only
slightly less than that of sulfite and the concentration of sulfite in
solution seems consistently higher than indicated by the calculated
thermodynamic curves.
In Tables XXVII and XXVIII,for runs with a calcium sulfite slurry
in the presence of oxygen, there is essentially no calcium ion lost in
either run. Therefore the negative sulfite balance should be largely
accounted for by the conversion of sulfite to sulfate, in the scrubber.
As the results show, the average total sulfate concentration exceeds the
sulfite loss by a considerable amount. At present there is no satis-
factory explanation of this anomaly but it may be due to difficulties with
the gravimetric determination. Alternative analytical methods are being
investigated for future work. The concentration of sulfate in solution
103
Arthur D Little. Inc
-------
is indicative of a saturated solution. Since the samples had been allowed
to equilibriate in the presence of CaSO, , supersaturated solutions would
not be expected.
DISCUSSION
Effect of Operating Variables on SO,, Removal Efficiency (No Oxygen)
Experiments have been carried out as described above to define the scrubbing
efficiency and scaling potential of calcium carbonate, limestone, and cal-
cium sulfite slurries as a function of various operating parameters such
as liquid to gas ratio (L/G) , stoichiometry of calcium carbonate or lime-
stone addition, temperature and slurry concentration. Derivative para-
meters such as steady state pH in the effluent hold tank purge stream and
S0£ removal rates were recorded.
The extent of scaling was observed primarily by the increasing pressure
drop across the packed bed scrubber, by visual observation of deposits in
the rest of the scrubber loop, and from mass balance results. These ob-
servations are discussed later.
In Figure 38 we display the proportion of sulfur dioxide removed from the
gas stream as a function of effluent hold tank pH. A feed of 0.5% calcium
sulfite slurry gave pH values below 4.3; an 0.5% calcium carbonate slurry
gave pH values between 4.3 and 6.4 and a 1% calcium carbonate slurry gave
operating pHs in excess of 6.4. Equipment limitations in the bench scale
scrubber prevent us from operating in the same regions of pH with these
different slurry feeds. However, we believe that the results represent
a continuous series of scrubber loop conditions across the pH range 3.7
to 6.6.
Superficial inspection of Figure 38 makes it clear that the kinetic and
equilibrium conditions in the scrubber are a complex function of pH (and
stoichiometry), temperature, and L/G. At different ranges of pH, dif-
ferent mechanism are in operation. Insights into the scrubber mechanisms
may be obtained by considering the effect of three variables on scrubber
performance.
The rate of sulfur dioxide removal is clearly a strong function of pH and
in all regions increases linearly with increasing pH. However, in the pH
region below 4.3 and above 6.6, the SO- removal rate is a stronger func-
tion of pH than it is in the region between 4.3 and 6.4.
In our experiments, the effects of stoichiometry cannot be distinguished
from the effects of pH, since changing stoichiometry also changes pH.
pH is recognized to be an important variable in equilibria and kinetic
104
Arthur D Little. Inc
-------
1.00
90
80
70
§ 60
0)
fM
O
t/1
50
40
30
1% CaCO,
T = 100° F
125° F
4-5 5.0 5.5 6.0
pH
FIGURE 38
EFFECT OF pH ON S02 REMOVAL FOR DIFFERENT SLURRIES
.(No Oxygen)
105
6.5
7.-
ArthurD Little Inc
-------
factors. The S(>2 removal rate is high at a stoichiometry down to about
0.8 which corresponds to almost complete carbonate utilization. The re-
moval rate drops off rapidly at lower stoichiometries and the pH also
falls rapidly. Stoichiometries greater than 1 increase the SO, removal
rate to a small extent. *
L/G
In'general, if absorption efficiency of a gas liquid contactor increases
with L/G it can be deduced that the reaction is kinetically limited by a
liquid phase reaction: the more liquid which is supplied to the scrubber
in comparison to the gas flow the less is this factor limiting on the
total rate of reaction. Referring again to Figure 38 it is clear that
L/G has a strong effect in some regions of pH and at some temperatures
but not at others. Thus, under some conditions, scrubber kinetics are
limited by liquid phase reactions; under other conditions other rate
limiting steps become important.
Temperature
In most instances the rate of chemical and diffusion reactions is expect-
ed to increase with temperature. Referring to Figure 38, we see that in
some regions of pH and at some values of L/G lowering the temperature of
the scrubber liquor from 125°F increases the SO- removal rate. The fac-
tor which may be responsible for this negative temperature effect is the
equilibrium partial pressure of sulfur dioxide above the scrubber liquor
which would be expected to decrease as the temperature falls. Thus in
those regions where a negative temperature effect is observed, we sup-
pose that the partial pressure of SO- above the liquor is in some way
rate limiting in the absorption of So? from the gas stream.
Kinetic Observations
The observed effects cannot yet be fully rationalized to yield a mech-
anistic understanding of the kinetics of the scrubber process. However,
it is possible to make some general observations on the mechanisms which
may be rate controlling in the various pH ranges.
pH Less than 4.3 (0.5% Calcium Sulfite Slurry)
In this pH range the overall mechanism of S02 removal from the gas stream
involves reaction with solid calcium sulfite to produce a calcium bi-
sulfite solution. This reaction results in a low steady state pH.
At an L/G of 100, a decrease in temperature from 125°F to 100°F increases
the rate of S02 removal. This inverse temperature coefficient is attri-
buted to a lower partial pressure of sulfur dioxide above the liquid
phase at the lower temperature. At 125°F, there is no L/G effect on the
total reaction rate indicating the absence of diffusional limitations in
the liquid phase over the pH range. However at 100°F, the higher SO-
interfacial reaction rate (due to the more favorable equilibrium) makes
the liquid phase resistance important and there is an increase in total
reaction rate by increasing the L/G from 50 to 100.
106 Arthur D Little, Inc.
-------
pH 4.3 to 6.4 (0.5% Calcium Carbonate Slurry)
In this region we observe both an effect of L/G at 100°F and 125°F and a
negative temperature effect upon the reaction rate. The steady state
kinetically limiting reaction mechanism must therefore involve a change
in sulfur dioxide partial pressure above the liquor and a subsequent
diffusion (or reaction, but this is unlikely) in the liquid phase.
Note that as the pH increases at an L/G of 100 and 50 the negative tem-
perature effect becomes less marked as the S02 partial pressure above the
liquor falls because of the rising pH. Also, in this pH range the bicar-
bonate concentration increases in the scrubber liquor as the pH rises.
Thus bicarbonate ion concentration may be contributing to the rise in SO
removal as a function of pH. In this pH region (at least up to pH 6), 2
equilibrium calculations show that calcium carbonate is not stable as a
solid phase if C02 is less than one atmosphere and all carbonate species
should be in solution.
pH Greater than 6.4 (1% Calcium Carbonate Slurry)
In this region we see no L/G effect and no temperature effect on the
reaction rate. However, the reaction rate is as strongly dependent on
the pH as it was below pH 4.3. In this region we again have a reaction
which is dependent on the dissolution of a solid phase (calcium carbon-
ate) as absorption of S02 occurs. Clearly the equilibrium SO, partial
pressure above the liquid is no longer rate controlling (the equilibrium
pressures are exceedingly low) and apparently liquid phase resistances,
despite the high rate of reaction, are absent:. Therefore^ the reaction
rate must be close to the gas phase limitation,(as is" observed with" high-
ly alkaline scrubbing media- see preliminary shakedown experiments),
and controlled by the reaction or production of a chemical species whose
concentration is solely and strongly dependent on pH (for example, car-
bonate or bicarbonate).
If the rate of dissolution of the carbonate is an important kinetic fac-
tor in determining the rate of reaction in this pH region, one might
expect the removal rate vs pH line to be shifted to lower pH levels with
more difficulty dissolving limestones.
Of particular importance to practical scrubber operation is the fact that
the results shown in Figure 38 indicate, just as the equilibrium calcu-
lations of Chapter II also do, that the pH range of normal scrubber op-
eration lies in a complex region of equilibrium and kinetic effects. In
the range of pH from 5.7 to 6.5, radical changes in the effect of all
the operating variables - L/G, temperature, and pH - are observed. Slight
changes in operating conditions can give very different results in regard
to S02 removal efficiency and, by implication, to scaling reactions.
Also, pH is a strong indication of overall scrubber operating conditions.
107
Arthur D Little, Inc
-------
Scaling Phenomena
Types of Scale
Several different types of scale were observed during the course of op-
eration of the bench-scale scrubber, corresponding to scale observed in
the field. Very early in the program, a thin hard sulfite scale with a
yellow surface coloration was observed to form on some of the glass sur-
faces of the apparatus. This type of scale was formed when scrubbing
with either calcium carbonate or limestone, although the color was slight-
ly different in the two cases.
Gas inlet deposits, consisting of a mixture of calcium sulfite, calcium
sulfate, and calcium carbonate, were formed, but in contrast to larger
units the deposits never built up to a troublesome level under the re-
latively low flow rate conditions used.
Sulfite scale built up gradually during runs made in the absence of
oxygen, as evidenced by a slow increase in pressure drop across the
packed bed, as well as negative calcium and sulfite mass balance results.
The sulfite scale build-up is more rapid at 125°F than at lOO^. At the
higher temperature, the pressure drop increase forced a shut-down of the
apparatus in 20 hours, while at 100°F, it was possible to run for 30
hours. The sulfite scale collected primarily on the packed bed and on
the walls below, but some scale also precipitated in the hold tank and
in the scrubber recycle loop. Scale deposition on the packed bed, be-
cause of increased liquid hold up (as shown by pressure drop increase)
tends to improve removal efficiency.
The presence of 4% oxygen, and hence the conversion of some sulfite to
sulfate, led to catastrophic scaling when scrubbing with limestone. At
125°F, the scrubber plugged up rapidly and without warning after 12 hours
of operating time. During that time, removal efficiency was unaffected
and the change in pressure drop across the column was minimal. In fact,
the severe scale deposition occurred mainly in the recycle lines, with
very little deposition on the packed bed itself. The material respon-
sible for the catastrophic plugging proved to be a mixture of calcium
sulfate, sulfite, and carbonate. Since oxygen was clearly implicated
in the scaling reaction, it is possible that sulfate deposition is re-
sponsible for initiating a cementing action.
Effect of Materials of Construction
In order to qualitatively evaluate the scaling potential for various op-
erating parameters, coupons of various materials were immersed in the
scrubber at six separate locations for a total of 37 1/4 hours of run
time. Major experimental parameters included 1/2% CaCO-, 4% 0_, 2300 ppm
S02 and liquid temperatures of 125°F and 100°F.
108
Arthur D Little, Inc
-------
Half-inch square by 1/16 inch thick samples were employed of Monel 400,
Teflon, Copper (OFHC), 316 Stainless, low carbon steel (1008) and boro-
silicate glass. One surface of the sample was polished and the other
was abraded on 240 grit SiC paper. Samples were linked together in a
chain with nylon thread. The individual coupons and the composite chain
were tared to the nearest tenth milligram.
Specimen chains were placed at six positions as follows:
T.. - In line from holding tank to makeup tank
T_ - In makeup tank
T, - In line from makeup tank to scrubber
T, - Top of scrubber
T - Just below bed (at gas inlet)
Tg - In holding tank
Examination of the sample chains after exposure showed no evidence of
any scaling, corrosion or staining of the teflon, glass or stainless
steel coupons. The monel coupons were tarnished in all cases, but
showed a weight change of no more than a few mg; there was no evidence
of scale formation. The copper coupons were all tarnished, some quite
heavily. Samples from positions 3 and 5 contained very small amounts
of white-colored scale.
Finally, the low carbon steel samples were all heavily corroded (oxi-
dized) . A white-to-dark gray scale was observed on samples at positions
1, 2, and 6. The other samples were very heavily corroded which may
have precluded .the formation of a scale. Weight changes for the low
carbon steel and copper samples are as follows:
Weight Loss - mg
Sample Position Low Carbon Steel Copper
173
2 7 21
3 24* 11
4 15* 23
5 372 10
6 18 4
*Weight gain - mg
109
Arthur D Little Inc
-------
X-ray diffraction patterns were obtained from the copper and low carbon
steel samples that appeared to have scaled; these samples included low
carbon steel, positions 1, 2 and 6, and copper, positions 3 and 6. In
all cases, the scale was found to be CaSO '1/2 H20. There was not evi-
dence for any transformation to sulfate.
It is difficult to account for the loss in weight of the copper coupons.
Presumably, the relatively low pH of the solution resulted in some dis-
solution of copper. The final weight of the low carbon steel samples,
of course, is a compromise between the weight loss due to formation of
a (mostly) non-adherent oxide versus the weight gain due to scale for-
mation.
In all cases the observed scale was very loosely adherent, being easily
removed with a scraper.
110
Arthur D Little Inc
-------
CHAPTER IV
POTENTIAL SCALE CONTROL METHODS
On the combined bases of the literature review, thermodynamic
analyses, and supplemental experimental work carried out over the past
year, we have assessed the potential effectiveness of scale control
methods that have been applied to wet scrubbers; we have identified some
of the critical factors which seem to influence scale formation; and we
have identified a number of approaches to scale control which merit
further exploration.
SUMMARY
Many of the basic principles involved in designing and operating
limestone wet scrubbers to minimize scaling problems are well recognized
and may be summarized as follows.
1. Good Engineering Design
In scrubber towers wherein a gas is to be in intimate contact with
a liquid flowing countercurrently, it is imperative to design and operate
the unit in such a manner so as to preclude any appreciable by-passing
of the two streams. Ordinarily this is accomplished by feeding both the
liquid and gas evenly over the tower cross section, and in some cases,
by collecting and redistributing the streams at one or more heights
within the column.
In the limestone tail-end addition systems of interest here, the
appreciable solids content in the liquid stream presents an additional
complication. With solids flow, a good distribution of the gas, liquid,
and solid portions is even more important since solids may become so
concentrated in certain zones as to form deposits that can grow and
disturb the efficient operation of the column, even to the extent of
complete blockage.
While the critical velocities needed to keep particles in suspension
are lower than those normally encountered in scrubbers, especially in
the case of vertical flow, areas of deceleration and possibly stagnation
may form behind sections of packing, at bends in the piping, and at any
areas of roughness in the pipe. It is in these areas that scaling could
well begin by growth of crystallites onto the surfaces of mechanically
deposited silt. Thus, any bends in the piping should be made with large
radii of curvature; valves other than gate valves should be avoided;
smooth piping should be used; obstacles in the flow path (uneven pipe
joints, unnecessary valves, probes) must be eliminated; and abrupt
changes in pipe diameter in the piping system should be avoided.
Ill
Arthur D Little; Inc
-------
2. Avoidance of Intermittently Wet and Dried Areas
Experience has amply demonstrated that all surfaces exposed to the
scrubber slurry should be well irrigated.
Gas inlet deposits, which have plagued most scrubber units, at
least in the initial phases of operation, can generally be controlled
by a combination of mechanical removal methods and inlet design modifi-
cation.
Demister and reheater deposits have proved to be a more difficult
problem. Effective demister wash systems are not always practicable,
and periodic mechanical cleaning seems at present to be unavoidable.
3. pH Control
The solubility of calcium sulfite increases sharply abe^e pH 6.
When CaS03»l/2H20 is formed by the absorption reaction (CaC03 +
S02 + 1/2H20 + CaS03-l/2H20) above pH 6, the potential for sulfite
scaling is very great. At pH levels below 6, a considerable amount of
the sulfite formed can remain in solution ' as bisulfite. The experiments
discussed in Chapter III suggest that S02 removal efficiencies of 80 -
90% can be achieved down to pH 5.7 or 5.8 at the bottom of the scrubber.
Efficiency falls off rapidly at lower pH levels. If for any reason,
there are local increases in pH above 6 in the scrubber proper, preci-
pitation of calcium sulfite, and eventual scaling may be anticipated.
A. High L/G
It has been found empirically that increases in L/G have a beneficial
effect on reducing the rate of scale deposition. At least two mechanisms
may be operative. High liquid flow rates might simply provide a mechanical
flushing of surfaces where deposits might otherwise form. In addition,
the make per pass is inversely proportional to L/G. The lower the make
per pass, the lower the potential degree of supersaturation, and hence,
the less liklihood there is for chemical scaling. In the work carried
out at ICI in the 1930's, the safe make per pass was found empirically
to be in the neighborhood of 4-5 millieq/1 of total sulfur. In order
to maintain such a low make per pass in modern scrubbing units, L/G
would have to be well over 100, an intolerably high slurry circulation
rate from the point of view of pumping load.
5. Seeding
The use of seed crystals to dissipate supersaturation is a venerable
scale control technique. The levels found by ICI to be effective in wet
scrubbing, 3-5% each of calcium sulfite and calcium sulfate in the slurry,
are far higher than the concentrations normally required. Since a high
solids content in the slurry does tax the pumps, we would recommend that
optimum seed concentrations be established empirically for typical
112
Arthur D Little, Inc
-------
modern scrubber operating conditions. While the effectiveness of gypsum
seed crystals for dissipation of calcium sulfate supersaturation has
been demonstrated in the laboratory, the usefulness of calcium sulfite
seed crystals has never been definitively established. In the review of
scale control techniques used in industrial processes, which concludes
this chapter, we discuss some of the fundamental factors related to
optimum choice of seed concentration and structure.
6. High Residence Time in Hold Tank
A delay tank or hold tank is normally introduced into the scrubber
loop to allow sufficient time to desupersaturate scrubber liquors
before they are returned to the scrubbing tower. Addition of fresh
calcium carbonate to the scrubber effluent liquor just before it enters
the hold tank further aids in dissipating any sulfite supersaturation,
since an increase in pH promotes precipitation of calcium sulfite. The
required delay times to assure desupersaturation under given scrubber
operating conditions have not been well-established. ICI provided for
a three to four minute delay time, but limited the sulfite-sulfate make
per pass to 3-4 millieq/1. Delay times of 10-20 minutes are more common
in modern practice, but the make per pass is typically much higher.
7. Temperature Control
Although the options for controlling scrubber liquor temperature
are limited in practice, our work suggests that scaling problems are
considerably less acute at 100°F than at 125°F. The rate of deposition
of calcium sulfite appears to be more rapid at the higher temperature.
In addition, the equilibrium transition point between gypsum and anhydrite
lies at 100°F. At higher temperatures, a deposit of the stable anhydrite
or the metastable hemihydrate (plaster of paris) on process equipment
surfaces, might initiate a cement forming or "setting" reaction that
could lead to catastrophic scaling. It thus may be advantageous to
operate wet scrubbers at temperatures below 100°F if at all practicable.
8. Control of Oxidation
While oxidation clearly plays a role in calcium sulfate scaling,
the optimum degree of oxidation for minimum scaling is unknown. On the
one hand, a high degree of oxidation might assure a steady supply of
seed crystals for dissipation of supersaturation. On the other hand,
sulfate scale is much more difficult to remove than sulfite scale, and
it therefore might be desirable to inhibit oxidation. The extent to
which oxidation can be catalyzed or inhibited, so that potential benefits
'might be determined, merits further investigation.
9. Use of Additives
Threshold additives of polyphosphates and polyacrylates have been
used to control scaling in many industrial processes, but have not
been explored for wet scrubbers. In the literature review which follows,
the available data on the use of additives for scale control are discussed.
Although none of the data were obtained under the precise conditions
characteristic of wet scrubbers, threshold additives do appear to hold
sufficient promise for reducing scale deposition rates in this application
to warrant some investigation.
113
Arthur DLittklnc
-------
SCALE CONTROL TECHNIQUES USED IN INDUSTRIAL PROCESSES
Many industrial processes have been afflicted by deposition of
calcium and magnesium scales from aqueous process streams. Research
has been conducted by a variety of groups to devise operating procedures
which minimize this problem. Advances have been made in many fields in
the understanding of scale deposition processes and hence in the develop-
ment of largely successful scale prevention techniques. While in many
cases the methods used are unique to the industrial process under con-
sideration, the general technology provides an essential background of
information for a rational approach to the solution of scaling problems
in wet scrubbers. Moreover, it appears that in some cases the methods
which have been found to be effective might be directly applied or
adapted to wet scrubber operation. We should note however, that in
many cases conditions of temperature and chemistry are not the same,
and experimentation will be essential to evaluate approaches.
Scaling problems arise in almost any process where heat is being
applied to solutions of calcium or magnesium salts. Such solutions may
be process liquors, or natural fresh, brackish or salt water often used
for cooling purposes. In cases where evaporative loss of water occurs,
deposition of dissolved salts is expected as their solubility products
are exceeded. However, this is by no means the only cause of salt pre-
cipitation which under appropriate conditions may lead to scaling on
the walls of containing vessels. For example, natural waters usually
contain quantities of dissolved calcium bicarbonate which, when heated
decomposes to calcium carbonate which is sparingly soluble and preci-
pitates:
2Ca(HC03)2 ->• 2CaC03 + C02 + H20
In wet scrubber operation, this is also a possible scaling reaction.
In addition the analogous sulfite reaction may occur:
2Ca(HS03)2 -> 2CaS03 + S02 + H20
particularly where calcium bisulfite solutions meet hot incoming gases.
Furthermore, some salts have inverse solubility curves; that is, their
solubility decreases with temperature. If such a salt is nearly sat-
urated and the solution meets a hotter surface such as a heat exchanger
tube in a boiler, the solubility of the salt at the surface may be ex-
ceeded at the surface and scale be deposited. Such a salt is calcium
sulfate dihydrate above about 55°C., and also anhydrous sodium sulfate
and sodium carbonate monohydrate. It seems unlikely in wet scrubber
operation, however, that the region of temperature of inverse solubility
for calcium sulfate is reached, unless perhaps in the region of the in-
coming hot flue gases.
The industrial processes in which scaling has been a problem are
mainly those where evaporation of calcium solutions takes place. These
include the following:
114
Arthur D Little, Inc
-------
Industry
Sulfur dioxide scrubbing from
flue gas.
Distillation of brackish and
sea water.
Sugar manufacture.
Salt manufacture.
Calcium sulfate manufacture.
Sulfur processing.
Paper manufacture.
Phosphate processing.
Sodium sulfate manufacture.
Process with scaling problems
Gas liquid contactors and subsequent
handling of liquors.
Brine evaporation. Cooling waters
for the steam.
Evaporation of sugar solutions.
Evaporation of concentrated brines
containing calcium sulfate.
Evaporation of concentrated calcium
sulfate solutions.
Evaporation of "black liquors" con-
taining calcium sulfite and sulfate.
Evaporation of saturated solutions
for recrystallization.
Some of the methods which have been developed for scale prevention
in these industries are either too specialized or otherwise inappropriate
for application to the wet scrubbing problem. These include pretreat-
ment of cooling water by softening or by acidification and probably the
addition of physical scouring agents (such as Taprugge balls) to recir-
culating liquors. We will not discuss these further. Two methods for
scale control of potential utility in wet scrubbing, however, are those
of seeding and threshold treatment.
Scale Control by Seeding
Scale deposition is often a slow process usually taking place from
supersaturated solutions. If supersaturation is minimized by the accelera-
tion of the precipitation process away from the walls of the containing
vessel, scaling can be avoided. It has been found that in some systems,
this can be achieved by seeding the solution with particles of a
particular type, preferably of the same polymorph as the precipitating
species. Nucleation of crystallites is thereby enhanced, and precipi-
tation is largely confined to the bulk of the solution. Even where crys-
tallization still occurs on the walls of the vessel, it is often found
that the deposits are softer and easier to remove.
Sodium Sulfate Manufacture - An early description of the evaluation
of seeding to control scale occurs in a review of evaporator design for
sodium sulfate manufacture^"'. Without seeding, scale of 3/8 inch thick
115
Arthur D Little, Inc
-------
accumulated in less than an hour and the evaporator ceased to function;
with recirculation of seed crystals, evaporator operation was unaltered
even after 5 1/2 days. However, special pumps had to be designed to
withstand the abrasive action of the seed crystals. Features included
an extra large impellor and shaft, very large oil ring bearings, a long
running fit casing between impellor housing and stuffing box, circulating
fresh water around the shaft in front of the packing .and a special grade
of packing.
Early Sulfur Dioxide Wet Scrubbing - This work inspired the use of
seeding to control scaling in one of the first sulfur dioxide wet scrubbing
plants to be installed at Fulham Power Station, England by Howden-ICI
19-21. in an earner installation at Battersea, Thames River water
was used to absorb the sulfur dioxide as sulfurous acid. This became
oxidized to sulfuric acid by the air and was returned to the river to
be neutralized by the "natural alkalinity" of the river water. At
Fulham*, such a simple process could not be used because the additional
acid would have exceeded the river's assimilative capacity. A lime
slurry was therefore used to scrub the gas. Lime improved the absorption
efficiency from 45% to 75% and eventually to over 98% compared to water
alone.
However, in preliminary experiments, "a serious difficulty" arose
in the form of deposits of calcium sulfate scales containing traces of
fly ash, calcium sulfite and calcium carbonate which completely blocked
the pilot washing plant after 72 hours. Studies were made thereafter
of the rate of desupersaturation of calcium sulfate solutions when seeded
with different amounts of calcium sulfate dihydrate crystals.
This work showed that if supersaturation of calcium sulfate were
to be limited in the scrubber and the total liquor capacity of the system
were to be of reasonable size, a high proportion of solids in the recir-
culating liquors was necessary to seed the crystallization adequately
(5% by weight of slurry or more). Actual times for the desupersaturation
of solutions which had the necessary slight degree of supersaturation
required to prevent scaling were longer than those shown in Figure 38,
but similar seed concentrations were found effective. Stagnation at
surfaces was found to encourage crystal attachment and was minimized,
and the crystalline shape of the calcium sulfate was found to seriously
limit the effectiveness of calcium sulfate seeding. Therefore, oxidation
of sulfite to sulfate was encouragedby maintaining the pH below 6.4 so
as to keep the sulfite in solution, while still maintaining high S02
scrubbing efficiency. The pH electrode had just been developed around
the time that ICI was initiated. Therefore the numerical values of pH
should probably not be taken too literally.
*Fulham Power Station was designed to liberate 1,000 million cu. ft. of
flue gas per day containing 0.9 grain of S02 per cu. ft. In 1937, this
was reduced by lime slurry scrubbing to only 0.0046 grains per cu. ft.
116
Arthur D Little Inc
-------
ICI found that calcium sulfite supersaturation was not a problem
because of solubilization by the dropping pH in the scrubber, the oxidation
of sulfite to sulfate and the proportion of make up liquor to liquor
rate. With wood grid packing it was found that a delay time to allow
precipitation external to the scrubber of 14 minutes per grain equivalent
of sulfur per gallon absorbed and oxidized within the scrubber to calcium
sulfate in one liquor cycle was adequate, when the suspended calcium
sulfate was over 2.5% by weight (the scrubber operated at pH 6.8 - 6.3).
A plant operated at Billingham under these conditions was found to be
scale-free (despite bird's nests in the scrubber!) over a period of
3,500 hours.
The optimum procedure was therefore to use a slurry of lime or
calcium carbonate containing 5% of calcium sulfate dihydrate, intermittent
flushing (to prevent deposition in stagnant regions), pH control, a
chember for oxidation of sulfite to sulfate, and settling chamber. This
process was patented in 1937.
Paper Industry
Several other successful uses of seeding to control scale were
reported subsequently in a variety of applications. In the paper industry,
the wide use of steam, and cooling and process water gives rise to the
usual boiler operation scaling problems. Calcium sulfite scales may
also form in sulfite digesters. In addition, increasing concern in
the 1950's over the discharge of the dilute waste sulfite liquors, which
contain 50% of the original cellulosic material, into rivers and streams
led to more recirculating systems, and the indirect heating and evapora-
tion of these spent liquors. Concentration of waste liquor by evaporation
prior to further treatment steps had been frustrated by gypsum scaling
on evaporator surfaces. The use of calcium sulfate seeding to control
this scaling was first reported in Sweden.24 and subsequently developed
there as the Escher Wyss gypsum process. No details are noted, but
the process depends on the addition of a gypsum slurry of "a certain
kind" and "in correct quantity and fineness" to the liquor. Whereas
cleaning was necessary every five days without seeding, operation for
more than 50 days was accomplished with little deterioration of heat
transfer performance. This approach would seem to be better than using
expensive sequestering agents for cleaning sulfate scales as recommended
elsewhere. Despite the promise of this approach, it did not appear to
have come into general practice following this earlier work.
27
Recently, French workers have reported successful experimental
and applied control of scaling by seeding in salt manufacture. In an
experimental unit (See Figure 27)scaling was followed by measuring the
Iieat transfer coefficient on a 90/10 Cu Ni tube heated from the inside.
Calcium sulfate hemihydrate was used as a seed with less than 100(j
particle size at a concentration 0.1 to 30 g/1. As the length of time
the seeds were circulated increased, the mean particle size fell to
below 5y. Also, after 6-8 hours the hemihydrate was changed into anhydrite
117
Arthur D Little, Inc
-------
APPii(A7>ON CE LA GtRliNATiQH A LA
&r:.' fltfO\<> J,V///S JW i'-iJ.'A
QRCt/i'T t>YtWtftt£ 4 (?.?
-------
the stable phase at the temperature of interest (100°C). Reasonable
success in prevention of scale was achieved with a seed concentration
of 30 g/1 and Reynolds number greater than 35,000 (rate of circulation
0.4 - 1.8 m/s which is equivalent to Reynolds number of 16,000 to 65,000).
It was shown that the deposition rate of scale depended both on the
hydrodynamic velocity and the concentration of seeds. Circulation velo-
cities had to be sufficient to cause turbulence. The optimum seed
crystals were anhydrite at a particle size less than 20y and concentration
of 20-30 g/1.
Since 1966 the method was successfully used to prevent calcium sulfate
scales in a NaCl salt factory at Varengeville under the following condi-
tions :
Concentrate composition
NaCl 309 g/liter
CaS04 4.08 g/liter
MgCl2 0.7 g/liter
CaCl2 0.2 g/liter
Seed composition (anhydrite + Mg(OH)_) 20-30 g/liter
Particle size 6-25
Circulation velocity 1.2 m/s
Reynolds number 100,000
The evaporator was a five-effect unit. It was found necessary to prevent
boiling from taking place in the tubes. No scale has been formed on
the heat transfer surfaces since start up of the plant in 1966.
Refining of Sugar
In the sugar industry, seeding has shown some promise for scale control
in evaporators. 28,29 ^8 tne cane juice is evaporated, the calcium sul-
fate and calcium phosphate rise in concentration and may precipitate.
In addition, other calcium salts are present which have inverted solu-
bility curves - acetate, citrate, maleate and oxalate. Sugar evaporator
scales are typically composed of silica and calcium sulfate, phosphate
and oxalate. An early Norwegian patent described the addition of scale
forming substances as seeds.*® This approach was subsequently supported
by experimental work and observation of operating plants.il»32 However,
no systematic application of this technique apparently resulted in this
industry either.
119
Arthur D Little Inc
-------
Salt Manufacture
In the manufacturing of salt, seeding has been more successful. Evapora
tion of the brines from which the salt crystallizes first causes preci-
pitation of calcium sulfate which tends to deposit scales on the heat
exchanger tubes. In the Alberger and Grainer processes for salt produc-
tion, the hottest tubes must be drilled out on a twice daily to weekly
routine. It has now become regular practice to return a small amount of
pan brine containing calcium sulfate crystals to the evaporator to act
as preferential nuclei for deposition of the calcium sulfate. 33 Suspen-
sions of 20 grams per liter have been found to be optimum, and have elim-
inated the scaling problem in vacuum pans. Even when running an unpuri-
fied brine, salt evaporators may now be operated for 1-2 years without
having to be drilled out, ^ and the process has been patented.
The control of scale in boilers has been advocated by the addition of
calcium carbonate slurries, ^o
Desalination
A slightly different method of scale control which closely follows the
approach of seeding and which showed initial success is that of contact
stabilization. ^ When studies showed that the supersaturation of CaC03>
Mg(OH>2 and CaSC>4 solutions in seawater was stable over considerable
periods, it was suggested that the brine in seawater evaporators should
be continually circulated through material which would nucleate the super-
saturated solutions and precipitate the salts thereby reducing supersatura-
tion in the heat exchanger. Such a unit is shown in Figure 40.
In a small experimental unit with a potential scale factor* of 1.9 Ibs/
1000 gal. of distillate, 0.87 Ibs./lOOO gal. were deposited in the evap-
orator without contact stabilization, but only 0.09 Ibs./lOOO gal. with
contact stabilization using limestone. In the larger unit illustrated
above (60 gallon/hour) contacting with sand reduced the scale from
0.84 Ibs. to 0.11 Ibs. from water with a potential scale factor of 1.1 -
1.3 Ibs./lOOO gallons distillate. This method might usefully be applied
to a wet scrubber if it is found that the liquid enters the scrubber un-
necessarily supersaturated.
duriJ^hfT intensi5ied in the use of seeding for scale control in the 1960 's
during the development of seawater evaporators; stimulated by the activi-
ties and investments of the U. S. government Department of the Interior
Office of Saline Water (OSW) . At the Wrightsville Beach OSW demonstra-
*"Potential scale factor" is a useful concept used in the desalination
industry to indicate the quantity in Ibs. of precipitated salt expected
as the brine is concentrated to produce a given quantity of distilled
water - usually 1000 gallons.
120
Arthur D Little. Inc
-------
(intuit H.ol
^JJT
L^&v-J
Cc->?r«**3r-'/ ^,
liiJYM
*._.../ \
J| Heoi E»chy>Q*r
I—» Bftne Ov«fflo«
JL
Co^ic,
UM
t ' 'I
jkj
Punp
3
-I
DitftllOK
Pi iip Q.
Heoi E»chc«Gfr I
-
c»cm
Pu
Source: Reference 34
Figure40. A Vapor Compression Seawater Distillation Unit
Fitted with a Contact Stabilization Unit
121
Arthur D Little Inc
-------
tion plant, a pilot plant long-tube-vertical (LTV) single-effect flash
evaporator was used 38 in an evaluation of seeding for scale control.
CaCOj, CaS04 and Mg(OH)2 seed crystals were evaluated under a variety of
conditions. Using a "backward feed" method of operating, in which in-
creasing concentrations of brine occur at increasing temperatures, CaCOo
seed at a concentration of 0.5 - 1% prevented scale formation from clean
seawater for week long runs at a feed temperature of 190°F. , boiling
temperature of 192°F. and overall temperature differences of 8-10°F.
CaC03 seed could not prevent CaSO^ scale at higher temperatures (above
200°F.), either alone or in combination with either anhydrous or calcium
sulfate dihydrate. This scale was found to be calcium sulfate hemi-
hydrate but this salt could not be used for seeding as it "set up to
plaster at every pocket where the liquid could cool slightly." "Forward
feed" operation was then initiated in which the calcium sulfate solubility
limit is never exceeded. At 250°F. , rapidly scaling Mg(OH)2 was not af-
fected by the presence of solid CaCO-j seed crystals, but a slurry of solid
(3-5 gram/liter) prevented such a scale for more than a week.
Following this work, seeding was proposed as a method of scale control at
both the Roswell and Freeport OSW demonstration plants. The problem at
Roswell, a vapor compression plant, was particularly severe ^' as the
calcium concentration in the Artesian well water feed was more than twice
that of seawater (918 ppm vs. 410 ppm) with the sulfate concentration
about the same (2,886 ppm vs. 2,710 ppm). An alarming scaling potential
of 20 Ibs./lOOO gallons of distillate or 10 tons per day was expected at
the designed brine concentration factor* of 4. Laboratory work indicated
that a 1% by weight slurry of calcium sulfate prevented scale during a
176 hour test in a laboratory unit (See Figure 41)at a brine concentra-
tion factor of 5 and steam at 6-7 psig; without the solid phase present
calcium sulfate deposited on the heat exchanger walls reducing the heat
transfer coefficient from about 650 to 225 (BTU/F2, °F, hour) in only
38 hours at a brine concentration factor of only 2.8.
It was believed that calcium sulfate slurry of 0.5% by weight would be
adequate, but experimental difficulties prevented such an evaluation.
Both synthetic water and real water samples gave the same results. Con-
clusions were confirmed in a two-stage 1600 gallons per minute pilot
plant evaporator at Wrightsville Beach, North Carolina.
In other work, a recycle system in a six-stage evaporator using calcium
carbonate seeds was operated for 200-300 hours without scaling •»** up
to 275°F. Calcium sulfate was ineffective. Settling and entrapment of
seed crystals is a continuing problem. Work in England showed that
calcium sulfate hemihydrate was partially effective in preventing sulfate
scale, and a comprehensive theory for seed recycle operation was presented.
*"Brine concentration factor" is the ratio of the volume of the incoming
water to the final brine volume after evaporation and at blow down.
122
Arthur D Little. Inc
-------
nASH TANK |14 IN. DIA.j
COHOCNSH
DISTIlUlf
l»t HUT
tUKANSlt Tims.
8IAWUSS SIfil
Vi IK. 0.0. It CACI
41 IN. IOKG;
Source: Reference 36
Figure 41.A Laboratory Forced Circulation Evaporator Used to Test
Calcium Sulfate Seeding as a Method of Scale Control
123
Arthur D Little. Inc
-------
Japanese workers have found that clay in suspension ("Kuriverter" was
its commercial name) acted as a scale prevention agent by nucleation of
scale forming precipitates. With 200 mg. per litre in the feed water
the amount of deposited magnesium hydroxide scale in a test unit was
reduced by one fortieth. Supersaturation of magnesium hydroxide in the
liquid was also reduced.
Most experience at OSW plants has been less encouraging. At Point Loma,
San Diego, seeds were, found to become incorporated into the scale rather
than preventing it. A grinder was subsequently found necessary at Roswell
to ensure that the particles remained small and therefore active. At the
Freeport, Texas, multistage flash evaporator (MSF) plant, magnesium hy-
droxide seeding was abandoned when silt build up problems were encountered,
seeds grew too large,^6 and carbonate scales continued unabated. ' In
general, OSW has found that acid, threshold (q.v.) or water presoftening
methods of scale control have been more successful than seeding.
Somewhat more recently, work for OSW by Baldwin, Lima and Hamilton (no
longer active In this technology) demonstrated effective scale control in
an exploratory test with calcium sulfite seeding in a 16-stage flash
distillation pilot plant operated at 275°F. and at a concentration factor
of 1.6 far into the scaling range.^ No sign of fouling of heat-transfer
surfaces occurred. The flow circuit is Illustrated in Figure 42.
More optimistic results on the use of seeding in scale control in sea-
water plants has been reported^ by Russian workers, and their interest
continues. Work on the use of CaC03 and CaSO^ seeds to prevent scales
in the production of table salt from seawater had shown them as early as
1955 that so long as boiling took place away from the heat transfer sur-
faces, the treatment was successful, even after 10-14 days, compared to
the severe scaling normally encountered within 9-11 hours. They showed
by electrically heating a stainless steel tube in calcium bicarbonate
solution that scaling is intensified at locations where vapor bubbles
are formed and in the paths taken by bubbles across the heat transfer
surface. In a 4 ton/hour experimental multiple effect plant, scaling
was followed by monitoring heat transfer coefficient. Finely-ground
natural chalk prevented scaling in the first evaporator at a boiling
temperature of 115°C., even though Mg(OH)2 was the principal precipitate.
However, in the subsequent efforts where boiling took place in the
heating tubes, scale was deposited. The minimum seed concentration
with elevated boiling zones was 5 Kg/m3 at 90°C., and 15 Kg/m3 at 115°C.
A 90 day run was completed free of scale.
Attempts to control calcium sulfate scale in a three-stage vertical tube
evaporator train at brine concentration factors of 4-5 were initially
less successful. Scale formed on the tubes and the interesting observa-
tion was made that while the scale had a definite anhydrite structure, the
slurry adopted an amorphous structure. The reason may be that the tempera-
ture cycling caused the crystals to partially convert from one hydrate
to another, but never to completion. When the slurry was circulated
separately in each unit at a constant temperature, prevention of scaling
was achieved for 15 days of operation. The following conditions were
used:
124 Arthur D Little. Inc
-------
D
BRINE HfATER
FLASH PILOT
OVERFLOW
MAKE-UP
r
i—x^±—iii: [>o—
1 AUXILIARY
HEATER
PRE-TPJEATING
SYSTEM
CLARIFIER
,"1
PRODUCT
BLOW DOWN
AGITATION
STEAM
I
\ SEEP STORAGE
SUPPLEMENTAL 8
RT UP SEEDS
-CXJ-
I SEEDS
L _ _
Source: Reference 45
Figure 42.Slurry Recycle Process Used in a Pilot Flash Distillation Plant
125
Arthur D Little Inc
-------
Boiling Temp. Slurry Concentra-
Stage °C. tion Ke/m3 Seed Composition
1 108-112 140-200 CaS04 + CaS04. 0.5 H20 + CaC03
2 65-70 30-40 CaC03
3 40-45 20-25 Ca(X>3
A vertical tube multiple atage distillation plant with elevated boiline
zone was constructed JU at Shevchenko in the USSR to make use of these
results: co-current flow was used - the first stage boiling temperature
being 90-95°C., the last 40°C. and the brine concentration factor being
4-4.5 (50-60 gram/liter of salt). These temperatures are low enough so
that calcium sulfate will not precipitate. The concentration of seed
crystals varied from 20-22 Kg/m3 in the first effect to 40-45 Kg/m3 in
the fourth effect (ground chalk). Circulation velocities varied from 0.9
to 2.5 ra/sec. Later (1965), the temperatures were raiaed-.to a range of
110°C. to 50°C. The plant operated from 1963 to 1967 'without deposi-
tion of scale. During operation the Mg(OH)2 content of the seed crystals
tended to increase to a limit of about 7.5 wt/% of Mg (Ca at 14.1 wt/%
and then stabilize.
Later, a seven stage horizontal flash plant was constructed 51 to further
study this method. Caspian seawater was used which, though less saline
than ocean seawater, contains similar amounts of scale forming materials.
Calcium ppm 400
Magnesium ppm 778.2
Ruthenium ppm 207.4
Sulfate ppm 3059.17
Seed crystals of common ground chalk were used, and heat transfer coef-
ficients were monitored with time. Although scale formation was de-
layed at temperatures of 125°C. and 115°C., a dense white scale was
eventually deposited with the composition
CaO 40.60
MgO 1.07
Ru03 20.16
Si02 1.12
126
Arthur D Little, Inc
-------
that is a calcium carbonate, calcium ruthenate mixture. At 105°C. only
small quantities of scale were formed and contained large propor-
tions of corrosion products. At 95°C. no scaling was detected.
Summary
1. Seeding with chalk, Mg(OH)2 and calcium sulfate has been
shown to prevent or reduce scaling from supersaturated solu-
tions under a variety of conditions.
2. Optimum concentrations of seed range from 0.5 to 5% by weight.
3. Several difficulties have been encountered which must be anti-
cipated in any new application of the technique, such as:
a. sludging and build up of seed particles in stagnant
zones - particularly with calcium sulfate hemihydrate
b. if temperature cycled, seeds of calcium sulfate may
lose their crystalline form and become less effective
c. the presence of calcium sulfite reduces the efficiency
of calcium sulfate seeds
d. effectiveness is a function of both seed concentration
and solution velocity
e. grinding may be necessary to keep recirculated seed
crystals small and active.
In practically all cases where successful inhibition of scaling
has been demonstrated by use of seeding, the system concerned had quite
different properties to that of wet scrubbing slurries. Temperatures
were frequently higher. No experience is reported for conditions in
which calcium sulfite is precipitated. The varied experience reported
by different workers for the same scaling system adds to the uncertainty
about the effectiveness of seeding in the wet scrubber system. Only
by experimentation will its usefulness be evaluated.
127
Arthur D Little; Inc
-------
Scale Control by Threshold Treatment
Other than water pretreatment to remove scaling agents and pH control,
the most common method for scale prevention used in contemporary desalina-
tion processes makes use of a poorly understood but fortuitous phenomenon
known as threshold treatment. Small quantities of certain polymeric elec-
trolytes inhibit scale growth when present in far less than stoichiometric
amounts. These substances are primarily polyphosphates, but other poly-
electrolytes have been shown to have the same effect.
Sodium hexametaphosphate (Graham's salt) was..first used to treat boiler
water. But here it was used as a sequestering agent to solubilize cal-
cium and magnesium salts which might otherwise form scale. As much as
14 parts of phosphate to 1 part of calcium was typically employed. Later
it was discovered that hexametaphosphate prevented deposits of calcium
carbonate in ammoniated irrigation water at a level of only 2 ppm, even
though the carbonate was present at 200 ppm, and was in fact no more effec-
tive at higher concentrations—hence the term "threshold treatment".
Since then, this method has been widely used in scale control in cooling
water systems, heat exchangers and boiler feed lines, often in proprietary
formulations. For example, a mixture of sodium tripolyphosphate, liquid
sulfonate and antifoaming agent was promoted*'for seawater evaporators.
Polyphosphates are degraded in aqueous solution by hydrolysis , the
high members trimetaphosphate and tripolyphosphate reverting to ortho-
phosphate. While at room temperature the rate is extremely slow, increas-
ing temperature and falling pH accelerates the decomposition. The times
for 5% hydrolysis at 50° C. are as follows:
salt time to 5% hydrolysis at 50° C.
hours
pH 4- 1 10
pyrophosphate 24 105 2000
triphosphate 9 60 300
long chain 60 900 900
Source: Reference 54.
Calcium Carbonate Deposition
The effect of the addition of polyphosphate on the precipitation of
calcium carbonate is shown in Figure 43.
128
Arthur D Little Inc
-------
01
£
I
.u
s
t—
a
-------
The threshold dose of 1 ppm indicated in Figure43has been found
to be dependent on the pH (which, of course, affects the bicarbonate
concentration in solution) but to be independent of the calcium concen-
tration. The presence of neutral salts tends to reduce the threshold
dose; potassium hydroxide improves the efficiency of the hemameta-
phosphate55 in suppressing precipitation despita the increased pH.
Effective materials include sodium hexametaphosphate, pyrophosphate, 55
tripolyphosphate56 and linear polymers. As is found in the case of the
polyacrylates described later, crystals formed in the presence of poly-
phosphates are distorted, but the crystals contain much less of the
additive. (Calcium to phosphate ratio about 320-350 to 1) than in
the case of the organic polymers (as much as 40% by weight(, q.v.).
Although hexametaphosphate addition has been found to greatly increase
the negative mobility of the cAlcium ion as though strong negatively
charged complexes were formed, and the amount of lanthanum or magnesium
ions needed to reverse this mobility is then much greater, the dose of
polyphosphate in threshold treatment is too low for such sequestering
mechanism to be tenable. It has been shown that surfaces treated with
hexametaphosphate retain their scale retarding properties when contacted
with supersaturated carbonate solutions not containing the phosphate.
This evidence combined with the observation of distorted crystals suggests
a surface ,absorption mechanism in which crystal nucleation and growth is
inhibited.
Calcium Sulfate Deposition -
Indications that polyphosphate could inhibit calcium sulfate dihydrate
deposition at room temperature was shown as early as 1954^58 although high
doses in greatly supersaturated solutions were needed. Strontium sulfate
supersaturation is also stabilized:^ a ten-fold increase in polyphos-
phate concentration reduced the nucleation rate by a factor of about 10 .
Magnesium Hydroxide Deposition -
Tr^nolyphosphate has been found much less effective on magnesium hydroxide
scales: a 2 ppm dose reduced calcium carbonate deposition by 99%, mag-
nesium hydroxide by only 40%. An increased dose was no better.
Other Inorganic Agents -
Many other inorganic materials have been tested^^ for use in threshold
treatment
tetraborates antimonates
metaborates arsenates
orthovanadates arsenites
meta vanadates potassium nitrate
5f> bismuthates
and:
nitrates metasilicates
iodates
but none showed a threshold scale prevention effect.
130 Arthur D Little: Inc
-------
Use in Seawater Evaporators -
Extensive study has resulted in the successful use of polyphosphates for
scale control in seawater evaporators. 'Hagevap' is a proprietory and
commonly used polyphosphate formulation which has been particularly
successful in the flash evaporators in Kuwait56 where it has been used
since 1955. The following conditions were ^ound to be satisfactory:
'Hagevap' concentration 4 ppm in feed water,
Water velocity greater than 1.7m/sec., in the condenser tubes,
Water temperature less than 91°C.,
No boiling in tubes and minimum tube surface temperature,
Maximum salinity 70,000 ppm with low iron content.
This success in Kuwait, permitted modifications to be made to another
plant in Ceuta, Spain to enable it also to run with Hagevap. Success-
ful use of Hagevap at a plant in Las Palmas, Spain was reported.^
Sludging using Hagevap has been controlled using Taprogge techniques
(circulating rubber balls).
At Ceuta, cost comparisons of various treatment chemicals, mostly based
on polyphosphates, were made yielding the following results: 61
Relative Cost
Product Name Chemical Type
Hagevap LP polyphosphate formulation
Hagevap PD-8 polyphosphate formulation
Evap. Treatment No. 5 polyphosphate formulation
Evap. Treatment No. 10 polyphosphate formulation
Darex 40* polyelectrolytes
AC-1 polyelectrolytes
of Water
1.00
1.00
1.19
1.06
1.05
promising
*was not tested at maximum permissible temperature.
The United States Government Office of Saline Water (OSW) has also found
polyphosphate treatment to be an effective method for alkaline scale
control. One of the most comprehensive early programs which found Hagevap
treatment successful was at the Point Loma flash evaporation plant in San Diego.
Two thousand five hundred hours of operation was achieved at 200° F.
without the need for cleaning. The polyphosphate was used at a level of
4 ppm; the calcium concentration being 400 ppm and the magnesium 1200 ppm.
Iron, which is known to interfere with the effectiveness of polyphosphates,
was always less than 0.1 ppra. Polyphosphate treatment was also found
to be successful at Freeport, Texas (using a dose of 4-6 ppm) and partially
at Guantanamo Bay, Cuba, although sludging was experienced—possibly due
to the presence of iron.
Thus, Hagevap polyphosphate treatment is effective and cheaper than poly-
electrolytes, but its high temperature instability limits its usefulness
in high temperature distillation plants.
131
Arthur D Little Inc
-------
In addition,the presence of the polymer makes the crystals much smaller
and distorted. The mechanism proposed is that the nucleation sites are
masked by absorption of the polymer. Indeed, scales formed contained as
much as 40% of the organic material, which indicates that in precipita-
tion conditions such as in scrubber operation considerable attenuation of
the polymer would be expected with assoicated economic penalties.
Dilatometry has also been used to follow the crystal growth of calcium
sulfate.- Once again induction periods were shown to be lengthened
and growth rates reduced by the addition of polyelectrolytes—0.87 ppm
of a copolymer of methylvinyl ether and maleic acid increased the induc-
tion period from 15 to 95 minutes, for example. The following generali-
zations were made as a result of this work:
1. Polymers containing carboxyl groups were most effective,
particularly if the groups were neighboring,
2. The higher the concentration the greater the effect.
3. A molecular weight in the region of 1600 was optimum.
4. With carboxyl groups the effectiveness increased with pH,
roughly following the degree of association of the carboxyl
groups.
Reports that acrylic based polymers yielded self-cleaning scales '
s'e found » to be probably dependent on the synergistic presence of
small amounts (M).2 ppm) of aluminum, magnesium or zinc. In addition,
difficulties in obtaining reproducible scales in experimental work was
found to be due to the occasional presence of a derivative of polyethylene
oxide, leached from the polythene storage bag. The authors noted that
it appeared to be an effective scale prevention agent. Storage in high
density polythene avoided this difficulty.
Scales formed in the presence of both polyelectrolytes and the metal ions
above were found to strip easily from the copper surfaces of a laboratory
spray evaporator on exposure to air. Once a'gain, the lower molecular
weight materials in the range 1000-1600 were most effective. This work
indicates how sensitive scale formation is to precise operating conditions
particularly with respect to trace impurities in the system. This could
explain different experiences in scrubber operation and in experimental
studies in scales.
In Kuwait also, following the success of Hagevap treatments, attempts
were made to find higher temperature polymers to permit operation at
200° F. and above. Both "beaker" and plant tests were carried out to
study suppression of alkaline scale. The following results were obtained:
132
Arthur D Little, Inc.
-------
Studies of Polyorganic Compounds;-
The success of polyphosphate treatment at low temperatures led to consider-
able efforts to find other materials with the same activity, particularly
with effects on calcium sulfate scales and with improved stability at high
temperatures. A range of acrylic acid homopolymers and copolymers was syn-
thesized and tested for alkaline scale suppression in a pilot plant evapo-
rator at around 3 ppm. A screening test was developed in which saturated
carbonate, hydroxide and sulfate solutions were boiled down in beakers in
the presence of various additives and the wuantity of scale produced mea-
sured. Promising results were obtained with polymers below a molecular
weight of 30,000. In particular one polymer composed of acrylic acid, ethyl
acrylate, and isopropyl acrylate in the ratio by weight of 61:10:29 with a
molecular weight of 19,000 showed particular promise for inhibiting calcium
sulfate scale, reducing the amount deposited in a typical test from 42 mg
o
°Q
x
UJ
o
o
o
I
o
u.
o
UJ
0_
Concentration in ppm
12
16 20
TIME (minutes)
28
Source: Reference 66
Figure44 Effect of Polymer Concentration on the Unhibition of Gypsum Crystallization
to 2.4 mg at a concentration of only 7 ppm. In the pilot plant tests, it
showed excellent scale prevention in respect to Mg(OH)2. Sulfate scale
deposition showed a minimum at an acrylic acid composition of about 60%
in the range of copolymer formulations evaluated.
Using,a polymer of this type containing 66% acrylic acid, it was shown fur-
ther by using conductance measurements to follow ionic concentration, that
133
Arthur D Little Inc
-------
the mechanism of scale suppression was an interference with crystallization
kinetics which results in modified crystal habit. During the boiling down
of calcium sulfate solutions in the beaker tests, specific conductance in-
creased as the concentration rose, but then fell rapidly as the scaling
rate exceeded the rate of concentration of the solution. Figure 44 shows
the effect of concentration at 100°C. Four ppm was sufficient to extend
the period for initiation of significant precipitation from 10 minutes to
£5 minutes.
Interestingly the second-order dependence of scaling rate with supersatu-
ration
(CT " CEquilib) = K f
(where CT is the concentration at time T, C . is the thermodynamic
equilibrium solubility and K is constant) e1ulllb is maintained but the
presence of the polymer introduces an induction period, and at 3 and 4 ppm
radically decreases the rate of crystallization. This effect is dramati-
cally illustrated in Figure 45.
it
o
ft)
~ 800
o
600
O
55
or
V)
cc
LJ
Q.
u.
o
o
o
o:
a
a
QL
400
200
Annotation denotes rate
max error
15%
40.4
to
18 22
TIME (minutes)
26
30
Source: Reference 66
Figure 45. Reciprocal of Supersaturation as a Function of Time
for Gypsum Solutions Crystallizing at 100°C in the
Presence of Polymer.
Top line:
2nd line:
0 ppm treatment
1 ppm treatment
3rd line: 3 ppm treatment
Bottom line: 4 ppm treatment
134
Arthur D Little Inc
-------
Compound
Darex 40
(W. R. Grace)
Composition
Test
Calnox 214
(Milchem)
Calnox 214
Vaptreat H
(Houseman and
Thompson)
Sodium polyetha- Plant
crylate (25%
solids)
Dosage
Temperature ppm Result
200° F
Copolymer of
acrylic acid
(33% solids)
Plus sulfuric
acid
Includes poly-
acrylic acid
(40% solids)
Beaker 212° F
Plant 200° and
225° F
Beaker
Plant 200° F
Beaker 212° F
3-9 Initial cleaning
action.
0-25 Maximum performance
at 5 ppm but Mg(OH)
not controlled.
1-18 Worked well at the
low temperature.
1-25 Optimum dose at
2 ppm.
0.75-1.2 Functioned when
greater than 1 ppm.
0-25 Optimum dose at
4-5 ppm.
Thus all the materials behaved well in beaker tests, and in plants at
200° F. where tested. The "threshold-like" dependence of scale preven-
tion on dose is typified in Figure 46 for beaker tests on Darex 40. All
the materials tested showed a similar strong reduction in scale deposi-
tion as a function of concentration reaching a maximum effectiveness at
4-5 ppm.
72
Another laboratory study compared seventeen polymers for the capability
in preventing the precipitation of alkaline scales and calcium sulfate
from saline and water. At a dose of 20 ppm, the most effective were the
anionic low molecular weight materials based on acrylic acid and its co-
polymers. Even in sodium chloride, the polymers were able to hold more
calcium sulfate in solution. For example, with polyacrylic acid (Molecular
weight = 5,000) at 22 ppm , the following precipitation points for calcium
sulfate were obtained.
water
3% NaCl
no additive
1,800
6,300
with additive
4,800
11,900
A sharp maximum of effectiveness was found at a pH of 10.3.
135
Arthur D Little. Inc
-------
Evaporation at 100°C (212°F)
to average concentration 2.0
Sea water C1 = 20,250 -
P.AIk. = 12 -
T.AIk. =
136 -
23,000 ppm
24 ppm CaCO,
160ppmCaCO,
Cross-hatched region encloses all data
points from batches of sea water of varying
alkalinitiesand turbidities
10 20
Darex 40 Dosage - ppm solids
30
FIGURE 4$. ANTISCALING ACTION AS A FUNCTION OF CONCENTRATION OF A POLYETHACRYLATE
(DAREX 40)
136
-------
Summary;
1. Polyphosphates have shown remarkable ability to control alkaline
scale at concentrations of 2-5 ppm up to temperatures of 190°F.,
although less effective for Mg(OH)2 than for CaCO..
2. Ample evidence exists to show that they will control calcium
sulfate scale, but this is of little significance in seawater
distillation technology since such scales only occur at temper-
atures above the stability limit of the polyphosphates. This
could of course be useful in scrubber operation.
3. The mechanism of action is thought to be interference with crystal
nucleation and growth by absorption at active sites. Nucleation
times are greatly extended and rate of growth may be reduced.
4. By synergism with small concentrations of certain metal ions,
self-cleaning (loosely adherent and brittle) scales may be formed.
5. Other polymeric anionic materials, particularly the low molecular
weight acrylic homopolymers and copolymers show similar activity,
and better temperature resistance. The crystal deposits have a
large organic content, however, and considerable attrition of material.
must be anticipated.
6. Their use in other applications, such as wet scrubbing, may involve
several difficulties. Traces of other materials may effect the
activity. They are chemically unstable particularly at lower pH.
Whether they can inhibit precipitation of one species (e.g., calcium
sulfate) while precipitation of another (e.g., calcium sulfite)
is occurring is doubtful. In systems where large quantities of
material are precipitating, attrition of the additive can be
expected.
7. Doubts about the mechanism of inhibition and the differences between
wet scrubbing slurries and other systems where their effectiveness
has been demonstrated points up the necessity of an experimental
evaluation of their usefulness in the scrubber system.
137
Arthur D Little. Inc
-------
APPENDIX I
METHOD OF CALCULATING IONIC EQUILIBRIA
Our method of calculation is to set up the mass law equilibrium equations
for all ionic, solid, and gaseous species. Equilibrium constants are cal-
culated for the temperature under consideration using literature data as
summarized in the Radian report. Preliminary initial values of the acti-
vity coefficients are assumed as follows:
ionic charge activity coefficient
zero 1.0
one 0.8
two 0.4
The electrical neutrality equation, which equates concentrations of posi-
tive and negative species, is then solved to give an approximate value
of the calcium activity. The activities and concentrations of all other
species are then computed, and fresh values of the activity coefficients
are calculated from the ionic strength by using the modified Debye-Huckel
expression. This process is repeated until the calculation cbnverges.
The electrical neutrality equations solved for each of the five systems
in Table III are quadratics of the form:
2
Aa + Ba + C = 0
ca ca
where a is the activity of calcium in solution. Values of the expres-
sions A, B, and C for each of the four systems are shown in Figure 1A.
The various symbols used in Figure iA are explained in Table IA.
The program logic and card decks may be obtained from Arthur D. Little,
Inc.
138
Arthur D Little. Inc
-------
In the form:
Ax + B + C = 0 where x is calcium ion activity.
x
Case
1.
2.
3.
A.
A
1 , Kll
f 9 ' P V
*13 "11 10*10
1 Kll
+
F13 2A11F10K10
1 Kll
— ± — + — ^— — ^— —
F13 2A11F10K10
1 Kll
H
F13 2A11F10K10
-B
Kll All
2A11F12 2F11
Kll K2A11 All
2A11F12 2W9 2F11
Kll All A11K2 . „ , A11SF14
2A11F12 2F11 2W9 " 2K15F15
Kll All A11K2
2A11F12 2F11 2KAK9F9
K
F
-C
Kl . A11K1
Fl 2K3F3
Kl K2 A11K1 AHK2"
Fl F2 2K3F3 2KAFA
Kl . K2 . A11K1 , A11K2
Fl F2 2K3F3 2KAFA
1 K2 AllKl + KU^-SlK14 >A2*lll
+ — — + + », • -- r ' ?K F
i + F2+2K3F3 F14 --'-15F15 'ZV4
Figure 1A. Quadratic Equations Solved in Equilibrium Calculations,
c
C"
-------
Table IA. List of Symbols
Equilibrium Constants*
KX CaS03-l/2 H20(s)
K2 CaS03(s)
S HS°3~
K, HCO ~
Activity Coefficients
F! S03=
F2 C03=
F3 HS03"
F HC03-
Other Symbols
A^^ activity H+
A13 activity Ca"4
S dissolved S(>4=
concentration
6
K7
K8
K9
K
K,,
2 3
CaS03°
CaC03°
CaKC03+
CaOff1"
H00
"14
K
3.5
S02(g) + H20(I)
H2S03(soln)
C02(g) + H20(I)
H2C03(soln)
CaS04-2 H20 (s)
HSO ~
Q
13
14
15
H2C°3
CaS0
CaC03
CaHC0
CaOff1"
Ca
S°4
HSO,
y products of solid species; dissociation constants of neutral
or ionic species
140
Arthur D Little. Inc
-------
REFERENCES
1. Nonhebel, G., Balten, W. E., Sayles, H. S., and Lewis, J. G., "Scrubber
for the Removal of Dust and Sulfur from Boiler Flue Gases," Ninth
Interim Report, No. 111,556, ICI, Ltd., 4 March 1935.
2. Lessing, R., J. of the Soc. of Chem. Ind., 'Transactions and Communi-
cations," November, 1938, pp. 373-388.
3. Gleason, R. J., and McKenna, J. D., "Scrubbing of Sulfur Dioxide from
a Power Plant Flue Gas," Cottrell Environmental Systems, 69th
National Meeting of AIChE, Cincinnati, Ohio, May, 1971.
4. Gleason, R. J., "Limestone Scrubbing Efficiency of Sulfur Dioxide in
a Wetted Film Packed Tower in Series with a Venturi Scrubber,"
Cottrell Environmental Systems, Second International Limestone
Wet Scrubbing Symposium, New Orleans, Louisiana, Nov. 8-12, 1971.
5. Saleen, A., Harrison, D., and Sekhar, N., "Sulfur Dioxide Removal by
Limestone Slurry in a Spray Tower," The Hydro-Electric Power
Commission of Canada, Toronto, Ontario, Second International Lime-
stone Wet Scrubbing Symposium, New Orleans, Nov. 1971.
6. Kelso, T. M., Williamson, P. C., Schultz, J. J., "Removal of Sulfur
Dioxide from Stack Gases by Scrubbing with Limestone Slurry,"
TVA Pilot Plant Tests, Second International Limestone Set
Scrubbing Symposium, New Orleans, Louisiana, November, 1971.
7. Craig, J. M., Bell, B., and Fayadh, J. M., "Mobile Pilot Plant Study
of the Wet Limestone Process for S02 Control," Zurn Environmental
Engineers, Second International Limestone Wet Scrubbing Symposium,
New Orleans, Louisiana, November, 1971.
8. McKenna, J. D., and Atkins, R. S., "The R C/Bahco System for Removal
of Sulfur Oxides and Fly Ash from Flue Gases," Cottrell Environ-
mental Systems, Second International Limestone Wet Scrubbing
Symposium, New Orleans, Louisiana, November 8-12, 1971.
9. Potts, J. M., Slack, A. V., and Hatfield, J. D., "Removal of Sulfur
Dioxide from Stack Gases by Scrubbing with Limestone Slurry:
Small Scale Studies at TVA," Second International Limestone Wet
Scrubbing Symposium, New Orleans, Louisiana, November, 1971.
10. Uno, T., Atsukawa, M., and Muramatsu, K., "The Pilot Scale R & D
and Prototype Plant of MHI Lime-Sypsum Process," Mitsubishi
Heavy Industries," Second International Limestone Wer Scrubbing
Symposium, New Orleans, Louisiana, November, 1971.
141
Arthur 1) Link-1 iv
-------
11. Hausbery, G., "The Bischoff Process" - Initial results from a full-
size experimental plant, Second International Limestone Wet
Scrubbing Symposium, New Orleans, Louisiana, November, 1971.
12. McCarthy, J. H., and Roosen, J. J., Detroit Edison Pilot Plant and
Full Scale Development Program for Alkali Scrubbing Systems -
A Progress Report, Detroit Edison, Second International Lime-
stone Wet Scrubbing Symposium, New Orleans, Louisiana, Nov. 1971.
13. Shapiro, J. L., and .Kuo, W. L., "The Mohave/Navaho Pilot Facility for
Sulfur Dioxide Removal," Second International Limestone Wet
Scrubbing Symposium, New Orleans, Louisiana, November, 1971.
14. Person, R. A., Allenback, C. R., Shah, I. S., Sawyer, S. J., A Pilot
Plant Test Program for Sulfur Dioxide Removal from Boiler Flue
Gases Using Limestone and Hydrated Lime, Second International
Limestone Wet Scrubbing Symposium, New Orleans, Louisiana,
November, 1971.
15. Phillip Lowell, Theoretical Description of the Limestone Injection
Wet Scrubbing Process, Vol. I, PB 193029, 9 June 1970.
16. Danckwerts, P. V., "Gas-Liquid Reactions," McGraw-Hill, N. Y. (1970)
17. Phillips, J. M., Radium Corporation, Austin, Texas, Project No. 200-
014 (1972)
18. Badger, W. C., Caldwell, H. B., "Studies in Evaporator Design-VI.
Evaporators for Salts with Inverted Solubility Curves." Trans.
Am. Inst. of Chem. Eng. 16 131 (1925).
19. Pearson, J. L., Nonhebel, G., and Uhlander, P. H. N., "The Removal of
Smoke and Acid Constituents from Flue Gases by a Non-effluent
Water Process." J. Inst. Fuel VIII (39) 119 (1935).
20. Pearson, J. L., Nonhebel, G. and Uhlander, P. H. N., "Flue Gas Cleaning
at Fulham Power Station." The Fuel Economist p. 87 (Nov. 1936)
and p. 108 (Dec. 1936) .
21. , "The Elimination of Sulphur from Flue Gases."
Engineering, p. 499, Oct. 28 (1938).
22. Nonhebel, G., and Pearson J. L., "Wet Purification of Gases," U.S. 2,090
142 (1937).
23. Stevens, C. F. B. and Birchard, W. H., Papertrade J. 103 (7) 78 (1936).
24. Scholander, A., "Scale Control by CaSO^ Seeding in Evaporation of Sulfite
Liquors." Svensk. Papperstid, _53_ 681 (1950).
142
Arthur D Little, Inc
-------
25. Hirzel, E., "The Economic Concentrations of Sulfite Waste Liquor."
Escher Wyss News 29 (1) 15 (1956).
26. Dunkak, J., "Control of Scaling in Calcium Sulfite Digester Heat
Exchanges." TAPPI 45 (4) 196A (1962).
27. Charuit, P., Marchand, R., Viard, M., Courvoisier, P. and Vignet, P.,
"A Study of the Prevention of Deposits of Calcium Sulfate by
Crystallization Correlation (sic) of the Experimental and Indus-
trial Results." 3rd Int'l Symp. on Fresh Water from the Sea
J. 405 (1970).
28. Fitzwilliam C. V. "A Review of the Evaporator Scaling Problem," Proc.
Brit. West Indies Sugar Technologists 13_ 55 (1947).
29. Kubelka, V., "Scale Control in Sugar Evaporators." Chemicke" Zvesti
VIII (6) 8_ 385 (1954).
30. Pat. No. 30367 (1920).
31. Claassen, In't. Sug. 7. 17 (1934).
32. Hamous, Chem. Abstr. 7688 (1935).
33. Hester, A. S., Diamond, H. W., "Salt Manufacture." Ind. Eng. Chem. 47
(4) 672 (1955). ~
34. Badger, W. C. and Standiford, F. C., "Engineering in Salt Manufacturing."
.62 (3) 173 (1955).
35. Bull, H. J. Scale Control in NaCl Plant. U.S. Patent 1,399,845.
36. Hartung, H. 0., "Scale Control in Boilers by CaCO- Solids'.1 J. Am.
Water Workers Assoc. 48 1523 (1957).
37. Langelier, W. F., Caldwell, D. H., and Lawrence, W. B., "Scale Control
in Sea Water Distillation Equipment: Contact Stabilization."
Ind. Eng. Chem. (Anal.) 42_ (1) 126 (1950).
38. Standiford, F. C. and Sinek, J. R., "Stop Scale in Seawater Evaporators."
Chem. Eng. Prog. _57 (1) 58 (1961).
39. Gainey, R. J., Thorp, C. A. and Cadwallader, E. A., "CaS04 Seeding
Prevents CaS04 Scale." Ind. & Eng. Chem. 55_ (3) 39 (1963).
40. Woodbury R. E. and Cadwallader, E. A., "Alkaline Scale Control by
Internal Contact Stabilization." Presented at Div. of Water,
Air and Waste Chemistry, Am. Chem. Soc. Detroit, Mich. (April 5,
1965).
143
Arthur I) Little. Inc.
-------
41. Tidball, R. A. and Woodbury, R. E., "Methods for Scale Control in
Flash Systems". Proc. 1st Int'l. Symp. on Water Desalin. 2^ 317,
Pub. by U.S. Dept. of Interior, Washington, B.C. (1965). ~
42. Elliot, M. N., "The Present State of Scale Control in Seawater Evaporators "
Desalination £ 87 (1969).
43. Simpson, H. C., and Hutchinson, M., "Calcium Sulfate Scale Deposition
in Seawater Evaporators." Desalination 2_ 308 (1967).
44. Matsuda, T., Akimoto, S., and Taniguchi, Y., "Scale Control in Seawater
Distillation." Proc. 1st Int'l. Symp. on Water Desalin. 2^ 25,
Pub. by U.S. Dept. of Interior, Washington, D.C. October 3, (1963).
45. Mulford, S. F., "Scale Control in Seawater Evaporators." OSW Research
and Development Progress Report No. 133, Nov. 1964.
46. Jebens, R. H. and Dykstra, D. I., "Advanced in Seawater Distillation."
Chem. Eng. Prog. 61 (8) 68 (1965).
47. McCutchan, J. W. et al., "Scale Control in Saline Water Evaporators -
A Review of Current Status." OSW Research and Development Progress
Report No. 411 (April 1969).
48. OSW Sal. Wat. Conv. Report (1969-1970) p. 388 pub. of U.S. Dept. of
Int. Washington, D.C.
49. Chernozubov, V. B., Zaostrovsky, F. P., Shatsillo, V. G. Golub, S. I.,
Novilov, E. P. and Tkach, V. I., "Prevention of Scale Formation
in Disillation Desalination Plants by Means of Seeding." Desalin-
ation ± 50 (1966), and Proc. 1st Int'l, Symp. on Water Desalination
2^ 539 (Oct. 3, 1969 Pub. by U.S. Department of Interior, Washington,
D.C.).
50. Zaostrovsky, F. P., Novikov, E. P., Shatsillo, V. G., Golub, S. I.,
Chernozubov, V. B. and Tkach, V. I., "Distillation Desalination
Plant in the City of Shevchenko. Layout, Equipment and Operating
Experience." Desalination 1 165 (1966).
51. Chernozubov, V. B., Novikov, E. P., Kartovski, Y. V., Golub, S. I.
and Shatsillo, V. G., Test of the Desalting Flash Pilot Plant
with Recirculating Seed. Desalination 5_ 34 (1968).
52. Rosenstein L., U.S. Patent 2,038,316 (1936).
53. Liddell, R. W., U.S. Patent 2,782,162 (1957).
54. Elliot, M. H., "Scale Control by Threshold Treatment", Desalination
8 221 (1970)
144
Arthur D Little, Inc
-------
55. Reitemeier, R. F. and Buehner, T. F., J. Phys. Chem. 44_ 535 (1940).
56. Raistrick, B., Discuss. Farad. Soc. .5 234 (1949).
57. Hatch, G. B. and Rice 0., Ind. Eng. Chem. H. 51 (1959).
58. Burick, E. J., Producer's Monthly 19_ 42 (1954).
59. Miura, M. and Naono, H., Bull. Chem. Soc. Japan 38_ 80 and 492 (1965).
60. Ali El-Saie, M. H., Secret of Success Using Hagevap (polyphosphate)
Treatment and Deaeration of Seawater Feed in Flash Distillation Plant.
3rd International Symposium on Fresh Water from the Sea 1 405 (1970).
61. Troyano, F., Mir, P., Oltra, F., and Ortega, J. M., "Operational
Experience with Autiscale Products in the Ceuta Desalination Plant
(4,000 m3/day Polyphosphate Treatment)," 3rd International Symposium
on Fresh Water from the Sea _1 537 (1970) .
62. Ali-Saie, M. H. and Suarez, J., "Las Palmas 20,000 m3/day Flash
Distillation Duel Water and Power Plant", International Atomic
Energy Author., Symposium on Nuclear Desalination, Madrid p.
SM-113/56 (Nov. 1968).
63. Mulford, S., Glater, J., and McCutchan, J. W., "An Appraisal of Scale
Control Methods at the OSW Flack Evaporator Plant in San Diego,
California." Proc. 1st Inc. Symp. on Water Desalination, U.S.
Dept. of Int. Wash. D.C. 2_ 139 (1965).
64. Soloman, D. H. and Rolfe, P. F., "Polymers that Inhibit the Deposition
of Calcium Sulfate," Dssalination .1 260 (1966).
65. Herbert, L. S., Rolfe, P. F. and Sterns, U. J., "Saline Water Dis-
tillation-Scale Prevention by Polymer Additives", Proc. 1st Int.
Symp. on Water Desalination, October 3, 1965, p. 39 Washington, D.C.
66. Rolfe, P. F., Polymers that Inhibit the Deposition of Calcium Sulfate:
Some Interesting Conductance Observations, Desalination 1 359 (1966).
67. McCartney, E. R. and Alexander, A. E., "The Effect of Additives upon
the Process of Crystallization. I, Crystallization of Calcium
Sulfate", J. Colloid Sci. JL3_ 383 (1958).
68. Smith, B. R., Scale Prevention by Addition of Polyelectrolytes.
Desalination ^ 263 (1967).
69. Herbert, L. S. and Sterns, U. J., Saline Water Converstion II. Am.
Chem. Soc. Series No. 38, 52 (1963).
145
Arthur D Little. Inc
-------
70. Sweett, F., Smith, B. R., Caramento, P., and Taylor, R. N., Scale
Control by Polyacoylates: The Effects of Molecular Weight and
of Traces of Metals. Desalination j[ 167 (1970).
71. Khalaf, K. A. and Watson, B. M., "Recent Kuwait Developments in
Elevated Temperature Scale Control", 34d International Symposium
on Fresh Water from the Sea 1^ 513 (1970).
72. Flesher, P., Streatfield, E. L., Pearce, A. S. and Hydes, 0. D.,
"Polymeric Additives as Scale Preventatives", 3rd International
Symposium on Fresh Water from the Sea 1^ 493 (1970).
146
Arthur D Little, Inc.
------- |