U.S. Environmental Protection Agency Industrial Environmental Research EPA~600/7-77- 00!
Office of Rus'Kirjh and Development Laboratory .-.__
Research Triangle Park. North Carolina 27711 January 1977
EVALUATION OF THE
GENERAL MOTORS' DOUBLE
ALKALI SO2 CONTROL SYSTEM
Interagency
Energy-Environment
Research and Development
Program Report
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S.
Environmental Protection Agency, have been grouped into seven series.
These seven broad categories were established to facilitate further
development and application of environmental technology. Elimination.
of traditional grouping was consciously planned to foster technology
transfer and a maximum interface in related fields. The seven series
are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy-Environment Research and Development
This report has been assigned to the INTERAGENCY ENERGY-ENVIRON>ENT
RESEARCH AND DEVELOPMENT series. Reports in this series result from
the effort funded under the 17-agehcy Federal Energy/Environment
Research and Development Program. These studies relate to EPA's
mission to protect the public health and welfare from adverse effects
of pollutants associated with energy systems. The goal of the Program
is to assure the rapid development of domestic energy supplies in an
environmentally—compatible manner by providing the necessary
environmental data and control technology. Investigations include
analyses of the transport of energy-related pollutants and their health
and ecological effects; assessments of, and development of, control
technologies for energy systems; and integrated assessments of a wide
range of energy-related environmental issues.
REVIEW NOTICE
This report has been reviewed by the participating Federal
Agencies, and approved for publication. Approval does not
signify that the contents necessarily reflect the views and
policies of the Government, nor does mention of trade names
or commercial products constitute endorsement or recommen-
dation for use.
This document is available to the public through the National Technical
Information Service, Springfield, Virginia 22161.
-------
EPA-600/7-77-005
January 1977
EVALUATION OF THE
GENERAL MOTORS' DOUBLE ALKALI
SO2 CONTROL SYSTEM
by
Edward Inter ess
Arthur D. Little, Inc.
20 Acorn Park
Cambridge, Massachusetts 02140
Contract No. 68-02-1332, Task 3
Program Element No. EHE624
EPA Task Officer: Norman Kaplan
Industrial Environmental Research Laboratory
Office of Energy, Minerals, and Industry
Research Triangle Park, NC 27711
Prepared for
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Research and Development
Washington, DC 20460
-------
TABLE OF CONTENTS
Page
I. SUMMARY 1
A. Results and Conclusions 1
1. SO- Removal 1
2. Filter Cake Properties 1
3. Solids Dewatering 3
4. Lime Stoichiometry 3
5. Carbonate Softening 3
6. Soda Ash Stoichiometry 3
7. Scaling in Scrubber Loop A
8. Oxidation 4
9. Entrainment 5
10. Reliability and General Status 5
C. Recommendations 5
1. System Operation 5
2. Double Alkali Process Technology 7
II. INTRODUCTION 9
A. Background 9
B. Test Program 11
1. Plan and Execution H
2. Data Acquisition . 12
3. Analytical Procedures 12
III. SYSTEM OPERATING HISTORY 15
IV. CHARACTERIZATION OF SYSTEM OPERATION 25
A. S02 Removal 27
B. Filter Cake Properties 28
C. Solids Dewatering Properties 28
D. Lime Stoichiometry 29
E. Carbonate Softening 30
F. Soda Ash Stoichiometry 30
iii
-------
TABLE OF CONTENTS (Cont'd)
Page
G. Scaling in Scrubber Loop 31
H. Oxidation 33
I. Entrainment 34
APPENDICES
A - SUMMARY OF SCRUBBER AND BOILER OPERATION 35
B - MATERIAL BALANCES 47
C - CHEMICAL ANALYSIS PROCEDURES FOR DOUBLE ALKALI PROCESS
STREAM SAMPLES 59
D - SYSTEM ECONOMICS 89
iv
-------
LIST OF TABLES
Table
1 Key Operating Parameters and Results 2
2 Projected Operating Parameters 6
3 Summary of Operating Parameters ..... 24
4 Mean Filter Cake Compositions and Indicated Operating
Parameters 26
5 Sodium Balance Data - Softening Vs. Solubles Loss 32
B-l GM Parma Plant - One-Month Overall Balance—Intensive
Test Period No. 1 49
B-2 GM Parma Plant - One-Month Overall Balance—Intensive
Test Period No. 2 50
B-3 GM Parma Plant - Long Term Overall Balance—Intensive
Test Period No. 3 52
B-4 Summary of Short Term Balances - August 27-30, 1974 .... 54
B-5 Summary of Short Term Balances - February 26-28, 1975 ... 55
B-6 Summary of Short Term Balances - April 26-30, 1976 57
B-7 Summary of Short Term Balances - May 3-7, 1976 58
D-l System Economics 91
-------
LIST OF FIGURES
Figure Page
1 Flow Diagram - General Motors Double Alkali System 10
2 Data Collection Points 13
3 Scrubber Availability 17
4 Operating Flow Diagram - 1st Intensive Test 20
5 Operating Flow Diagram - 2nd Intensive Test 21
6 Operating Flow Diagram - Non-Intensive Test 22
7 Operating Flow Diagram - 3rd Intensive Test 23
B-l Definition of Terms - GM System 53
C-l Apparatus for Sulfate Determination 69
C-2 Apparatus for Carbonate in Solids 82
vi
-------
I. SUMMARY
The purpose of this test program was to obtain data to characterize the
GM system from the standpoint of SOo removal, process reliability, sul-
fate control, waste characteristics, degree of closed loop operation
(i.e., sodium makeup versus losses), lime and other chemicals utilization,
and material balances on the overall system and the important component
parts. Both GM and ADL had responsibilities for data acquisition, while
ADL, under contracts to GM and EPA, performed the chemical analyses. The
program consisted of three one-month intensive tests and 18 months of
lower level non-intensive testing, including six months of formalized
process data acquisition.
During the test program which covered the period August 19, 1974, to May
14, 1976, the General Motors Double Alkali system at Chevrolet-Parma, Ohio,
showed significant improvement in both process and mechanical performance.
A number of operating modes were investigated by GM in accomplishing this
improvement. The mode of the final intensive test, April 17, 1976, to
May 14, 1976, showed the greatest promise for a viable long term operation.
A. Results and Conclusions
The important results of the test program are summarized in Table 1. The
key results excluding SC^ removal are based on the average cake composi-
tions over each entire run and therefore are the best indication of the
actual performance for the extended periods.
1. SCs Removal
Despite the low S02 concentration in the flue gas of 800-1300 ppm (vol.)
caused by the high excess air used by the boilers, the S02 removals have
been excellent, in the range of 90%. This is well above the average
statutory requirement of about 60-70% for the coals burned (1.5-3.3% S,
11,000-13,500 Btu/lb). In one experimental operating mode, which has now
been abandoned, the average removals fell to the 65% range.
2. Filter Cake Properties
The filter cake properties as measured by percent total solids (56% solids)
and solubles content (0.03 mole Na2/mole SOX in cake) improved dramatically
in recent operation. Waste properties wera better than target levels for
dryness (minimum of 50% solids) and for cake wash efficiency (solubles
recovery equivalent to a maximum loss of 0.05 mole Na2/mole SOX in the
cake). The drier solids have resulted in a lower sludge hauling require-
ment while the decreased solubles loss improves the environmental accept-
ability of the cake for disposal; It should be noted that although cake
washing to a low solubles loss rate has been demonstrated, this rate does
not correspond to the expected long term soda ash stoichiometry of 0.05
mole Na2/mole SOX. At steady state, the sodium content of the cake must
match the carbonate (as soda ash) used for softening the regenerated liquor,
as further discussed in Section 6. Soda Ash Stoichiometry.
-------
Operating Period
S02 Removal
Recycle pH
Scrubber Feed Location
Lime Stoichiometry mole Ca/mole S in cake
Soda Ash Stoichiometry mole Na9/mole SO removed
£• X
Solubles Loss mole Na_/mole S in cake
Oxidation mole CaSO,/mole CaSO in cake
" X
Ca-H- in Scrubber, ppm
Scaling?
Na+ = Total, M in Scrubber Feed
Na+ - Active, M in Scrubber Feed
Cake Washing?
Total Solids, wt.%
Soluble Solids, wt.% of Wet Solids
Insoluble Solids, wt.% of Wet Solids
Soluble Solids/Total Solids
Table 1
KEY OPERATING PARAMETERS AND RESULTS
8/19/74-9/13/74.
90-95% (rpt. by GM)
5.5-6
Top Tray
1.90
0.11
60%
305
Yes(CaC03)
0.58
0.087
No
47.1
2.4
44.7
0.05
2/17/75-3/14/75
65%
7-7.5
Recycle Tank
1.65
0.08
0.12
47%
465
Yes(CaSO-)
J
0.66
0.13
No
41.4
2.4
39.0
0.06
4/19/76-5/14/76
90%
6.0
Recycle Line
1.32
0.12
0.03
83%
490
No
0.96(1>
0.12
Yes
56.0
1.0
55.0
0.02
(1) Includes
Na+ as NaCl
(2) No data taken on S0~ removal
-------
3. Solids Dewatering
The ease of solids dewatering (settling and filtration) also improved
dramatically as mirrored by the improvement in cake dryness discussed
above. Elimination of underflow recycle to the clarifier center wells
and the reactor system lowered the gross load on the clarifier and solved
the temporary problems of solids overflow and poor solids concentration.
Increased total oxidation (from 50-60% to 83%) by air sparging in the
scrubber bleed lines was significant in improving the solids filterability
as measured by dryness.
4. Lime Stoichiometry
The lime Stoichiometry improved consistently over the test program from
1.90 to 1.32 mole Ca/mole SOX in the final test period. This improve-
ment reflects: better control of lime feed; the use of hydroxide concen-
tration to monitor and control excess lime fed; and the increased con-
centration of sodium ion in the liquor which permits hydroxide solubility
to reach the target level of about 0.1M. Although the lime usage is still
somewhat high, we expect that with continued close control, the Stoichio-
metry can be appreciably further reduced. The minimum level will be
limited only by reactor residence time which is about one hour when two
scrubbers are operated.
5. Carbonate Softening
Two methods were used for softening of the regenerated liquor: sparging
of gaseous carbon dioxide into the liquor; and addition of soda ash feed
in the center well of the softening clarifier. Both methods are effective
in reducing the soluble calcium from saturation to levels of concentra-
tion below 400 ppm. However, carbon dioxide forms an acid in solution
neutralizing two moles of hydroxide per mole of calcium precipitated and
making the job of 862 removal more difficult in the scrubber. Further,
lime Stoichiometry must increase in order to provide the additional
hydroxide. Other drawbacks of G02 softening are the relatively poor-
settling calcium carbonate it produces and the filter cloth blinding by
the finely divided CaCO^. Therefore, soda ash softening, to the extent
that long term sodium balance can be achieved, is the only viable approach.
6. Soda Ash Stoichiometry
Soda ash is fed to the system for two purposes: (1) to provide sodium
ion to replace that lost in the cake and (2) to provide carbonate ion to
precipitate soluble calcium from the regenerated liquor. With adequate
cake washing, the softening requirements set the minimum soda ash addi-
tion rate.
During the early months of operation when the filter cake was not washed,
the sodium loss was above 0.1 mole Na2/mole SOX in the cake. After cake
washing was initiated, the loss was reduced to as low as 0.028 mole
mole SO for an extended one-month average.
A
-------
At this low sodium loss rate soda ash must be added to the system in the
same proportion of SOo removal, i.e., 0.028 mole Na2C03/mole 862 removed.
If more were added, there would be an accumulation of Na2SOx in the sys-
tem over time.
This quantity of carbonate addition is sufficient only to reduce the
soluble calcium in the regenerated liquor about 85 ppm from the saturation
level of about 800 ppm. Without having attempted a demonstration at such
a low softening level, General Motors estimates that a soluble calcium
reduction of 150 ppm (to 650 ppm total) can be a viable operational
approach. This corresponds to a soda ash makeup rate of 0.05 mole/mole
862 removed which would subsequently appear in the cake as a soluble
sodium loss of 0.05 mole Na~/mole SO in the cake.
2 x
7. Scaling in Scrubber Loop
There have been two sets of circumstances leading to solids deposition in
the scrubber loop. The first was carbonate scaling on the top tray due
to poor liquid mixing on the tray with resulting localized high pH and
carbonate scrubbing. This problem has been eliminated by effective mix-
ing of the high pH regenerated liquor with the more acid recycle liquor
before entering the top tray.
The second was calcium sulfite plugging which resulted from a combination
of factors in an experimental operating mode, since abandoned.
Throughout the test program, including the above situations of solids pre-
cipitation as well as the problem-free periods, the soluble calcium level
in the scrubber loop has been in the range of 350-550 ppm. This compares
with a saturation level of about 800 ppm. The problems have been solved
and no such scaling has occurred in the last 12 months. Importantly, the
degree of softening (soluble calcium reduction) usually performed has re-
sulted in a calcium ion concentration reduction of 250-450 ppm below
saturation level. Since there has been no extended operation below this
range, it is not clear how little softening can be performed before en-
countering calcium sulfite or other scaling problems, although some re-
duction can probably be tolerated. GM estimates that operation at a
soluble calcium level of about 650 ppm would be viable.
8. Oxidation
The oxidation rate in the scrubbers has been about 45-60% of the 862 re-
moval rate, owing to the high excess air in the boilers and the resultant
high oxygen content of the flue gas relative to the S02 concentration.
Further oxidation to improve solids properties was attempted in two ways:
(1) by sparging air into chemical mix tank #1 while solids were recycled
or while lime was added to the same vessel; and (2) by sparging air into
the scrubber bleed liquor upstream of the chemical mix tank. The latter
was the only effective method, increasing oxidation beyond that in the
scrubbers by about 30% of the total S02 removal.
-------
The system has demonstrated its capability of removing (precipitating) sul-
fate at even the highest oxidation rates, i.e., greater than 90%.
9. Entrainment
Liquor and soluble solids entrainment in the scrubbers has occurred
periodically. In a test.authorized by GM, an average of about 0.019 m3/
hr (5 gal/hr) per scrubber was found to be entrained in the treated flue
gas. Although somewhat high, this was within the design specification.
Assuming an average scrubber liquor containing 0.5M Na2SO^, we project
that about 1.4 kg/hr of soluble solids entrainment at the full capacity
of a single scrubber.
10. Reliability and General Status
The system was developmental throughout the test program and as such did
not generally perform in accordance with design criteria. It did, how-
ever, show substantial improvement over time.
Because of the continuing developmental nature of the system, many of the
outages, particularly the longer ones, were for mechanical modifications
t'o accommodate new modes of operation. In all, the total scrubber avail-
ability to the boilers over the test program was 77.9% excluding four
long-term planned shutdowns for system modifications. Therefore, the
General Motors Double Alkali Process has not. yet been proven over an ex-
tended period to be a commercially viable process for 862 control. Be-
cause of recent process improvements demonstrated in the four-week test
in April-May 1976, we believe the system is'Aow capable of proving its
long term reliability.
The expected long term operating parameters for the General Motors Double
Alkali system are summarized in Table 2. These are based on demonstrated
operation as well as projected steady state criteria.
C. Recommendations
1. System Operation
a. General Motors should continue tc operate the plant in much the
same mode that proved successful during April-May 1976.
b. GM should commit to an extended demonstration of system reliability
to show that the plant is capable of operating at these optimal
levels for long periods of time with a high availability factor.
A test of 6-12 months i? desirable during which operating hours,
S02 removals, chemical makeup rates (based on feeds and apparent
stoichiometries from cake), and filter cake properties would be
monitored regularly. In the early stages of this reliability
demonstration, particular attention should be given to reducing
-------
Table 2
PROJECTED OPERATING PARAMETERS
GENERAL MOTORS DOUBLE ALKALI SYSTEM
SO- removal
Softening, reduction of Ca-H-
Soda ash addition, as Na?
Solubles loss, as Na-SO,
Lime stoichiometry
Solids properties
dry solids
solubles as Na^SO, (wet basis)
90%
150 ppm
0.05 mole/mole SO removed
0.05 mole/mole SO in cake
<1.3 mole/mole SO removed
55%
2.8?,
-------
excess lime feed, based on a reactor pH of about 12, and to clos-
ing the material balance on sodium. Later, as it becomes possible,
further oxidation of the bleed liquor should be attempted to deter-
mine if further improvement of cake dryness is feasible.
2. Double Alkali Process Technology
Double alkali technology should continue to be pursued as a viable option
for industrial boiler S02 control.
In the operation of its system, General Motors has acquired considerable
practical knowledge of both the process variables and, particularly, of
the mechanical considerations involved in such a system. This knowledge
should be used as a valuable resource by those contemplating the implemen-
tation of double alkali process technology.
-------
II. INTRODUCTION
A. Background
In March 1974, General Motors started up its demonstration S02 control sys-
tem at the Chevrolet Parma (Ohio) Plant using a double alkali wet scrubbing
technique developed in their prior pilot work at the same facility.
The steam plant at the Parma facility contains four boilers, two with a
nominal steaming capacity, two at 27,300 kg/hr (60,000 lb/hr) and two at
45,500 kg/hr (100,000 lb/hr). They are spreader stoker-fired with travel-
ing grates and operate with variable excess air rates in the 100% range.
Normally burning high sulfur (2-3%) eastern coal plus occasional lower sul-
fur waste oil, the raw flue gas generally contains 800-1300 ppm volume of
S02- The two larger boilers are equipped with economizers with resultant
lower flue gas temperatures than the smaller boilers. There already were
existing mechanical dust collectors on each boiler for primary particulate
control.
Each scrubber was installed to control the emissions only from its respec-
tive boiler, i.e., no provision was made for crossing over from one scrub-
ber to a parallel one while operating the same boiler. Booster fans were
installed upstream of the scrubbers where the flue gas is dry.
Since the scrubbers were required to provide only a modest reduction in
particulate loading from 0.73 to 0.12 gm/Nm3 dry (0.3 to 0.05 gr/scf d) ,
they were to be principally S02 removal devices. As such, they were de-
signed as three- tray columns (not high energy scrubbers such as Venturis)
with quench spray sections at the inlets and unwashed, full-diameter wire
mesh demister pads at the outlets. Provision was made for indirect reheat
of the flue gas using 175 psig steam to heat a fresh air stream which is
then mixed with the treated flue gas.
Figure 1 is a generalized flow schematic of the General Motors system
(Figures 4-7 in Chapter III show the specific schemes employed during the
test program). In the scrubber, S02 removal is effected by contact, at an
L/G of about 2.7 £/m3 (20 gal/1000 cu. ft.) of saturated flue gas, and
reaction with a circulating sodium solution by the following reaction
sequence :
2NaOH
S02 -> 2NaHS03
Simultaneously, some of the oxygen present in the flue gas participates
in the oxidation reaction
NaS0 + 1/2 0 ->• NaS0.
-------
Lime Water
Chemical
Mix
Tank #2
Chemical
Mix
Tank #1
Scrubber
H
Solids Contact
Reactor #1
Sludge
Blend
Tank
Filtrate
Soda Ash
Solids Contact
Reactor #2
Cake Wash Water
Filter Cake
Water
Surge
Tank
FIGURE 1 FLOW DIAGRAM - GENERAL MOTORS DOUBLE
ALKALI SYSTEM
-------
A portion of the circulating liquor containing a mixture principally of
sodium sulfite, sodium bisulfite and sodium sulfate is bled to a reactor/
clarifier system where the sodium value is regenerated for recycle to the
scrubber by reacting the solution with lime:
2NaHS03 + Ca(OH)2' + Na2S03
Na2S03 + Ca(OH)2 -»• 2NaOH
Na2S04 + Ca(OH)2 ->• 2NaOH
A residence time in the lime reactor of five minutes was provided at full
design flow. In practice, however, it has been more nearly 30-60 minutes
at the lower operating rates and sometimes using two reactors (chemical
mix tanks).
After the precipitated solids are further reacted and concentrated in a
18.3 m diameter, 4.3 m high clarifier, the liquor is softened in a second
equal-size clarifier to minimize the opportunity for deposition of calcium
salts in the scrubber. The clarifiers provide for a surface rise rate of
1.1 m/hr (0.46 gpm/ft2) at the design flow rate. Slurry is withdrawn as
underflow at about 15-20% solids.
Softening is accomplished by reaction with fresh soda ash feed which serves
also to replenish sodium to the system.
Ca++ + Na2C03 -»• CaCO^ + 2Na+
Alternatively, if soda ash is not required to the full extent of soften-
ing, carbon dioxide may be added also.
Ca++ + C02 + H20 -»• CaC03 + 2H+
However, since this has the effect of increasing the lime requirements to
maintain the regenerated liquor pH at a constant value, it is the less
preferred softening route.
The slurry underflows from both clarifiers are bled to a surge tank for
subsequent rotary vacuum filtration with '..he filtrate recirculated to the
regeneration system and the filter cake sent to disposal. There are two
full-size filters installed, each with 46.5 square meters (500 square feet)
of cloth area. It was found that nylon cloth afforded better cake release
than the originally installed polypropylene although it was less durable.
B. Test Program
1. Plan and Execution
The EPA test program of the GM system as planned and performed by ADL was
scheduled to be one year long and was to consist of three one-month intensive
11
-------
test periods and nine months of non-intensive testing. In the intensive
tests the system was to be monitored closely with ADL representatives
present and with samples of important process streams withdrawn daily for
analysis. In the remaining non<-intensive periods, this close day-to-day
observation would not be required and frequency of sampling would be re-
duced to weekly. Field support for the program was to be supplied by ADL
to coordinate and supervise the test program, and to observe the system
operation, and was to be supplied by GM to log operating data, to with-
draw samples and to analyze samples. (Before the start of the test pro-
gram, manpower availability changed at GM and the analytical work was con-
tracted to ADL by GM for the first intensive test and by EPA for the re-
mainder of the program. GM retained responsibility for system operation,
logging of operating data, flue gas analyses and withdrawing of necessary
samples.)
In program execution, the test periods were scheduled according to system
availability to ensure sufficient data for analysis. As a result, the
program was stretched to 21 months during which three one-month periods
of intensive testing and six months of formalized non-intensive testing
were performed. The remaining 12 months consisted of a lower level of
effort to keep informed of the operation and modifications to the process.
The first intensive test was viewed as a period of orientation to be used
to plan the remaining tests. As such, the planned data acquisition was
modest requiring no routine gas sampling or sludge weighing. In the
second test, data acquisition was more complete. We hoped to observe
three scrubbers operating at peak winter loads, but unexpected calcium
sulfite plugging in the scrubbers restricted SC^ removal and ultimately
scrubber operability. An average of only one scrubber was run for the
month. The third intensive test, in which data acquisition was again
relatively complete, went smoothly with two boilers and two scrubbers
operated for most of the time.
2. Data Acquisition
Figure 2 shows a generalized flow schematic of the GM plant indicating
the sampling and measuring points used in monitoring the process opera-
tion. Much of the data was reduced to the material balances described in
Appendix B.
3. Analytical Procedures
During the test program a large number of liquid, solid and slurry samples
were withdrawn from the system for determination of the relative quantities
of liquid and solid present and for analyses of the chemical species of
interest. Appendix C describes in detail the step-wise procedures used in
performing these chemical analyses.
12
-------
Flue Gas
Lime Water
0-i KD
©
Chemical
Mix
Tank #2
Chemical
Mix
Tank #1
Scrubber
No. 1
(Typ.)
L/G = 2.78/m3
(20 Gal/1000 cu. Ft.)
W
60
pH = 12-12.3
<1%Solids
Flow
\J) Temperature
PH
Weight
S02 Concentration
Sample Point
Solids Contact
Reactor #1
Sludge
Blend
Tank
800-1300 ppm
Flue Gas From
Boiler No. 1 Only
(Typ.)
O2«=10%
15-20% Solids
Filtrate
pH= 12.2-12.5
Soda Ash CO2
©
Solids Contact
Reactor #2
Cake Wash Water
Filter Cake
50% Solids
0
Water
t®
Surge
Tank
FIGURE 2 DATA COLLECTION POINTS AND NORMAL OPERATING CONDITIONS
-------
III. SYSTEM OPERATING HISTORY
This chapter presents a brief history of the operation of the General
Motors Double Alkali S02 Control System at Parma, Ohio, from August 19,
1974 to May 14, 1976.
Appendix A shows graphically when the four boilers were in operation and
also when the corresponding scrubbers were on. Comments indicating pro-
cess and mechanical operating difficulties, equipment modifications and
the timing of tests are noted on the appropriate dates. Discussion of
the important process parameters pertaining to these operating difficul-
ties is included in Chapter IV.
A summary of scrubber availability by month is presented in Figure 3.
Superimposed on the total operating hours for each boiler are the hours
in which the scrubber actually ran and additionally the percent of hours
in which the scrubber was ready but not run. From August 19, 1974, to
May 14, 1976, the scrubbers ran a total of 37.5% of the boiler operating
hours and were available for an additional 23.5%. The maximum potential
then was 61.0% over this period. When down time for major system modifica-
tions is excluded, the total was 77.9%.
These major modifications involved three periods:
• October 7, 1974, to November 12, 1974, when a number of changes
agreed among GM, EPA and ADL were implemented. Included were instrument
recalibration, cleaning of control valves, installation of sample petcocks,
installation of cake wash nozzles and investigation of chemical mix tank
and clarifier overflow plugging problems.
• March 14, 1975, to April 15, 1975, when GM was investigating the
sulfite plugging problem in the scrubbers. Ultimately, the scrubber
operating mode was changed requiring some repiping of the recirculating
and regenerated liquor lines.
r • June 27, 1975, to September 8, 1975, when open trough lines were
installed in the overflows of both chemical mix tanks and clarifier //I,
all of which had plugged seriously at various times.
• March 5, 1976, to March 29, 1976, when a section of the regenerated
liquor line was replaced along with a new orifice plate, both of which are
intended to afford accurate flow monitoring at low flow rates relative to
the original system design. At the same time, the instrumentation on the
lime feed line was also being replaced to permit better control of the
lime stoichiometry.
The test periods during which we monitored the system closely were:
• Intensive—August 19, 1974, to September 13, 1974
• Intensive—February 17, 1975, to March 14, 1975
15
-------
PAGE NOT
AVAILABLE
DIGITALLY
-------
• Non-Intensive—October 1975, to April 1976
• Intensive—April 19, 1976, to May 14, 1976.
In these various periods the operating flowsheets were different, notably
in the locations at which the regenerated liquor entered the scrubber sys-
tem and in the reactor system configuration. Figure 4 shows the original
flowsheet when the regenerated scrubber feed entered the scrubber on the
top tray while the recycle entered the scrubber separately, also on the
top tray. Later (Figure 5), the regenerated liquor entered the recycle
tank to mix with the liquor from the bottom tray. Next (Figure 6), the .
scrubber feed was mixed with the recycle liquor in the pipe prior to enter-
ing the scrubber to assure good mixing. Also during this period the fil-
ter cake was washed regularly, solids were recycled to chemical mix tank
//I from clarifier //I, and air was sparged into chemical mix tank #1 in an
effort to improve solids properties by oxidizing the sulfur species to
sulfate. Most recently (Figure 7), the scrubber feed system was unchanged
and cake washing continued. However, the solids recycle was discontinued,
lime was added to chemical mix tank //I instead of //2, and oxidation by air
sparging was performed in the scrubber bleed line upstream of the reactor.
Table 3 summarizes the important system operating parameters for the in-
tensive test periods.
19
-------
KJ
o
Chemical
Mix
Tank #2
Chemical
Mix
Tank #1
Solids Contact
Reactor #1
Sludge
Blend
Tank
Filtrate
Soda Ash CO2
I I
Solids Contact
Reactor #2
Filter Cake
Water
Surge
Tank
FIGURE 4 OPERATING FLOW DIAGRAM - 1ST INTENSIVE TEST
-------
Lime Water
I 1
Chemical
Mix
Tank #2
Chemical
Mix
Tank #1
Solids Contact
Reactor #1
Sludge
Blend
Tank
Filtrate
Soda Ash
Sol ids Contact
Reactor #2
Filter Cake
Water
Surge
Tank
FIGURE 5 OPERATING FLOW DIAGRAM - 2ND INTENSIVE TEST
-------
N>
.ime
I
Water
_L
Lime
Slaker
1
Chen
M
Tanl
nical
X
<#2
1
t I"
Chemical
Mix
Tank #1
^ J
(
)
^
Air
Scrubber
Solids Recycle
Solids Contact
Reactor #1
Sludge
Blend
Tank
Filtrate
Soda Ash
Solids Contact
Reactor #2
Cake Wash Water
Filter Cake
Surge
Tank
•€)
FIGURE 6 OPERATING FLOW DIAGRAM - NONINTENSIVE TEST
-------
Chemical
Mix
Tank #2
Lime Water
Chemical
Mix
Tank #1
Solids Contact
Reactor #1
Sludge
Blend
Tank
Filtrate
Soda Ash
Solids Contact
Reactor #2
Cake Wash
Water
Filter Cake
Surge
Tank
FIGURE 7 OPERATING FLOW DIAGRAM - 3RD INTENSIVE TEST
-------
Test Period
Regenerated Liquor
Na+, total, M
Na+, active, M
TOS*, M
OH~, M
Ca-H-, M
Cl~, M
Scrubber Bleed Liquor
TOS*, M
H+, M
Suspended solids
Chemical Mix Tank #2
Suspended solids, wt.%
OH~, M
Clarifier #1 Underflow
Suspended solids, wt.%
Clarifier #2 Underflow
Suspended solids, wt.%
Table 3
SUMMARY OF AVERAGE OPERATING PARAMETERS
8/19/74-9/13/74 2/17/75-3/14/75
4/19/76-5/14/76
0.58
0.087
0.004
0.080
0.0081
N.A.
0.021
0.025
0.1-2.0
0.1
18
35
0.66
0.13
0.010
0.085
0.012
0.12
0.37
0.02
0.3
0.09
16
13
0.96
0.12
0.009
0.084
0.013
0.18
0.033
0.031
0.8
0.07
13
29
* Total oxidizable sulfur.
-------
IV. CHARACTERIZATION OF SYSTEM OPERATION
For intermediate or longer term operation (i.e., greater than about two
times the system holdup), a reliable measure of several important system
operating parameters, such as overall oxidation and lime and soda ash
stoichiometries, is the filter cake composition. This is because vir-
tually all chemicals fed to the system as well as the S02 removed from the
flue gas leave via the filter cake.
Reflecting the load-following nature of the GM boilers, the operation
during each test period was not steady because the number,of scrubbers
operated, S02 removal, the feed forward rate, chemicals makeup rate and
solids filtration rate were variable over the periods. Consequently, the
cake compositions were also variable, but less so because of the damping
effect of the relatively long residence time of solids in the clarifiers.
Therefore, although less reliable than in a steady period, filter cake
composition is the most reliable basis available for quantifying the
important parameters of operation.
Table 4 summarizes the average filter cake analyses for each test. The
tolerances shown account for the variability in composition experienced
over the one-month periods and reflect a 95% confidence level (+ two
standard deviations of the mean). There are two components of variability
which are included in the magnitude of the tolerances:
• actual changes in process operation affecting cake compositions;
and
• analytical error.
Our experience with reproducibility of duplicate analytical results of
this type indicates that most analyses are accurate to within 5% (95%
confidence). Therefore, the major portion of the variability probably
lies in the process operation. This confirms the conclusion of the over-
all material balance calculations (Appendix B) that in each test period
there was no extended steady state operation.
The variability during the first intensive test was clearly greater than
during the second. This is principally because during the first test the
system operating rate was higher (with two scrubbers operating most of the
time) and because solids were removed from the system much faster than
they were produced (i.e., clarifier inventory was depleted). Therefore,
with less opportunity for solids backmixing within the clarifiers, the
composition of filter cake during the period was less uniform than during
the second test when an average of one scrubber was on and solids were re-
moved at a rate considerably lower than they were produced.
During the non-intensive test and again in the final intensive test, the
variability of cake composition decreased. This was due in large part to
the greater attention given to the system chemistry during the latter
months of the test program. We found over the first week of the third
25
-------
Table 4
MEAN FILTER CAKE COMPOSITIONS AND INDICATED OPERATING PARAMETERS
Test Period
Intensive/Non-Intensive
(8/19/74-9/13/74)
Intensive
Composition of Total Cake, millimoles/gram dry cake
Ca
so3
S°4
OH
co3
Na
Composition of Insoluble Solids, millimoles/gram dry cake
Ca
so
OH
co3
Na
c
Composition of Soluble Solids, millimoles/gram dry cake~
Ca
so3
SO,
OH
co3
Na
Lime Stoichiometry, Ca/SO in cake
Soda Ash Stoichiometry, Na2SOx in cake
Total Oxidation, SO,/SO in cake
' 4' x
Solids in Cake - Total Solids, wt.%
Soluble, wt.% as Na,,SO,
Insoluble, wt.%
Reported Tolerances are +_ 2 Standard Deviations (95% confidence)
Assumed equal to zero—i.e., assumed all Na ultimately leachable.
2Assumed equal to zero—too few analyses were performed.
Assumed all Na2SO^—no filtrate analyses were taken.
**Reported for last three weeks of four week test—see text.
5Based on filtrate analysis.
0.02 + 0.002
0.005 + 0.0008
0.329 + 0.06
0.12 + 0.012
O2
0.8 + 0.10
1.90 + 0.42
0.11 + 0.02
0.60 + 0.08
46.7 + 3.0
2.5 + 0.4
44.2 (by diff.)
(2/17/74-3/14/75)
Intensive
(10/22/75-4/1/76)
Non-In t ens ive
0.047 + 0.04
0.019 + 0.017
0.423 + 0.046
0.163 + 0.012
O2
0.98 + 0.09
1.65 + 0.14
0.12 + 0.02
0.47 + 0.04
40.9 + 0.2
2.4 + 0.3
38.5 (by diff.)
0.443
0.883
1.69 + 0.16
0.11 + 0.021
0.49 + 0.065
39.0 + 3.57
2.4
36.6 (by diff.)
(4/19/76-5/14/76)
Intensive
7.03
1.64
2.37
0.85
3.00
0.81
+ 0.
± °-
± °-
(by
± °-
± °-
34
42
40
diff.)
72
12
6.
2.
1.
1.
2.
1.
84 + 0.
23 + 0.
95 + 0.
66 + 0.
33 + 0.
01 + 0.
10
32
24
74
48
14
6.
2.
2.
0.
2.
0.
66 + 0
06 + 0
09 + 0
66 + 0
70 + 0
88 + 0
.18
.28
.33
.33
.51
.23
6.
0.
3.
0.
1.
0.
13
63
79
73
48
26
+ 0.24
+ 0.20
+ 0.17
+ 0.38
+ 0.34
+ 0.11
im dry cake
6.84
1.58
2.07
0.72
2.69
+ 0.
+ 0.
± °-
+ 0.
± °-
O1
41
37
66
31
75
6.
2.
1.
1.
2.
80 + 0.
07 + 0.
45 + 0.
17 + 0.
08 + 0.
O1
19
08
32
42
52
6.66
2.06
1.65
0.66
2.70
Ql
6.
0.
3.
0.
1.
10
63
67
70
48
+ 0.24
+ 0.20
+ 0.17
+ 0.38
+ 0.34
O1
0.01 + 0.001
0.001 + 0.001
0.12 + 0.01
0.03 + 0.01
O2
0.26 + 0.02
1.32 + 0.1061*
0.028 + 0.0141*
0.83 + 0.0414
56.0 + 2.51
1.0 + 0.4
55.0 (by diff.)
-------
intensive test that the cake properties changed rather rapidly to a point
where the analyses remained relatively constant for the remainder of the
test period. This indicates that the solids which were present in the
clarifiers at the beginning of the test were probably different from those
produced during the test. Therefore, in Table 4 the key operating para-
meters, lime and soda ash stoichiometries and oxidation, are summarized
for only the last three weeks of the test.
Generally, the stability of these tests was sufficient to permit a charac-
terization of the operating parameters of interest.
A. SOp Removal
Immediately prior to the test program, the system underwent a one-week
test for SC>2 removal. The performance was excellent with removals in the
range of 94-99% with relatively low inlet S02 levels of 600 and 1200 ppra,
reflecting the high excess air, and some outlets as low as 20 ppm or less.
The periodic carbonate scaling on the top trays of the scrubbers in the
initial mode of operation was attributed to poor mixing of regenerated
and recycle liquors with resultant locally high pH's (Section IV-G). To
counteract this problem, GM modified the piping so that the regenerated
liquor would enter the recycle tank to ensure good mixing. This resulted
in poor 862 removals ranging from about 20% to 85% with an overall average
of 65%. 862 removal suffered because of the lower pH in the scrubber and
the general loss of countercurrency (overall lower driving force). Fur-
ther, there was calcium sulfite plugging,1 which had not previously
occurred.
As a result of this experience, GM reverted to essentially the initial
flow scheme but with more efficient mixing of the liquors in a pipe out-
side the scrubber. 862 removals improved to the 90+% range immediately
as the pH rose.
During the final intensive test we noted that the 862 removal in scrubber
//I was consistently somewhat lower than in scrubber #4. Number four is
the older boiler with higher excess air and a consequently higher oxida-
tion rate. Thus, a larger fraction of the S(>2 pickup is present in the
liquor as SOA rather than SOo. At the same controlled pH (6.0 on the
bottom tray), the partial pressure of 862 above the solution is lower
with higher oxidation, resulting in better 862 removal. General Motors
performed an abbreviated test to determine if lowering the recycle rate
of low pH material to scrubber #1 would have a noticeable effect on SO-
removal by effectively increasing the mixed stream pH. Since the frac-
tional change (about 20% reduction) in the recycle rate has a very small
effect on the mix pH, the observed 862 removal was virtually unchanged.
1This plugging of scrubber trays, discussed later in Section IV-G, probably
also contributed to the poor 862 removals as there was probably some gas
channeling through the scrubber which inhibited mass transfer.
27
-------
B. Filter Cake Properties
The quality of filter cake produced is measured in two ways:
9 the total solids content of the cake which ultimately translates
to the total weight of the sludge which must be hauled from the plant for
disposal; and
• the soluble solids content of the cake which impacts directly on
the environmental acceptability of disposing of the sludge as a landfill
material.
The total solids (soluble plus insoluble) content in the filter cake for
the periods summarized in Table 4 averaged 46.7%, 40.9%, 39.0% and 56.0%,
respectively. For those produced in the first three periods, the fraction
of solids in the cake was considerably lower than a reasonably expected
level of 50%. Soluble solids in the cake of about 2.4-2.5% (wet basis)
reflected the fact that no cake washing was performed during the first
two test periods.
During the extended non-intensive test period, several other factors con-
tributed to the solubles loss:
• the sodium concentration in the liquor loop increased substan-
tially, to approximately the design level, by an average factor of almost
two as compared with the first two intensive tests;
9 at times the cake was not washed; and
o the low percent solids indicates a large moisture and solubles
loss.
In the final period, improvement in solids properties was substantial.
The cake contained 56% total solids of which only 1% was soluble on a wet
basis. These are attributable to continuous cake washing, maintenance of
low excess lime feed, and effective oxidation of the scrubber bleed.
C. Solids Dewatering Properties
During the first two tests the solids content of the filter cake was below
expectation because of the excess calcium hydroxide in the cake and the
relatively low concentration of gypsum (CaSO^ * 2H20). In an effort to
counteract these effects, General Motors later adopted the flowsheet of
Figure 6 for the non-intensive test period in which air was sparged into
chemical mix tank (CMT) #1 to oxidize the sulfite/bisulfite which was not
already oxidized in the scrubber system. Since this oxidation procedure
lowers the pH by increasing the bisulfite/sulfite ratio, the stream must
be at least partially neutralized to protect the mild steel reactors. The
underflow from clarifier //I was chosen for recycle to CMT //I where both
the alkaline solution and the excess hydroxide in the solid could neutral-
ize the acidity. In addition, the presence of seed crystals would minimize
28
-------
the possibility of .sulfate supersaturation and subsequent scaling in CMT
#2 and the clarifiers.
Unfortunately, poor solids settling and frequent solids overflow from the
clarifiers resulted. Since the clarifiers were adequately sized to handle
the flow from five scrubbers—and seldom were more than two operating on
an extended basis—solids settling had never been an important problem,
although flocculant was periodically used. Because solids recycled to
the chemical mix tank from the clarifier underflow were carried in a three
inch line in which a minimum velocity was required to prevent plugging,
the total solids load .on the clarifier increased several-fold as the solids
concentration through the reactor-clarifier system increased to about 5%
from the less-than-1% design level while the total volumetric flow was
simultaneously increasing. The clarifiers, now experiencing hindered
settling, were unable to handle this large duty, and solids carryover—
to the extent that large quantities of flocculant were of no value—was
the result.
In the third intensive test the solids recycle was eliminated. Oxidation
upstream of the chemical mix tank apparently was effective. Slaked lime
was added continuously to the first of two equal-size chemical mix tanks
on the basis of periodic hydroxide analyses to avoid feeding excess lime.
Thus, with good-settling gypsum crystals and less unutilized lime, the
solids produced in the third intensive test were consistently the driest
observed. And for the first time, GM discontinued the use of flocculant.
D. Lime Stoichiometry
The consumption of lime during the entire test program, while improving,
has been significantly greater than theoretical. Lime Stoichiometry
values of 1.90, 1.65, 1.69 and 1.32 moles Ca per mole S02 removed were
observed for the four individual test periods. Because of difficulties
with the automatic lime feed control system (plugged control valve,
scaled magnetic flowmeter, etc.), the operators have been forced to feed
lime to the system under manual control.
The method of controlling the lime feed rate coupled with the low sodium
level in the system contributed to the poor utilization in the first two
tests. Lime was fed to the chemical mix tank on the basis of the target
hydroxide level of 0.1M OH in the scrubber feed. At an Na2SO, concen-
tration of 0.33M, the solubility of hydroxide is somewhat less than this
target value. Thus, overfeeding of lime was common, since a large in-
crease in the feed rate results in a small, if any, increase in the hydrox-
ide level.
The lime feed control method during the non-intensive test was to manually
adjust the lime feed rate in proportion to the rate of regenerated liquor
sent, to the scrubbers assuming the target 0.1M hydroxide was met. However,
considerable difficulty was experienced, not because the principle was not
sound, but because the scrubber feed rate—the critical variable—could not
be measured accurately. Finally, before the third intensive test, the flow
orifice and a section of pipe were replaced to obtain good flow indication.
29
-------
In the third intensive test the lime feed (again manually controlled) was
set to give a [OH~] in chemical mix tank #2 of about 0.08-0.09M. This has
worked best of any method to date at GM, resulting in a lime stoichiometry
of 1.32 for the period, somewhat higher than the target of 1.2 (set by GM,
EPA and ADL for this program).
E. Carbonate Softening
As discussed earlier, some degree of softening is necessary before the
regenerated liquor is returned to the scrubber loop to avoid the potential
for solids deposition there. The soluble calcium level for the liquor
entering the softening clarifier has been about 800 ppm by weight, and
was reduced to the 350-550 ppm range depending on the operating mode, the
evidence and recent history of plugging, and the source of carbonate for
softening.
By the end of the test program, when most operating parameters were within
expectation, the softening was performed to reduce Ca-H- concentration to
a level of 500 ppm continuously, resulting in no evidence of plugging. It
is not yet clear what maximum level of soluble calcium can be tolerated.
However, since the degree of softening directly impacts on the consumption
of soda ash (Section TV-F), it is important that softening be minimized
to reduce operating costs and reduce the soluble sodium salt content in
the waste solids.
The General Motors system has provisions for using either of two carbonate
softening agents, carbon dioxide or soda ash. The use of gaseous carbon
dioxide was attempted to various extents for the first year of system
operation. The principal drawback of this approach was found to be the
fact that CC>2 addition neutralized two moles of alkalinity per mole of
calcium precipitated. This lowered the capacity for SC>2 removal un-
less the scrubber feed rate was correspondingly increased. Further, GM
reports that the dewatering properties of the (^-precipitated solids
were inferior to those produced from soda ash. As a result, GM has dis-
continued regular use of CC>2 and has it available only as a contingency.
Soda ash softening is far more effective. No neutralization of liquor
alkalinity occurs and the solids settle and filter well. One area re-
quiring caution is that of controlling the sodium balance as discussed
below.
F. Soda Ash Stoichiometry
As reported in Table 2, the solubles loss for the early test periods was
in excess of 0.1 moles Na2 per mole SOX removed. This is relatively high,
reflecting the fact that the filter cake was not washed. Despite wash-
ing which had begun by October, 1975, the soda ash stoichiometry remained
excessive because of: (1) high total sodium (>1.2M) in the liquor requir-
ing more washing; (2) occasionally unwashed samples which hurt the overall
average; and (3) high liquor content in filter cake increasing the mois-
ture and solubles losses.
30
-------
During the last month of the test program, there was a very low solubles
loss, about 0.03 moles of Na2 per mole of SO removed, which is lower than
the General Motors design level of 0.05. This demonstrated the system's
ability to achieve a significant reduction in soluble solids by cake
washing.
However, there was a buildup of sodium in the system over this test
period as reflected by the higher rate of sodium input (as soda ash) than
the rate of removal (as soluble solids losses in the filter cake), and by
the increase in concentration of sodium ion over the period. Therefore,
this is not a steady state mode of operation.
System operation to date indicates that the filter cake can be washed to
levels resulting in sodium loss rates which are lower than minimum sodium
makeup rates required for soda ash softening (Table 5). Therefore,.it
would be misleading to indicate that a soda ash stoichiometry of 0.03 has
been achieved based only on filter cake data. Since the minimum level of
softening has not yet been determined, the theoretical minimum soda ash
stoichiometry cannot yet be projected. However, if the softening require-
ment can be lowered to a Ca-H- reduction of 150 ppm, the soda ash stoich-
iometry would drop to about 0.05. The system has not yet been operated
at a Ca-H- reduction of less than 250 ppm for an extended period.
G. Scaling in Scrubber Loop
Operation of the General Motors Double Alkali system during testing has
encountered two sets of circumstances leading to buildup of solids in the
scrubber loop. Prior to the program, GM noted on several occasions the
buildup of calcium carbonate scale on, the top tray of the scrubber. The
flowsheet at these times was as the scheme in Figure 4 with the scrubber
feed entering the top tray directly and mixing with the recycle stream
upon the tray. The fact that CaCO^ scale was in fact formed on the top
tray is indicative of incomplete mixing of the liquid streams and locally
high pH levels capable of scrubbing C02 from the flue gas.
After some experimentation to correct the carbonate scaling, GM adopted
the flow scheme of Figure 5 (Chapter III). In this configuration, solids
plugged the trays with calcium sulfite. Plugging occurred because:
(1) the pH in the recycle tank was about 7.5 (versus 6.0 in the other
tests), therefore, the TOS was mostly sulfite; (2) the soluble calcium
level was 400-500 ppm; (3) the total sodium concentration of about 0.33M
Nao, about 30-40% below the design concentration, resulted in a low ionic
strength of about 1.0; and (4) the concentrations of these species ex-
ceeded the apparent solubility product for calcium sulfite of about
4 x 10~5 in the solution.
Supetsaturation of the solution in the recycle tank resulted in precipita-
tion in the scrubber system. Plugging was predominant in areas
31
-------
Table 5
SODIUM BALANCE DATA
SOFTENING VS. SOLUBLES LOSS
Date 4/26-30/1976
2
Regenerated liquor flow, m /hr(gpm) 44(193)
Soluble calcium - unsoftened, ppm 880
softened, ppm 520
SO removal rate, kg-mole/hr (Ib-mole/hr) 3.00(6.59)
A
Calculated calcium precipitation
mole Ca/mole SO removed
Filter cake solubles, mole Na2/mole
SO in cake
0.13
Soda ash makeup rate, mole Na^/mole
SO removed 0.12
0.056
5/3-7/1976
44(193)
800
440
3.10(6.81)
0.13
0.11
0.01
32
-------
of maximum SC^ absorption; i.e., in the vicinity of the valve slots on
the trays. Plugging also occurred in the quench nozzle where the relatively
high temperature (350°-550°F) of the untreated flue gas may have played a
role.
After this sulfite scaling experience, GM reverted to the flow scheme
where regenerated liquor enters the top tray along with recycle liquor—
with one major modification. The streams mix entirely outside the scrubber
in a length of recycle pipe. This assures good mixing and avoids the local
carbonate scale problem of the earlier operating days. This mode has
worked well showing no tendency to scale at normal operating pH.
H. Oxidation
The overall oxidation rates apparent from the filter cake analyses were
about 60% in the first period and 47% in the second. This is as expected
since in the first test period the scrubbers operated principally were
//I and //3 and in the second //I was operated predominantly. Boiler #3 is
one of the older boilers and operates with considerably more excess air
than //I resulting in greater oxygen pickup by the scrubber liquor and
hence more oxidation.
Another factor which may have contributed to lower oxidation rates in the
second test period was that the sulfur content of the coal was probably
higher. The average sulfur value of the coal purchased in August and
September 1974 was 2.36%. In February 1975 it was 2.48%. In the latter
period, flue gas S02 analyses were taken almost daily and indicated that^
on average, the coal actually burned was nearer 3.0% sulfur. Thus, the
oxidation rate as a fraction of S02 removal decreased as the S02 removal
increased. Flue gas analyses were not performed during the August-Sep-
tember 1974 period.
In the non-intensive period the average oxidation rate was about 49%,
consistent with that of the second intensive test. Since boilers //I and
//2, the newer ones, carried most of the load for this six-month period,
scrubber oxidation was in the lower range.
During this period, air was being sparged into chemical mix tank //I while
solids were being recycled as well (Figure 5). This proved rather ineffec-
tive for a combination of several possible reasons:
e there was no provision for measuring the total air flow which may
in fact have been quite low;
• although the sparger design (two perforated pipes) appeared to be
adequate, the gas-liquid interface for oxygen mass transfer may not have
been adequate; and
• with the presence of recirculated solids containing a considerable
quantity of unutilized alkalinity, the sulfite and bisulfite may have been
substantially precipitated before having sufficient opportunity to oxidize.
33
-------
In the final test period the cake analyses indicated an average overall
oxidation rate of about 83%. From the material balance data of Appendix
B, a scrubber oxidation rate of about 55% was apparent, which was quite
consistent with earlier periods. The remaining oxidation occurred in the
newly installed sparger in the scrubber bleed line upstream of the chemi-
cal mix tanks.
I. Entrainment
Isolated instances of high entrainment from the scrubbers were observed
during the first intensive test as evidenced by considerable dense mist
emerging from the scrubber stacks and carrying down from the roof to the
ground. As a result, it was decided to measure entrainment in the scrubber
flue gas effluent. This entrainment test, performed by Crobaugh Labora-
tories on January 14-16, 1975, showed entrainment of 19.5 and 16.8 liters
per hour (5.16 and 4.43 gallons per hour) from scrubbers #1 and #4, res-
pectively.
Although entrainment of scrubber liquor is within the design specification,
the total potential entrainment of soluble solids (assuming 0.5M Na^SO,),
is about 1.4 kg/hr or 0.013 gm/Nm3 (3 Ib/hr or 0.012 grains/scfj. Based
on a heat input of 20 x 106 Kcal/hr (80 x 106 Btu/hr) to the boiler, the code
restricts the partlculate emissions to about 7.3 kg/hr (16 Ib/hr), depend-
ing on the boiler operated. Therefore, soluble solids lost as entrainment
from this system can contribute significantly to the allowable particulate
emissions.
A contributing factor to the periodic entrainment problem is the location
of the demister relative to the transition piece at the top of the scrubber.
Immediately above the full-diameter demister pad, is a square reducer
through which the outlet gas passes into a high velocity duct. Inspec-
tions during shutdowns of solids buildup on the pad indicate that the gas
is to a large extent not "seeing" the full diameter but is making the
transition upstream of the demister, thereby rendering much of it ineffec-
tive.
Over the past few months the stainless steel wire mesh entrainment separators
have been replaced as necessary with plastic mesh pads of the same size.
These polypropylene units are expected to outlast the original equipment
which deteriorated seriously over time. Although there is no reason to
believe that these plastic mist eliminators would work more effectively
than the original wire ones, there has been little visual evidence in re-
cent months of any entrainment problems. Further, the sodium balances
have closed well indicating little, if any, loss via entrainment.
34
-------
APPENDIX A
SUMMARY OF SCRUBBER AND BOILER OPERATION
SCRUBBER AND BOILER OPERATION
Bl SI B2 S2 B3 S3 B4 S4
8-19-74
20
21
22
23
24
25
26
27
28
29
30
31
9- 1-74
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
10-1-74
2
3
4
5
6
7
8
9
10
11 —
12
13
14
Recently Installed weir block on top tray of scrubber #3 to avoid
local scaling. Began first intensive test.
Sump pumps kicking on causes CMT #1 overflow. Entrainment problem-
mist elim. requires testing. Some instruments require recalibra-
tion, new relays.
Ended first intensive test.
System down for modifications agreed among EPA, GM and ADL.
Scrubber feed control (on pH) - changed from multiplier to adder -
should be log not linear. SI SCR turbine under repair. CMT //I
overflow plugged - probably during intensive test period. Scrubber
feed pressure regulator had not been operating properly - has just
been cleaned.
35
-------
SCRUBBER AND BOILER OPERATION
Bl SI B2 S2 B3 S3 BA S4
Still down-plugged Hi SCR overflow. Open trough overflow from SCR //1-//2
contemplated. Trying to add scrubber feed to recycle tank. Valve In
CMT #1 overflow line removed and line replped to avoid future plugging.
10-15-74
16
17
18
19
20
21
22
23
24
25
26
27
28
29 -
30
31
11- 1-74
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19 Clarlfler ill overflow plugged.
20
21
22
23
24
25
26
27
28
29
30
12- 1-74
2
3
4
5
6
7
8
9
10
11
12
13
36
System has been run with no lime
addition only Na.CO- added to bring Na+ level up.
-------
SCRUBBER AND BOILER OPERATION
Bl SI B2 S2 B3 S3 B4 S4
12-14-74
15
16
17
18
19 Limited operation during holiday period.
20
21
22
23
24
25
26
27
28
29
30
31
1- 1-75
2
3
4
5
6
7
8
9
10
11
12
13 Entrainment test with Koch present. Reported scrubber Hi - 5.16
, gal/hr. Reported scrubber 112 - 4.43 gal/hr. Running with Na-CO,
makeup only.
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
2- 1-75 — Scrubbers down for weekend.
2
3
4 All scrubber feeds repiped - recycle tank.
5
6
7
8
9
10
11
12 37
-------
SCRUBBER AND BOILER OPERATION
Bl SI B2 S2 B3 S3 B4 S4
Began second intensive test.
#2 shutdown - some pluggage of trays.
Scrubbers down for weekend—manpower shortage.
Scrubber #2 still down - #3 and 04 plugging on lower trays. Filter
wash not resumed: poor vac., may need new pumps - belt pulls apart,
solids in underflow quite low - poor release properties.
2-13-75
14
15
16
17
18
19
20 — 03 scrubber - plugged trays.
21
22
23
24
25
26
27
28
3- 1-75
2
3
4
5
6
7
8 Scrubbers down for weekend—manpower shortage.
9
10
11
12
13
14 Solids analysis of scrubber 02 bottom of tray 1 - 24.6% CaSO^, 1/2
15
16
quench nozzle - 70.8% CaSO., 1/2 HjO. Ended second Intensive
test—scrubbers down during investigation of scrubber plugging
problem.
17
18
19
20
21
22
23
24 —
25
26
27
28
29
30
31
-1-75
2
3 —
4
5
6
7
8 —
9
10
11
12
13
14
Attributing poor SO. removal to CaSO, scaling in scrubbers.
seems time delayed.
Frecip.
01 up at 10:30 4/1 - down 4/3 12:30; attempting to run low liquid
level in recycle tank and cavltated pump.
This and following run of 4/15/75 somewhat experimental In new
mode of scrubber pH control.
38
-------
SCRUBBER AND BOILER OPERATION
Bl SI B2 S2 B3 S3 B4 S4
4-15-75
16 Overflowed solids Into system. Running on bottom tray pH
17 • Some cake washing. Still trying to get Na+ up.
18
19
20
21
22
23
24
25
26
27
28
29
30
5 -1-75
2
3
4
5
6
7
8
9
10
11
12 .
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30 Plugged; chemical mix tank 02 overflowed.
31
6 -1-75
2
3
4
5
6
7
8
9
10
6.3.
39
-------
SCRUBBER AND BOILER OPERATION
Bl SI B2 S2 B3 S3 B4 S4
6-11-75
12
13
14
15
16 System restarted after chem. mix tank 02 overflow bypassed.
17
18
19
20
21
22
23
24
25
26
27 System down for Installation of open-trough overflow lines.
28
29
30
7-1-75
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25 Annual shutdown.
26
27
28
29
30
31
8-1-75
2
3
4
5
40
-------
SCRUBBER AND BOILER OPERATION
Bl SI B2 S2 B3 S3 B4 S4
8-6-75
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
9-1-75
2
3
4
5
6
7
8 System restarted with open trough overflow lines.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27 #2 ID fan vibrating.
41
-------
SCRUBBER AND BOILER OPERATION
Bl SI B2 S2 B3 S3 B4 S4
9-28-75
29
30
10 -1-75 — 01 scrubber plugged due to solids recycle.
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16 03 bleed line plugged—3 hours.
17
18
19
20 Solids carryover.
21
22
23
24
25
26
27
28
29
30
31
11 -1-75
2
3
4
5
6
7
8
9
10
11 Solids carryover.
12
13
14 Solids carryover.
15
1.6
17
18
19
20
42
-------
SCRUBBER AND BOILER OPERATION
Bl SI B2 S2 B3 S3 B4 S4
11-21-75
22
23
24
25 Solids carryover.
26
27
28
29
30
12 -1-75 - 113 quench line-leaking - cannot start up.
2 Solids carryover.
3
4
5
6
7
10 Solids carryover.
11
12
13 Solids carryover.
14
15
16
17
18
19 Down for holidays; catching up on solids filtration.
20
21
22
23
24
25
26
27
28
29
30
31
1 -1-76
2
3
4
5
6
7
8
9 Solids carryover.
10
11
12
43-
-------
SCRUBBER AND BOILER OPERATION
Bl SI B2 S2 B3 S3 B4 54
1-13-76
14
15
16
17
18
19
20
21
22 Solids carryover.
23
24
25
26
27
28
29
30 Solids carryover.
31
2 -1-76
2
3
4
5
6
7
8
9
10
11
12
13 04 scrubber .plugged.
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
3 -1-76
2
3
4
5 Down for scrubber feed and lime feed instrumentation changes.
6
44
-------
SCRUBBER AND BOILER OPERATION
Bl SI B2 S2 B3 S3 B4 S4
111
11
3 -7-76
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
4 -1-76
2
3 --- Leak in til scrubber bleed line.
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
61 booster fan vibrating; foundation cracked; 62 demist, plugged,replaced.
Intensive test no. 3 began.
—- Took 5 hours to start booster fan—electrical problem.
45
-------
SCRUBBER AND BOILER OPERATION
Bl SI B2 S2 B3 S3 B4 S4
5 -1-76
2
3
4
5
6
7
8
9
10
11
12
13
14 End test program.
46
-------
APPENDIX B
MATERIAL BALANCES
47
-------
APPENDIX B
MATERIAL BALANCES
A. Overall and Component Material Balances
We ran material balances for the overall system and for component systems
to check the validity and consistency of the various measurements of flow
and composition made during the three intensive test periods. In planning
the data acquisition program we viewed the first intensive test period as
one of orientation. Only the instrumentation initially incorporated into
the plant design was recommended for use. If the test results showed that
additional or more accurate data collection were necessary or that minor
equipment or instrumentation revision were required for the success of
future tests, such modifications would be undertaken.
Although no flue gas measurements were made in the first test—they had
been made one month earlier—and the filter cake was not weighed, the
material balances performed (and discussed below) appeared consistent.
For the second and third intensive tests these measurements were made re-
gularly and provide the definitive material balance basis upon which to
draw appropriate conclusions.
Ideally, to minimize the difficulties in closing material balances, the
system is best tested at steady state. This, however, is not practical
from the standpoint of the constraints of meeting the variable steam re-
quirements of a manufacturing complex and weekend shutdowns for energy
conservation. Therefore, it was not unexpected that the system experienced
no long steady-state periods during the tests.
1. One-Month Intensive Test Balances
Over each intensive test period a total (liquid + solids) balance was
attempted. Thus, ignoring the oxygen for oxidation and assuming no net
C0~ absorption,
S02 absorbed + Lime + Water + CO™ addition + Na2CO = Wet Cake + Evap.
For Intensive Test Period No. 1 (8/19/74 through 9/13/74), this balance
was not performed as all the necessary data was not collected for reasons
discussed above. With appropriate estimates, the balances shown in Table
B-l were made for calcium, sodium, and carbon. These balances show that
during the 26-day period, the system inventory was not constant as more
calcium, sodium and carbon was removed than was fed.
For Intensive Test Period No. 2 (2/17/75 through 3/14/75), Table B-2
shows that the system seemed to be accumulating solids and sodium,1 but
to be closer to overall balance than in the previous test period.
1 This apparent buildup of sodium is at least partly offset by overflow
losses from the chemical mix tanks and clarifiers, which although known
to have occurred are not quantifiable.
48
-------
Table B-l
GM PARMA PLANT
ONE-MONTH OVERALL BALANCE—INTENSIVE TEST PERIOD NO. 1
8/19/74 through 9/13/74
Measured Inputs
CaO
H20 (liquid)
CO 2
Na2C03
Measured Outputs
Wet Cake:
Cake Composition:
Balances
111,700 kg
5,352,100
14,600
3.300
5,481,700 kg
995,700 kg (est. from Clarifier Underflows)
47.1% Solids Containing
7.03 m Moles of Ca++/g Dry Solids
1.64
2.37
3.00 CO 3
0.81 Na+
0.85 OH~
Ca
Na
C
Inputs (I)
79,800 kg
1,440
4,350
Outputs (0)
131,900 kg
8,740
16,900
(0 - I)/I
+ 65%
+508%
+288%
Note: Elemental C balances exclude C03 in lime which
analyses showed to be negligible.
49
-------
Table B-2
GM PARMA PLANT
ONE-MONTH OVERALL BALANCE—INTENSIVE TEST PERIOD NO. 2
2/17/75 through 3/14/75
Measured Inputs
S02
CaO
H20 (liquid)
C02
Na2C03
84,800 kg
76,300
3,364,400
3,700
11,800
3,541,000 kg
Measured Outputs
S02: 29,400 kg
Wet Cake: 309,600
339,000 kg
Cake Composition: 41.4% Solids Containing
6.84 m Moles of Ca+ /g Dry Solids
2.23 SOs
1.95 SO"
2.33 COs
1.01 Na+
1.66 OH~
Balances
(1)
1. Evaporation by Overall Balance
= 3,541,100 - 339,100 = 3,202,000 kg (22.7 gal/min)
2. Elemental Balances
Ca
S
Na
(2)
Inputs (I)
54,500 kg
42,400
5,100
2,360
Outputs (0)
35,100 kg
31,900
3,000
3,580
(0 - I)/I
-36%
-25%
-42%
+52%
(1)
(2)
Agrees with psychrometric calculations.
Elemental C balances exclude C03 in lime which
analyses showed to be negligible.
50
-------
For Intensive Test Period No. 3 (4/19/76 through 4/15/76), Table B-3
shows an overall net accumulation of solids in the system, mainly sul-
fur salts in clarifier //I while carbonate was being depleted from clari-
fier #2. As shown in the next section, the middle two weeks of the test
program were relatively steady material balance periods. Net accumulation
of solids occurred mainly during the final week of testing at reduced
filtration rates. The accumulation of sodium in the system agrees with
the measured increase of total sodium concentration in the liquor from
.935 M to .987 M over the two one-week short-term material balances.
2. Short-Term Intensive Test Balances
The approach taken in reducing the raw data for short-term intensive
test material balances was to determine the relatively steady periods of
operation by appropriately plotted and tabulated operating data and analy-
ses. These were three to five day periods in each case. For a selected
portion of an intensive test period, measurement of all the streams noted
in Figure B-l would allow complete balances to be run on all three systems
shown in Figure B-l.
To complete the material balance for the selected steady period of
Intensive Test Period No. 1 (8/27/74 through 8/30/74) we made several
assumptions. The scrubber feed rate was estimated at 23 m3/hr (100 gpm—
the flow was normally in this range and the estimate corresponded to
reasonable evaporation and S02 pickup rates). The scrubber bleed flow
was then estimated by assuming a sodium balance across the scrubber.
Since the wet cake weight was estimated from the clarifier underflows and
there is no filtrate flow meter, balances were estimated. SO^ removal and
evaporation could be estimated based on liquor flows as shown in Table B-4.
Table B-4 confirms that the system was probably losing inventory but that
the loss was not as serious over the 4-day material balance period as it
was over the month-long test period.
The balances for the selected portion of Intensive Test Period No. 2
(2/26/75 through 2/28/75), as expected, were more complete. The only
assumptions required were that:
• the outlet water content of the flue gas was estimated based on
saturation at scrubber gas outlet temperature; an adiabatic satura-
tion line on the air-H^O psychrometric chart was used to estimate
inlet humidity;
e again, the scrubber bleed flow was estimated from a sodium balance
across the scrubber after a reasonable estimate of the scrubber
feed; and
0 the filtrate flow and solids content were estimated from the cake
and clarifier underflow measurements.
The balances for the second period are summarized in Table B-5. They con-
firm a small accumulation of material inventory during the test period.
51
-------
Table B-3
GM PARMA PLANT
LONG-TERM OVERALL BALANCE—INTENSIVE TEST PERIOD NO. 3
4/19/76 through 5/14/76
Measured Inputs
SO 2
CaO
H20 (liquid)
Na2C03
Measured Outputs
S02:
Wet Cake:
96,800 kg
92,800
3,481,200
20,000
3,690,800 kg
10,600 kg
309,400
320,000 kg
Cake Composition: 56.0% Solids Containing
6.13 m Moles of Ca++/g Dry Solids
0.63 SOg
3.79 SO"
1.48 COa
0.26 Na+
0.73 OH~
Balances
(1)
Evaporation by Overall Balance
= 3,691,000 - 319,900 = 3,371,100 kg (24.7 gal/min)
Elemental Balances
Ca
S
Na
c(2)
(1) Agrees with psychrometric calculations.
(2) Elemental C balances exclude C0$ in lime which
analyses showed to be negligible.
Inputs (I)
66,300 kg
48,500
8,670
2,260
Outputs (0)
43,800 kg
24,500
1,040
3,080
(0 - I)/I
-36%
-50%
-88%
+36%
52
-------
"Overall System"
Gas In
Lime
Water
Soda Ash
C02 (if used)
Scrubber System
Scrubber Bleed
Caustic (if used)
Reactor/
Clarifier
System
Underflow #1
Under-
flow #.
~l
Scrubber Feed
Filtrate
Filtration System
Gas Out.
Filter Cake
_l
FIGURE B-1 DEFINITION OF TERMS - GM SYSTEM
53
-------
Table B-4
SUMMARY OF SHORT TERM BALANCES - AUGUST 27-30, 1974
Scrubber System
Total Flows
H20
kg/hr
Total
Input (I)
38,400
36,900
370
920
500
12
Total
Output (0)
38,500
36,900
370
1,050
500
12
Difference
(0 - I)/I, %
0.3
o.o(1)
o.o(2)
14.3(3)
o.o(4)
0.0
Na
Ca
Entire Regeneration System
Total Flows 42,100 43,300 2.8
H20 40,400 39,700 -1.7
S 370 550 48.6
S04 1,050 1,410 34.3
Na 500 540 8.0
Ca 140 700 400.0
(1) Difference is zero because output includes evaporation which is
calculated to give an exact H20 balance across the scrubber system.
(2) Difference is zero because input includes S absorbed which is cal-
culated to give an exact S balance across the scrubber system.
(3) Serves as an estimate of scrubber oxidation rate.
(4) Difference is zero because the output flow is calculated to give an
exact Na balance across the scrubber system.
54
-------
Table B-5
SUMMARY OF SHORT TERM BALANCES - FEBRUARY 26-28, 1975
kg/hr
Total
Input (I)
23,400
22,300
250
570
350
10.9
29,100
27,800
270
670
380
130
2,400
2,010
61
110
37
90
Total
Output (0)
23,200
22,100
246
600
350
10.5
25,700
24,300
260
680
380
105
2,400
2,040
55
100
33
86
Difference
(0 - I)/I, %
-0.8
-0.9
-1.6
5.3
o.o(1)
-4.2
-11.9
-12.5
-3.7
1.5
0.0
-19.2
o.o(2)
1.5
-9.6
-9.1
-10.8
-4.4
Scrubber System
Total Flows
H20
S
so4
Na
Ca
Reactor/Clarifier System
Total Flows
S
so4
Na
Ca
Filter System
Total Flows
H20
S
so4
Na
Ca
(1) Difference is zero because the output flow is calculated to give
an exact Na balance across the scrubber system.
(2) Difference is zero because the filtrate flow was estimated to
give an exact material balance across the filter system.
55
-------
The estimated S02 absorption was equivalent to 52.7 kg S/hr (116 Ib S/hr),
the average scrubber evaporation rate was 1930 kg H20/hr (4,240 Ib/hr), and
the average scrubber oxidation rate was equivalent to production of 34 kg
of SO^/hr (74 Ib/hr). These figures compare with 60.5 kg S/hr, 6,280 kg
H20/hr, and 131 kg SO^/hr in the earlier test. The evaporation estimate
for the first intensive test is based on a water balance for the period.
That for the second test is based on what is now an apparently inaccurate
scrubber outlet temperature because of a buildup around the thermocouple in
the stack.
Tables B-6 and B-7 summarize the material balances of two five-day periods
of Intensive Test No. 3 (4/26-30/1976 and 5/3-7/1976). The comparable S02
removal rates were 95.9 kg S/hr while the water evaporation averaged 6,386
kg H20/hr and 6,697 kg H20/hr, respectively. Oxidation of sulfur resulted
in sulfate formation rates in the scrubber system of 155 kg SO^/hr (54% of
the S02 removed) and 162 kg S0,/hr (54%), respectively.
56
-------
Table B-6
SUMMARY OF SHORT TERM BALANCES - APRIL 26-30, 1976
kg/hr
Total
Input (I)
45,600
42,900
620
1,540
820
23
51,200
48,200
630
1,770
890
140
5,160
4,500
105
280
61
154
Total
Output (0)
47,800
45,200
600
1,680
820
22
49,300
46,200
620
1,810
880
170
5,160
4,490
130
310
62
150
Difference
(0 - I)/I, %
4.8
5.4
-3.2
9.1
o.o(1)
-4.3
-3.7
-4.2
-1.6
2.3
-1.1
21.4
0.0<2>
-0.2
23.8
14.1
1.6
-2.6
Scrubber System
Total Flows
H20
S
so4
Na
Ca
Reactor/Clarlfier System
Total Flows
S
so4
Na
Ca
Filter System
Total Flows
H20
S
so4
Na
Ca
(1) Difference is zero because the output flow is calculated to give
an exact Na balance across the scrubber system.
(2) Difference is zero because the filtrate flow was estimated to
give an exact material balance across the filter system.
57
-------
Table B-7
SUMMARY OF SHORT TERM BALANCES - MAY 3-7, 1976
Scrubber System
Total Flows
H20
S
so4
Na
Ca
Reactor/Clarifier System
Total Flows
H20
S
so4
Na
Ca
Filter System
Total Flows
Na
Ca
kg/hr
Total
Input (I)
45,500
42,700
650
160
865
19
50,700
47,600
680
1,900
950
130
5,820
4,770
145
375
73
280
Total
Output (0)
47,900
45,100
630
177
865
19
49,900
46,300
700
1,980
940
300
5,820
4,990
148
385
74
180
Difference
(0 - I)/I, %
5.3
5.6
-3.1
10.6
o.o(1)
0.0
-1.6
-2.7
2.9
4.2
-1.1
131.0
o.o<2>
4.6
2.1
2.7
1.3
-3.6
(1) Difference is zero because the output flow is calculated to give
an exact Na balance across the scrubber system.
(2) Difference is zero because the filtrate flow was estimated to give
an exact material balance across the filter system.
58
-------
APPENDIX C
CHEMICAL ANALYSIS PROCEDURES FOR
DOUBLE ALKALI PROCESS STREAM SAMPLES
59
-------
LIST OF PROCEDURES
Method Number
1
2
3
11
12
12a
13
14
15
16
16a
17
18
19
41
42
51
52
52a
53
54
55
55a
56
57
58
59
60
62
Category
Separations
V
Solution
Analyses
Slurry
Analyses
Solids
Analyses
Title
Separation of Solids from Slurries
Separation of Solutions Containing
Suspended Solids
Solids in Filter Cakes
TOS (Total Oxidizable Sulfur)
Sulfate
Total Sulfur
Hydroxide
Acidity
Carbonate
Calcium (titrimetric)
Calcium (atomic absorption)
Sodium
pH
Chloride
Calcium
Hydroxide
TOS
Sulfate
Total Sulfur
Hydroxide
Carbonate
Calcium (titrimetric)
Calcium (atomic absorption)
Total Sodium
Acid-Insoluble Materials
Alkalinity in Lime
Calcium in Lime (titrimetric)
Carbonate in Lime
Carbonate in Soda Ash
60
-------
SAMPLES
The sample sizes and amounts of reagents are specified for process
streams having the following nominal composition:
Solutions: 0.04M TOS, 0.5M sulfate, 0.02M carbonate,
0.02M calcium, 0.1M hydroxide, O.M chloride
Solids: calcium (7 millimoles/g), TOS (2 millimoles/g),
sulfate (3 millimole/g)
For streams varying significantly from these levels, aliquots and/or
reagent volumes should be changed accordingly.
61
-------
REAGENTS
All of the tltrants listed below may be purchased either as prepared
solution or as concentrates which are diluted to volume before use.
These items can be purchased from many reliable manufacturers. We have
used Fisher Scientific and E. Merck reagents.
- 0.11J Iodine solution; should be stored in automatic buret to
minimize volatilization losses. Titer should be checked
daily versus the 0.1N_sodium thlosulfate solution.
- 0.11? Sodium Thlosulfate solution.
- 0.1N Sodium Hydroxide solution, should be stored in automatic buret
or plastic reservoir protected from atmospheric carbon dioxide.
- 0.1N Hydrochloric Acid solution
- 0.1M Ethylenediamine tetraacetic Acid (EDTA, disodium salt)
solution.
- 0.025M EDTA solution, prepared by volumetric dilution of the O.IK
solution.
- 1M Barium chloride solution, dissolve 210 g of barium chloride
crystal in 1000 ml of distilled water.
- IN Hydrochloric Acid.
- 10% Hydrochloric Acid; mix 1 volume of concentrated Reagent hydro-
chloric acid with 9 volumes of distilled water. Note: the IN
reagent may be substituted for this. ~
- IN Sodium Hydroxide solution, store in automatic buret or plastic
reservoir protected from atmospheric carbon dioxide.
- 1M Perchloric acid, prepared by diluting 1 volume of 70-72% (w/w)
perchloric acid with 10 volumes of distilled water.
- 0.1M Perchloric acid, prepared by diluting 1 volume of 1M perchloric
acid with 9 volumes of distilled water.
- 30% Hydrogen Peroxide solution - Note: we have found that Fisher
//H325 and Mallinkrodt #5240 (not stabilized) give low blank
values consistently.
- 11J Potassium Hydroxide solution, protected from atmospheric car-
bon dioxide; solution may be prepared from carbonate-free
solids or concentrated stock solution.
- 1M Sodium Perchlorate solution, prepared by dissolving 140 g of
NaCJKVH20 in distilled water and diluting to 1 liter with
same.
62
-------
Reagents (continued)
- 0.2000M Sodium Sulfate standard solution, prepared by dissolving 28.41 g
of Reagent grade Na2SOit (previously dried at 110° - 120° C for
3 hours) in distilled water and diluting to 1 liter with same.
- 0.1M_ Lead Perchlorate solution, prepared by dissolving 46.0 g of
Pb(CJlOil)2' 3H20 (Catalog No. 22118, Alfa Products Division,
Ventron Corp., Beverly, Mass.) in C02~free distilled water
(either boiled or N2 purged) and diluting to 1 liter with same.
This solution should be protected from atmospheric C02- Daily
standardization of this reagent by titration of a 15-20 m£
aliquot with the standard Na2SOit solution using the procedure
in Method #6 is required both as a check on the free lead
concentration and as a confirmation that the electrode system
is performing properly.
- 10 mg/l Manganese Chloride solution, prepared by dissolving 0.04 g of
Mn CX,2-4H20 in 1000 ml of distilled water.
- 0.015M Mercuric Nitrate solution, prepared by dissolving 5.15 g of
Hg(N03)2 H20 in 100 m£ of distilled water containing 0.5 ml
of concentrated Reagent nitric acid and diluting to 1 liter
with distilled water.
- Barium Nitrate solution, saturated; mix approximately 15 g of barium
nitrate with 1 liter of distilled water
- Calcium Sulfate. solution, saturated; mix Reagent grade calcium
sulfate dihydrate (gypsum) with tap water (room temperature) at the
rate of 3 g per liter, mix well and allow solids to settle out before
using supernate.
- Acetic acid, glacial, Reagent grade.
- Hydrochloric Acid, concentrated Reagent grade.
- Methyl Alcohol, Reagent grade.
- Phenolphthalein indicator solution, 0.1» in ethanol.
- Thymolphthalein indicator solution, 0.05% in ethanol.
- Bromocresol Green indicator solution, 0.40% in water, neutralized.
- Starch solution, either 5% or 1% in water.
- Calcein indicator solution, 0.02% in water; add potassium hydroxide
solution (5N) dropwise to dissolve indicator solids. A 0.1% dry mix-
ture of calcein in KCS, powder may also be used.
63
-------
Reagents (continued)
- Methyl red indicator solution, 0.02% in water.
- Cation exchange resin, Dowex 50Wx8, Rexyn 101x8 or equivalent, 20/50
mesh. The resin should be slurried and washed with distilled water
to remove all colored species, and the washings checked for the
absence of soluble sulfate by adding barium chloride solution to
some of the washings. No precipitate or turbidity should be present
after a 10-minute reaction time.
- Calcium chloride solution, 2.5 g of CaCH2.6H20 in 100 mil distilled
water.
- Ammonia Buffer for Magnesium Titrations; mix 6.75 g of ammonium
chloride with 57 m£ of concentrated Reagent ammonium hydroxide and
dilute to 100 mi with distilled water.
- Erlochrome Black T Indicator Solution; 0.5% in ethanol.
64
-------
SPECIAL APPARATUS
For carrying out the lead titratlons for determination of sulfate, the
following apparatus is required:
- Lead-Sensitive Specific Ion Electrode, Model 94-82, Orion Research.
Inc., Cambridge, Mass.
- Double-junction Reference Electrode, with sodium nitrate outer bridge
solution so that no chloride or sulfate can flow into the titration
solution; Model 90-02 from Orion Research Inc. is recommended.
- pH/millivolt meter, readable to ± millivolt.
- Microburet, 10 ml capacity, calibrated at each 0.1 mS, increment for
addition of the 0.2M sodium sulfate titrant.
Note: If available, an automatic titrator or motor-driven buret
assembly which can be set up to record the titration curve
will be extremely useful.
SAMPLE PREPARATION AND SEPARATION
Note: All solids samples taken from slurries are used as a measure of
solids content in the slurry as well as for composition of the
solids. The solids sample should be taken from the slurry
before the solution sample is order to insure that a representa-
tive portion is obtained.
Method #1 - Solids in Slurries
1. Rapidly transfer a 250 mi (±2 mi) portion of the well-mixed reactor
slurry (or 50 mi of settler underflow) into a 250 mi graduated
cylinder. Pour this portion onto a fritted-glass filter crucible
(coarse porosity), and start the filtration. The amount of slurry
taken may be changed, if necessary, in order to yield at least 2 g
of dry solids.
65
-------
2. Transfer 25 mi of wash solution (saturated calcium sulfate solution)
into the sample graduate. Just as the liquid level in the filter
reaches the top of the solids bed, pour the wash solution onto the
solids. In order to avoid possible channeling and inefficient
washing, it is extremely important that the liquid level not be
allowed to go beneath the surface of the solids.
3. Repeat the displacement wash in Step 2 three times more.
4. Transfer the solids quantitatively to a tared drying dish and dry
at 83-85°C to constant weight. This drying step usually requires
3-4 hours in a forced draft oven.
5. Allow the dried solids to cool, then weigh to ± 0.01 g and transfer
to a resealable plastic bag for storage. The solids must be ground
well prior to analysis.
6. Calculations:
Reactor Slurries: % Solids (g/100 mi) = 0.4 x g of dry solids
Settler Slurries: % Solids (g/100 mi) « 2 x g of dry solids
Method #2 - Solutions Containing Suspended Solids
1. Transfer portions of the slurry into 4 or 6 15 ml centrifuge tubes
(to give a total volume of 50-60 mi) and spin until the supernate
is clear. This requires approximately 5 minutes in a small labora-
tory centrifuge (3600 rpm).
2. Carefully decant the clear supernate into a 60 m£ polyethylene
bottle for storage.
Method #3 - Solids in Filter Cake
1. Dry approximately 20 g (weighed to ± 0.01 g) at 83-85°C to constant
weight. This usually requires 3-4 hours in a forced draft oven.
2. Allow the dried solids to cool, then weigh to ± 0.01 g and transfer
to a resealable plastic bag for storage. The solids must be ground
well prior to analysis.
3. Calculations:
% Solids = * drl 8011?8 x 100
g wet sample
66
-------
Method #11 - Total Oxidlzable Sulfur (TOS)
1. Transfer 25 mi (measured to ±.02 mi) of 0.1 I2 solution into a
125 m£ Erlenmeyer flask and cover with a watchglass.
Note: Experience may show that some sample streams require
less iodine.
2. Add approximately 20 mi of distilled water and 2 mi of glacial
acetic acid. In lieu of acetic acid, 15 mi of 0.1N HC£ may be
used.
3. Pipet 10 mi of sample (volumetric pipet) directly into the acidified
iodine solution while swirling the solution to assure rapid mixing
of the sample. If all iodine color is discharged, repeat the pro-
cedure using more I2 solution and adjust calculations accordingly.
4. Back titrate with 0.1N Na2S203 to pale yellow color. Add starch
solution to give blue color and continue dropwise addition of
titrant to disappearance of blue color.
5. Calculations:
TOS (moles/0 = ^ (M ^ ( N I2) ]-[(m£ S203) ( N S203) ]
(2) (10 mi)
where: If I2 is the normalty of the I2 solution
N, S203 is the normality of the Na2S203 solution
2 is the equivalence factor for moles of oxidizable sulfur
67
-------
Method #12 - Sulfate
Note: If only an approximate value for sulfate Is needed for samples
containing low TOS and carbonate, the determination may be
carried using the sample aliquot plus 60 m£ of distilled water
starting with Step 5.
1. Set up sparging/boiling apparatus as shown in Figure C-l and add
30 ma of distilled water and 10 mj, of 1M perchloric acid to the
Erlenmeyer f lask(s) .
2. Heat the solution nearly to boiling and purge with a gentle
nitrogen flow for 3 minutes to remove dissolved oxygen. Allow to
cool slightly.
3. Pipet a 2 mj, aliquot of sample into the flask.
4. Heat to boiling and purge with N2 for 15 minutes. Note; A Pasteur
pipet or bubbler tube may be substituted for the gas dispersion
tube in which case, the N2 purge should be continued for 30 minutes,
5. Allow to cool and transfer purged solution into a 150 m£ beaker.
Add 2 drops bromoctesol green indicator solution and neutralize
solution with IN NaOH adding one drop of 1M HCjj.0^ after indicator
has turned green.
6. Add 10 m'£ of 1M NaC£04 solution and enough methanol to make a 50%
(v/v) solution.
7. Under rapid stirring, add 25 mj, of standardized (0.1M) lead perch-
lorate solution. The amount of lead added should be such that an
excess of lead remains in solution. If necessary readjust pH of
solution to 3.5-4.5 with ECi.Ok or NaOH.
8. Back titrate the excess lead with 0.2M NazSO^ solution using the
EMF between the lead-selective specific ion electrode and the
double- junction reference electrode to indicate the endpoint.
Titrant increments . at the endpoint should be 0.1 mi,. The end-
point region can be seen from the daily standardization curve, and
may vary due to changes in the condition of the lead electrode.
The endpoint can be determined graphically or by means of the
second derivative technique.
9 . Calculations :
Sulfate (moles/£) =
t(m£ lead perchl.)(M lead perchl.)] - [(ml ^280^) (M Na2SOit) ]
m£ of sample
68
-------
N2 manifold
clamp
Pasteur plpet or
dropping pipet tip
r
I
screw clamp
vinyl tubing
rubber stopper, notched to allow
gas to escape
50 mfc Erlenmeyer flask
Hot plate
or hot water bath
Figure C-l
APPARATUS FOR SULFATE DETERMINATION
69
-------
Method //12a - Total Sulfur (Gravimetric)
1. Transfer 2 m& of IN NaOH into a 100 mSL beaker, and add approximately
10 mil of distilled H£0 and a drop of phenolphthalein solution.
2. Plpet a 2 mi aliquot of sample into the NaOH solution. If necessary,
add more NaOH to give a pink color to solution.
3. Add (graduated cylinder 'Or pipet) 5 mi of 30% hydrogen peroxide solu-
tion. Swirl to mix, cover beaker and allow to stand for 15 minutes.
4. Transfer to a 400 mi beaker and dilute to approximately 250 mi with
distilled water.
5. Add 10% HCJl dropwlse until disappearance of the pink color. Then add
3-4 drops of methyl red indicator and continue dropwise addition of
10% HCJl until the solution turns red.
6. Heat on hot plate nearly to boiling, and then add 25 mil of 1M barium
chloride solution while stirring the solution rapidly. Make certain
that the solution is still acid (red) to methyl red Indicator.
7. Allow precipitate to settle and test for completeness of precipita-
tion with a few drops of the barium chloride solution. If necessary
add more barium chloride solution in 5 mi Increments, testing for
completeness of precipitation after each addition.
8. Digest sample hot (below boil) for at least 3 hours.
9. Gravity filter the hot solution and precipitate through Whatman //42
(or equivalent) filter paper. Use a rubber policeman to effect
quantitative transfer of precipitate.
10. Test for completeness of washing by adding a drop of 0.1N silver
nitrate solution to a few drops of wash liquor from the tip of the
funnel stem. If only a very faint turbidity appears, washing is
complete.
11. Transfer the filter paper and precipitate into a preignited crucible.
Place in a cold muffle furnace and ignite at 850°C for three hours.
12. Cool crucible in dessicator and weigh.
13. Calculations:
Total S (moles/i) = (8 BaS(V
(2 mO (0.2334)
where: 0.2334 is the millimolecular weight (g/millimole) of BaSOi*
70
-------
Method #13 - Hydroxide
1. Pipet a 10 m£ aliquot of sample into a 125 m£ Erlenmeyer flask. Add
approximately 25 ml of distilled water, 10 mi of 2.5% calcium
chloride solution, and 2-3 drops of thymolphthalein.
2. Titrate with o.lN HC£ to the disappearance of the blue color.
Note; if too much indicator has been added, the endpoint is seen
as a marked decrease in the intensity of the blue color, rather
than complete disappearance.
3. Calculations:
Hydroxide (moles/£) = <"* HC&> CS HC£)
10
Method #14 - Acidity
1. Transfer 15 m£ (± ,01m£) of O.lN NaOH solution into a 125 m£ Erlen-
meyer flask, and add 2-3 drops of thymolphthalein indicator solution.
2. Pipet a 10 mi aliquot of sample into the NaOH solution and mix well.
3. If solution remains blue, titrate with O.lN HC£ to disappearance of
blue. Note; if too much indicator has been added, the endpoint is
seen as a marked decrease in intensity of the blue color when 1-2
drops of HC£ are added.
A. If indicator color disappears during addition of sample and does not
reappear upon mixing, repeat the determination using a larger aliquot
of O.lN NaOH.
5. The exact IJ of the NaOH is determined by titrating a known volume
(15m£, ±.01 m£) with O.lN HC£ to the thymolphthalein endpoint. Then
N NaOH = __ x (N Q£
- m£ NaOH -
6. Calculations:
Acidity (moles/£) - [.<>£ NaOH) (N NaOH) ]- [(m£ HC£)( N HC£) ]
10 m£
where: N NaOH - the exact normality determined above.
71
-------
Method #15 - Carbonate
Note; Once the determination is started, it must be carried through
Step 7 (acid to bromocresol green) in order to avoid error from
absorption of atmospheric C02.
1. Transfer 25 mi of 0.1N NaOH into a 125 mi Erlenmeyer flask and add
1 drop of phenolphthalein indicator solution.
2. Pipet a 10 mi aliquot of sample solution into the sodium hydroxide.
3. If the phenolphthalein color disappears immediately, add an addi-
tional 10 mil of NaOH solution. The pink (alkaline) indicator color
should persist for at least 30 seconds.
4. Add 5 mi of 30% hydrogen peroxide, mix well and allow to stand for
10 minutes. Add 1 mi of the 10 ing/liter manganese solution, and
boil until effervescence ceases.
5. Cool quickly in water bath. If pink color has faded, add 1 drop of the
phenolphthalein solution to ascertain that solution is still alkaline.
6. Titrate with 0.1N ECi to disappearance of pink color. Note; If
too much indicator has been added, the endpoint is seen as a marked
decrease in intensity of the red color. Note the volume (level) of
0.1N HCJ, in the buret, and call this "A".
7. Add 3-4 drops of bromocresol green indicator solution to the titra-
tion solution, and continue titration with O.IN^ HCS, until a perma-
nent yellow color is seen. Then add 3 mi of titrant in excess.
Note the reading of the HCJt buret and call this "B".
8. Quantitatively transfer the titrated solution to a 150 m£ beaker and
boil (uncovered) for 10 minutes. Note; It may be necessary to add
small amounts of distilled water during the boiling in order to
avoid spattering losses.
9. Cool the solution, and back-titrate with 0.1N NaOH to a green end-
point. Call this volume "C".
10. Run at least one blank determination on all reagents.
11. Calculations:
Millimoles carbonate = [ (B-A) x N HCJ,] - [C x N NaOH]
_ . , , ... (millimoles COo in sample) - (millimoles COq in blank)
Carbonate (moles/£) = r
10 in*
where: 1? HC£ = normality of ECi
N. NaOH = normality of NaOH to bromocresol green
72
-------
Method //16 - Calcium via EDTA Titration
1. Pipet a 25 mi aliquot of sample into a 250 mi Erlenmeyer flask, and
add 35 mi distilled water.
2. Add 10 ml of 10% HCj, and mix well. Then either boil or sparge with
nitrogen gas for 10 minutes.
3. Add 0.1N_ I 2 solution until an excess (permanent pale yellow color)
is present. Then add 0.1IJ Na2S203. solution dropwise until the yellow
color is just discharged. Avoid adding an excess of
4. Add 15 mi of IN KOH followed by 15 mi of 1% KCN. Note; The potas-
sium hydroxide must be added before the cyanide solution.
5. Add a small amount of the calcein indicator (either 1-2 drops of
solution or a few milligrams of dry mixture) sufficient to give a
visible green fluorescence. Illumination with longwave (360 run)
ultraviolet light will greatly enhance the fluorescence. As with most
most indicators, the amount added should be only the minimum neces-
sary to observe the fluorescence. A large excess of indicator will
result in a poorly defined endpoint.
6. Titrate with 0.025M EDTA solution to the disappearance of the fluo-
rescence. If a large amount of calcium is present, the fluorescence
may reappear after a few seconds due to dissolution of precipitated
calcium hydroxide. The true endpoint is stable for at least 2
minutes. Note; Use of an ultraviolet light allows this indicator
to be used for extremely turbid solutions, if necessary.
7 . Calculations :
„ , , , , /nN (ma EDTA) (M EDTA)
Calcium (moles /£) = _ _ ' *— _ '
25
Method #16a - Calcium via Atomic Absorption Spectrometry
Note; This method can be used as the routine method if the analyst has
easy access to an AA spectrometer. The details of operation of
the instrument are found in the instruction manual for that
instrument.
1. Dilute a 5 mi aliquot to 250 mi with distilled water. Call this
Solution I;
2. Dilute a 5 mi aliquot of Solution I to 50 mi with distilled Water.
Call this Solution II.
3. Dilute 5 mi of Solution II plus 1 mi of 1% LaCl3 (in 2.5% HC1)
to 10 mi with distilled water. Call this Solution III.
73
-------
3. Calcium calibration standards should be prepared (including the lan-
thanum solution) and run in order to bracket the sample concentration.
The usual range for the measured solution is from 0-10 milligrams of
calcium/ liter.
4. Measure the absorbances of Solution II and of standards on the atomic
absorption spectrometer at the most sensitive calcium wavelength of
422.7 nanometers.
5. Determine the calcium concentration (in milligrams/A) in Solution II
from the calibration data.
6. Calculations:
Calcium (milligram/it) = D x - x - x ^ = D x 100
CaUl- <-°le/«> * " * ¥ ' " ' %5
where: D is the concentration (milligram/&) of calcium in Solution II,
40 is the atomic weight (gram/mole) of calcium
Method #17 - Sodium via Atomic Absorption Spectrometry
Note; (a) All glassware used in these measurements must be cleaned
in dilute (4M) nitric or hydrochloric acid before use in
order to minimize sodium contamination.
(b) In order to lessen the effect of contamination with extra-
neous sodium, the measurements are made at 330 nanometers,
a relatively insensitive line.-
(c) The operational details of the particular instrument are
found in the instrument instructional manual.
1. Dilute a 5 ma aliquot of sample to 100 ma with distilled water. Call
this Solution I.
2. Dilute 10 ma of Solution I to 100 ma with distilled water. Call
this Solution II.
3. Prepare appropriate sodium calibration standards in distilled water.
The usual range if from 0-200 milligrams of sodium/liter.
4. Measure the absorbances of Solution II and of the standards on the
atomic absorption spectrometer at the 330 nanometer (secondary)
sodium line.
74
-------
5. Determine the sodium concentration in Solution II from the calibration
data.
6. Calculations:
Sodium (moles/ ) = (23) (i000) * ^ * ^ = HJ
where: E is the concentration (milligram/A) of sodium in Solution II,
23 is the atomic weight (gram/mole) of sodium
Method #18 - Measurement of pH
The pH can be measured with any glass-reference electrode set provided
that the following points are observed:
1. The glass electrode must be clean and free of cracks and other marks.
A wide range (0-14 pH) industrial-quality electrode is best. A new
glass electrode should be soaked in water (or pHA buffer solution)
for 1/2 hour before use.
2. The reference electrode must be filled with the appropriate electrolyte.
Electrical conductivity through the liquid junction must be confirmed.
A sleeve-type junction is best for this purpose, although the electro-
lyte level must be checked frequently since the electrolyte leakage
rate tends to be high with this junction.
3. The electrode pair must be set (i.e. standardized) using a standard
buffer reasonably near the expected sample pH, and cross-checked
against a buffer beyond the expected sample pH. This latter check
is extremely important to prove the linearity of electrode response
in the region of interest. Provided that fresh buffers are used,
the indicated pH should be within 0.1 pH unit for the check (second)
buffer.
Method #19 - Chloride
1. Pipet a 10 mj, aliquot of sample into a 125 mfc Erlenmeyer flask.
2. Add 10m£ of distilled water, 2 drops of phenolphthalein indicator
solution and sufficient IN NaOH to give a red color.
3. Add 2 ml 30% H202, mix and let stand for 10 minutes.
4. Add 1 m£ of the 10 mg/fc solution of manganese. (The amount of chloride
added is neglible, equivalent to only 4 x 10~5M chloride in the sample.)
75
-------
5. Heat solution to boiling for approximately 15 minutes to destroy the
peroxide, adding more distilled water if necessary to maintain liquid
level. Note; Absence of peroxide is indicated by change in the
boiling (gas evolution) character, or by failure to bleach the color
when another drop of indicator is added to the solution.
6. Cool the solution to room temperature, add 3-4 drops of bromocresol
green indicator and bring just to the green color with 1M perchloric
acid (HCfcOi^). Then make yellow-green with 0.1M perchloric acid (and
0.IN NaOH, if necessary).
7. With rapid swirling, add 20 mJl of saturated Ba(N03)2 solution and
sufficient bromocresol green indicator to see the color. Add 0.1M
perchloric acid and 0.1N NaOH to achieve a yellow-green color.
8. Add 10 drops diphenylcarbazone indicator solution and titrate with
0.015M Hg(N03)2 solution to a faint purple endpoint which is stable
for at least 30 seconds.
9. Calculations:
Chloride (moles/A) - (m£ Hfi Titrant)(M of titrant)(2)
10 mi
10. Standardization of Hg(N03)2 Titrant: Titrate a 2 mi aliquot of
0.1N_ HCA using the procedure outlined above.
M of n^^i- - (2 mA) (N HCA)
M or titrant = -7 .' ^ *. ,LN
— (m£ titrant)(2)
Method /Ml - Calcium In Lime and Limestone Slurries
Note; Sample aliquot should be weighed if analyses are to be expressed
as weight percent. The method is set up for 10% (w/v) slurries;
sample size and/or dilution may have to be modified for other
concentration.
76
-------
1. Pipet a 10 mi aliquot of well-mixed slurry into a 250 mi beaker. Add
50 mi distilled water and 1-2 drops of phenolphthalein solution.
2. Add sufficient 10% HCi to neutralize the alkalinity. Then add 10 mi
more of 10% ECi for lime slurries, or 25 mi more of 10% RCi for lime-
stone slurries. Warm the solution, if necessary, to dissolve all
solids (some silica may remain undissolved).
3. Dilute to 250 mi in a volumetric flask with distilled water.
4. Pipet a 25 mi aliquot into a 100 mil beaker, and boil for 15 minutes
adding distilled water to maintain liquid level.
5. Continue tltration as in Method #16 using 0.1M EDTA tltrant.
6. Calculations:
Calcium (mole/0 = /«x / , /»%
Ca(OH)2(mole/£) =
(5 mi) (2)
77
-------
Solids Analyses
Note: Solids analyses are always carried out on dried, ground solids
unless otherwise noted.
Method #51 - Total Oxidizable Sulfur (TOS) (Iodine Titration)
1. Weigh out from 0.18 g to 0.2 g (weighed to ±0.0001 g) of solids
on a weighing paper or boat.
2. Transfer 25.00 m£ of 0.1N-I2 solution into a 250 m£ Erlenmeyer
flask. Add approximately 75 m£ of distilled water and 10 ml of
10% HCfc.
3. Quickly transfer the weighed solids to the iodine solution and
cover the flask with a watchglass.
A. Stir the suspension Until all of the solids have dissolved. A
magnetic stirrer is useful for this operation. Break up any lumps
with a glass rod. Note; In solids with a very high CaSO^ content,
not all of the solids will dissolve in the acid. In general all
solids should dissolve in 30 minutes.
5. As soon as all solids have dissolved, back titrate the excess I2
with 0.1N Na S203. When the solution has reached a pale yellow
color, add starch solution to give a blue color. Continue the
titration dropwise until the blue color is discharged.
6. Carry at least one reagent blank through the process, using the
same length of stirring time as for the sample. This will help to
ascertain whether air oxidation is causing an interference in the
method. The blank titer (mfc Na2S203 per mi of I2 taken) should be
within 0.005 of the standardization titer.
7. Calculations:
TOS (millimoles/g) = (m& T2 x 11 T2> - 0»* S203 x N S203)
(2)(g sample)
where: JN 12 and JJ 8263 are the normalities of the I2 and Na2S203
solutions.
2 is the equivalence factor for moles of oxidizable sulfur.
78
-------
Method #52 - Sulfate In Solids
1. Set up the sparging/boiling apparatus as shown in Figure A-l, and
add 30 rox, of distilled water to the Erlenmeyer f lask(s) .
2. Heat the water to boiling and purge with a gentle nitrogen flow for
3 minutes to remove dissolved oxygen.
3. Transfer 0.2 g to 0.3 g (weighed to ±0.0001 g) of sample to the
purged water.
A. Purge with nitrogen for 3 minutes more, then quickly add 10 mi of
1M perchloric acid (restopper flask quickly) to dissolve sample.
5. When all solids have dissolved, bring to a boil and purge for at
least 30 minutes (no less). Then cool solution.
6. Transfer the solution to a 150 mSL beaker, add 2 drops bromocresol
green indicator and neutralize (to within 1 drop on the acid, yellow,
side) with 1M NaOH and 1M HC«V
7. Add 5 grams of washed cation exchange resin, and mix (magnetic
stirrer) for 30 minutes.
8. Filter the solution and resin through a one (1) inch thick bed of
the same resin contained in a sintered glass filter crucible, and
wash through with distilled water. The total volume of solution
plus washes should be less than 100 mi. Note; This ion-exchange
operation can also be carried out on a column containing 10 gm of
resin.
9. Allow to cool and transfer purged solution into a 150 'mi beaker.
Add 2 drops bromocresol green indicator solution and neturalize
solution with IN NaOH adding one drop of HCiO^ after indicator has
turned green.
10. Add 10 mi of 1M NaCiO^ solution and enough methanol to make a 50%
(v/v) solution.
11. Under rapid stirring, add 25 mi of standardized (0.1M) lead perchlo-
rate solution. The amount of lead added should be such that an
excess of lead remains in solution. If necessary readjust pH of
solution to 3.5-4.5 with HCiOi, or NaOH.
12. Back titrate the excess lead with 0.2M Na2SOit solution using the EMF
between the lead-selective specific ion electrode and the double-
junction reference electrode to indicate the endpoint. Titrant
increments at the endpoint should be 0.1 mi. The endpoint region
can be seen from the daily standardization curve, and may vary due
to changes in the condition of the lead electrode. The endpoint
can be determined graphically or by means of the second derivative
technique .
79
-------
13. Calculations:
Sulfate (millimoles/g) =
[(m£ lead perchl.)(M lead perchl.)] - [(m£ Na2S01+)(M Na2S0lt) ]
g sample
Method f?52a - Total Sulfur (gravimetric)
•1. Weigh out from 0.13 g to 0.15 g (weighed to ± 0.0001 g) of solids
and transfer them into a 150 m£ beaker. Add 25 to 50 mj, of dis-
tilled water and 1 drop of phenolphthalein indicator solution.
2. While stirring the suspension, add O.IN^ NaOH dropwise until a per-
manent pink color is achieved.
3. Add 5 m£ of 30% hydrogen peroxide solution, mix well and allow to
stand for 15 minutes.
4. With constant stirring add 0.1N HC£ dropwise until the pink of
phenolphthalein disappears. Then add 2-3 drops of methyl red
indicator solution.
5. Continue the stirring and the dropwise addition of 0.1N HCA until
the solution turns red. Then add a few drops more of the 0.1N
HC£ slowly.
6. If solids do not dissolve after 10 minutes, add a few drops more of
the 0.1N HC£. Continue stirring until all solids have dissolved.
7. Add 2 g of the washed ion exchange beads, and allow the beads to
remain in the solution for 1 hour with frequent stirring.
8. Decant as much liquor as possible into a 400 mi beaker, and filter
off the beads on a sintered glass Buchner funnel. Wash the beaker
and beads with distilled water, and add the washings to the decanted
liquor.
9. Dilute the solution to 250 ml with distilled water, and carry out
the precipitation and ignition as in Method #6a, starting with
Step 6.
10. Calculations:
Total Sulfur (millimoles/g) - 8 BaSO|t
(g sample)(0.2334)
where: 0.2334 is the millimolecular weight (g/millimole) of BaS04
80
-------
Method 053 - Hydroxide
1. Weigh out from 0.50 g to 0.55 g (weighed to ±0.0001 g) of sample
and transfer to a 125 mfc Erlenmeyer flask containing 40 mfc of
distilled water and magnetic stirring bar.
2. Add 2-3 drops of thymolphthalein indicator solution and 10 m£ of
2.5% calcium chloride solution.
3. Titrate with 0.1N HC£ to the disappearance of the blue color. The
indicator color may reappear on continued stirring. The titration
should be continued until the endpoint holds for at least one
minute. Note; If too much indicator has been added the endpoint
is seen as a marked decrease in the intensity of the blue color.
4. Calculations:
,j / .,11., T / \ (m* HC£)(fl HCfc)
Hydroxide (millimoles/gram) = TV— ' —
g sample
Method #54 - Carbonate
Because calcium is present in the solids, the carbonate is separated
and measured as C02 after first oxidizing all sulfite species to sulfate.
Note; Once the determination is started, it must be carried through
Step 13 (acid to bromocresol green) in order to avoid error from absorp-
tion of atmospheric C02.
1. Set up the apparatus shown schematically in Figure C-2.. Three midget
impingers or midget bubblers in series will give a 98% collection
efficiency for C02.
2. Just prior to performing the analysis, place 50 m& of 0.114 HCfc and
3 m£, of methyl red indicator solution into the addition funnel.
Also transfer 15 mfc of 0.5N NaOH into each of the midget impingers.
3. Transfer between 1.0 g and 1.1 g (weighed to ±0.0001 g) of the solids
into the reactor, and add a magnetic stirring bar, and 50 mfc of dis-
tilled water.
4. Add 1 drop of phenolphthalein solution, and then add 0.1N NaOH drop-
wise until a permanent faint pink color is seen in solution.
5. Add 5 ms, of 30% hydrogen peroxide and immediately seal the system.
Then start the magnetic stirrer.
6. Start the sparging system and adjust to give an air flow of approxi-
mately 2 liters/minute.
7. Add the 0.IN HCfc slowly into the reactor unitl...the yellow color (fiom
the methyl red) turns red.
81
-------
Flow Meter
To vpaucjpm
125ml
Addition Funnel
Midget Impinger
250 ml Flask
Gas Dispersion Tube
0""
Magnetic Stirrer
Figure C-2
APPARATUS FOR CARBONATE IN SOLIDS
82
-------
8. Continue stirring and sparging until all the solids dissolve. It may
be necessary to add more the of 0.1N HC£ dropwise in order to achieve
complete solution.
9. Once all solids have dissolved, continue the purge for 15 minutes more.
10. Transfer the impinger solutions to a 150 mfc beaker washing the im-
pinger with distilled water to effect quantitative transfer.
11. Add 1 mjl of 30% H202, mix and let stand 15 minutes (covered).
12. Add 1 m£ of the 10 mg/liter manganese solution and boil the solution
(covered with a watchglass) for 5 minutes. Cool the solution to near
room temperature quickly (to minimize exposure to atmospheric C02).
13. Add to the impinger solution, 1 drop of phenolphthalein solution and
5 mSL (pipet) of IN ECU.
14. Titrate with 0.1N HC£ to disappearance of pink color. Note; if too
much indicator has been added, the endpoint is seen as a marked
decrease in intensity of the red color. Note the volume (level) of
0.1N HC in the buret, and call this "A".
15. Add 3-4 drops of bromocresol green indicator solution to the titra-
tion solution, and continue titration with 0.1N HC£ until a permanent
yellow color is seen. Then add 3 mi of titrant in excess. Note the
reading of the ECU buret and call this "B".
16. Quantitatively transfer the titrated solution to a 150 mi beaker and
boil (uncovered) for 10 minutes. Note; It may be necessary to add
small amounts of distilled water during the boiling in order to
avoid spattering losses.
17. Cool the solution and back-titrate with 0.1N NaOH to a green endpoint.
Call this volume "C".
18. Run at least one blank determination Including sparging for the same
amount of time.
19. Calculations:
Millimoles carbonate - [(B-A) x N HCi] - [C x N NaOH]
Carbonate (millimoles/gram =
(millimoles 0)3 in sample) - (millimoles COa in blank)
g sample
where: N HCS. = normality of HC£
N NaOH = normality of NaOH to bromocresol green
83
-------
Method #55 - Calcium (titrimetric)
1. Weigh out 0.13 g to 0.14 g (weighed to ± 0.0001 g) of sample and
transfer into a 250 mfc beaker. ;
2. Add 10-15 mil of distilled water and 5 m£ of 10% HC£, and mix well
to dissolve solids. Then add 60 m£ more of distilled water.
3. Boil or purge with nitrogen gas for 10 minutes.
4. Add O.DJ I 2 solution until an excess (yellow color) is present.
Then add 0.1N Na2S20s solution dropwise until the yellow color is
just discharged. Avoid adding an excess of
5. Add 15 m«, of IN KOH followed by 15 m* of 1% KCN. Note; The potas-
sium hydroxide must be added before the cyanide solution.
6. Add a small amount of the calcein indicator (either 1-2 drops of
solution or a few milligrams of dry mixture) sufficient to give a
visible green fluorescence. Illumination with longwave (360 run)
ultraviolet light will greatly enhance the fluorescence. As with
most indicators, the amount added should be only the minimum
necessary to observe the fluorescence. A large excess of indicator
will result in a poorly defined endpoint.
7. Titrate with 0.1M EDTA solution to the disappearance of the fluo-
rescence. If a large amount of calcium is present, the fluorescence
may reappear after a few seconds due to dissolution of precipitated
calcium hydroxide. The true endpoint is stable for at least 2
minutes. Note; Use of an ultraviolet light allows this indicator
to be used for extremely turbid solutions, if necessary.
8 . Calcula tions :
„ , .,,, , / N (ma EDTA) x (M EDTA)
Ca (millimoles/g) = —
g sample
Method #55a - Calcium (Atomic Absorption Spectrometry)
Note; This method can be used when the analyst has easy access to an
AA spectrometer. The details of operation of the instrument
can be found in the particular instrument manual.
1. Weigh out 0.10 g to 0.11 g (weighed to 0.0001 g) of sample, transfer
into a 150 mj, beaker and add 25 m£ of 10% HC i.
2. Stir to dissolve all solids and then dilute to 250 m£ in a volumetric
flask with distilled water. Call this Solution I.
-------
3. Mix 5 m£ of Solution I with 5 mfc of 1% LaCJh (in .2.5% HC*0 and
dilute to 50 m£ with distilled water. Call this Solution II.
4. Calcium standards should be prepared (including the lanthanum
solution) and measured in order to bracket the sample concentration.
The usual range in the measured solution is from 0-10 milligrams of
calcium per liter.
5. Measure the absorbances of Solution II and of the standards on the
atomic absorption spectrometer at the most sensitive calcium wave-
length, 422.7 nanometers. Either an air/acetylene or a nitrous oxide
flame may be used, although the latter gives a more linear calibration.
6. Determine the calcium concentration (in milligrams/liter) in
Solution II from the calibration data.
7. Calculations:
Calcium (millimoles/g) »
fg sample \ (0.040) (100 mg/g) (1000 m£/£) /5_\
250 m ) (SO)
B E(0.0625)
(g sample)
where: 0.040 is the milliatomic weight (g per millimole) of calcium
E is the measured calcium concentration (mg/A)
Method #56 - Total Sodium
Notes; (a) All glassware used in these measurements must be cleaned
in dilute (3 or 4 M) nitric or hydrochloric acid before use
in order to minimize sodium contamination.
(b) In order to lessen the effect of contamination with extra-
neous sodium, the measurements are made at 330 nanometers,
a relatively insensitive line.
(c) Operational details of the particular instrument are found
in the instrument instruction manual.
1. Weigh out 1.0 g to 1.2 g (weighed to ± 0.001 g) of sample and
dissolve in 25 m«. of 10% HC£ plus approximately 100 mi of distilled
water. Then dilute to 250 mfc with distilled water.
2. Prepare appropriate sodium calibration standards in distilled water.
The usual range is from 0-200 milligrams of sodium/lifer.
85
-------
3. Measure the absorbances of the diluted sample and of the standards
on the atomic absorption spectrometer at the 330 nanometer sodium
line.
4. Determine the sodium concentration (C) in the sample solution from
the calibration data.
5. Calculations:
Total Sodium (weight percent) = / C \ / 250 \ / 1 \ /
(g sample ) (lOQo) (^lOOoJ ( 10°
where: C is the concentration (milligram/liter) of sodium in the
diluted sample
Total Sodta 0-lli-Wg) - (j^jj) (i|jf) (o^Tso) (lo5o)
where: C is the concentration of sodium in the measured solution
0.0230 is the milliatomic weight (g/millimole) of sodium
Method #57 - Magnesium (titrimetric)
Note; This test is an empirical test designed to provide an estimate of
insoluble materials (principally silicates) coming both from the
lime and from the flyash.
1. Transfer between 3 g and 7 g (weighed to + 0.001 g) of ground
sample into a 250 m£ beaker.
2. Add 100 ml of 10% HC1, and stir for 30 minutes.
3. Filter through a medium porosity sintered glass filter funnel that
has been previously dried and tared under the same conditions as
used for the sample.
4. Wash the acid-insolubles with 50 m£ of distilled water.
5. Dry to constant weight (2-3 hours) at 105° - 110° C and obtain
(net) weight of dried acid-insolubles.
6. Calculations:
IT j I... T, *. A jj T i ui g dry insolubles ,nn
Weight Percent Acid Insolubles = •a z : x 100
6 g sample
86
-------
Method #58 - Available Alkalinity in Lime
1. Weigh put between 0.3 g and 0.35 g (weighed to 0.0001 g) of well-
mixed sample and transfer to an Erlenmeyer flask containing approxi-
mately 100 atS, of distilled water.
2. Add 1 drop of phenolphthalein indicator solution and titrate with
11* HCS. to the permanent disappearance of the pink color. The solu-
tion should remain colorless for at least 3 minutes at the endpoint.
Note; It will be necessary to stir the slurry well in order to
dissolve all of the lumps of lime. Addition of 1-2 drops of thymol-
phthalein will result in a violet to pink color change just prior
to the disappearance of the pink phenolphthalein color, which may
aid the observation of the endpoint.
3. Calculations:
/ \ » ..ii-i 11 ,j j , .,-, * i «»-/ N (mfc HC«,)(N of HC£)
(a) Available alkalinity (as millimoles OH /g) = _ — _ '
g sample
(b) Available alkalinity (expressed as
%Ca(OH)2) = (m£ HC&) (N HCA) (.0741)
(g sample)(2)
where: 0.741 is the millimolecular weight of Ca(OH)2
Method #59 - Total Calcium in Lime
1. Transfer between 0.10 g and 0.12 g (weighed to ± 0.0001 g) of well-
mixed sample into an Erlenmeyer flask containing 50 m£ of distilled
water.
2. Add 5 mx, of 10% HCS, and stir to dissolve all solids.
3. Continue determination as in Method f/55, Step 5, except that the
KCN may be omitted.
4. Calculations:
n-ii f mi- i i \ 0°* EDTA)(M of EDTA)
Calcium (millimole/g) - ^__ 7 v— '
g sample
87
-------
Method #60 - Carbonate in Lime
Use Method #54 (Carbonate in Solids) with a 1.0 to 1.1 g sample,
except that Step 5 (addition of H202) may be omitted. Calculations
are the same.
Method #62 - Carbonate in Soda Ash
1. Transfer between 0.11 g and 0.12 g (weighed to 0.0001 g) of well-
mixed sample into an Erlenmeyer flask containing 60-70 mSL of distilled
water.
2. Add 1 drop of phenolphthalein indicator solution and titrate with 0.1N
HC£ to the disappearance of the pink color. Call this volume of tit-
rant "A".
3. Add 2-3 drops of bromocresol green indicator solution, and continue
the titration with 0.1N HCjl until a yellow color is reached. Then add
2-3 ran more of titrant. Call this volume of titrant (from the phenol-
phthalein endpoint on) "B".
4. Transfer the titrated solution to a beaker and boil for 10 minutes
or so. Cool the solution.
5. Back titrate with 0.1N NaOH to a green endpoint. Call this volume
of NaOH "C".
6. Calculations:
Total carbonate in soda ash (as % Na2C03) =
[(B x N HCjl) - (C x N NaOH)] x 0.106 x 100
g sample
where: 0.106 is the millimolecular weight of Na2C03
Excess alkalinity in soda ash (as % NaOH) =
{(A x N HC*) - [(B x N HC ) - (C x N HaOH)]} x MQ x 1Q(J
g sample
where: 0.040 is the millimolecular weight of NaOH
88
-------
APPENDIX D
SYSTEM ECONOMICS
89
-------
APPENDIX D
SYSTEM ECONOMICS
Based on operating results to date, General Motors has provided Table D-l
which summarizes the costs of constructing and operating the Chevrolet-
Parma Double Alkali S02 control facility.
90
-------
Table D-l
SYSTEM ECONOMICS
Total Capital Cost = $3.2 Million (1975$)
Operating Costs/Ton Coal = $11.70
Scrubber Operating Costs (1975$)
$/Ton of
Component
Lime
Soda Ash
Water
Electricity
Solid Waste Handling
Labor
Maintenance
Capital Charge
Total
Total Annual Costs = $643,500
11.73
Mils/Kwh
0.98
0.82
0.16
1.32
0.81
2.04
1.05
4.55
0.41
0.34
0.07
0.55
0.34
0.85
0.44
1.90
4.90
(1)
(2)
(3)
Based upon burning 55,000 TPY coal (i.e., approximately 55%
load factor).
Supplied by General Motors.
Based on equivalent heat rate of 10,000 Btu/kwh and firing 12,000
Btu/lb coal.
91
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-600/7-77-005
2.
3. RECIPIENT'S ACCESSION-NO.
4. TITLE AND SUBTITLE
Evaluation of the General Motors'
Control System
Double Alkali SO2
5. REPORT DATE
January 1977
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
Edward Interess
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Arthur D. Little, Inc.
20 Acorn Park
Cambridge, Massachusetts 02140
10. PROGRAM ELEMENT NO.
EHE624
11. CONTRACT/GRANT NO.
68-02-1332, Task 3
12. SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development
Industrial Environmental Research Laboratory
Research triangle Park, NC 27711
13. TYPE OF REPORT AND PERIOD
Task Final; 3/74-8/76
COVERED
14. SPONSORING AGENCY CODE
EPA-ORD
15.SUPPLEMENTARY NOTES IERL-RTP Task Officer for this report is Norman Kaplan, 919/-
549-8411 Ext 2915, Mail Drop 61.
16. ABSTRACT
The report is an evaluation of the double alkali flue gas desulfurization
(FGD) system, installed to control SOx emissions from the coal-fired industrial boiler-
complex at General Motors' Chevrolet plant in Parma, Ohio. It describes the boiler
and FGD systems. It addresses performance with respect to SO2 removal, filter cake?
properties, lime stoichiometry, carbonate softening, soda ash stoichiometry, scaling,.
oxidation, and reliability. The evaluation is presented in terms of three 1-month-long'
intensive test periods and a longer-term non-intensive test period. System material 5
balances are presented for some of these periods. A general history of the operation I
is also presented. j
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
COSATI Field/Group
Air Pollution
Desulfurization
Sulfur Oxides
Flue Gases
Scrubbers
Coal
Boilers
Calcium Oxides
Carbonates
Sodium Carbonates
Scaling
Oxidation
Air Pollution Control
Stationary Sources
Double Alkali System
13 B 13 A
07A,07D
07B
21B
21D
18. DISTRIBUTION STATEMENT
Unlimited
19. SECURITY CLASS (This Report)
Unclassified
21. NO. OF PAGES
99
20. SECURITY CLASS (Thispage}
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
93
------- |