RESEARCH REPORT
to
NATIONAL AIR POLLUTION CONTROL ADMINISTRATION
CONTRACT NO. PH 86-68-8U
Task Order No. 17
November 30, 1969
BATTELLE MEMORIAL INSTITUTE
COLUMBUS LABORATORIES
-------
THE COLUMBUS LABORATORIES of Battelle Memorial Institute comprise the original
research center of an international organization devoted to research.
The Institute is frequently described as a "bridge" between science and industry a role it
has performed in more than 90 countries. As an independent research institute, it conducts
research encompassing virtually all facets of science and its application. It also undertakes
programs in fundamental research and education.
Battelle-Columbus with its staff of 3,000 serves industry and government through
contract research. It pursues:
research embracing the physical and life sciences, engineering, and selected social
sciences
design and development of materials, products, processes, and systems
information analysis, socioeconomic and technical economic studies, and manage-
ment planning research.
505 KING AVENUE* COLUMBUS, OHIO 43201
-------
FINAL REPORT
on
INVESTIGATION OP THE LIMESTONE-SO WET SCRUBBING PROCESS
to
NATIONAL AIR POLLUTION CONTROL ADMINISTRATION
CONTRACT NO. PH 86-68-8U
Task Order No. 17
November 30, 1969
R, W. Coutant, R. H. Cherry, H. Rosenberg,
J. Genco, and A. Levy
BATTELLE MEMORIAL INSTITUTE
Columbus Laboratories
505 King Avenue
Columbus, Ohio 43201
-------
TABLE OF CONTENTS
Page
MANAGEMENT SUMMARY i
INTRODUCTION 1
SUMMARY 2
EXPERIMENTAL WORK AND DISCUSSION 5
SO -Uptake 6
Results 12
Analytical Model for the Wet Lime-SO Scrubbing Process. . 37
Hydration of Burnt Lime kl
Apparatus and Procedure. ..... ... kl
Results kk
Dissolution $6
Procedure. 56
Results 58
Analysis of Liquors 60
CONCLUSIONS 63
RECOMMENDATIONS 6k
APPENDIX
IDENTIFICATION AND COMPOSITION OF SAMPLES A-l
-------
MANAGEMENT SUMMARY
The wet lime-SO scrubbing process was investigated on a laboratory
scale in support of full-scale prototype studies being undertaken by the National
Air Pollution Control Administration (NAPCA).
This investigation consisted of laboratory scale experiments in the
following areas:
1. Measurement of the overall rate of uptake of SO in a stirred-pot
reactor,
2. Measurement of the relative rates of hydration of selected
limestones and dolomites,
3. Measurement of the relative rates of dissolution of selected
limestone and dolomite materials,
k. Chemical analysis of selected dolomite- and limestone-based
liquors prepared at three temperatures.
These experiments were designed to yield qualitative indications of the
importance of individual physical and chemical processes to the overall limestone-
SO wet scrubbing process.
The results of the current experiments indicate that lime in particulete
form reacts readily with various sulfur species or carbonate in solution to yield
a coating which inhibits utilization of the bulk of the lime. Fine grinding of
the lime might alleviate this problem to some extent. However, because of the
observed tendency for particles to become cemented together forming large clusters,
a high degree of utilization of the lime may not be possible as long as the lime is
admitted to the scrubber in particulate form. It is therefore recommended that
-------
further consideration be given to the importance of particle size in the overall
scrubbing process. It is further recommended that consideration be given to the
possibility of predissolving the lime, or limestone, in the feed water to the
scrubber through the use of excess CO or other solubilizing agents.
Further development is also needed in the area of modeling of the
overall reaction system. The model given in this report is only a first attempt
at description of the scrubbing process, and as such does not give adequate re-
presentation of the mechanical and chemical factors Involved. For instance, the
dependence of the equilibrium partial pressure of SO on solution composition
is not explicit in the model given. Also, a detailed analysis of these data will
yield only rudimentary information on the various mess transfer resistances in
this system, which may not be directly applicable to large-scale systems.
ii
-------
INVESTIGATION OF THE LIMESTONE- SO WET SCRUBBING PROCESS
R. W. Coutant, R. H. Cherry, H. Rosenberg,
J. Genco, and A. Levy
INTRODUCTION
One of the major air pollutants in the United States is sulfur dioxide
produced by burning fuels containing sulfur. As part of a national program to
develop air pollution control processes, the National Air Pollution Control
Administration is undertaking prototype studies of the lime/limestone scrubbing
process. These studies will be carried out in three scrubbing systems, each
capable of handling 30,000 acfm of flue gas.
Aqueous lime scrubbing to control SO from combustion flue gas dates
back to the early 1930fs when several pilot and large scale projects were con-
ducted in England to develop a cyclic process in which all of the scrubbing
liquor would be recycled and only a solid waste product would be formed. This
work was never completed partly because of the interruption of World War II and
partly because of the development of the non-cyclic process still in use at
Battersea and Bankside near London. These units require large quantities of
Thames River water for scrubbing on a once-through basis. Little information
on the cyclic lime process is available as a basis for current studies.
A major departure from the earlier English practice involves the use
of the power-plant boiler as a calciner to produce the lime. In this procedure
limestone or dolomite is pulverized and injected into the furnace. The calcined
lime is subsequently collected in the scrubber as the reactant for removing SO .
In this system several factors could be key to a successful process design. For
-------
example, the availability of the lime for reaction in the scrubber could be
dependent on its rate of hydration or rate of dissolution. Any tendency for
insoluble carbonates or other reaction products to form dense layers on the
otherwise active reactants as veil as on scrubber equipment could be important.
Little is known about these or other potentially important reactions end
reaction rates.
This study was conducted under Contract No. PH 86-68-8^ (Task Order
No. IT) to provide preliminary information on the chemistry and relative re-
action rates of some of the more important reactions as a basis for estimating
their probable importance in the more complex large scale lime/limestone
scrubbing process. Its purpose was to investigate several of the individual
chemical processes which occur during the overall limestone-SO wet scrubbing
process: takeup of SO by lime solutions and slurries, hydration of burnt lime,
and dissolution of hydrated lime. This Final Report covers work done during
the period of April 15 through July 15, 1969.
SUMMARY
The wet-lime-SO scrubbing process was investigated on a laboratory
scale in support of full-scale prototype studies being undertaken by the National
Air Pollution Control Administration (NAPCA).
This investigation consisted of laboratory scale experiments in the
following area:
1. Measurement of the overall rate of uptake of SO in a
stirred-pot reactor,
2. Measurement of the relative rates of hydretion of
selected limestones and dolomites,
-------
3. Measurement of the relative rates of dissolution of
selected limestone and dolomite materials,
U. Chemical analysis of selected dolomite- and limestone-based
liquors prepared at three temperatures.
»
The SO uptake experiments were performed with simulated flue gas in
a stirred-pot reactor at 125 F operated batchwise with respect to the liquid
charge. Preliminary runs were made to determine the effects of gas-bubble
size, and gas-flow rate on the observed rate of uptake of SO by the liquid.
The experiments were limited to a brief exploration to assess the magnitude of
the effects and did not involve a detailed study of these variables. Within the
range of operations undertaken the effects of variation of these experimental
parameters were judged to be not significant.
The two primary variables studied were the flue ges and the scrubber-
liquor compositions. The change of pH of the scrubber liquor during a run
generally follows a trend which includes a sharp initial drop as CO is sorbed.
The pH remains reasonably constant at values in the range 6-7 until the re-
act ants of the liquid phase become exhausted or are rendered unavailable. At
this point a sharp drop in liquor pH is accompanied by SO breakthrough. SO
breakthrough occurred at liquor pH values in the range of 3-U when no NO was
Jt
present in the simulated flue gas; with NO , breakthrough occurred in the pH
X
range of If-5.
Little or no SO appears in the scrubber outlet gas until the reactant
is either exhausted or rendered unavailable. The stoichiometry at the point of
SO breakthrough generally corresponds to the formation of the sulfite from the
available reactants. After breakthrough, the SO concentration in the outlet
gas increases rapidly and asymptotically approaches the value of the inlet
concentration of SO .
-------
The rates of hydretion of several samples of burnt limes supplied by
NAPCA were determined by a temperature-rise procedure similar to ASTM Test C-110.
The results indicate a strong dependence of hydretion rate on the conditions of
calcination of the stone and particle size. As particle size is increased,
or as temperature of calcination is increased, the rate of hydration of the
resultant lime decrease's. The results indicate that the rate of the hydration
process is limited by diffusion of water through the lime particle, and that over-
burning of the lime markedly decreases hydration rate. Other experiments, using
partially sulfated limes or solutions containing S07, showed that hydretion rates
are severely limited by the presence or formation of a layer of sulfate on the
lime particles.
Rates of dissolution of hydrated lime were determined in a well-stirred
system using a calcium ion-specific electrode and a pH electrode for monitoring
H-
the concentrations of Ca and OH in solution. The results indicate that dis-
solution in water is rapid, only a few seconds being required for saturation
with Ca(OH) . However, when the solvent contains sulfite, sulfate, or carbonate,
the rate of dissolution of Ca(OH) is greatly decreased.
Analyses of solutions saturated with respect to CaSO , CaSO, , MgSO ,
and MgSOi showed that these solutions contained primarily MgSO, . These enalyses
also indicated a buffering effect of these salts on the pH of the solutions.
-------
EXPERIMENTAL WORK AND DISCUSSION
The limestone-SO wet-scrubbing process may be considered in terras
of a series of physical and chemical interactions involving: (l) transfer of
the SO to the surface of the liquid sorbent, (2) dissolution of SO and
*
reaction with dissolved components of the liquor, and (3) regeneration of
active liquor through hydration and dissolution of solid lime. It is generally
expected that the homogeneous solution reactions will be relatively fast and
not important in determining the overall rate of the scrubbing process.
Hence, this program has been concerned with an exploratory examination of the
heterogeneous processes; SO uptake, hydration of the lime, and dissolution
of the lime. As will be seen from the results, the rates of these processes
will not "be independent of each other in any real scrubbing system. These
results are therefore, at best, qualitative indications of the importance of
individual segments of the overall scrubbing process.
-------
SO -Uptake
The apparatus shown schematically In Figure 1 was developed to perform
laboratory-scale experiments to study the wet lime-SO scrubbing process.
Simulated flue gas was obtained by mixing gases from two cylinders, one contain-
ing 90 mole percent nitrogen and 10 mole percent oxygen end the other containing
75.5 mole percent nitrogen, 2^ mole percent CO , end 0.5 mole percent SO . Other
cylinders were available, so that SO or CO could be deleted from the mix, or
NO could be added to the mix. The gases from the cylinders were passed through
calibrated rotemeters equipped with dial thermometers and pressure gauges. The
rotameters were operated at 15 psig, and flow controllers were used to reduce
the pressure to atmospheric conditions. The line from the cylinder containing
SO had a check valve after the flow controller to prevent backflow from the
Np/0 cylinder. A nitrogen line (not shown) was connected to each of the gas-
cylinder lines immediately downstream from the shut-off valves and was separated
from the latter lines by check valves. The N was used for purging and standardizing
purposes. The line from the N /O cylinder was constructed of copper end the line
from the SQ^CQ^IX cylinder was constructed of stainless steel. However, down-
stream from the flow controllers, the entire flow system, with the exception of
the SO analyzer, was constructed of Pyrex glass.
After leaving the flow controller, the Hp/°o stream passed through s
preheat loop in a water bath and then through a water-bubbler humidifier, where
the stream was saturated with HO. The SO /CO_/N stream also passed through
-------
FIGURE 3. SCHEMATIC DIAGRAM OF LABORATORY APPARATUS
vent
FOR STUDY OF WET LIME-SO SCRUBBING PROCESS.
CTC
LEGEND
A - Analyzer, S02 P
CTC - Constant temp, chamber, 130°F R
CV - Check vclve S
F - Klowmeters SC
FC - Flow controller SO
H - Hunidifier T
M - Mixing chamber WB
Pressure gauges
- Reactor
HpOp scrubber
Stopcocks, 3-woy
Shut-off vnlves
Thermometer
- Water bath, 125°F
-------
8
a mixing chamber which vas immersed in the water bath. From the mixing chamber,
the gas stream flowed to a manifold containing 3 three-way stopcocks so that the
flow could be directed either to the reactor end then to the SO analyzer or vice
verse. The stopcock manifold permitted checking of the SO content of the inlet
gas to the reactor periodically during the course of a run. The SO content
of the outlet gas from the reactor was monitored continuously except for short
periods when the inlet concentration was being checked. After returning to the
stopcock manifold, the gas stream was routed to an HO scrubber to remove residual
SO before being vented to the hood. The reactor was also immersed in the water
bath, which was maintained at 125°F, and the entire flow system, between the flow
controller and the HO scrubber, was enclosed in a plastic cabinet maintained et
130°F.
A detailed sketch of the reactor is shown in Figure 2. The gas entered
the reactor through a hollow glass stirring rod and was dispersed through a
coarse glass frit at the bottom of the rod. Two sets of three glass propeller
blades were located just above the frit, to insure good mixing. In several runs,
a six-bladed Teflon impeller, having small holes drilled through the bottom, was
substituted for the glass frit-propeller blade combination. The gas left the
reactor through a small packed bed of glass beads in order to remove any en-
trained liquid. The reactor was equipped with a thermocouple well and fittings
for sealing in pH and calcium-ion electrode systems.
-------
Electrode
Thermocouple well
Stirring rod, hollow glass
Packing, 6 mm dia.
Pyrex beads on
Teflon support
Reaction vessel,
125 mm dia. Pyrex
Frit for gas dispersion
FIGURE
REACTION VESSEL FOR STATIC
CHARGE OF SCRUBBER LIQUOR.
-------
10
The SO analyzer was a Model 315 A Beckman Infrared Analyzer which
was calibrated with known mixtures of SO -containing dry gases. This calibration,
shown in Figure 3> vas double- checked by using gases directly from several cylinders
containing known amounts of SO . It was found that CO had a negligible effect on
the analyzer readings (10 mole percent CO was equivalent to about 10 ppm of SO ),
but that HO had a considerable effect on the readings. Figure 3 also shows a
curve for analyzer reading versus HO composition in mole percent. The theoretical
calibration curve for SO in gas saturated with HO at 125 P vas obtained by
adding the analyzer reading for 13-2 mole percent HO (gas saturated at 125°F)
to the analyzer readings for SO in dry gas. However, it was determined that the
effects of SO and HO on the analyzer were not additive. Therefore, calibration
curves had to be faired in for each run, as shown in Figure 3> based on analyzer
readings for the inlet gas to the reactor, with and without the
stream. That is, a zero reading was obtained from the saturated N /O streem,
and another point on the curve was obtained from the reading for the N /O streem
and the SO /CO 2/W stream mixed together after correcting this stream to saturated
conditions. The above two points were then used to fair in the curve.
The SO analyzer readings, reactor pH, and HO pH were recorded contin-
uously on strip-chart recorders. An attempt was also made to monitor the calcium-
ion concentration; however, difficulty was encountered with the calcium-ion
electrodes because of the changing pH of the scrubbing liquor. Use of these
electrodes was abandoned during the experiments. However, the calcium ion elec-
trodes were used to determine the initial concentration of the scrubbing liquors,
and two batches of these liquors were analyzed by atomic absorption as a check.
Good agreement was obtained between the atomic absorption and calcium-ion
electrode analyses.
-------
FIGURE 3. CALIBRATION OF INFRARED ANALYZER FOR SO .
120
100
80
c
H
T3
a
H
I
60
20
SO in saturated gas at
125 F (theoretical)
in dry gas
S _ SO in BeturfjXages et 125 F (typical)
HO in SO free gas
0
0 1000
2000
u
3000 4ooo
ppra
6 8
mole percent HO
5000 6000 7000
10 12 It
-------
12
Results
Sixteen experiments were performed to study the SO uptake "by various
scrubbing liquors. The simulated flue-gas compositions for the experiments ere
listed in Table 1, and the experimental parameters are summarized in Table 2.
Runs A, B, end 2A were performed in addition to the runs listed in the original
work statement and Runs 10 end 11 from this statement were not performed, in
accordance with verbal agreements made with Mr. J. Phillips of NAPCA. Curves of
SO concentration versus time for the inlet gas to the reactor end the outlet gas
from the reactor, and curves of pH versus time ere shown for each run in Figures
k through 19. All of the runs yielded essentially the same type of curve for the
SO concentration in the gas leaving the reactor, i.e., little or no SO in the
outlet gas until breakthrough time, at which time the SO concentration began
increasing and asymptotically approached a value close to the inlet value. Even
when the scrubber liquor becomes saturated with SO , the SO concentration in the
outlet gas should be slightly lower than that in the inlet gas because of the added
moisture picked up by the gas in passing through the reactor. It is fairly safe
to assume that the gas leaving the reactor is saturated with HJ) end, thus,
contains 13.2 mole percent HO. In most of the runs, the gas entering the re-
actor contained 7.06 mole percent HO. Calculated values for the SO concentration
in the outlet gas at liquor saturation ere listed for each run in Table 2 on
line 3. In the runs in which the outlet SO concentration had leveled off, in-
dicating that liquor saturation had occurred, there is good agreement between
the experimental value and calculated value of SO concentration in the outlet
gas at liquor saturation. It is estimated that the SO concentrations are accurate
to within ±1056 for all runs except A, B, 6, 7, 8, and 9 in which cases
-------
13
TABLE 1. COMPOSITION OF SIMULATED FLUE GAS FOR SO,
UPTAKE EXPERIMENTS (INLET GAS TO REACTORJ
Concentration, mole percent
Run No.
A
B
1
2
. 2A
3
4
5
6
7
8
9
12
13
14
15
so2
0.22
0.22
0.22
0.22
0.22
0.22
0.22
0.22
0.22
0.22
0.26
0.25
0.089
0.044
0.00
0.25
N2
75.76
. 75.76
75.76
75.76
75.76
75-76
75.76
75.76
75.76
75.76
73-31 .
73.31
77.09
77.62
77.00
88.04
°2
4.65
4.65
4.65
4.65
4.65
4.65
4.65
4.65
4.65
4.65
4.65
4.65
7.03
7-87
4.65
4.65
C°2
12.31
12.31
12.31
12.31
12.31
12.31
12.31
12.31
12.31
12.31
14.68
14.68
5.^08
2.49
11.29
0.00
H20
7.06
7.06
7.06
7.06
7.06
7.06
7.06
7.06
7.06
7.06
7-06
7.06
10.71
11.98
7.06
7.06
NO
X
0.00
0.00
0.00
o.oo
0.00
0.00
0.00
0.00
0.00
0.00
0.046
0.046
0.00
0.00
0.00
0.00
-------
TABLE 2. SUMMARY CF S02 UPTAKE EXPERIMENTS AT 125°F AND 1 ATMOSPHERE
Run Number
A 6 1
Inlet gas flow 10,760 10,760 1,076
rate, cc/r.in
at STP
S02 in Inlet 2,200 2,200 2,200
gas, ppm
S02 in outlet 2,060 2,060 2,060
gas at liquor
saturation'3'.
ppm
Impeller TI TI TI
type!*)
Impeller 1,500 1,500 1,500
speed,
rpn
Scrubber li- 1,OOO 1,000 1,000
qucr volume,
cc
Scrubber li- 1.43xlO~3M Dls- LOOxlO"1!1*
quor con- NaOH tilled HaOH
position H,0
Experimental <1 0 >180 '
breakthrough
tine, ir.in
Calculated 0.81 0 566
breakthrough
tirce'^J. rain
Ratio:
experimental/
calculated
r, moles of re- 2 2
acta.it con-
sumed per mole
of. S02 con-
Buned
2 2A 3 4 5 6
1,076 1,076 1,076 1,076 1,076 1,076
2,700 2,700 2,200 2,200 2,200 2,200
2,050 2,060 2,060 2,060 2,060 2,060
TI GDT TI GDT GDT GDT
1,500 1,500 0 1,500 1,500 1,500
1,000 1,000 1,000 1,000 1,000 1,000
0.975xlO~3M 1.05xlO'3M 0.925xlO~3M 1.20xlO~2M 6.06xlO~3M 9.BOxlO~3M
NaOH NaOH NaOH calcium doloroitic calcium
hydrate hydrate hydrate +
4.00 gra
solid
calcium
hydrate
55 8 H9 70 480
5.51 5.94 5.24 136 63.6 722
0.9 0.85 1.5 O.S8 1.02 0.66
222 1 11
7 8 9 12
1,076 1,076 1,120 1,136
2,700 2,500 2,500 890
fc
2,060 2,340 2,340 865
TI GDT GDT GDT
1,500 1,500 1,500 1,500
1,000 1,000 1,000 1,000
<1.42xlO"3M 4.42xlO~3M 6.85xlO~3M 1.09xlO"2M
dolorrutic calcium dolorsitic calcium
hydrate + hydrat* hyiirate hydrate
4.00 gm
solid
dolomitic
hydrate
650 76 105 155
841 43,8 67.8 290
0.77 1.7 1.5 0.53
(e) 1 11
13 14 15
1,136 1,076 1,076
440 0 2.500
434 0 2,340
GDT GDT GDT
1,500 1,500 1,500
1,000 1,000 1,000
1.02xlO~2M 1.0
-------
Ua
FOOTNOTES FOR TABLE 2
(a) Assuming reactor outlet gas is saturated with H?0.
(b) TI = Teflon impeller; GDT = gas dispersion tube.
(c) Run was terminated at 180 min and no breakthrough had occurred.
(d) Based on simplified process model.
(e) r = 1 for dolomitic "hydrate solution and r = 0.5 for solid dolomitic hydrate,
-------
OJ
W
§
a.
Inlet ges flow rate: 10,760 cc/min STP (wet bpsis)
Stirrer speed: 1500 RPM (Teflon impeller)
Scrubbing liquor: 1.1*3 x 10"3 M NsOH (l liter)
Temperature: 125 F
Theoretical breakthrough: 0.8l min
Rerctor pH
16 18 20 22 24 26 28 30 32 3^ 36
0 2
FIGURE U. S0g UPTAKE, RUN A.
-------
16
2400
2200
2000
1800
1600
C\J
g 1200
1000
Inlet gas flow rate: 10,j60 cc/min STP (wet basis)
Stirrer speed: 1500 RPM (Teflon impeller)
Scrubbing liquor: distilled HO (l liter)
Temperature: 125 P Inlet SO,
Theoretical breakthrough: 0 min e-
13
12
11
10
9
8
1
6
5
it
3
0
8 10 12
t, minutes
Ik 16 18 20
22
FIGURE 5. SO UPTAKE, RUN B.
-------
-------
Inlet SOp
Outlet S02
Inlet gas flow rete: 1,076 cc/min STP (vet besis)
Stirrer speed: 1500 RPM (Teflon impeller)
Scrubbing liquor: 0.975 x 10" 3 M NeOH (l liter)
Temperature: 125
Theoretical breakthrough: 5-51
Reactor pH
20
60
80 100 120
t, minutes
FIGURE 7. S02 UPTAKE, RIM 2,
11*0
160
180
200
-------
Outlet SO
Inlet gas flow rete: 1,076 cc/mln STP (wet basis)
Stirrer speed: 1500 RPM (ges dispersion tube)
Scrubbing liquor: 1.05 x 10~3 M NaOH (l liter)
Temperature: 125
Theoretical breakthrough: 5-9U- min
Reactor pH
iO 100
t, minutes
iou
FIGURE 8. SO UPTAKE, RUN 2A.
-------
C\l
(X,
Inlet gas flow rate: 1,0?6 cc/min STP (wet bpsis)
Stirrer speed: 0 RPM (Teflon libeller)
Scrubbing liquor: 0.925 x 10'3 M NeOH (l liter)
Temperature: 125 F
Theoretical breakthrough: 5-2^ min
0
100
t, minutes
120
140
lfc>0
100
FIGURE 9. S02 UPTAKE, RUN 3
-------
21*00
2200
2000
1800
1600
itoo
1200
1000
800
600
Uoo
200
0
Inlet SO.
M-
Ca concentration
(uncorrected)
Reactor pH
Outlet SO,
Inlet gas flow rete: 1,076 cc/mln STP (wet basis)
Stirrer speed: 1500 RPM (ges dispersion tube)
Scrubber liquor: 1.20 x 10-2 M calcium hydrate (l liter)
Temperature: 125 F
Theoretical breakthrough: 136 min.
Jr
60 80 100
t, minutes
FIGURE 10. SO UPTAKE, RUN k.
120
160
180
ro
-------
o
co
I
OJ
2600
2^00
2200
2000
1800
1600
1200
1000
800
600
uoo
200
0
Inlet gas flow rate: 1,OJ6 cc/min STP (vet basis)
Stirrer speed: 1500 RPM (gas dispersion tube)
Scrubber liquor: 6.06 x 10"3 M dolomitic
hydrate (l liter)
Temperature: 125 F
Theoretical breakthrough: 68.6 min
Inlet SO,
0 20 40 60 80 100 120
t, minutes
FIGURE 11
1UO
160
180
200
220
13
12
11
10
9
8
1
6
5
l*
3
2
1
0
SO UPTAKE, RUN 5.
ro
ro
-------
21*00
Inlet SOp
Inlet gas flow rete: 1,076 cc/min STP (wet besis)
Stlrrer speed: 1500 RPM (gas dispersion tube)
Scrubber liquor: 9.80 x 10~3 M celcium hydrete (l liter) +
U.OO CMS solid calcium hydrate
Tempereture: 125
Theoretical breakthrough: 722 min
0
20
22
FIGUBE 12. SO UPTAKE, RUN 6
ro
-------
o
CO
2400-
2200
2000
1800
1600
1400
cvi 1200
1000
800
600
400
200
0
0
T
Inlet S02
Inlet gas flow rate: 1,076 cc/min STP (wet bssis)
Stirrer speed: 1500 RPM (Teflon impeller)
Scrubber liquor: 4.42 x 10~3 M dolomitic hydrate
(l liter) + 4.00 CMS solid
doloroitic hydrate
Temperature: 125 F
Theoretical breakthrough: 84l min
Outlet SO,
12
11
10
9
8
7
6
5
1
0
ex
ro
FIGURE 13. S02 UPTAKE, RtM 7.
-------
o
w
2600
2400
2200
2000
1800
1600
11*00
1200
1000
800
600
1*00
200
Inlet SO,
Inlet gas flow rate: 1.0?6 cc/mln STP (vet basis; U60 ppm NO )
Scrubber liquor: ^.^2 x 10~3 M calcium hydrate (l liter)
Temperature: 125
Theoretical breakthrough: 1*3.8 min
20
60 80 100
t, minutes
FIGURE Ik. SO UPTAKE, RUN 8
120
160
13
12
11
10
9
8
U
3
180
-------
2600
2200
2000
l800
1600
g 1200
1000
800
600
400
200
0
Inlet gas flow rate: I,0j6 cc/min STP (wet basis; lj-60 ppm NO, )
Stirrer speed: 1500 RPM (gas dispersion tube)
Scrubber liquor: 6.85 x 10"3 dolomitic hydrate (l liter)
Temperature: 125 F
Iheoreticel breakthrough: 67.8 min
Inlet S02
Outlet SO,
100
t, minutes
FIGURE 15. S02 UPTAKE, RUN 9.
ro
-------
1200
1000
Inlet gas flow rete: 1,120 cc/mln (wet basis)
Stirrer speed: 1500 RPM (Ges dispersion tubes)
Scrubber liquor: 1.09 x 10~2 M calcium hydrate (l liter)
Temperature: 125 F
Theoretical "breakthrough: 290 ndn
Inlet SO.
800
8 6oo
6
a
UOO
200
100
200
1*00
t, minutes
FIGURE 16. SO UPTAKE, RUN 12.
ro
-------
TOO
600
500
,400
C/3
B
Pi
300
200
100
0
Inlet gas flow rate: 1,136 cc/mln STP (wet basis)
Stirrer speed: 1500 RPM (gas dispersion tube)
Scrubber liquor: 1.02 x 10~2 M calcium hydrate (l liter)
Temperature: 125 F
Theoretical breakthrough: 5^2 min
Inlet SO
2 _
100
200
t, minutes
FIGURE 17. S02 UPTAKE, RUN 13.
400
-------
0
Inlet gas flov rate: 1,076 cc/min STP (wet basis)
Stirrer speed: 1500 RPM (gas dispersion tube)
Scrubber liquor: I.Ok x 10~2 M calcium hydrate (l liter)
Tempereture: 125 F
20
60
80
100
t, minutes
120
140
Beactor pH
H2°2
Inlet end outlet SOp
zero
I I
13
12
11
10
9
8
5
U
3
2
160
180
200
FIGURE 18. CO UPTAKE, RUN lU.
-------
2800
o
w
PL,
OJ
1,076 cc/mln STP (wet bests
Inlet gas flow rate:
Stirrer speed: 1500 RPM (gas dispersion
tube)
Scrubber liquor:
calcium hydrate
(l liter)
Temperature: 125F
Theoretical breakthrough: 102 min
1.03 x Kf2 M
Reactor pH
100 120 lUO 160 180 200
t, minutes
220
240
260 280
FIGURE 19. S02 UPTAKE, RUN 15.
-------
31
the SO concentrations are estimated to be accurate to ±20 percent. This
latter error was caused by experimental difficulties which will be discussed
later.
The breakthrough time for SO was also reflected by a sharp decrease
in the reactor pH and by decreasing pH in the HO scrubber. *n the simulated
flue gas not containing NO , breakthrough of SO occurred at a liquor pH in
Jv £.
the range 3-^j with NO , the range was
JL
The scrubber liquors used in this study included distilled water, NaOH
solutions, and lime solutions. The lime solutions were prepared by saturating
# *#
HO with calcium hydrate or dolomitic hydrate in a well- stirred beaker for
about 2k hours at 125 F and filtering out the excess solids. In the case of
Run 8, the liquor was stirred for 66 hours prior to use. The lower calcium
concentration of this liquor is probably due to reaction with CO from the air,
resulting in the precipitation of CaCO . Chemical analysis of the dolomitic
solutions by atomic absorption revealed that there was no magnesium in solution,
indicating that MgO does not readily hydrate and dissolve. Therefore, the
dolomitic hydrate apparently was a mixture of Ca(OH) and MgO in a 1:1 mole
ratio; the mole ratio was calculated from the chemical analysis of the lime-
stone material used to prepare the hydrate (see Appendix). Thus, the only
scrubbing liquor that contained magnesium was the one in Run 7> for which k
grams of solid dolomite hydrate was added to the reactor Just before start
of the run.
Some of the chemical reactions that can occur between the flue gas
and the various scrubber liquors can be represented as follows:
* Supplier's notation for hydrated high-calcium lime.
** Supplier's notation for hydrated lime derived from dolomitic limestone.
-------
32
S02 + H20 « H2S03 (l)
HaOH + H SO = HaHSO + HO (2)
2NaOH + H SO = Na SO + 2H20 (3)
Ca(OH)2 + H2S0
Ca(OH)2 + C02 =
CaCO + HSO = CeSO + HO + CO (6)
MgO + H2S03 = MgS03 + H20 (7)
MgO + C02 = MgCO (8)
+ H0 + C0 (9)
Reactions involving NO complicated further an already complex
J\.
sitxiation in Runs 8 and 9> vhere this component was present. In Run lU, the
absence of SO eliminated reactions (l), (k) , and (6), and in Run 15, the
absence of CO eliminated reactions (5) and (6). Reactions (?)> (8), end (9)
are applicable only in Run 7-
It is postulated that the time until breakthrough of SO in the outlet
gas represents the time until exhaustion of reactant in the scrubber liquor, and
that the increasing segment of the SO -concentration curve represents the effect
of mass transfer on the dissolution of SO in the liquor. The theoretical break-
through time can be calculated front the following equation:
*B '
-------
33"
where
t£ = theoretical "breakthrough time, min
if - initial amount of reactant in scrubber liquor, moles
3
q ss inlet gas flow rate, cm per min
A
C s± concentration of SO in inlet gas, moles per cm
r = moles of reactant consumed per mole of SO consumed.
The term r can best be understood by comparing reactions (2) and (3). In
reaction (2) r has a value of 1 and the reaction product is NaHSO , while in
reaction (3) r has a value of 2 end the reaction product in Na SO . Theoretical
break-through times were calculated for each run by using the value for r
appropriate to the sulfite rather than the bisulfite reaction product (r = 1
and 2, respectively, for Ca(OH)2,end NaOH solutions). It is recognized that
continued sorption after the breakthrough point probably involves formation
of bisulfite, but the effect of this reaction on the equilibrium partial pressure
of SO was not explicitly considered. In the case of Runs 6 and 7, where solid
lime was added to the scrubber liquor, the breakthrough times were calculated by
assuming that all of the solid material would eventually dissolve. Experimental
breakthrough times were obtained from the curves of SO concentration in the
outlet gas Versus time and are compared with the theoretical breakthrough
times in Table 2. In some cases the agreement is within experimental error end
in other cases the agreement is poor; the breakthrough time comparison is dis-
cussed for each run in the following paragraphs.
Runs A and B were performed at a gas flow rate of 10 liters per minute,
_-3
with 10 J M NaOH and distilled water, respectively, as scrubbing liquors. Diffi-
culties were encountered in interpreting the data from the SO enelyzer because of
-------
excessive pressure drop needed for the high flow rate. Consequently, all other .
runs vere performed at a flow rate of 1 liter per minute. In Run Af it was
impossible to determine the experimental breakthrough time accurately, but it
was less than one minute. This compered favorably with the theoretical break-
through time of 0.8l min. Except for the slight break in the curve at the start
of Run A, Runs A and B exhibited the same shape of curve for SO concentration,
indicating that, once the NaOH reactant is exhausted, the scrubber liquor behaves
the same as distilled water at this low concentration of reaction product.
Runs 1, 2, 2A, and 3 were performed with NaOH solutions as the scrubbing
liquor. The concentration in Run 1 was 10" M while in the other runs it was
10 M. The theoretical breakthrough time for Run 1 is 566 min and, since the
run proceeded for only l8o min, no breakthrough would be expected to occur, as
indeed was the case. In Runs 2 end 2A, the experimental end theoretical break-
through times agreed very well, indicating that the reaction product was indeed
NapSO . In Run 2A, a gas dispersion tube, resulting in somewhat smaller gas
bubbles, was used for stirring, rather than the Teflon impeller. The SO con-
centration curves were about the same for both runs, indicating that the mass
transfer rate is not a strong function of bubble size. In Run 3, the Teflon impeller
was used but the stirring motor was not turned on; however, the SO concentration
curve was- still about the same as in Run 2, for which the stirring speed was 1500
rpm. This observation was taken to indicate negligible resistance to mass transfer
in the liquid phase. In Run 3> the experimental breakthrough time was greater than
the theoretical breakthrough time, possibly indicating the formation of some
WaHSO .
-------
35
Runs 4 end 5 vere performed with saturated solutions of calcium hydrate
and dolomitic hydrate, respectively, es the scrubbing liquors. There was very
good agreement between the experimental and theoretical breakthrough times for
both of these runs. The calcium ion electrode was used in Run h, in an attempt
to monitor the Ca concentration continuously. However, as the pH of the liquor
decreased, the calcium i-on electrode output became erratic and increased rapidly,
thus negating its usefulness.
Runs 6 and 7 were duplicates of Runs h end 5, except that ^.00 gms of
calcium hydrate or dolomitic hydrate was added to the respective liquors immediately
prior to the start of each run. Difficulty was encountered because of pressure
buildup caused by clogging of the gas dispersion tube in Run 6, so that the Teflon
impeller was used in .Run 7- However, even the relatively large holes (l/l6-in
diameter) in the Teflon became clogged with solid material during the course of
the run, and the run had to be terminated because of excessive pressure buildup
in the system. At the conclusion of Runs 6 and 7, both liquors contained a
solid crystalline material that was grossly different in appearance from the
solid hydrates that were added to the liquors. The crystalline materials were
not analyzed, but it is assumed that they are CaSO and MgSO . The experi-
mental breakthrough times in both runs were significantly lower than the
theoretical times, indicating that the crystalline material may have deposited.
around some of the solid hydrate particles, thus preventing their further reaction
with SO . Also, the presence of solid particles added another resistance to the
reaction mechanism in the form of finite rate of dissolution of the hydrates.
This added resistance was not accounted for theoretically. However, the break-
through times were only about 30 percent lower than expected.
-------
36
Runs 8 end 9 were also duplicates of Runs U end 5> except that U50 ppm
of NO vas present in the flue gas. The presence of NO caused the infrared-
*v "X
analyzer readings to drift upward with time and, consequently, made it more diffi-
cult to interpret the results. However, the experimental "breakthrough times were
50 to 70 percent longer than the theoretical times.
Runs 12, 13> and lU were similar to Run 4, except that the SO concen-
tration in the inlet gas was lowered to 890, MtO, and 0 ppm, respectively, from
the 2,200 ppm in Run 4. The CO concentration was decreased and the HO concen-
tration was increased in the simulated flue gas for Runs 12 and 13 because of
limitations in the cylinder gases available. The experimental breakthrough times
were about kO percent lower than the theoretical times in Runs 12 and 13-
Probable experimental error can account for only about a 25 percent difference.
At first glance it appears that the lower CO concentration might offer an ex-
planation for the discrepancy. However, in Run 15> in which no CO was present,
there was excellent agreement between the experimental and theoretical breakthrough
times. Therefore, it is difficult to explain the discrepancy between the break-
through times for Runs 12 and 13, especially in light of the results from the
other runs. The reactor pH-versus-time curve for Run 14, in which no SO was
present, indicates rapid formation of CaCO , followed by attainment of the carbonate
equilibrium at a pH of about 6.6.
The absence of CO in Run 15 had a significant effect on the shape of
the outlet SO concentration curve. Soon after breakthrough, the curve tended
to level off at about 500 ppm of SO at a pH of 3.8 and then began increasing and
asymptotically approached a value close to the inlet concentration.
-------
3T
Analytical Model for the Wet Liir.e-S0p
Scrubbing Process
An attempt has been made to fit the data to an analytical model based
on chemical reaction followed by dissolution controlled by gas-phase mass trans-
fer. The model is only a first attempt at an analytical description of these
experiments. For simplicity, it was assumed that the chemical-reaction end mass-
transfer aspects of this problem can be uncoupled and treated independently. The
chemical-reaction portion of the model has already been discussed in terms of
experimental and theoretical breakthrough times for SO concentration in the
gas leaving the scrubber. The increasing portion of the SO concentration curve
has been treated in terms of a gas-phase resistance to mass transfer present in
a thin film at the surface of each gas bubble passing through the scrubber liquor.
A material balance for SO in the scrubber liquor in the reactor leads to the
following equation:
dC^
where
C = unreacted SO concentration in scrubber liquor, moles per liter
C = SO concentration in outlet gas, moles per liter of dry gas
O e-
H = Henry's law partition coefficient at liquor saturation with SO ,
C = SO concentration in outlet gas at liquor saturation with SO ,
G £- <~
moles per liter of dry gas
CL = unreacted SO concentration in scrubber liquor at liquor saturation
with SO , moles per liter
-------
38
t = time, min
K ±= system constant having units of min" defined by
6k q L H
* = g (12)
v d V^^
where
k = gas-phase mess transfer coefficient, cm per min
o
3
q -gas flow rate, cnr of dry gas per min
L = average path length traveled by a gas bubble in the liquor, cm
v = average velocity of gas bubble, cm per min
d = average diameter of a gas bubble, cm
V e volume of liquor, cm .
A material balance for SO over the entire system yields
cg - * cg + vi
where
C = SO concentration in inlet gas (equal to the previous C )
6 f- S
moles per liter of dry gas.
From the definition of the breakthrough time,
so that, at t = t,,, Equation (13) becomes
a
at
-------
39
Equations (ll) and (13) can be solved simultaneously for C and C yielding
-* o
q C° <*(t-tj
6
u . V
and
C = C° [1 - «-0f t.) (17)
g g ~ B
where
(18)
K V-j^ *
J_ T 'r TT
From Equation (l?)
C
In (l - -S- ) = - «(t-tB). (19)
C
«
Therefore, a plot of In (l -- |) versus (t-t~) should yield a straight line vith
C
a slope equal to - ce. A comparison of the experimental data with the analytical
model is shown in Figure 20 for Runs 2A and 5« Considering the assumptions
involved, such as constant bubble diameter end constant residence time for the
bubbles in the liquor, the agreement between experiment end theory is good. It
is interesting to note the similarity between the date for Runs 2A and 5. The
experimental parameters are the same for these runs, but different scrubbing
liquors were used (0.975 x 10~^ M KaOH in Run 2A and 6.06 x 10"^ M dolomitic
hydrate in Run 5), resulting in different breakthrough times. However, as far
as the dissolution portion of the model is concerned, the only difference between
o
the runs would be different partition coefficient (H) for 10 M Ne SO end
_o
6 x 10 M CaSO . Evidently, at these low concentrations, the partition coefficients
are about the same, and about equal to that for distilled water.
2 1
The value of ot for Runs 2A end 5 is approximately 2. 76 x 10 min
as obtained from the slope of the line in Figure 20. This value of ff, together
-------
1.00
1 -
s
0.10
0.01
0
ORun 2A
A Run 5
20
Slope = -2.76 x 10~2 mirf1
I
60
t -
80
t,,, mln
100
120
FIGURE 20. COMPARISON OF EXPERIMENTAL DATA WITH ANALYTICAL MODEL
FOR S02 UPTAKE BY SCRUBBING LIQUORS.
ifco
-------
41
with the calculated theoretical breakthrough time of 68.6 min end C° of 2370
g
ppm of SO on a dry basis, was used to construct the curve of SO concentration
in the outlet gas versus time for Run 5 in accordance with Equation (l?). A
comparison of the experimental date with the calculated curve is shown in
Figure 21 j the agreement is very good, considering the simplified analytical
model used.
It is possible to calculate partition coefficients (H) for each run by
graphically integrating the SO concentration curves. Once values of a and H
ere known, one can calculate K from Equation (18) and, ultimately, the gas-phase
mass-transfer coefficient (k ) from Equation (12) if the average bubble diameter
D
and average residence time for the bubbles in the scrubbing liquor are known.
This was not done insofar as such an analysis did not fall within the scope of
this work. It must be realized that the parameters in the analytical model
are highly dependent on the scrubber design, and that values derived from the
current vork are not necessarily applicable to other types of scrubbers.
Hydration of Burnt Lime
Apparatus and Procedure
The rates of hydration of burnt limes were determined using apparatus
similar to that prescribed for ASTM Test C-110. A sample charge containing one
mole of calcium oxide was added to one liter of water (or other appropriate
liquor as specified by the Contract) at 125 F in a well-stirred and well-insulated
system and the rate of temperature rise in the system was monitored with a thermo-
couple connected to a recorder. During these runs, solution pH was also monitored.
-------
2400
O Experimental data
Calculated curve
2000
1600
en
M
to
<
PQ
1200
O
CO
CJ
800
400
0
80
120 160
t, minutes
200
240 280
ru
FIGURE 21. COMPARISON OF EXPERIMENTAL DATA WITH CALCULATED CURVE
FOR SO CONCENTRATION IN THE OUTLET GAS VS. TIME.
2
-------
The experimental temperature-rise data were corrected for thermal loss
by the system to obtain the true temperature change due to hydretion. It wes
assumed that the thermal loss rate of the system could be represented as
£2 - -k (T - T )
dt - k { e V
where T is the experimentally observed system temperature, T is the
temperature of an arbitrary heat sink in the system, end k is a constant which
includes the heat capacity of the solution, the thermal conductivity of the system
walls, geometry of the system, etc. Values of k and T vere found by determining
the thermal loss rate in the absence of hydration reaction at two different values
of T , and solving the above equation. The corrected temperature chenge in the system
was then calculated as the sum of the observed temperature rise and a correction
for thermal loss: . ;
AT = AT + k I (T - T )dt
Jt e °
= AT + k £ (T . - T ) At.,
e el o i
where the summation is taken over sufficiently smell time intervals to keep the
effects of local curvature in the temperature-rise curve to a negligible level.
In practice, it was found that T did not remain constant during e run. The
error resulting from this procedure was considered insignificant for runs requiring
less than 1 to 1-1/2 hours for completion, but for runs lasting for longer periods
of time the error became comparable with observed temperature changes. Therefore,
an arbitrary time limit of 83 min (5000 sec) was set for duration of hydration
experiments.
All of the samples used in the hydretion experiments were supplied by
NAPCA. These included limes derived from calcitic and dolomitic limestones which
-------
were prepared with en array of particle sizes and calcination temperatures. The
notation used to designate samples is that used in the work statement of the
contract: L is the calcitic lime sample calcined at the lowest temperature and
having the largest particle size, and D is the dolomitic lime calcined at the highest
temperature and having the smallest particle size. A listing of specific particle
sizes end calcination temperatures is given in Table A-l of the Appendix. Analyses
of the original limestones are given in Table A-2 of the Appendix.
tesults
Curves showing the observed temperature rise versus time for the various
samples investigated ere shown in Figures 22 through 31- For both L and D
J J O-J
there was very little apparent difference between hydration in tap water end
hydration in saturated lime water. However, with partially sulfated L _
(Fig. 31), hydration in lime water was severely limited; giving a temperature
rise of only 0.7°F in the first 30° seconds. Hydration runs were also performed
using L__ and D__ and a liquid phase consisting of a saturated Ca,l% - SO ,SO.
liquor (Fig. 30)« With these runs there was a rapid initial temperature rise
of about 5 F followed by extremely slow increase over a period of several hours.
As with the partially sulfated lime, the rate was too slow to be defined. Pre-
sumably the initial temperature rise was due to reaction of the sulfite-sulfete
liquor with the lime. The reaction product probably coated the lime particles
and prevented hydration. In fact, there was a strong tendency for agglomeration
of the reacted lime, with resultant formation of large clumps (up to 1 cm in
diameter) of solids.
-------
CM
o
0
t, seconds x 10
FIGURE 22. HYDRATION OF LIME IN WATER.
-------
'. 12
1000
2000
3000
4000
5000
t,sec
FIGURE 23. HYDRATION OF LIMESTONE-BASED LIMES IN SATURATED LIME WATER.
ON
-------
600
700 800
100 200 300 kOO 500
t, seconds
FIGURE 2U. HYDRATION OF HIGH CALCIUM LIMES IN SATURATED LIME WATER.
-------
8
10 12 Ik 16 18 20 22 2k 26 28 30 32 3k 36 - 38 kO
-2
t, seconds x 10
FIGURE 25. RERUN OF HYDRATION OF HIGH CALCIUM LIME IN LIME WATER.
kk k6
-------
24
22
20
18
16
D33
12
10
8
6
4
2
0
I
I
I
1
I
I
I
I
I
I
8 10 12 14 16 18 20 22 24 26 28 30 32 31
t, seconds x 10"
FIGURE 26. HYDRATION OF DOLOMTTIC LIME IN WATER.
I I I I I I
36 3"8 40 42 44 46
-------
0
1000
2000
3000
4000
5000
t,sec
FIGURE 27- HYDRATION OF DOLOMITE-BASED LIMES IN SATURATED LIME WATER.
-------
51
28
26
2U
22
20
18
16
12
10
8
6
k
2
0
Dll
I I I I I I I
8 10
12 14 16
t, seconds
18 20 22 2k 26 28 30
FIGURE 28. HYDRATION OF DOLOMITIC LIMES
IN SATURATED LIME WATER.
-------
pt,
I I I I I
I I I ! I
10 12 14 16 Id 20 22 24 26 2« 30 32 34 36 3o 40 42 44 46 48 50
0
v/i
ro
t, seconds x 10
FIGURE 29. RERUN OF HYDRAHON OF DOLOMITIC LIME IN LIME WATER.
-------
20
18
16
12 -
10 _
8
6
U
2
0
0
OJ
8 10 12
.16
FIGURE 30.
18 20 22 2k 26 28 30 32 3k 36 38 UO U2 kk U6 U8 50
t, seconds x 10"^
HYDRATION OF LIME IN SATURATED SOLUTION OF
CsSO^, MgSO , CaSO AT 125°F.
-------
20
18
16
12
fc
-. 10
S3
8
6
2
0
L (Sulfeted)
0 2
8 10 12 14 16 18 20 22 2^ 26 28 30 32 3^ 36 38 ^0 k2 kk k6 k8 50
«2
t, seconds x 10
FIGURE 31. HYDRATION OF SULFATED LIME IK LIME WATER.
-------
55
The data shown in Figures 23 end 27 illustrate distinct trends in the
dependence of hydration rate on particle size and calcination temperature.
With the dolomitic limes, increasing particle size or increasing calcination
temperature clearly decreases the rate of hydration. With the calcitic-based
limes, a similar trend was observed. However, the data for L end L are con-
sidered suspect since a considerable amount of uncslcined stone was found in these
samples. This material was a different color from the bulk of the lime end
evolved gas when reacted with HC1. With L , it is estimated that 30 to hO
percent of the sample was uncelcined. For the remaining calcite-based materials,
increased particle size or calcination temperature decreases the hydretion rate.
It should be noted that, for the limestone series, with only one
sample, L , did the total temperature change approach the theoretical value
for reaction of one mole of calcium oxide in one liter of water, 23.4 F. This
may be due to incomplete calcination of the particles, or the rate of reaction
in the later stages of hydration may have been too slow to be detected. With the
dolomitic limes, the temperature change for two samples, D end D , exceeded the
value of 23. k F, indicating that some of the magnesium oxide in the samples may
have been hydrated in these experiments. Upon calcinetion of the carbonate,
magnesium oxide sinters quite rapidly and becomes resistant to hydration. It
is therefore expected that contribution to the temperature rise by hydretion of
magnesium oxide would occur only with the relatively soft-burned dolomitic limes.
*
Boynton suggests that the rate of hydration of a lime is controlled
largely by the permeance of the lime to water and the rate of diffusion of water
* R. S. Boynton, "Chemistry and Technology of Lime and Limestone", Interscience
Publishers, New York, (1966), p 297-
-------
56
through the leyer of hydrated lime which is formed around the particles during
the initial stages of the reaction. This type of mechanism is consistent with the
results of the current work. Figure 32 shows a plot of 1 - (l - F) ' versus
time (F = fraction of lime reacted) for D . Such a plot should yield e
*
straight line if the reaction is chemically controlled; a curve of the type
shown is indicative of control "by diffusion within the particles. It is reason-
able, then, that formation of e sulfate coating, either by exposure of the lime
to SO prior to contact with the hydrating liquor, or "by reaction between the
lime end sulfate or sulfite in the liquor, might further inhibit permeation of
water into the lime.
Dissolution
Procedure
The rates of dissolution of hydrated calcite- and dolomite-based
hydrated limes were studied in water and in a liquor saturated with CeSO ,
CaSO, , MgSO , and MgSO, . Certain of the solutions were also saturated with CO
at e partial pressure of 77 torr. In these experiments, a two-gram charge of
hydrated lime was added to a well-stirred isothermal vessel containing one liter
of solvent (water or sulfite-sulfate liquor), and the pH and calcium-ion activity
were monitored continuously using appropriate electrodes. Calcium ion was detected
using an Orion liquid-liquid Junction-type electrode which was specific to calcium.
* 0. Levenspiel, "Chemical Reaction Engineering", Chapter 12, John Wiley and Sons,
Inc., New York, 1962.
-------
57
1.2
1.1
1.0
.9
.8
.7
.6
i
rH
.3
.2
.1
0
AT.
F =
= 10.5°C
oo
I
I
I
I
o 100 200 300 iioo 500 600 700 800 900 1000 1100 1200
1500
t, seconds x 10
-2
FIGURE 32. HYDRATION OF D
.
-------
With most of the runs, dissolution ves essentially complete within a few tens
of seconds, and except for the case of dissolution in tap water, the celcium ion
activity in solution did not change appreciably during the experiments.
Results
Dete summarizing the essential characteristics of the dissolution
experiments are shown in Table 3- I*1 those experiments involving the saturated
sulfite-sulfate solutions es the solvent, only the OH-activity changed during the
experiment. This is probably due to precipitation of sulfite or sulfate on the
surface of the lime, i.e.,
Ca(OH) (s) + SO" -* CeSO (s) + 20H~ .
C. X Jn.
As in the hydration experiments involving the use of the sulfite-sulfate liquor,
there was a strong tendency for agglomeration of the lime particles. Even with
tap water as the solvent, the dissolution of celcium was inhibited significantly
by contaminants in the water; the calcium ion activity never became es high as
would be predicted on the basis of the observed OH" activity.
Rates cited in Table 3 ere average initial rates of dissolution computed
by dividing the half-time for reaction into the OH~ activity change up to thet
point. The rate of dissolution in tap water was significantly greater than that
in the liquor. The rate of dissolution in the liquor was further suppressed by
the presence of ,CO . The latter fact could be due to deposition of carbonate as
well as sulfate on the surface of the particles. The rate of dissolution of the
dolomitic lime did increase by en order of magnitude when the temperature wes
increased from 55 to 125 F. The presence of fly ash lowered the celcium ion
activity, but hed only a minor effect on the apparent rate of dissolution.
-------
TABLE 3. DISSOLUTION OF HYDRATED LIMES
Hydrated Lime
Dolomite
Dolomite
Limestone
Dolomite
Dolomite
Limestone
Limestone
Solvent
Tap water
Liquor(d)
Liquor
Liquor
Liquor + CO
Liquor + CO
(f)
Liquor + fly aahv '
T,°F
55
55
55
125
125
125
125
pOH
Initial
6.72
6.45
6.58
6.85
7.78
7.84
7.01
Final
1.47
4.69
4.76
5.88
5.88
6.03
6.44
pCe
Initial
3-35
2.76
2.43
1.80
1.65
1.65
2.68
Final
2.47
2.76
2.43
1.80
1.68
1.65
2.68
Va- M
16
. 15.5
7.8
1.3
11.5
18.5
1.8
Rete(c)
1 x 10~3
4.2 x 10"8
1 x 10'6
4.5 x 10~7
5.6 x 10"8
2.5 x 10"8
4.6 x 10'8
(a) One liter at 77 F.
(b) Time required to reach one-half final OH" reactivity.
(c) Average rate = net change in OH" activity/2 t.. / , moles/liter-sec, for 2 grams of
(d) Saturated solution of CaSO^, CaSO , MgSO^, MgSO .
(e) PCQ = 77 torr.
(f) 5 grams fly ash per liter of solution.
\r\
-------
60
As in the case of hydration, the rate of dissolution of lime eppeers
to be severely limited when a layer of reaction product can form eround the lime
particles. This inhibition of hydration and dissolution mey well be the major
factor in reducing the efficiency of utilization of lime in an SO scrubber which
receives the lime in particulate form.
Analysis of Liquors
*
The compositions of saturated sulfite-sulfate-cerbonate solutions were
determined at 55> 90, and 125 F» These solutions were prepared by addition of
amounts of the components considerably in excess of the published solubilities in
water. After thorough mixing for a period of several hours, the solutions were
4-4- 4J4- ...
filtered at temperature, and were analyzed for Ca , Mg , SO", SOT, CO" HCO"
j *» j j>
and pH. Standard procedures were used in the chemical analyses; calcium was
determined atomic absorption, magnesium by gravimetry as the pyrophosphete, sulfite
by iodimetry, sulfate by gravimetry as the barium salt, and carbonate end bi-
carbonate by double-endpoint titration using phenolphthalein end methylorange
indicators. Results of these determinations ere shown in Table k. As can be
seen from the Table, the solutions are essentially magnesium sulfate solutions.
Analyses were also made of selected liquors recovered from the SO -
sorption experiments. These analyses were carried out using procedures similar
to those listed above, with the exception that atomic absorption was employed to
verify the absence of magnesium in the spent liquors. Results of these analyses
are shown in Table 5-
-------
61
TABLE 4. COMPOSITION* OF SULFITE-
SULFATS-CARBONATE LIQUORS
Lime
Additive
Limestone
Dolomite
Limestone
Dolomite
Limestone
Dolomite
T°F
55
55
90
90
125
125
_ ++
Ca
0.
0.
0.
0.
0.
0.
03
03
03
03
Ok
03
MB"
4.97
4.90
6.11
6.18
6.07
6.13
SV
0.
0.
0.
0.
0.
0.
4o
15
43
16
37
056
SV
19.
19.
23.
24.
24.
24.
3
1
5
1
0
1
cv
0.
0.
o.
o.
0.
0.
50
18
43
16
26
10
HCO " pH
0.25 8.90
0 8.75
o 8.30
0.16 8.10
0.11 7.85
0.04 7.65
* Weight percent.
-------
62-
TABLE 5. COMPOSITION* OF SPENT LIQUORS
Run Ko.
k
5
12
13
Ik
15
Ca"^ Mg"^
0.2VT 0
0.167 o
0.2^3 0
0.235 0
0.082 o
0.315 o
SO; SOJ CO; HCO-
0.079 1.34 0 0
0.073 0.92 0 0
0.076 1.29 0 0
0.063 1.09 o o
0 <0.1 0 0,23
1.00 0.69 0 0
* grains per liter.
-------
63
CONCLUSIONS
\
Although it is difficult to draw meaningful conclusions from
one-of-a-kind type experiments, the following facts have been brought to light
as a result of this program,
1. Under the experimental conditions employed, the wet lime-SO
scrubbing process can be represented by an analytical model based on chemical
reaction until exhaustion of reactant, followed by dissolution controlled by
gas-phase mass-transfer. For the scrubber investigated experimentally in this
program, there is little or no SO in the outlet gas from the scrubber until the
reactant is exhausted. The SO concentration in the outlet gas then increases
and approaches the inlet gas concentration asywptotically.
2. The reaction product between SO and the scrubbing liquor at the
breakthrough point is approximately represented by the sulfite rather than the
bisulfite.
3. The presence of solid lime caused serious clogging of the gas inlet
openings to the experimental scrubber. The SO breakthrough time is slightly
reduced, presumably because of coating of the limestone particles by sulfite
(or sulfate) crystals. This observation suggests that full-scale scrubber
operation might be expected to experience both scaling and reduced utilization
of any undissolved solids present in the scrubber liquor,
U, Uptake of CO causes the rapid initial drop in the pH of the
scrubber liquor during a run (by formation of CaCO ), but appears to have little
or no effect on the SO breakthrough time.
-------
5. For the SO uptake runs with slurries of hydrated material,
"breakthrough occurred significantly before the reactant was exhausted, indicating
solid-liquid mass transfer as a limiting step under these conditions.
6. The presence of NO in the flue gas significantly increases the
ratio of observed to calculated breakthrough times. It has been suggested that
oxidation of sulfite to sulfate is the cause of this effect.
7. Rates of hydration of burnt limes are controlled by diffusion of
water into the particles and hence are decreased by everburning of the lime
or use of large particles. The effect of particle size on hydration rate is
apparently much stronger for hard-burned than for soft-burned lime.
8. Deposition of sulfete on the lime hinders the hydration process.
9« Deposition of sulfate or carbonate on the lime hinders the dis-
solution of hydrated lime.
10. The process of dissolution can be accelerated by increasing the
solution temperature.
RECOMMENDATIONS
The current experimental work is largely of an exploratory nature,
and, as such, suggests several areas of need for better understanding and develop-
ment of the limestone SO vet-scrubbing process.
The results of the current experiments indicate that lime in particu-
late form reacts readily with various sulfur species or carbonate in solution
to yield a coating which inhibits utilization of the bulk of the lime. Fine
grinding of the lime might alleviate this problem to some extent. However,
because of the observed tendency for particles to become cemented together
-------
65
forming large clusters, a high degree of utilization of the lime may not be
possible as long as the lime is admitted to the scrubber in particulate form.
It is therefore recommended that further consideration be given to the importance
of particle size in the overall scrubbing process. It is further recommended
that consideration be given to the possibility of predissolving the lime, or
limestone, in the feed water to the scrubber through the use of excess CO
or other solubilizing agents.
Further development is also needed in the area of modeling of the
overall reaction system. The model given in this report is only a first attempt
at description of the scrubbing process, and as such does not give adequate
representation of the mechanical and chemical factors involved. For instance,
the dependence of the equilibrium partial pressure of SO on solution composi-
tion is not explicit in the model given. Also, a detailed analysis of these
data will yield only rudimentary information on the various mass transfer re-
sistances in this system, which may not be directly applicable to large-scsle
systems.
*****#*#*****
The data on which this report is based are recorded in Battelle
Laboratory Record Books No. 2175k, pp 1-^2, and No. 26980, pp. 1-57.
-------
APPENDIX
IDENTIFICATION AND COMPOSITION OF SAMPLES
-------
A-l
TABLE A-l. IDENTIFICATION OF
HYDRATION SAMPLES
Sample
°11
\3
D23
D31
D32
D33
Ln
*13
L23
L31
I|
^^^iO
L.0
Type
Dolomitic
Dolomitic
Dolomitic
Dolomitic
Dolomitic
Dolomitic
Calcitic
Calcitic
Calcitic
Calcitic
Calcitic
Celcitic
Particle Size
(mesh range)
-325
-100 +
-100 +
-325
-200 +
-100 +
-325
-100 +
-100 +
-325
-200 +
-100 +
200
200
325
200
200
200
325
200
Calcine tion
Temperature
°F
1700
1700
2200
2700
2700
2700
1700
1700
2200
2700
2700
2700
-------
A-2
TABLE A-2. CHEMICAL ANALYSES
OF LIMESTONES*
Component
CaO
MgO
Si02
Fe2°3
A1203
Loss on ignition
(iooo°c)
Dolomitic
Limestone
#*
30.39
21.54
1.05
0.28
0.10
46.63
High Calcium
Limestone
54.40
0.48
1.22
0.19
0.44
43.13
* Supplied by G&WH Corson, Inc.
** Weight percent.
------- |