-------
12. Correlation of jSO« .Reduction to Pore Volume for Different
Calcined Limestones
I40O-
I
O
I-
u
I2OO-
1000-
800-
600'
0
CM
O
UJ
Q.
4OO
2OO--
A aragonite
C calcite
D dolomite
K chalk
M marl
0.6
0.2
0.4
0.8
IX)
T
1.2
-t
1.4
PORE VOLUME, cc/g
194
-------
REMOVAL OF SULFUR DIOXIDE FROM STAC'. GASES
BY SCRUBBING WITH LIMESTONE SLURRY:
SMALL-SCALE STUDIES AT TVA
By
J. M. Potts, A. V. Slack, and J. D. Hatfield
Division of Chemical Development
Tennessee Valley Authority
Muscle Shoals, Alabama
Prepared for Presentation at
Second International Lime/Limestone Wet Scrubbing Symposium
Sponsored by the Environmental Protection Agency
New Orleans, Louisiana
November 8-12, 1971
195
-------
REMOVAL OF SULFUR DIOXIDE FROM STACK GASES
BY SCRUBBING WITH LIMESTONE SLURRY;
SMALL-SCALE STUDIES AT TVA
By
J. M. Potts, A. V. Slack, and J. D. Hatfield
Division of Chemical Development
Tennessee Valley Authority
Muscle Shoals, Alabama
ABSTRACT
Results from tests with a spray scrubber (two-stage, counter-
current, k-in. diameter, k-Q ft3/min gas flow) are presented. The
following parameters were studied.
'Surge tank volume was important; SOo removal increased with tank
volume up to about 900 gal/(lb S02)(min). Apparently the effect
was due to providing retention time for completing reactions be-
gun in the scrubber.
'increase in inlet gas temperature reduced absorption significantly,
presumably because of increased S02 vapor pressure over the liquor
leaving the scrubber.
'High liquor;gas ratio was required for good absorption. The
ratios used, on the order of 80 gal/Mcf of gas, were probably
higher—because of relative inefficiency of the spray scrubber—
than needed in practice. The beneficial effect of high liquor
rate indicates the importance of high "effective stoichiometry"
(moles CaC03 to which a mole of S02 is exposed during passage
through the scrubber).
*Use of excess limestone (beyond stoichiometric) was effective;
however, under the scrubber conditions used the stoichiometric
amount was adequate (about 90$ removal). Under less favorable
conditions, excess limestone might have a major beneficial effect
by increasing the steady-state effective stoichiometry in the
scrubber recirculating loop.
'increase in inlet S0g concentration gave higher exit concentration
and slightly lower percentage removal.
'Decrease in limestone particle size improved absorption but not
to a major degree.
196
-------
*Limej>tone type was not an important variable. Dolomitic
limestone was somewhat less effective, but performance of
high-calcium limestones was generally similar when particle
sizes were comparable.
'increase in soj.ids content of the slurry improved absorption
when other conditions were such as to give relatively low
scrubbing efficiency.
*Ionic strength of the scrubber solution was shown to be sig-
nificant; increase in ionic strength by addition of magnesium
sulfate improved absorption.
*Use of an oxidation inhibitor was effective in reducing sulfite
oxidation.
Calcium sulfite, the preponderant solid species formed, crystallized
in very small platelets that settled slowly and tended to'blind limestone
surfaces when conditions were such as to limit limestone dissolution rate and
thus promote reduction in pH. Oxidation to sulfate was an effective means
of avoiding these problems.
In tests of alkali scrubbing followed by regeneration with lime or
limestone, alkali bisulfite was easily regenerated to sulfite with limestone.
Sulfate purging is still a problem, but if ammonia is used as the alkali the
sulfate can be separated and either sold or regenerated with lime.
197
-------
REMOVAL OF SULFUR DIOXIDE FROM STACK CASES
BY SCRUBBING WITH LIMESTONE SLURRY;
SMALL-SCALE STUDIES AT TVA
By
J. M. Potts, A. V. Slack, and J. D. Hatfield
Division of Chemical Development
Tennessee Valley Authority
Muscle Shoals, Alabama
A test program on limestone scrubbing in a small (5-cfm) con-
tinuous scrubber has been under way at TVA for about 2 years.1 The
main purpose of the tests has been to support the pilot plant program
(2000-3000-cfm scrubber) being carried out at TVA1s Colbert Steam Plant,
which in turn is aimed at supplying information for design of a full-scale
550-mw installation at TVA1s Widows Creek Steam Plant. Much of the work
with the small-scale test loop has been concerned with roughing out the
effect of variables as a guide for the pilot plant; for this the unit has
served the purpose well. However, difficulty in controlling the continuous,
loop-type system closely, coupled with the small size of the unit, has made
it difficult to obtain precise data. Nevertheless, the data obtained can
be used to throw some light on the chemical and kinetic factors involved.
The scrubber loop assembly is shown in Figure 1. Two important
changes in operating procedure have been made during the past year. Easier
startup and better combustion control have been attained by burning straight
natural gas to provide stack gas instead of the previous admixture of gas
with oil containing carbon disulfide; sulfur dioxide is fed to the combustion
chamber from cylinders to provide about ^000 ppm in the stack gas. Also,
commercial spray nozzles are now used instead of the previous impingement
type, making it possible to use lower and more conventional liquid:gas ratios.
The stack gas is fed upward through the spray scrubber, which con-
sists of two separate stages each with its own circulating slurry system.
Fresh limestone slurry fed to the surge tank for the upper slurry system
causes overflow to the lower surge tank. Spent slurry overflows from this
tank and is filtered; the filtrate is used in making fresh limestone slurry
and the spent solids (wet) are discarded.
Slack, A. V. "Absorption of Sulfur Dioxide in Limestone Slurry: Small-
Scale Tests at TVA." Paper presented at International Symposium on Lime/
Limestone Scrubbing for S02 Control (sponsored by NAPCA), Pensacola, Florida,
March 16-20, 1970.
198
-------
STACK
t
CONE NOZZLE
(2nd stage) ,
1(
LIQUOR ,
SCREEN
TC
CONE NOZZLE —k
>"
r
(1st stage) t
12"
LIQUOR 1
SCREEN
TC IR
1 1
6U raw \
OBBATOR' S
/
TRAP
^~i~~
\ * I
u
x^
k j
v
V
f^
1 '
k J
v
TC
IH
GLASS TUBE
k" dia.
>
,1
SLURRY
MAKEUP
)
1 If
V^ >s
$ti
\
1'
SURGE —
, TANK
1
SURGE
TANK
^»
T
L JL
FIGURE 1
Small-Scale Scrubbing System Used in Study of Sulfur Dioxide Removal
from Stack Gas by Limestone Slurry Scrubbing
199
-------
In operation, the unit is started up with the surge tanks full
of limestone slurry of the prescribed solids content. The stoichiometry
is quite high in the beginning, of course, but the slurry is circulated
without limestone makeup until the S02 removal has declined to the ex-
pected level (from previous experience). The limestone feed is then
started. The S02 removal adjusts to the test conditions and usually levels
out quickly. The scrubbing is then continued until continued uniform S02
removal indicates that steady-state operation has been attained. By this
means, a datum point at steady state can be obtained in a day's operation.
Absorption Mechanism
Recent work by various investigators has clarified the situation
somewhat in regard to the nature and extent of reactions occurring in the
scrubber, although there appears to be some disagreement still. There is
some support for the following set of reactions:
S02(g) £ S02(aq) (l)
S02(aq) + H20 £ H2S03 £ H^ 4- HS03- (2)
HS03~ «» H+ + S03= (3)
CaC03(s) «»CaC03(aq) 00
CaC03(aq) ji Ca^ 4- C03= (5)
Ca4"4" + S03= + 0.5H20 «»CaS03-0.5H20(s) (6)
C03= + H+ ^ HC03~ (7)
H+ + HC03" £ H2C03(aq) «* C02(g) + H20 (8)
A diagram of the complicated set of relationships involved is
given in Figure 2.
For such a set of reactions there are several resistances, both
diffusional and chemical, that can affect the rate of the overall reaction:
(s) + s°2(g) +
CaS03-0.5H20(s) + C02(g) (9)
These include (l) diffusion of S02 to and through the gas- liquid film, (2)
dissolution of S02 and CaC03, (3) hydrolysis of S02, (4) diffusion through
the liquid film and into the droplet interior, (5) reaction of H*~ and CaC03,
and (6) reaction of Ca++ and S03=. Added to these are auxiliary rates such
as oxidation rate of dissolved sulfite species and crystallization rate of
CaS03-0.5H20 and CaS04-2H20 (which have a strong tendency to supersaturation. )
200
-------
ro
o
CaC03, . 6
3(aq) -,—
DISSOCIATION CONSTANTS AT 50°C
I. 0.78 x 10-2
2. O.42 x IO~I
3. 0.30 x ID'3
4. 0.84 x IO-~
5. 0.18 x IO-8
6. 0.47 x 10 -3
7. 0.46 x 10"'
8. 0.67 x 10-'°
9. 0.52 x I0~f
10. 0.36 x 10-'
II. 0.57 x I0-5
12.0.48 x IO'2
13. 0.35 x IQ'J
14. 0.22 x|0'4
HENRYS LAW CONSTANT, 50°C
15. 0.54
16. 0.019
Ca
Ca
Ca
I
+H20
~H2O
CaS03-0.5H20(s)
+ co;
OH
S04"
+H20
1S^ CaOH* ll " Ca(OH)2(s)
13
\
(aq)
12 +H
•»-H20
"IT* CaS04- 2H20(S)
-H20
HS04"
FIGURE 2
Equilibria in the System CaO-SOP-SO.q-COg-H20
Added MgO gives similar equilibria, with Mg"1^ replacing C
Added Na20 and N205 give the species CaN03+, NaOH, NaC03M
NaHC03, NaS04-, and NaN03. Added HC1 gives Cl" ions only.
-------
Work by Boll1 has indicated that both diffusion and chemical
reaction are rate controlling, and that rate of CaC03 dissolution and of
gas phase diffusion are likely to be the important parameters. He also
concluded that the chemical reaction is first order with respect both to
the limestone surface area per unit of liquid volume and to the S02 partial
pressure in equilibrium with the liquid.
Distribution of Reactions in Scrubber Loop
There is a further consideration in operation of a countercurrent
scrubber, in that pH can drop as the slurry flows down the scrubber and the
concentration of dissolved sulfite species can thereby be increased. This
not only increases rate of CaC03 dissolution in the lower part of the
scrubber (because the equilibrium partial pressure of S02 over the liquid
is increased) but also may leave some "free S02" in solution that will react
with CaC03 outside the scrubber if given time before return of the slurry
to the scrubber.
This is analogous to use of Mg(OH)2 for S02 absorption in the
paper pulp industry, where all the S02 absorbed during a scrubber pass can
be kept in solution (no crystallization in scrubber) because of the high
solubility of Mg(HS03)2 at low pH. The main difference between this and
the calcium system is that in the latter the solubilities of the various
sulfite species are much lower, so that all the S02 absorbed in a scrubber
pass (under normal conditions of liquor rate and S02 inlet concentration)
cannot be held in solution (except perhaps by supersaturation). Therefore
part of the absorption in a given pass must be accounted for by crystalli-
zation in the scrubber and the remainder by increase of sulfite concentration
in the solution. The magnitude of the latter will depend on the pH level
at the scrubber exit, which in turn depends on inlet S02 partial pressure,
liquor flow rate, and type of scrubber (fully countercurrent vs partially
backmixed).
Calculated data for the effect of pH on solubility of calcium,
sulfite, and sulfate in the system CaO-S02-S03-H20 are given in Figure ~$.
Sulfite solubility increases rapidly with decrease in pH. Therefore the
net amount of dissolved sulfur species picked up in the scrubber depends
on the net change in pH, which depends on many factors. A calculation was
made based on the following assumptions:
pH of inlet slurry 6.0
pH of slurry leaving scrubber 5-0
L/G, gal slurry per Mcf of gas 40
Inlet S02 concentration in gas 2500 ppm
Outlet S02 concentration in gas 250 ppm
1 Boll, R. H. "A Mathematical Model of S02 Absorption by Limestone Slurry."
Paper presented at International Symposium on Lime/Limestone Scrubbing for
S02 Control (sponsored by NAPCA), Pensacola, Florida, March 16-20, 1970.
202
-------
30,000
20,000
10,000
50OO
3000
2000
0.
o.
1000
500
300
200
(00
50
30
20
T T
0 Ca
D SO2 (SULFITE)
A S03 (SULFATE)
— SATURATED WITH BOTH CaS03+ CoS04
' SATURATED WITH CoSOi
(NO SULFATE PRESENT)
SATURATED WITH CoS04
(NO SULFITE PRESENT)
2.5
PH
Solubility Relationships in the System CaO-:SOg-S03-IfeO at 50°C (l22°F)
203
-------
Assuming saturation equilibrium at both scrubber inlet and outlet, about
20$ of the total S02 absorption can be accounted for by formation of
dissolved species, which at the pH involved is mainly bisulfite (HS03~);
the remaining 80$ can be assumed either to have crystallized in the scrubber
as CaS03'0-5H20 or to be held in solution by supersaturation.
The assumed decrease of one pH unit (from 6.0 to 5.0) may or
may not be close to the actual value. The pH at the scrubber outlet is
a transient value, because as soon as the slurry is no longer exposed to
the high driving force of the S02 and C02 in the inlet gas, these gases
can start escaping from the liquid phase. Moreover, dissolved sulfite
species continue reacting with CaC03 and therefore raise the pH. The value
needed is the instantaneous pH at the point where incoming gas and effluent
liquor part company, which is somewhat difficult to measure. In the ICI-
Howden work in England,1 the pH drop in the scrubber was reported to be
0.4 pH unit (from 6-5 to 6.l) for limestone and 0.8 unit (6.9 to 6.l) for
lime. In the TVA pilot plant tests with limestone slurry, the pH as measured
runs about 6.0 in the feed to the scrubber and about 5.7 in the receiving
tank from the scrubber. In the small-scale TVA scrubber, tests were run
in which the pH probes were inserted into the scrubber drain line to give
a continuous reading, thereby avoiding the possible increase in scrubber
outlet pH caused by the delay between sampling and measuring when a grab
sample is taken. For an L/G of 88 gal/Mcf, 97$ removal of 3200 ppm S02,
1.0 limestone stoichiometry, and inlet gas temperature of 150°F, the pH
was 6.06 at the scrubber liquor inlet and 5-22 at the outlet. The drop of
0.8 pH unit is close to the 1.0 assumed in the above calculation.
The foregoing does not take into account the possible effect of
supersaturation on distribution of crystallization in the scrubber loop.
To the extent that calcium sulfite - sulfate forms in the scrubber and does
not crystallize because of supersaturation, the crystallization must take
place outside the scrubber. Calculated data on relationship between pH,
degree of calcium sulfite supersaturation, sulfate:sulfite ratio in solution,
and S02 vapor pressure for the system CaO-S02-S03-H20 are given in Figure i*-.
These indicate that S02 vapor pressure does not increase much with super-
saturation; thus a very high degree of supersaturation would be required to
interfere with absorption in the pH range normally encountered.
Figure 4 also indicates that under the conditions encountered in
limestone scrubbing (pH of about 6 and S04=:total S ratio in solution of 0.60
to 0.90), the solution may be highly supersaturated. In the reports of the
English work, it is indicated that at the circulation rates necessary to avoid
scaling, the supersaturation developed (over saturation concentration) was
about 3*1-0 ppm each for sulfite and sulfate. At the pH level involved (6.0 to
6.3), this would be a high degree of sulfite saturation. In the TVA work,
analyses of the scrubber effluent solution has indicated a considerable super-
saturation of both calcium sulfite and sulfate, considerably higher than had
been expected. The situation is being explored further.
1 Pearson, D. L., Nonhebel, G., and Ulander, P. H. ». J. Inst. Fuel VTII
(39), 119-156 (February 1935).
204
-------
400
200 -
SATURATED
UNSATURATED
ALL SOLNS. SATURATED WITH CoS04-2H20
NUMBERS ON CURVES REPRESENT
PERCENT OF S PRESENT AS SULFATE
NUMBERS ON CURVES REPRESENT
VAPOR PRESSURE OF S02
O.I
gIGURE
Supersaturation of CaS03° 0.
in System
205
-------
The effect of pH on liquid phase composition, assuming a given
degree of supersaturation for calcium sulfite and sulfate, is given in
the computer printouts shown in Table I.
Surge Tank Retention Time; In view of the foregoing, it appears
that the surge tank in the scrubber loop must serve the functions of (l)
completing reactions between limestone and dissolved acidic sulfite species
and (2) crystallizing calcium sulfite and calcium sulfate formed in the
scrubber but not yet precipitated. The size of this tank and the resulting
retention time obviously are important design considerations.
The effect of surge volume on S02 removal in the small-scale
scrubber is shown in Figure 5- The data indicate a major benefit from
increase in surge tank volume up to about 500 gal/Mcfm (or about 1000 gal/
Ib S02*min); further increase had only a small effect. At this surge volume:
gas ratio, the retention time in the surge tank at the 37 gal/Mcf rate was
about 7 ™in as compared with 3 min at the 100 gal rate. Presumably the lower
reactive sulfite concentration at the higher liquor rate required less time
for sufficient limestone to dissolve and react.
A retention time on the order of 2 to 3 min was specified in the
ICI-Howden work (about 1000 ppm S02 and 115 gal/Mcf) to control scaling.
This corresponds to a surge volume of about 1500 gal/lb S02«min.-
Gas and Solution Temperature; One of the more important variables
in the TVA work has been inlet gas temperature (Fig 6). Presumably the
adverse effect of high temperature is due to stripping of dissolved sulfite
species not yet reacted with calcium, thus reducing the amount of reaction
that can be accomplished in the surge tank.
The magnitude of the effect appears too large even for the maximum
amount of unreacted sulfite species that could be present in the solution
(on the basis of calculated values). A possible explanation is that the high
temperature heats the liquid film on droplets and thus interferes with gas-
liquid transfer. In other words, the stripping may not involve dissolved
equilibrium species, which would require both heating the entire drop and
also a relatively low pH in the drop to develop any significant back pressure
of S02. Instead, the dissolution of S02 in the liquid film and the hydrolysis
to HgS03 may be inhibited, which would reduce not only the amount of dissolved
unreacted species but also the supply of sulfite for the Ca4"1"/SQ^ reaction
in that part of the scrubber affected. For the equilibrium
S02(g) *S02-%0(aq) (10)
The effect of temperature on the Henry's law constant (h = aS02-H20/Pg0 )
is as follows:
206
-------
TABLE I
Effect of pH on Liquid Phase Compound at Given Degrees
of Sulflte and Sulfate Supersaturatlon
i>U:
b.O
SI* 4,0
S2» 1,56
PC02» 0,0067
SPECIES ACTIVITY KOLAI.1TY ACT.COCFf
h20
H
OH
H303
CAS03
CA
CAOM
C,\r!C£>3
CAC03
C A304
HS04
SO 4
r!CO.J
COS
0,1000f:-04
o,
0,liU9ti-01
P,'.311C-04
0,11906-04
o!
0.99917
0.7741U-02
0.1102E-08
0.10S6E-05
0.7100E-09
0,110Utf-03
Ci.ll22t-02
0.1095001
0.1409E-OB
0,793flE*PO
0.1012E+01
0.7926G-MIO
0,3947L: + 00
O.KH2E + 01
0,7015^-09 0,1012H*01
O.V193E-05 0,1158c-04
0, J2P2C—03
0,64536-05
0.1252E-03
0.8142E-0&
0.1102E-09
0.3563E-0.0
rt,1000E*01
0,39476*00
CALCjOfl
CAS.T3
C A
CACIH
C/HC03
CAC.M
CA30«
,*,8
63 , rt
0,9
0,0
0,0
32,4
Sl'LFITF. SULFAie
HZSOi
HSOX
SOX
CAS03
0.1
91,2
0,8
8,0
CAS04
HS04
S04
43.7
0 1 1
56,3
C02
CAIICfl.f 1,0
CAC03 fi.O
H2C03 P?,9
HC03 0,0
COS 0,0
f»- *tO Sl= 4,0 S2* 1,56 PC02»
SPECIES
H20
H
OH
HZS03
HS03
SOS
CAS03
CA
CAOH
CAKC03
CAC01
CAS04
KS04
S04
H2C03
HCOS
C03
ACTIVITY MOLALITY
0.10UOt-05. 0.118BE-05
0.547ttC-07 0.6eiflf--07
0.1570E-06 0.1554L-06
O.l??9fe-02 0.1531E-02
.0.5237E-04 n,1259t-03
0.1136F-02 O.J124E-02
0,6416F-i)2 O.l^OSE-Ol
n,9796C-OB 0.12I9F-0/
n,9015C"05 O.Jj?2K-04
0,*>8fl5E-n7 1,5824t-07
0.0753F-T2 0.9652I--OR
O.U CAS03 4P.4
n,u
37,5
ACT.COEFF
0,99939
0,8417E:*00
0 , 8034G>f»0
0,101()F»01
0,^031^*00
n , 4ii>9t* oo
o.ioion*ai
0 , 42"2£- + 09
0,flQ34!;«nO
0 , ft034B*00
0,1010F*Ol
n,ioioF*oi
o,«o34fc*no
0 , 3fllftr.»03
n , i oioii*oi
0,8031^*00
0,4159F:*00
STKJd'lT IOM, %
SULFAT6
CAS04 40,7
HS04 0,0
S04 59,3
> 0,0067
C02
CAHCOJ 5.2
CACD3 0.0
H2C03 *>7.7
HC03 37,1
C03 0,3
* Degree of eupersaturation: for sulflte, h.O times saturation; for aulfate,
1-56.
207
-------
SURGE VOLUME: S02 RATIO, GAL /(LB S02)(MIN)
374 748 1122 1496 1870
80
70
60
50
UJ
O
§
40-
(A
30
20
IO
I
1
2244
I
O LIQUID: GAS RATIO, 100 GAL/MCF
A LIQUID: GAS RATIO, 37 GAL/MGF
ALL TESTS WITH 2% SLURRY OF
COARSE (76% -200 MESH) SPRING
VALLEY LIMESTONE. 3000 PPM S02
IN GAS. GAS TEMP: I25-I35°F
TWO-STAGE, CLOSED LOOP OPERATION
I
I
1
I
> 200 400 600 800 1000
SURGE VOLUME: GAS RATIO, GAL /MCFM
FIGURE 5
Limestone Slurry Scrubbing of S02 from Simulated Stack Gas-
Effect of Surge Volume on SOp Removal
208
J200
-------
TWO-STAGE, CLOSED LOOP SCRUBBING
WITH 2% SLURRY OF SPRING VALLEY
LIMESTONE(76% -200 MESH ; 51.5% CaO).
STOICHIOMETRY:I.O
40
50 60 70
LIQUOR: GAS RATIO, GAL/MCF
80
90
FIGURE 6
Limestone Slurry Scrubbing of S0g from Simulated Stack Gas--
Effect of Liquor:Gas Ratio and Inlet Gas Temperature on S0£ S.emoval
209
-------
Temperature, "F h
150 0.3^3
200 0.169
250 0.092
300 0.055
At the point of initial gas-liquid contact, the low pH in the liquid film
increases the S02 back pressure and the increase in film temperature de-
creases the Henry's law constant, thus reducing transfer through the film
and into the droplet.
The effect of inlet gas temperature may not be as significant in
pilot or full-scale scrubbers as in the small-scale unit. Presumably the
stripping effect of the hot gas would be limited to only a small portion of
the scrubber height, since mass transfer of water vapor and consequent
cooling of the gas is rapid. Since the height of the small-scale scrubber
was quite small as compared with pilot and full-scale sizes, the portion
affected by the hot gas would represent a much greater percentage of the
total height in the small-scale unit as compared with larger ones.
Tests were run also in which the overall solution temperature was
varied and the inlet gas temperature held at 150-200°F (Fig 7). The adverse
effect of higher solution temperature on removal was quite significant.
It has been generally considered that large-scale scrubbers would
be operated at the wet bulb temperature of the gas, which is on the order
of 120-125°F in power plants. However, there may be special situations in
which the wet bulb temperature would be higher or lower; the data in Figure 7
indicate the likely result.
Sulfite Oxidation: The degree of sulfite oxidation in and outside
the scrubber is important in several ways.
1. Degree of oxidation in the scrubber affects the degree of calcium
sulfate supersaturation in the effluent liquid phase, which is
an important factor in scaling (see paper in this symposium:
"Removal of Sulfur Dioxide from Stack Gases by Scrubbing with
Limestone Slurry: Operational Aspects of the Scaling Problem"
by A. V. Slack and J. D, Hatfield). There should be no other
effect on solution composition, however, as the presence of
excess sulfite and sulfate crystals should keep the solution
at least saturated at all times (as compared with oxidation of
dissolved sulfite in a crystal-free slurry, in which case the
sulfite content could drop below saturation and thus affect pH
and species distribution).
210
-------
80
UJ
ec
g
x
o
u.
-I
3
(0
60
50
40
L/G : 79 GAL /MCF
INLET GAS TEMP: 150-200°F.
STOICHIOMETRY: 1.0
PARTICLE SIZE:
72% MINUS 200 MESH
SOLIDS IN SLURRY: 2%
90 100 120 130 140
AVERAGE TEMP IN SLURRY SURGE TANKS, °F
ISO
FIGURE 7
Limestone Slurry Scrubbing of S0g from Simulated Stack Gas:
Effect of Slurry Temperature
211
-------
2. There is some evidence that under upset scrubber conditions
which promote low pH, the increased sulfite concentration
in solution resulting from the low pH causes sulfite cry-
stallization on limestone surfaces and consequent blinding.
Increase in oxidation rate would reduce the sulfite content,
which perhaps would decrease the blinding effect.
J. Degree of oxidation both in and out of the scrubber affects
the sulf ate: sulfite ratio in the solid phase, which is
important in solids settling because the sulfate crystals
grow larger and thus settle better (Fig 8).
k. Oxidation throughout the system, including the waste pond,
reduces the possibility of water pollution by sulfite.
5- If a recovery process were operated with CaO-CaC03 as the
absorbent, oxidation would not be desirable because difficulty
in decomposing sulfate (as compared with thermal decomposition
of sulfite to give a rich stream of S02 and to regenerate CaO
for recycling) would likely make it better to discard the
sulfate and replace it with makeup limestone.
Data on degree of oxidation in the small-scale tests have been
somewhat erratic, ranging from 25 to 100$ depending on type of spray nozzle
and other factors; with the present atomizing nozzles, the range is 70 to
100$. The tests are not applicable to the actual power plant situation
anyway, because of the absence of fly ash (which may or may not promote
oxidation), absence of phenolic impurities in the gas . (which are known to
inhibit oxidation), and possible differences in N02 content (which is known
to promote oxidation).
Data on the composition of solids from pilot plant runs are given
in Table II.
The calculated species composition for run B is given in Table III.
There was no opportunity in these runs to study the effect of
operating variables on oxidation. Work reported in the literature indicates
the following general relationship for sulfite oxidation,
- d(SOg) = g[HSO~"]
dt
where the lefthand term is rate of sulfite oxidation, the term g includes
the effects of free radicals and other effects characteristic of the system
under study, and [H] is the activity of the hydrogen ion. Thus a decrease
in pH may or may not reduce oxidation, depending on how rapidly the HS03~
concentration increases as the pH is reduced. In the limestone system,
212
-------
FIGURE 8
Electron Micrograph or Solids Taken from Limestone Scrubber Slurry
The large blocky crystal in the upper lefthand corner is Ca304-2H20> the
small spheres art fly ash, and the flat platelets are CaS03-0.5H20, Some
limestone is also present. The rosette in the center is made up of calcium
sulfite crystals. (Photograph from work by Dr. G. H. McClellan, TVA, )
213
-------
a drop in pH from 6.0 to 5.0, in solutions saturated with CaS03'0.5H20,
increases the value of the [HS03~]/[H]1 2 term from 2.10 to 2.35, indicating
that pH reduction in this range promotes oxidation. The effect of pH on
the term under various conditions of saturation is given in Table IV. The
data indicate that pH increase inhibits oxidation except when the solution
is saturated with both calcium and sulfite.
TABLE II
Composition of Solid Phase from Pilot Plant Limestone Slurry Tests3
Weight percent in various runs
Constituent ABC
Ca 2k.Q 26.6 23.5
Mg 1.0 0.8 0.8
Total S 11.7 10.1 8.0
S as sulfite 9-8 8.2 6.8
S as sulfateb 1.9 1.9 1.2
Degree of oxidation, % 16 19 15
a Limestone composition, %: CaO, 52.3; MgO, 1.2. S02 in inlet
gas, 2800 ppm. L/G = 60 gal/Mcf. Solids content of slurry:
b 15-20*.
By difference.
TABLE III
Species Distribution in Solid Phase
a
from Limestone Slurry Scrubbing
Species Weight percent
CaS03-0.5H20 33.0
CaMg(C03)2 6.2
CaS04-2%0 10.2
CaC03 31.5
Fly ash 19-0
Stoichiometry: 1.5. S02 removal,
about 65$.
214
-------
TABLE IV
Effect of pH on [HS03~ KH"!12 in System CaO-S02 H20
At constant Ca
moles/kg H20
£H_
6.0
5.0
4.0
6.0
5.0
4.o
6.0
5.0
4.0
Ca (. total) SOP (total)
0.00152p
0.001525
0.001525
At
0.001525
0.001314-2
0.001258
0.001525
0.0014.520
0.01664
0.002649
0.003005
0.003188
constant S02
0.002649
0.002649
0.002649
At constant
0.002649
0.008725
0.03342
Saturation [HS03~ "]/ [H4] V2
1.00
0.15
0.017
(total sulfite)
1.00
0.12
0.012
saturation
1.00
1.00
1.00
2.096
0.864
0.291
2.096
0.766
0.244
2.096
2.353
2.670
The preceding discussion refers only to oxidation rate in solution
assuming that oxygen is already present. The rate of oxygen absorption
should also be important in determining overall rate in a limestone slurry
scrubber. Chertkov1 proposed a generalized empirical equation for oxygen
absorption based on data from several scrubbers and differing absorbents.
GO = 0.8Q°-7.a.(S/C)6 (12)
Y ' M-
where GQ = g 02 absorbed per hour per m2 of liquid/gas contact
surface (considered to be equivalent to rate of oxidation)
Q = liquid flow rate, m3/m2-hr
a = t ) where t is the average solution temperature in °C
50
S/C = mole ratio of sulfite to absorbing species in solution
Y = solution density, kg/m3
jj, = solution viscosity, kg. sec/m2
1 Chertkov, B. A. J. Appl. Chem. USSR 3J4_ (4), 743-47 (l96l).
215
-------
The data currently available from TVA work are not sufficient to evaluate
this equation. However, Q is the only variable that can be changed easily,
and the minimum level of Q is set by the S02 removal requirement.
Another factor is the ratio of 02 and S02 partial pressures in the
gas phase, which affects the overall oxidation rate by changing the amount
of 02 absorbed per unit of S02. Thus a relatively low S02 content in the gas
should give relatively high oxidation rate. Comparison of current pilot plant
tests on gas from western coal (if-00-600 ppm)--plus results from the ICI-Howden
work (1000 ppm)--with the TVA tests (2500-3000 ppm) indicate that this is true.
The situation in regard to the various promoters involved (fly ash,
N02, limestone constituents) and inhibitors (phenolics) is somewhat confused.
More work must be done before the relative importance of these factors can
be determined.
The effect of added inhibitors was tested in the small-scale TVA
work. Under conditions that gave 50$ oxidation of the solids, the oxidation
was reduced to 1$ by addition of 0.1$ hydroquinone. At the lower pH level
resulting from use of benzoic acid to promote S02 removal, use of the same
amount of hydroquinone reduced oxidation from 95 to 20$. Presumably the
larger amount of sulfite available in solution for oxidation at the lower
pH increased the percentage oxidation of the solids--both with and without
inhibitor. This assumes that oxidation involves the steps of sulfite disso-
lution, oxidation of sulfite in solution, and crystallization of sulfate.
The higher solubility at low pH is taken into account in Table IV.
The discussion thus far has been concerned with oxidation in the
scrubber, which may or may not be practicable to change in actual operation.
It should be feasible to oxidize outside the scrubber, however, if oxidation is
indicated as a means of coping with settling, blinding, and pollution problems.
This could be done by passing air or oxygen through the slurry, either through
the recycled slurry or through the sidestream diverted for solids separation.
The problem is getting an adequate rate of oxygen absorption in the
slurry, which in Japan has led to development of a special "spinning cup"
technique to increase absorption. This could be used, or the operation could
be carried out under pressure, or both.
Exploratory tests with the spinning cup oxidizer have been made in
the TVA small-scale work. Oxidation was much faster than with introduction of
the air through fritted glass into a stirred solution; about four times as much
air throughput was required with the fritted glass to accomplish complete oxi-
dation. Activated charcoal and titanium increased the rate somewhat but not
significantly. Use of ultrasonic energy was not effective.
In the pilot plant, one exploratory test was made in which air was
bled into the scrubber (airrgas volume ratio about l:l) to increase oxidation
rate. Oxidation was increased by a major degree, as indicated by increase in
calcium sulfate content of the slurry solids.
216
-------
An oxidizing tower has been installed in the pilot plant and
will be tested on the solids-separation sidestream when time is available.
Tests have also been made to determine whether sulfite will
oxidize on standing in the waste pond. Samples of moist product slurry
solids-were stored in 1-in. vertical tubes, unstoppered but. taped to pre-
vent exposure to light. No oxidation occvirrad in storage over a period
of 16 weeks.
Surface Area Holdup in Scrubber
Since rate of limestone dissolution appears to be. a controlling
factor in S02 absorption, it is important to have a large limestone surface
area per volume of slurry and to hold this area in contact with the gas in
the scrubber as long as possible. Hence the requirement can be expressed
as surface area of limestone exposed to each unit of S02 passed through the
scrubber. There are several variables that affect this.
Solids Content of Slurry; A simplified flowsheet for the scrubber
loop is shown in Figure 9. A large body of crystals, mainly CaC03 and
CaS03*0.5H20 (plus smaller amounts of CaS04«2H20, CaMg(C03)2, and fly ash),
circulate around the loop, to which is added a relatively small stream of
CaC03 and from which is withdrawn a relatively small stream made up of
calcium sulfite, unreacted calcium carbonate, and the other solid con-
stituents. Thus the gas is exposed to an amount of limestone in the recy-
cling solids that can be much larger (per unit of S02) than the amount of
limestone fed. The concentration of limestone in the recycling slurry
therefore is an important variable, with a linear effect on the "effective
surface area" (limestone surface area in scrubber per unit of S02 per unit
of time)o
Only limited data on effect of solids content in the slurry have
been obtained in the small-scale tests; examples are given in Table V.
The improvement from increasing solids concentration under conditions that
gave a low level of S02 removal (single-stage test) was much larger than
for those that gave a fairly high level (two-stage). This has been generally
true for all variables; more benefit was realized, even on the basis of
percentage reduction of unremoved S02, when there was more room for improvement.
The increase in limestone surface area per volume of slurry at
steady state brought, about, by doubling the solids content of the slurry
should have been about 35$ for the one- stage tests and rf^% for the two-stage.
In the pilot plant, a much higher solids content is used ( 12-
than is feasible for the nozzles in the small-scale spray tower. The high
level is used mainly to provide sulfite-sulfate crystals for scaling control,
but some improvement in S02 removal has been noted also.
217
-------
SO2
t
GAS TO STACK
CaSO3
< CaC03
CaS03i CaC03
FIGURE 9
Simplified Flowsheet for Limestone Scrubber Loop
Particle Size; The limestone surface area is inversely proportional
to the diameter of the particles; for example, breaking 200-mesh particles
into 325-mesh particles (W^) should increase surface area by about
An example of the effect of particle size on S02 removal is given
in Table VI. Some improvement resulted from finer grinding but the effect
was not a major one.
Stoichiometry; Feeding of excess limestone beyond the stoichiometric
amount should have a considerable effect on limestone content of the slurry
in the circulating loop at steady state. For a Stoichiometry of 1.0 and S02
removal of 80$, 20$ of the calcium (molar basis) in the loop, at steady state,
should be in the form of unreacted CaC03. For 1.2 Stoichiometry, assuming
the same removal, 33% should be CaC03. Thus a 20$ increase in external
Stoichiometry gives a 33$ increase in "internal" Stoichiometry.
218
-------
TABLE V
Effect of Limestone Slurry Concentration on S0g Removal
a
Slurry concentration, % S0g removal,
One- stage, L/G = 37
1 38
2 59
Two- stage, L/G = 72
1 75
2 78
5 79
Feed stoichipmetry : 1.0. Closed- loop operation.
Gal/Mcf to each stage.
TABLE VI
,a
Effect of Limestone Type and Size on SOo Removal
Calcitic Dolomitic,
Hard Chalk hard
Particle size, mesh 76$ -200 92$ -325 96$ -325 92$ -325
S02 removal, % 78° 83 78
* L/G = 72. Stoichiometry: 1.0. Two-stage closed loop.
Three different quarries.
Data on effect of Stoichiometry are given in Figure 10. For tests
at less than stoichiometric, the removal increased almost linearly with amount
rof limestone fed. Above 1.0 the effect was small; however, test conditions
were such that excellent removal was obtained at 1.0 and there was not much
room for further improvement.
219
-------
100
80
O
5 60
a:
UJ
o
x
§40
U.
_J
20
TWO-STAGE, CLOSED LOOP
SOLIDS CONTENT, 2%
LIQUOR/GAS RATIO, 77 GAL/MCF
INLET GAS TEMP, I40-I75°F
1
i
1
0 1/3 2/3 '
LIMESTONE FED, PROPORTION OF STOICHIOMETRIC
FIGURE 10
Effect of Stoichiometry on S02 Removal by Limestone Slurry Scrubbing
220
-------
In the pilot plant work (see paper in this symposium: "Removal
of Sulfur Dioxide from Stack Gases by Scrubbing with Limestone Slurry:
TVA Pilot Plant Tests — Part I, Scrubber-Type Comparison" by T. M. Kelso,
P. C. Williamson, and J. J. Schultz), there has been some indication
that operation with excess limestone (perhaps as high as 1.5 stoichio-
metric) may be advisable to ensure that conditions do not develop which
cause the pH to become unstable. If this is borne out by further tests,
then the effect of stoichiometry on S02 removal becomes a secondary con-
sideration.
Liquid:Gas Ratio: The amount of slurry holdup in the scrubber,
which is the major factor in surface area holdup, depends on scrubber type
and on circulation rate. Very few data are available regarding effect of
scrubber type on holdup and, in any event, scrubber type is likely to be
dictated by scaling considerations rather than S02 removal efficiency.
Data on effect of slurry circulation rate (l/G) are given in
Figure 6. The effect is major; in fact, slurry rate and inlet gas tempera-
ture have been the only variables with a major effect in the TVA small-scale
work. This has also been the situation in the pilot plant tests. Again,
high circulation may be essential for avoiding scaling, which makes its
effect on S02 removal secondary.
Other Rate Factors
Limestone Type; It might be expected that the physical and chemical
properties of the limestone, and their effect on rate of limestone dissolution,
would be important rate factors. Such properties are important in the dry
sorption method. Tests in the wet system with various limestones, however,
have not indicated any major effect of limestone type—although the test work
has not been extensive.
Examples of data from such tests are given in Table VI. The chalk
used was a soft limestone that had seemed in preliminary tests to give better
results than hard limestones. However, because of its softness the chalk
tends to break down into very small particle size during the scrubbing operation,
which may have contributed to its activity. When hard calcitic limestone was
also ground to very small particle size, the S02 removal efficiency was the
same.
Dolomite was somewhat less effective although not greatly so. This
may be associated with the reportedly slow reaction of MgC03 minerals with
weak acids.
The pilot plant work has supported these conclusions. No significant
difference has been noted between the various hard limestones tested.
221
-------
Blinding: Since pilot plant results had indicated the possi-
bility of limestone blinding during periods of unstable pH, small-scale
tests were made to explore ways for avoiding the problem. Blinding had
never occurred in the previous small-scale work, presumably because
conditions were such as to ensure steady pH (perhaps resulting from the
relatively high liquor rate necessary in the small-scale equipment).
The pilot plant results indicated that if for some reason the
pH in the loop decreases to a low level, the slurry becomes unresponsive—
that is, further addition of limestone does not raise the pH to the normal
level of 6.0-6.1. In tests with the small-scale scrubber, it was found
that a similar unresponsive slurry could be developed by operating with a
deficiency of limestone. In further studies with batches in a gas bubbler,
a solution of CaS03-0.5H20 in H2S03 solution (pH, 2.6) was found to be also
unresponsive; additions of limestone would not raise the pH above $.6. An
electron micrograph of solids from such an unresponsive slurry (Fig ll)
shows that blinding of the limestone surface by calcium sulfite apparently
had taken place.
In further tests, the unresponsive slurry was oxidized by bubbling
air through it for ~$ hours, which precipitated about 97$ of the dissolved
calcium as CaS04-2H20. The slurry was then responsive; addition of limestone
to the oxidized slurry (in which the pH had decreased to 1.9) brought the
pH up to 6.2.
A possible explanation for the blinding is that if pH for some
reason decreases during operation, presumably because the rate of limestone
dissolution lags behind S02 absorption, the sulfite content of the solution
increases rapidly because of higher sulfite solubility at low pH and the
reservoir of sulfite present as sulfite crystals. Then when a limestone
particle dissolves at its surface and forms a film around it of higher pH
solution, the pH increase causes rapid sulfite crystallization at the surface
and consequent blinding.
One difficulty with this postulation is that adding S02 to a
saturated solution of CaS03'0.5H20 canjnake the solution unsaturated. The
effect is shown in Table VII. The S03~ concentration decreases rapidly with
pH by shifting back to HS03-, with which it is in equilibrium (see Fig 2).
Thus if the Ca++ concentration remains constant, or even if it increases
somewhat, the Ca++/S03~ solubility product is not exceeded because of the
very low S03~ concentration.
Hence if limestone dissolution were inhibited and the pH were
lowered by S02 absorption, the solution might well become unsaturated and
would be less likely to precipitate CaS03-0.5H20 than under normal operating
conditions.
222
-------
FIGURE 11
Electron Micrograph of Solids Taken from Limestone Scrubber Slurry
The particle at the right is a crystal of calcium sulfite
(CaS03-0.5Hao). At the left is a fly ash particle and in
the center is a limestone particle apparently blinded by
calcium sulfite crystals growing edgewise on its surface.
The scale shown is 1 micron. (photograph from work by
Dr. G. H. McClellan, TVA).
Another possibility is that the unsaturation developed at low pH may
actually cause dissolution of CaS03-0.5 H20 crystals and reduce the surface
area available for crystal growth. Then when CaC03 is .added in the delay tank
and a surface film of relatively high pH developed, the sulfite supersaturation
induced by the higher pH may cause nucleation on the CaC03 surface rather than
on the sulfite crystals that would ordinarily have been present. More data
must be gathered before any conclusions can be made.
223
-------
TABLE VII
Concentration of Sulfite Species at Constant Calcium
Concentration (0.001525 m/l) but with Varying pH
Concentration, moles/1
H2S03 HS03- S03= CaS03 Total S02 m/l
pH x 1Q6 x 10+3 x 10+ 5 x 105 x IP3
6.0 0.27 2.2i»- 11.80 28.37
5-0 3.49 2.94 1.57 fc-33 3.005
^•0 37.2 3.14 0.17 o.Vf 3-188
The rapid undersaturation developed at lower pH may also explain
why sulfite crystals are usually small. Even the pH variation in normal
operation may cause wide swings from rapid dissolution to rapid precipitation
from supersaturated solutions.
If an insufficient rate of limestone dissolution is the initial
cause of the problem, then anything that increases the rate should be helpful.
In the pilot plant, increase in limestone stoichiometry has been effective.
Length of operating period also seems to be a factor, presumably because the
ionic strength builds up and makes calcium more soluble (see later discussion).
It may be possible to ensure steady pH at lower limestone stoichi-
ometry by feeding a little lime or ammonia along with the limestone. Tests
of this procedure are planned.
There may be some blinding even at normal operating pH; adequate
data for evaluating this have not yet been obtained. If so, grinding of
the recycling slurry to renew the limestone surface, as practiced by Bahco
/see paper by K. A. Gustavsson, "Bahco S02 Scrubber," presented at Internatioal
Symposium on Lime/Limes tone Scrubbing for S02 Control (sponsored by NAPCA),
Pensacola, Florida, March 16-20, 1970/ may be indicated.
Ionic Strength: Increase in ionic strength by buildup in the
solution of constituents such as Mg, Cl, Na, K, and N03 should make Ca more
soluble and therefore increase the driving force for dissolution. Calculated
solubilities for various ionic strengths are given in Table VIII. The data
indicate an optimum level for calcium solubility.
224
-------
TABLE VIII
Solubility in the System CaO-SOg-HgO
as Affected by Ionic Strength and pH
Ionic strength
0.004-0. Oi<.a
ppm
PH
0.004-0.04*
ppm
Ca
SOo
1.0°
ppm
Ca
S02
4. 0&
ppm
Ca
SOo
6.0
5-0
4.0
61
181
667
170
558
2139
135
353
1077
360
1093
3446
92
231
701
236
711
2334
a
In the system CaO-S02-H20.
Ionic strength increased by adding noncom-
plexing ions such as Cl~ and N(CH3)4+.
In the small-scale work, the simulated nature of the stack gas—
and the consequent absence of ionizable impurities such as carried in actual
stack gas—keeps the ionic strength from building up. Tests were run,
therefore, in which the ionic strength was varied by adding various amounts
of magnesium sulfate (Fig 12). An optimum level of ionic strength is indi-
cated at about 1.0.
The pilot plant tests have given the opportunity of determining
the degree to which ionic strength will build up in actual operation, con-
sidering the fact that part of the solution is purged with the wet solids
and replaced with water. Extended round-the-clock runs have indicated that
steady state in regard to ionic strength is attained in a day or so; there
is little further change in the concentration of constituents such as Mg,
Fe, Na, Mn, K, and Cl. There has also been little variation in Ca and total
S, although these are subject to change with minor variations in pH, Average
composition over a 2-week run is given in Table IX.
The ionic strength of this solution is about 0.120, considerably
below the indicated optimum. The only apparent way to increase it is to bleed
less liquid phase with the solids, but the 60$ liquor content of the purged
solids obtained in the current tests may be difficult to reduce.
Inlet S02 Concentration: In the usual gas scrubbing situation, a
decrease in inlet concentration of the constituent to be absorbed would be
expected to decrease the percentage absorption because of the lower driving
force for gas-liquid transfer. For limestone slurry scrubbing of S02, however,
the role of solid-liquid transfer complicates the situation.
225
-------
IONIC STRENGTH
1.77
5 10 15 20
MgS04 CONTENT OF SLURRY, %
25
FIGURE 12
Effect of MgS04 Addition on S0g Removal
in Closed-Loop Limestone Slurry Scrubbing
226
-------
TABLE IX
a
Average Filtrate Composition Over 2-Week Pilot Plant Run
Constituent Concentration, gpl
Ca
Mg
Total S
Fe
Na
Mn
K
Cl
Total N
1.06
0.46
0.91
0.001
0.038
0.002
0.059
1.32
o.ok
Limestone stoichiometry: 1.5- L/G: About 40 gal/Mcf.
Solids content of slurry: 12-15$* Water content of
purged solids: About 60%. Samples taken at scrubber
outlet.
Data from tests on effect of inlet S02 concentration are given
in Figure 13. The percentage removal was actually higher for the lower
concentrations, probably because of the smaller amount of limestone
dissolution required per unit volume of slurry (the liquor rate was the
same for all of the tests). The effective surge tank volume (gal/lb
SOg-min), which has been shown to be important, was much higher at the
lower S02 concentrations (range from 1240 to 4900 gal/lb S02-min).
Thus it appears that low exit S02 concentrations, on the order
of 50 PP™ or less, can be attained with high liquor rate and surge tank
retention time.
Waste Disposal
As noted earlier (Fig ll), the calcium sulfite crystals formed
in limestone scrubbing are very small, thin platelets. When the slurry is
allowed to stand, the crystals settle rapidly for a short time—giving a
fairly clear supernatant layer—but the settling rate then drops off sharply.
It is postulated that the decreased rate begins when the crystals start
touching each other; a voluminous, gel-like structure develops in which
settling is relatively slow.
227
-------
PO
ro
oo
100
I I I
%-200 MESH LIMESTONE
O 95
A 85%-200MESH LIMESTONE
500
Effect of
1000 1500 2000 2500 3000 3500
S02 IN INLET GAS, PPM
FIGURE 13
Inlet SO, Concentration on SO. Removal with Two-Stage Closed-Loop Limestone Slurry Scrubbing
The liquortgas ratio was 80 gal/Mcf, the solids content of the slurry 2*,
and the inlet gas temperature 150-200°F.
-------
In small-scale tests with the slurry, it was found that the
average rate of free settling was about 2 in./hr; this rate was fairly
uniform as long as the free settling phase lasted. A sample of the gel
formed (l6$ solids) after compressive settling began was tested with a
gelometer; gel strength increased with time, as follows.
Time elapsed after
stirring, min Yield point, g-cm
0 0
30 6
60 9
1080 25
o
Settling of the gel could be accelerated by periodic slow stirring
to destroy the gel structure. A gel that had settled to 38$ solids and
showed no indication of further significant settling was settled to 48$
solids by three stirrings about a day apart.
Oxidation was quite effective in promoting settling. Further
work will be done on this approach.
Alkali - Lime-Limestone
The many problems involved in lime-limestone scrubbing make it
quite desirable to find a better throwaway method. Several organizations,
including TVA, are working on a process that involves scrubbing with an
alkali solution (NH4, Na, K.) and regenerating the alkali with lime or lime-
stone. The waste product is calcium sulfite (plus calcium sulfate, hopefully)
as in straight lime-limestone scrubbing. Keeping the lime-limestone out of
the scrubber should greatly reduce the problems of scaling, erosion, silting,
wet-dry deposition of solids, particulate entrainment, and high pumping cost.
Of the various alkalies, Na and K have the advantage that they are
not volatile as is NH.a -which simplifies scrubber design and operation. How-
ever, sulfate formed by oxidation in the scrubber is quite difficult to regen-
erate; there have been efforts over the last several decades to convert byproduct
sodium sulfate to caustic soda without success. Regeneration in an S02 scrubbing
loop may be feasible, however, because only a partial conversion per pass is
acceptable, but it is likely that the resulting NaOH solution would be quite
dilute and that circulation rate might therefore have to be high.
•
If ammonia is used as the alkali, regeneration may be somewhat
easier since (NH4)2S04 can be crystallized from a sidestream and sold (whereas
Na2S04 would have a quite limited market). Or the (NH4)2S04 can be reacted
with lime to regenerate the ammonia, as in the Kuhlmann Electricite de France
process.
229
-------
In the TVA small-scale work, limestone rather than lime has
been used as the regenerant since it is cheaper. It reacts well with
alkali bisulfite (Fig 14) but not with sulfite; however, the sulfite can be
recirculated to the scrubber as is the common practice in alkali scrubbing
processes of the recovery type.
The data in Figure 14 indicate that the regeneration temperature
should be somewhat elevated; fortunately, in power plant stack gas scrubbing
heat from the stack gas gives an adequate liquor temperature level, which
may not be true for other types of stack gas.
Since the calcium sulfite formed in the tests settled well, no
difficulty is expected in separating and washing the waste solids. Further
work will be concerned with separation of (NH4)2S04. Data on the ammonia
scrubbing step is being obtained in the current EPA-TVA ammonia scrubbing
pilot plant project. Testing of the integrated process is planned in a
TVA pilot plant now under construction.
230
-------
B
110
•I fV*\ •^» ^__ ^^_ ^^M ••••§ *mmm M^ «••
ro
CO
_ TESTS AT 75 F
1*
Theoretical maximum
/^reaction
LEGEND (for both figures)
Q Sodium salts
A Potassium salts
Q AmtoDnium salts
Salts in solutions con-
taining 16$ "bisulfite and
2% each of sulfite and
sulfate.
110
- 100
1.25 0.5 0.75
Ca:2HS05 mole ratio
FIGURE Ik
Effect of Calcium Carbonate Proportion on Degree of Reaction
with Solutions of Alkali Bisulfite-Sulfite-Sulfate Salt Mixtures
1.0
1.25
-------
Absorption Studies of Equimolar
Concentrations of NO and NCL In
Alkaline Solutions
By: L, H. Garcia
Division of Control Systems
Laboratory Research Branch
Special Projects Section
233
-------
ABSTRACT
ABSORPTION STUDIES OF EQUIMOLAR CONCENTRATIONS OF NO & NO IN
ALKALINE SOLUTIONS
Equtnolar concentrations of NO and NO- in flue gas were passed
countercurrently to various alkaline solutions in a bench-scale
packed column. The concentrations of NO and NO. studies were from
250 ppm to 750 ppm each. Liquid velocity, gas velocity, size of
packing, and liquid temperature were the parameters varied.
The results of this experimentation indicates that at a concentration
of 250 ppm each of NO and NO. in flue gas the percent removal for
NO and NO., expressed as total NO was between 12 to 14 percent; at
500 ppm each of NO and NO. in flue gas, the percent removal was between
17 to 19 percent; at 750 ppm each of NO and NO. in flue gas, the percent
removal was between 27 to 29 percent.
234
-------
Absorption Studies of Low Concentrations
of NO and VK>2 in Alkaline Solutions
BACKGROUND
The Federal Government has enacted legislation to regulate and control
the levels of nitrogen oxides in the ambient air and to limit their emission
from specific new stationary sources. The control of nitrogen oxides is
important because these air pollutants are Involved in the complex photo-
chemical reactions in the atmosphere which result in smog formation and
because, by themselves, they have adverse physiological effects for all
forms of life. Although, a number of oxides of nitrogen are recognized,
the most common forms are nitric oxide (NO) and nitrogen dioxide (NO-) on
account of their widespread production and their relatively high atmospheric
stability. In this paper the term "nitrogen oxides (or NO )" refers to
either or a combination of these two gaseous air pollutants.
Mobile sources, the largest single source category, contribute over 40
percent of all the man-made NO emitted in the United States. Current
knowledge on the methods of control are covered in detail in AP-66, "Control
Techniques for Carbon Monoxide, Nitrogen Oxide, and Hydrocarbon Emissions
from Mobile Sources." The next largest source is electric power generation,
which is responsible for nearly 20 percent of all man-made NO . About
40 percent of NO emitted from stationary installations is attributed to
electric generating power plants. Control of NO from stationary sources
is discussed in detail in AP-67, "Control Techniques for Nitrogen Oxide Emissions
from Stationary Sources."
About 1 percent of the total man-made NO emitted to the ambient air
of the United States is formed by chemical sources, mainly related to the
manufacture and use of nitric acid. Concentration from these sources are
usually much greater, however, than those from combustion sources, and as a
consequence, these sources often give rise to a highly visible, brown-red gas.
THEORETICAL CONSIDERATIONS
Under proper conditions, nitrogen and oxygen tend to combine in
accordance with the following equation:
N2 f °2"* 2N°
235
-------
The equilibrium concentration of NO varies with temperature; it ib
negligible below 1,000°F but quite significant above 3000°F. In addition,
it is influenced by gas composition; at a given temperature, for example,
the equilibrium concentration of NO in air exceeds that of NO in flue gas
of 3 percent oxygen content by a factor of approximately 3.
Nitric oxide tends to react with oxygen as follows:
2NO + 0
This equation implies the coexistence of NO and NO-* Calculated equilibria
indicate that the stability of NO. decreases with Increasing temperature.
Nevertheless, from an equilibrium standpoint, the absolute concentration
of NO. increases with temperature while the ratio of its concentration
to that of NO decreases with increasing temperature.
Chemical equilibria depend on the initial and final conditions and
not at all on reaction mechanisms or Intermediate reaction steps. Equilibrium
concentrations are obtained after the lapse of sufficient reaction time;
therefore, they are not necessarily observed experimentally. Because of
the simplicity of the molecules involved In the previous equations, their
thermodynamic properties are accurately known and equilibrium calculations
are made easily.
It has been suggested that equimolar mixtures of NO and NO-
might be removed by scrubbing with alkaline solutions. The use of the
alkaline materials (NaOH and Na.CO-) for NO scrubbing was studied
originally by Sherwood and Pigford ' ^ wh5 found, using a concentration
of one percent NO in the gas phase, that a 46 percent NaOH solution absorbed
nitrogen oxides from a gas containing equimolar quantities of NO and NO-
at rates as much as ten times faster than if NO. were present alone.
When they compared the NO + N0~ absorption rates with the calculated
equilibrium concentration of N^O- (NO + N02 -»• N-0-), it was found that
the rate was approximately proportional to the First power of the N-O-
concentration, gindlcating that this was the reacting species. Ganz and
co-workers performed similar studies using sodium carbonate solutions
and reached similar conclusions. Andrew and Hanson™ describe the dynamics
of nitrous gas absorption into water as the result of a number of separate
reaction mechanisms, and state that relative importance of these mechanisms is
primarily dependent on the gas composition.
In order to determine the applicability of aqueous alkaline scrubbing
of equimolar NO/NO, mixtures at typical flue gas concentrations, the EPA
laboratories decided to study the absorption of NO and NO. at concentrations
in the range of 250 ppm to 750 ppm each in flue gas.
Although control of most of the NO emissions from combustion sources
is expected to be achieved by various combustion modification or control
methods, additional flue gas or process stream treatment methods may be
required for specific problem areas. Consequently, a limited effort has been
236
-------
directed to research, leading to the identification and development of effective
flue gas treatment techniques for NO control.
Rates of formation can be calculated by kinetic equations that rely
heavily on experimental measurements. The reaction products depend to a
large measure on the relative speeds of the reactions that actually occur.
The rate of oxidation of nitrogen to NO is highly temperature-dependent;
it is very slow at 5QO°F, but fast at 4000°F. The underlying reason is that
a high level of energy is needed to break the N-N (dissociation energy 225.3
kcal/mole) triple bond of molecular nitrogen so that oxygen can react.
Conversely, a smaller but still relatively large amount of energy is needed
to break the N-0 bond to permit decomposition of nitric oxide into its
elements. This means that high temperatures are required to form NO; once
formed, it resists any breakdown into its elements. Breakdown becomes more
and more unlikely as temperature decreases, because the energy available for
thermal breakdown diminishes rapidly with, decreasing temperature. Thus, an
Initially high temperature followed by quick cooling, even to a relatively
high temperature level, produces large amounts of NO.
At temperatures above 2000°F, both NO and NO, are formed, but the amount
of NO, Is usually less than 0.5 percent of the total NO . The oxidation of
NO to NO2 ky oxygen, however, is peculiar in that the rate of reaction
decreases with increasing temperature. This is one of the few known reactions
that exhibit an inverse temperature relationship. The resultant slow
oxidation rate at high temperatures accounts in part for the negligible
amounts of NO, frequently found in hot combustion gases. Another
characteristic of the oxidation of NO to NO, is the fact that the rate
varies with, the square of the NO concentration. The rate of oxidation of NO
by oxygen in air falls off rapidly, therefore, with dilution of the NO.
A long period of time may be required to oxidize trace quantities of NO
by this mechanism.
EXPERIMENTAL
A laboratory bench-scale packed scrubber was fabricated (see Figure
1). The scrubber consisted of a glass (Pyrex) column 2 1/4 inch I.D. packed
with eighteen inches of 3/8-inch Berl saddles. The scrubber was equipped
with an inlet liquid port and an outlet gas port at both the top and bottom
of the scrubber. Because of this, the laboratory bench-scale scrubber was
capable of passing scrubbing liquor cocurrently as well as countercurrently
to the gas flow over a wide range of liquid to gas (L-to-G) flow rates.
In. the bencft-scale packed scrubber, the following parameters can be
varied and measured: gas and liquid flow rates, concentrations of the gaseous
pollutants, packing height, type of packing, temperature, and concentration
of alkali.
237
-------
FIGURE I
FLOW DIAGRAM OF EQUIPMENT
USED FOR ABSORPTION OF NO 8 N 02
pQ
X-<
J FLOiV i
J McTER
1 -C
y
i
1
.
y
i ,
_»•
>
|
*
ri
*
RgCORDSR
'INFRARED
c"rt *"'
2
RECCROuR
— ..
f
IKrRAfJED
KO
t —..«/*
RECORDER \
v
V
— i
1
ULTRAVIOLET
ANALYZER
N02
r-r
1
* C
r~^-^
c*i i/->f i f /^r^
/v
\
< CH/}
«•— •
•« ' AiH
'4 i
_^
^ "I
A A p.
ri L J
> FLUu GAS N'°2 _ NP ^s°2
YT »
R
^
it
A ^ i
' - • • H X-X
| PACstED FEED
••••^ Htf J ~-*"i"Cwwi3»ir» irsi ^- \
• .
CATCH
238
-------
The main objective of these studies was to determine the rate of
mass transfer for equimolar concentrations of NO and NO. in the range of
250 ppm to 750 ppm each. The parameters affecting the rate of mass transfer
were determined by a study of the following: gas and liquid velocitiies,
types (i.e., CaCOHK, Mg(OH)2, and NaOH} and concentrations of scrubbing
solutions, temperature, and mole ratios of HO and N02, size of packing,
and height of packing.
TESTING
These studies were performed using flue gas from the combustion
of natural gas (see Figure 1.). Air, natural gas, and NO were blended
and burned, (N0» was added prior to .entry to the preheater to minimize
thermal decomposition of NO. to NO), to give a standard flue gas having
the following approximate composition: C02, 6.4%; 0,, 3.0%; HjO, 12.7%;
N2, 74.9%; NO, 0.025 to 0.075%; and NO,, 07025 to 0.075%. The two gas
flow rates used were 20 SCFH and 30 SCFH. The two liquid flow rates were
170 ml/min and 350 ml/min. These combinations of gas and liquid flow rates
correspond to a range of L-to-G ratios from 9.3 to 29.2. A flue gas
preheater was used to raise the temperature of the gas entering the scrubber
to 350°F. After passing the gas countercurrently to the liquid flow through
the scrubber, the gas was pulled through two condensers to remove water
since water interferes with the monitoring of NO and N02. Several monitors
were used to follow the concentration changes of the gaseous components of
primary interest. A non-dispersive infrared spectrophotometer, that
provided continuous SO. measurement, was installed in case of SO. fluctuations
in the natural gas supply. During the course of these measurements, no
appreciable content of SO. was observed in the flue gas. Another infrared
instrument provided continuous NO measurement, and an ultraviolet
spectrophotometer provided continuous NO. measurement. The specific
instruments used in these tests were BecKman Model 315A Infrared Analyzers
for SO. and NO and a DuPont Model 400 Ultraviolet Analyzer for the NO.
measurement. By proper selection of flow paths using a three-way valve
(see Figure 1), the gas could be monitored by the instruments prior to
entry to the scrubber or after passing through the scrubber and condensers.
One feed tank and one catch tank were used in these experiments. Scrubbing
liquor, i.e., (_Ca(OH)2, Mg(OH)2, or NaOH solutions, were fed on a once-
through basis from the feed tank through the scrubber and into the catch
tank. The temperature of the solutions in the feed tank was maintained by
means of a 1000-watt electric heater which was thermostatically controlled
to ±5°F. The range of temperatures over which this heater operated was
from 100°F to 200°F.
A run was performed in the following manner. The feed tank was charged
with fifteen gallons of distilled water and heated to 100°F, 125°F, or 150°F,
while stirring. After the water reached the desired temperature, the solid
alkaline material was added to the feed tank. The amount of Ca(OH>2 or Mg(OH)2
239
-------
added was equal to that amount necessary to make fifteen gallons of solution
saturated with the alkaline material at the particular operating temperature.
The concentrations of NaOH solutions were from 10""* N solutions to 5%
solutions. After the pH of the solution reached a maximum. value, i.e., about
fifteen minutes, the alkaline solution (of cither Ca(OH)», Mg(OH)2, or NaOH)
was then pumped into the top of the scrubber at either 1/0 ml/min or 350 ml/min/
As soon as the temperature in the packed bed of the scrubber reached within
± 2 F of the liquid temperature of the feed tank, the three-way gas feed valve
was switched and the NO^ and NO^ containing flue gas was fed to the bottom of
the scrubber. The gas leaving the top of the scrubber was monitored for NO,
NO 2 and S02 as mentioned previously. The time for the S0~ and NO to go
from the furnace to the monitoring instruments was twelve seconds at 20 SCFH,
and six seconds at 30 SCFH. The time for the NO. to go from the preheater
to the instruments was approximately ten seconds at 20 SCFH, and five
seconds at 30 SCFH.
ANALYTICAL PROBLEMS OF SCRUBBING STUDY
A. Gas Analysis
The results of engineering studies can be no better than the analytical
tools used to obtain these results. In the study of NO scrubbing with a
packed bed reactor, a proper interpretation of NO and NO- concentrations
Cas obtained from the non-dispersive infra red analyzer and the ultraviolet
spectrophotometer, respectively) is dependent upon the specificity of the
instrument for the species being studied. NO chemistry is very complex
and a large number of equilibria occur simultaneously, i.e.,
NO + N02^N20
2NO + 02^T
3N02 + HjO .? 2NOJ + 2H+ + NO
N2°3 + H2° ^ 2HN02
3HN023T & + N0~ + 2NO + ty
and others.
During the course of packed bed scrubbing studies, scrubbing efficiency
is determined by comparing the data for the scrubbed gases with those obtained
from by-passing the scrubber and passing the NO containing flue gas into
the analyzers directly. It was found that the ?otal NO determined by the
analyzers was greater than the sum of the entering NO and NO. concentrations.
It was subsequently deduced that the NDIR was sensing N^O, as well as the
NO which it was designed to determine. This interference was eliminated by raising
the temperature of the gas going to the NDIR, thus, decomposing any NO
240
-------
to N0~. It was believed advisable also to raise the temperature of the
gas sample seen by the UV NO. analyzer. This was done to avoid any possibility
of interference by N20> or N-05 with the determination of NO-.
It was decided that, because of possible interference problems with the
process monitors, a basic standard method was needed to establish the accuracy
of the monitors when used under various experimental conditions. The first
approach to a solution of this problem involved gas chromatography. A
method was developed in which the gas sample was separated on a Porapak
Q column (i ft x 1/8 in.) at 30°C. With this procedure it is possible to
determine NO in NO-^ mixtures if the NO concentration exceeds 8000 ppm.
This lack of sensitivity causes the gas chromatographic approach to be
inappropriate for levels of NO normally found in flue gas.
The second approach at a standard method to check out the accuracy
of the NO monitors involved the use of chemiluminesence. An instrument
built by the Division of Chemistry and Physics was borrowed and comparisons
made to the process monitors. After the modifications of the process
monitors previously mentioned were completed, the concentrations determined
by the monitors were identical to results determined by chemilumenesence.
This calibration yielded satisfactory results for the process monitors
in that material balances in the gas phase could then be achieved.
B. Liquid Analysis
8
A liquid analysis for N0~ and NO., described by Walters and Uglum
was performed on the scrubber solution when water or NaOH solutions were
used. This liquid analysis did not yield acceptable results with Ca(OH).
and MgCOH^ solutions. The chief problems which were encountered were
caused by precipitation from the saturated solutions being used for scrubbing
in the cuvette causing erroneous results. No liquid analyses were reported
in this paper due to the variation in instrument readings caused by this
precipitation.
DISCUSSION AND RESULTS
Figures 2 through 18 are plots of all the data. The ordinate of these
figures is the rate of absorption of NO and N0_ per hour per scrubber cross-
sectional area. The abscissa is the mole percent of NO and NO- in the flue
gas stream.
Many of the figures indicate a slurry flow or liquid rate of 170 and
350 ml/min. Two figures could have been drawn showing that the data points
would have been identical to each other at the two slurry or liquid flow
rates. It was found during this work that slurry flow rate had no effect
on the adsorption rate of NO . Consequently, many of the figures indicate
flow rates of both 170 and 350 ml/min where the other parameters and
resultant adsorptions are replicated. Instead of making two different
241
-------
figures, it was decided to make one figure.
Figures 2 through 8 show the rate of absorption of NO and NO. in flue
gas with use of saturated solutions of -magnesium hydroxide under various
conditions. The parameter changed in Figures 2 and 3 was the gas flow rate
i.e., from 20 SCFH to 30 SCFH. As can be seen from these figures, there is
no apparent change in the rate of absorption of NO and NO. with gas flow
rate. Also, it should be pointed out that changing the slurry or liquid
flow rate from 170 ml/min to 350 ml/min produced no change in absorption.
In Figures A and 5 the parameters changed were liquid or slurry temperature
and gas flow rate. The data on Figures A and 5 indicate no apparent change
in absorption rate due to variation in slurry or liquid temperature and
gas flow rate. Figures 6 through 8 shown one gross change relative to Figures
1 through 5, and that is a change in the size of packing. Four millimeter
Berl saddles were used instead of the 3/8-inch Berl saddles. This change
was made to see if a change in interfacial area would affect the rate of
absorption.
Another change made within the data in Figures 6 through 8 was the
slurry or liquid temperature. No statistical difference could be found
due to a temperature effect. Comparing Figure 6 through 8 to Figures 2
through 5 where the gross change was the packing size, no statistical
difference was found between the slopes of the best fit lines. Intuitively,
one might conclude that increasing the interfacial area would yield an
increase in the rate of absorption but there is no statistical evidence for
this.
Figures 9 through 12 shown the rate of absorption of NO and NO. from
flue gas using saturated solution of calcium hydroxide under various conditions.
The main parameter changed in Figures 9 and 10 was the gas flow rate,
i.e., from 20 SCFH to 30 SCFH. No change in the absorption of NO and NO.
was noted. Similar to the results with the magnesium hydroxide solutions,
no change in absorption was noted due to the change in slurry or liquid flow
rates, i.e., from 170 ml/min to 350 ml/min.
Again one main parameter was changed between Figures 11 and 12 and that
was the gas flow rate i.e., 20 SCFH to 30 SCFH. No significant change
in absorption was noted.
Comparing Figures 9 and 10 to Figures 11 and 12, where the main parameter
changed was temperature (i.e., 125°F to 150°F}» no statistical change was
noted in the rate of absorption.
Figure 13 shows the results of the use of calcium hydroxide as a
scrubbing agent while the column was packed with four-millimeter Berl saddles.
It was not long before the column plugged and started to flood, yielding
limited data. The plugging was caused by the reaction of calcium
hydroxide with the carbon dioxide in the flue gas.
242
-------
FIGURE 2
RELATIONSHIP BETWEEN RATE OF AQUEOUS ALKALINE
SCRUBBING OF NO AND M>2 AND INITIAL CONCENTRATIONS
OF EQUIMOLAR AMOUNTS IN FLUE GAS
o
E
—
1
z
o
t
or
o
CO
GO
**
u.
0
UJ
<[
or
-3
10X10
9
8
7
6
5
4
evj
.1 3
\
2
-4
10X10
9
8
7
e
5
4
3
-5
1 f\ VO R
aCRUBBINO LIQUOR - MgfOHL
BERL SADDLE 3.IZE - 3/3"
LIQUID TEMPERATURE - •
» I faO r
UIOUID FLCW RATE -
1708 350 ml/mln
GAS FLOW RATE - 2Q SCFH O
D
D
_
0
o
-
o
D
0
O a
a
O
-
-
a
_
-
O NO
••
a NO,
2
_
i i i . . » i i I
IX10"
5 6 7 8 9 10X10
-2
INITIAL CONCENTRATION of NOand of N02 , MOLE PERCENT
243
-------
FIGURE 3
RELATIONSHIP BETWEEN RATE OF AQUEOUS ALKALINE
SCRUBBING OF NO AND N0g AND INITIAL CONCENTRATIONS
OF EQUIMOLAR AMOUNTS IN FLUE GAS
-3
10X10
9
8
7
6
SCRUBBING LIQUOR
BERL SADDLE S.IZE
LIQUID TEMPERATURE
LIQUID FLOW RATE
6AS FLOW RATE
I
z
O
E
or
O
CO
•**
u.
O
UJ
H
or
10X10
9
8
7
6
-5
•
~ Mg(OH)2
- 3/8"
I25*F
"" I7O ft 350ml/min
3O SCFH
O
O
ixlO
-2
O
a
O NO
a NO«
i i
5 6 7 B 9 10X10
.2
INITIAL CONCENTRATION of NOandof NO2 , MOLE PERCENT
244
-------
FIGURE 4
RELATIONSHIP BETWEEN RATE OF AQUEOUS ALKALINE
SCRUBBING OF NO AND N02 AND INITIAL CONCENTRATIONS
OF EQUIMOLAR AMOUNTS IN FLUE GAS
-3
10X10
9
8
7
6
5
2
O
E
tc
O
O
li.
O
Ul
-4
10X10
9
8
7
6
-5
10x2.5
SCRUBBIN8 LIQUOR
BERL SADDLE 3.1 ZE
LIQUID TEMPERATURE
LIQUID FLOW RATE
6AS FLOW RATE
Mg(OH)2
n
3/8
I50*F
I7OS 350m|/mln
20 SCFH
a
O
a
O
a
0
a
O
O
D
O NO
NO,
.2
I xlO
INITI
5 e 7 8 9 10X10
AL C ONCE NT RATION of NOand of NO2 , MOLE PERCENT
245
-------
FIGURE 5
RELATIONSHIP BETWEEN RATE OF AQUEOUS ALKALINE
SCRUBBING OF NO AND N02 AND INITIAL CONCENTRATIONS
OF EQUIMOLAR AMOUNTS IN FLUE GAS
-3
10X10
9
8
7
6
5
I
z
o
fc
or
o
CD
QC
-4
10X10
9
8
7
6
SCRUBBING LIQUOR
BERL SADDLE S.IZE
LIQUID TEMPERATURE
LIQUID FLOW RATE
GAS FLOW RATE
MgtOH)2
3/8*
I50*F
I7O8 350 ml/min
30 SCFH
a
O
O NO
a NO,
5 6 7 Q 9 10X10
-2
INITIAL CONCENTRATION of NOandof NOg, MOLE PERCENT 246
-------
FIGURE 6
RELATIONSHIP BETWEEN RATE OF AQUEOUS ALKALINE
SCRUBBING OF NO AND N0g AND INITIAL CONCENTRATIONS
OF EQUIMOLAR AMOUNTS IN FLUE GAS
V
i
I
5
1
2
O
E
ft
o
tft
o
*
u.
0
iii
h
e
-3
10X10
9
8
7
6
c
4
N
£
L:
2
_4
10X10
9
8
7
6
5
4
3
-5
in*9 R
SCRUBBIN0 LIQUOR - Mg(OH>2
SADDLE 3JZE - 4 mnii _
LIQUID TEMPERATURE - Iftn9e °
•. IW r
. L10U1° FLOW RATE - .7oml/mln e
6AS FLOW RATE - 20SCFH
a
-
8
9
o
O NO
a wo2
• ,.,.••
ixlO
3.
5 6 7 8 9 IQXIO
INITIAL CONCENT RATION of NOandof N02 , MOLE PERCENT
247
-------
FIGURE 7
RELATIONSHIP BETWEEN RATE OF AQUEOUS ALKALINE
SCRUBBING OF NO AND N0g AND INITIAL CONCENTRATIONS
OF EOUIMOLAR AMOUNTS IN FLUE GAS
-3
10X10 T SCRUBBING LIQUOR -
BERL SADDLE 3.IZE -
8p
LIQUID TEMPERATURE -
7
6
-
LIQUID FLOW RATE -
GAS FLOW RATE
Mg(OH)2
4mm.
I25*F
I70ml/min
20SCFH
or
0
m
**
u.
0
UJ
^
or
-4
IOXIO
9
8
7
6
-5
10X2.5
o
O
D
O
o
D
o
D
O NO
D NO,
56789 10X10
IXIO
-2
248
INITIAL CONCENTRATION of NOandof NO2 , MOLE PERCENT
-------
FIGURE 8
RELATIONSHIP BETWEEN RATE OF AQUEOUS ALKALINE
SCRUBBING OF NO AND NOg AND INITIAL CONCENTRATIONS
OF EQUIMOLAR AMOUNTS IN FLUE GAS
-3
10X10
9
8
7
6
SCRUBBING LIQUOR
BERL SADDLE 3,1 ZE
LIQUID TEMPERATURE
LIQUID FLOW RATE
6AS FLOW RATE
Mg(OH)2
4mm.
150° F
170 ml/min
2OSCFH
o
O
z
0
E
a
o
in
CD
<
u.
O
Ul
fc
a:
-4
10X10
9
8
7
6
-5
10X2.5
O NO
a NO,
IxlO
-2
6789 10X10
.2
INITIAL CONCENTRATION of NOandof NO2, MOLE PERCENT
249
-------
FIGURE 9
RELATIONSHIP BETWEEN RATE OF AQUEOUS ALKALINE
SCRUBBING OF NO AND NOg AND INITIAL CONCENTRATIONS
OF EQUIMOLAR AMOUNTS IN FLUE GAS
-3
10X10
9
8
7
6
I
z
o
t
or
O
CO
o
UJ
10X10
9
8
7
6
5
4
3
-5
10X2.5
SCRUiiINC LIQUOR
BERL SADDLE S.IZE
LIQUID TEMPERATURE
LIQUID FLOW RATE
6 AS FLOW RATE
- CO (OH)
M ^
- 3/8
~ I25*F
"~ 1708-350
~ 20 SCFH
Oo
q,
o o
D
O NO
a NO,
ixlO
"2
5 6 7 8 9 IOXIO
.2
INITIAL CONCENTRATION of NOandof NO , MOLE PERCENT
25C
-------
FIGURE 10
RELATIONSHIP BETWEEN RATE OF AQUEOUS ALKALINE
SCRUBBING OF NO AND N0g AND INITIAL CONCENTRATIONS
OF EQUIMOLAR AMOUNTS IN FLUE GAS
1
V
i,
J)
75
E
^
1
z
o
E
te.
O
-------
FIGURE II
RELATIONSHIP BETWEEN RATE OF AQUEOUS ALKALINE
SCRUBBING OF NO AND N02 AND INITIAL CONCENTRATIONS
OF EQUIMOLAR AMOUNTS IN FLUE GAS
V
Si
0
E
.0*
•M>
1
O
t
or
o
00
**
u.
o
tu
^
or
-3
10X10
9
8
7
6
5
4
CM
£
L:
^
2
_4
10X10
9
8
7
e
4
3
-5
10X2.5
1
SCRUBBINC LIQUOR - CflfoH)
BERL SADDLE SJZE - 3/9"
LIQUID TEMPERATURE - ,•„
- ISO F
LIQUID FLOW RATE -
170 ft 350 ml/mln
GAS FLOW RATE - 2Q SCFH
-
-
a
D
D
O
D 0
O
D
O
D
O
-
0
O NO
•M
D NO,
i i i i . » i i I
xlO"2 2 3 4567891
.2
INITIAL CONCENTRATION of NOandof NO2,MOLE PERCENT
' 252
-------
FIGURE 12
RELATIONSHIP BETWEEN RATE OF AQUEOUS ALKALINE
SCRUBBING OF NO AND N02 AND INITIAL CONCENTRATIONS
OF EQUIMOLAR AMOUNTS IN FLUE GAS
-3
10X10
9
8
7
6
SCRUBBIN8 LIQUOR - Cfl(OH)
BERL SADDLE S.IZE •
LIQUID TEMPERATURE
LIQUID FLOW RATE '
GAS FLOW RATE
z
o
E
a:
o
00
^
u.
o
Ul
^
tr
^
1 Ox 10
9
8
7
e
•
3/8
•
150 F
170 & 350 ml/min
30 SCFH
-5
10x2.5
o
O
D
D O
D
O
O NO
a NO,
IxlO
5 678 9 10*10
.2
INITIAL CONCENTRATION of NOandof NO2 , MOLE PERCENT
253
-------
FIGURE 13
RELATIONSHIP BETWEEN RATE OF AQUEOUS ALKALINE
SCRUBBING OF NO AND NOg AND INITIAL CONCENTRATIONS
OF EQUIMOLAR AMOUNTS IN FLUE GAS
-3
10X10
9
8
7
6
I
z
O
fc
or
o
oo
u.
O
UJ
.4
10X10
9
8
7
-5
10X2.5
8CRUSBIN8 LIQUOR - Ca(OH)
BERL SADDLE SJZE
LIQUID TEMPERATURE
LIQUID FLOW RATE '
GAS FLOW RATE
4mm.
125* F
170 ml/min
20 SCFH
IxlO
-2
FLOO DIN 6
REGION
O
a
O NO
a NO,
56789 10X10
INITIAL CONCENT RATION of NO and of NO2,MOLE PERCENT
254
-------
The data of Figures 14, 15, and 17 are for runs made with sodium
hydroxide. Two concentrations of sodium hydroxide used were 10"^ molar
and a 52 solution. Comparing these figures, no statistical difference
in the rate of absorption of NO and N02 relative to changes of temperature,
solution or slurry concentration, and packing size were observed.
Figure 16 shows the rate of absorption of NO and NO- when a 10~ molar
solution of sodium hydroxide is spiked with sodium chloride to 2/3 saturation
at the temperature of 1.25°F. The reason for this run was that the literature
quoted the possibility of HNO- mist formation during absorption of NO and NO .
The 2/3 saturated NaCL solution's H20 vapor pressure is well below that of
the inlet gas stream, thus providing a nucleating site for HNO- formation.
This approach resulted in a small decrease in the rate of absorption of NO
and N02. The rate of absorption was found to be statistically different
(lower; from the rates of absorption using magnesium and calcium hydroxide.
Figure 18 shows the rate of absorption of NO and NO- when distilled
vater is used as the scrubbing agent. One can see that the absorption
data are very similar to the other data with the hydroxide solutions.
A statistical analysis of the data Ci-e. data of Figures 2 through 18)
indicates the following:
(a) The concentration of NO and NO- in the flue gas stream signifi-
cantly affects the rate of absorption; NO and N02 at 250 ppm
in flue gas is absorbed at a lower rate that 750 ppm NO and N02.
(hi The type of hydroxide significantly affects the rate of NO and
NO, absorption, i.e., NaOH is more effective than Mg(OH>2, Mg(OH)2
is better than CaCOHK, Ca(OH)2 is better than water. On this
particular point we should like to mention that we believe this
notable difference was probably associated with the presence or
absence of precipitate formation mentioned previously. Precipitate
formation during a run changes the flow patterns and causes channeling
which produces changes in the rates of absorption.
Ccl Over the ranges employed, changes in slurry or liquid temperature,
changes in slurry or liquid flow rates, and changes in packing produced
no statistical differences in the absorption of NO and N02«
The percentage removals of NO and N0« expressed as total NO^ removed
(shown in Figures 2 through 18) are as follows for the feed concentration
studied:
255
-------
FIGURE K
RELATIONSHIP BETWEEN RATE OF AQUEOUS ALKALINE
SCRUBBING OF NO AND NOg AND INITIAL CONCENTRATIONS
OF EQUIMOLAR AMOUNTS IN FLUE GAS
-3
10X10
9
8
7
6
I
O
fc
O
tn
oo
u.
o
UJ
10X10
9
8
7
-5
10X2.5
SCRUBBIN0 LIQUOR
BERL SADDLE 3.IZE
LIQUID TEMPERATURE
LIQUID FLOW RATE
6AS FLOW RATE
-4
10 M NaOH
3/8-
IOO°F
170 ml/mln
20 SCFH
O O
D
9
O
D
O NO
NO,
IXIO
56789 10X10
,2
256
INITIAL CONCENTRATION of NOandof N02 , MOLE PERCENT
-------
r iouK£ I b
RELATIONSHIP BETWEEN RATE OF AQUEOUS ALKALINE
SCRUBBING OF NO AND NOg AND INITIAL CONCENTRATIONS
OF EQUIMOLAR AMOUNTS IN FLUE GAS
-3
10X10
9
8
7
6
SCRUBBING LIQUOR
BERL SADDLE SJZE
LIQUID TEMPERATURE
LIQUID FLOW RATE
6AS FLOW RATE
5 -
.
z
O
E
tt
0
(A
CD
*
U.
0
III
^
Q:
^
10X10
9
8
7
e
-5
10X2.5
-4
10 M NaOH
it
3/8
125 F
I70ml/min
20 SCFH
a
O
a
O
a
O
D
O
a
O
e
O NO
D NO,
I xlO
56789 10X10
.2
INITIAL CONCENTRATION of NOandof NO2, MOLE PERCENT
257
-------
FIGURE 16
RELATIONSHIP BETWEEN RATE OF AQUEOUS ALKALINE
SCRUBBING OF NO AND NOg AND INITIAL CONCENTRATIONS
OF EQUIMOLAR AMOUNTS IN FLUE GAS
\
^
JW
o
E
.d
****
1
z
o
fc
or
o
m
^
•a
II.
0
UJ
K
or
-3
10X10
9
8
7
6
5
4
(VI
J= 3
\
2
^
10X10
9
8
e
5
4
3
-5
-4
aCRUSSINfl LIQUOR- |o M NoO H spiked to2/3 saturoted
BERL SADDLE S.IZE - 3/8"
LIQUID TEMPERATURE - ,ocrec
125 r
_ LIQOI° FLOW RATE - .70 ml/mln
GAS FLOW RATE ~ SCFH
a
O
^
O
_. a
o
-
O
a
a
O
.
O
a
•
0
O NO
-
a wo2
"
_
10X2.5 . _o „ , >i R e 7 B 9 1C
» -2
I X4O
INITIAL CONCENTRATION of NO and of NOg , MOLE PERCENT
258
-------
FIGURE 17
RELATIONSHIP BETWEEN RATE OF AQUEOUS ALKALINE
SCRUBBING OF NO AND NOg AND INITIAL CONCENTRATIONS
OF EQUIMOLAR AMOUNTS IN FLUE GAS
-3
I0x|0
9
8
7
6
SCRUBBINO LIQUOR - 5% NoOH
BERL SADDLE S.IZE
LIQUID TEMPERATURE
LIQUID FLOW RATE
GAS FLOW RATE
4mm
100 F
l7Oml/mln
20 SCFH
0
E
a:
o
OQ
^
U.
O
Ui
Jj
(C
^
10X10
9
8
7
e.
-5
10X2.5
I xlO
n
O
O NO
a NO,
56789 10X10
-2
INITIAL CONCENTRATION of NOandof NO2 , MOLE PERCENT
259
-------
FIGURE 18
RELATIONSHIP BETWEEN RATE OF AQUEOUS ALKALINE
SCRUBBING OF NO AND NOg AND INITIAL CONCENTRATIONS
OF EQUIMOLAR AMOUNTS IN FLUE GAS
0
e
£
|
I
z
0
t
or
o
OQ
**
U.
O
UJ
F-
or
-3
10X10
9
7
6
5
4
M
H-
\
2
•A
10X10
9
8
7
e
5
4
3
-5
I/N vO R 1
SCRUBBINO LIQUOR - DISTILLED H.,0
- u ^
BERL SADDLE 3.IZE - 3/8
" o
LIQUID TEMPERATURE - |25 F
LIQUID FLOW RATE - ,70 a 350 ml/min
GAS FLOW RATE ~ '£> SCFH
rf
9
O
O
O
a
-
a
0
0
-
a
-
»
-
O NO
~
a NOO
2
i i i t . t i t I
I xlO"
5 6 7 8 9 10X10
-2
INITIAL CONCENTRATION of NOandof N02, MOLE PERCENT
260
-------
NO and NO, Concentration in Flue Gas % Removal
250 ppm 12 to 14
500 ppm 17 to 19
750 ppm 27 to 29
CONCLUSIONS
The main conclusion drawn at present from our test is that equimolar
concentrations of NO and NO- in flue gas between 250 ppm and 750 ppm are
scrubbed by aqueous solutions of CaCOH),,, Mg(OH)_, and of NaOH with poor
overall scrubbing efficiency. Since no change in absorption was observed
due to changes in liquid flow rate, and the large effect of NO partial
pressure on scrubbing efficiency was found, this may be indicative of gas
phase reaction limitations.
A study of the kinetics of the reactions taking place in the gas
phase might yield the rate limiting mechanism. Should such a mechanism
or mechanisms be found, then steps to catalyze these reactions possibly
might improve the alkaline scrubbing efficiency sufficiently to make alkaline
scrubbing an acceptable abatement process for low concentrations of equimolar
quantities of NO and NO,.
261
-------
BIBLIOGRAPHY
1. Bartok, W., et al, Final Report, "Systems Study of Nitrogen
Oxide Control Methods For Stationary Sources" - Volume II,
Contract NO. PH-22-68-55, November, 1969.
2. Sherwood, T. K. and Pigford, R. L., "Absorption and Extraction",
Chem. Eng. Series, McGraw Hill, 1952.
3. Ganz, S. N. and Mamon, L. I., Zhur Priklad Khim, 30, p. 369, 1957.
4. Ganz, S. N. and Kuznetzov, I. E., Tr Dnepropetr Khim - Takhol.,
Inst, 16, p. 17, 1963.
5. Ganz, S. N. and Lokshin, M. A., Zhur Priklad Khim, 30, p. 1525,
1957.
6. Ganz S. N. and Dravchinskaya, S. B., Zhur Priklad Khim, 28,
p. 145, 1955.
7. Andrew, S. P. S., and Hanson, D,, Chem. Engr. Sci., 14, pp. 105-113,
1961.
8. Vfettera and Uglum, J. Analytical Chemistry. 42, (3), pp. 335-340,
March 1970.
262
-------
REMOVAL OF SULFUR DIOXIDE FROM STACK GASES
BY SCRUBBING WITH LIMESTONE SLURRY:
USE OF ORGANIC ACIDS
By
J. D. Hatfield and J. M. Potts
Division of Chemical Development
Tennessee Valley Authority
Muscle Shoals, Alabama
Prepared for Presentation at
Second International Lime/Limestone Wet Scrubbing Symposium
Sponsored by the Environmental Protection Agency
New Orleans, Louisiana
November 8-12, 1971
263
-------
REMOVAL OF SULFUR DIOXIDE FROM STACK GASES
BY SCRUBBING WITH LIMESTONE SLURRY;
USE OF ORGANIC ACIDS
By
J. D. Hatfield and J. M. Potts
Division of Chemical Development
Tennessee Valley Authority
Muscle Shoals, Alabama
ABSTRACT
A survey of weak organic acids showed that many of them are
capable of dissolving CaC03 and MgC03 and therefore are of potential
value in improving the scrubbing of S02 by limestone slurries.
The physicochemical properties of four selected acids—benzoic,
phthalic, glycolic, and adipic—and their calcium and magnesium salts were
determined with respect to stability, solubility, effect on the oxidation
rate of sulfite to sulfate, and the nature of aqueous species.
Small-scale tests were made to demonstrate the effectiveness of
some of the organic acids in improving absorption of S02. The proportion
of benzoic acid required to effect an improvement was determined in closed-
loop spray tower operation under conditions designed to remove about 50$
of the S02 without additive. The addition of 10$ of the-stoichiometric
amount of benzoic acid to solubilize the feed limestone increased the S02
removal by about 10 percentage points; 50$-°f tne stoichiometric acid gave
a 22-percentage point increase, but further improvement with added acid
was 'small. The use of benzoic acid resulted in a higher degree of oxidation
of the product solids in the scrubbing operation, although laboratory tests
with oxygen indicated that oxidation of dissolved sulfite is inhibited by
the acid. The high oxidation in the scrubber tests presumably resulted
from the lower pH brought about by the benzoic acid addition and the con"
sequent larger amount of dissolved sulfite available for oxidation.
. Organic acids not only improve S02 removal but also offer the
opportunity of getting complete utilization of limestone by using enough
acid to dissolve all the feed limestone, in which case the only solid
phases are calcium sulfite and sulfate. The increased oxidation in the
scrubber should also improve slurry settling rate and avoid the danger of
limestone surface blinding. Work on these possibilities is continuing.
264
-------
REMOVAL OF SULFUR DIOXIDE FROM STACK GASES
BY SCRUBBING WITH LIMESTONE SLURRY:
USE OF ORGANIC ACIDS
By
J. D. Hatfield and J. M. Potts
Division of Chemical Development
Tennessee Valley Authority
Muscle Shoals, Alabama
The dissolution of limestone is one of the limiting reactions in
the removal of sulfur dioxide from stack gases by scrubbing with a lime-
stone slurry. The equilibrium solubility of limestone in water (Table l)
does not give a great driving force to effect solution; consequently, it
has been found advantageous to have a surge vessel to permit dissolution
of limestone and precipitation of calcium sulfite - sulfate before sending
the slurry to the scrubber. The low concentration of basic components in
the solution also necessitates a higher ratio of liquid to gas (L/G) in
the scrubber than is required for other scrubbing agents such as alkali
metal salts or ammonia.
TABLE I
Solubility of CaC03 in Water at 50°C (l22°F)
Ca
ppm
9-5
16.8
30.7
57.7
111.1
221.7
1020.3
PC02
atm
0.000136
0.000803
o.oo466
0.0271
0.159
0.950
36.3
PJL
8.5
8.0
T..5
7.0
6.5
6.0
5.0
a Calculated using Radian1s equilibrium pro-
gram (PB 193029) for the system CaO-C02-H20.
The solubility of limestone increases rapidly with decrease in
pH as shown in Table I. A similar increase can be achieved by adding another
acid, provided that:
(l) the calcium salt of the added acid is soluble, and
(2) the ionization constant of the added acid is greater
than that of carbonic acid, H2C03 (K! = lj-,4 x 10~7)
265
-------
By stipulating also that the added acid be weaker than sulfurous acid, H2S03
(KX = 1.3 x 10"2), the limestone slurry scrubbing of stack gases might well
be improved significantly by the addition of a small amount of an extraneous
acid (an acid stronger than H2S03 would volatilize S02 from solution and
prevent its precipitation as the calcium salt).
This paper describes a survey of possible acid additives, labora-
tory tests of properties of selected acids and their salts, and small-scale
tests of selected additives to determine their effectiveness in S02 removal
and other operational characteristics.
Survey of Acids and Exploratory Tests
Of the 132 acids- (intermediate in strength to H2S03 and H2C03)
that are listed in chemical handbooks, many are organic acids whose physi-
cochemical properties in aqueous systems have not been studied. Many can
be eliminated from consideration as additives to limestone slurries because
of cost or difficulty to produce; others are of too high molecular weight
to be soluble in water; and others, such as oxalic acid, have extremely
insoluble calcium salts. The ideal additive would be
(l) soluble in water and limestone scrubber slurry;
(2) nonvolatile and stable under scrubbing conditions;
(3) not precipitated by Ca, Mg, or other metals likely
to be found in the solution; and
commercially available and of reasonable cost.
Solubility of Carbonates at 25°C: Measurements were made of the
ability of several organic acids to dissolve calcium carbonate and magnesium
carbonate at room temperature (~78°F). The acids were those available in
our laboratory at the time and consisted of both aliphatic and aromatic
acids (monobasic and polybasic) and covered a wide range in molecular weight.
The procedure consisted in adding 1 gram of the organic acid (or 1 ml if the
acid was a liquid) to 100 ml of water and then adding 5 grams of reagent
CaC03 or MgC03 to the solution or mixture. The mixtures were allowed to
stand for 1 week with occasional agitation by hand to promote dissolution.
Aliquots of 10 ml of the clear solution were taken after 3 days and after
1 week from each mixture and were analyzed for the dissolved metal. The
pH was measured before addition of the carbonate and after the 7-day sample.
The results are presented in Table II.
266
-------
TABLE II
Use of Organic Acids in the Dissolution of Calcium and Magnesium Carbonates
pH Of
Acid
acid
Formula solution
Sol., Ca, vt. %,
After 7 days Sol., MR, wt. 1>,
after mole ratio
3 days
7 daye
acld:Ca
After 7 days
after mole rat^o
pH 3 days
Aliphatic monocarboxyllc acids
Formic
Acetic
Chloroacetlc
(Rycolic
ftenylacetic
Proplonlc
Lactic
Butyric
Caproic
Qluconic
Acrylic
Oleic
HCOOR
C%COOH
CICHaCOOH
HOCHaCOOH
C.aHsCHaCOOH
CaH5COCH
CHgCHOHCOOH
CaHTCOOH
CsHnCOOH
CHeOHCCHOH)*
COOB
CHs=CHCOOH
CH3(CHa)7C=C
(CHZ)TCOOH
2.30
2.30
2.15
2.1*0
2.70
2.90
2.50
2.90
3.00
2.86*
2.6O
5.80*
O.U8
0.32
0.21
0.22
0.13
0.28
0.21
0.22
0.17
0.06
0.31
0.01
0.5U
0.37
0.21
0.25
0.15
0.28
0.23
0.21
0.17
0.06
0.31
0.01
1.7
1.9
2.0
2.1
2.0
1-9
1.9
2.2
2.0
3.U
1.8
16.7
7.00
7.20
7.15
7.>»0
7.^0
7.20
6.90
6.65
6.15
7.00
6.60
6.60
0.1*0
0.31
0.20
0.2U
0.11*
0.27
0.21
0.23
0.15
0.06
0.25
0.03
7 days
O.Ul
0.31
0.20
0.27
0.13
0.21
0.21
0.21
0.12
0.08
0.21
0.03
acld:Mg"
l.l*
1.1*
1.3
1.2
1.1*
1.6
1.3
1.3
1.7
1.6
1.6
3.fc
Price
pH dollars/lb.
7.60
7.65
7.85
7.50
8.15
7.75
7.70
7.50
7.90
8.00
7.70
8.1*5
0.11*7
0.09
0.21
0.10
0.68
0.1U7
0.275
0.33
0.1U5
0.31
0.23
Aliphatic polycarboxyllc acids
Oxalic
Succinlc
lartaric
Malic
Pumarlc
Malelc
Adipic
Citric
COOHCOOH
COOHCHaCHaCCOH
COOHCHOHCHOH
COOT
COOHCHOHCHsCOOH
COOHCH=CHCOOH
HOaCCH=CHCOaH
COQH(CH2)4COOH
(COOHCH2)eOOH
CCOH
1.75
2.30
2.25
2.1*0*
2.35a
1.95
2.80
2.25
0.01
0.27
0.01
0.09
0.25
0.23
0.28
0.02
0.01
0.32
0.01
0.09
0.28
0.31
0.28
0.02
1*1*. I*
1.1
26.7
3.3
1.2
1.1
1.0
10.
7.70
7.35
7.70
7.60
7.30
7.25
7.20
7.80
0.13
0.29
0.21*
0.22
0.28
0.31
0.26
0.2U
0.08
0.30
0.26
0.2U
0.27
0.30
0.21
0.21*
3-1*
0.69
0.62
0.76
0.78
0.70
0.79
0.53
8.35
7.80
7.85
8.1*0
8.00
7.70
7.80
8.50
0.22
0.62
O.U15
0.315
0.225
0.1*8
0.18
0.33
Aromatic acids
Benzole
Salicylic
P-Amino
benzole
Dinitro
benzole
GaUic
ftthalic
crKaphthoic
CeHgCOOH
HOCaRtCOOH
NHaCeBeCOOH
(HOaJaCeHg
COOH
(HO)3CeH7
COOH
CeRe(COOH)2
CioH7COOH
2.75*
2.60*
5.50s
2.80a
2.85
2.l*5B
3.75a
O.Ik
0.11
0.15
0.09
0.11
0.12
0.02
0.17
O.ll*
0.15
0.10
0.11
0.06
0.01
1.9
2.1
1.9
1.9
2.1
1*.0
23
7.35
7.35
7.M)
7.1*5
6.20
7.60
6.70
O.OU
0.21
o.it*
0.09
0:12
0.22
0.08
0.13
0.13
O.lU
0.09
0.11
0.21
0.06
1.5
l.U
1.3
1.3
1.3
0.70
2.3
8.20
8.15
8,00
8.10
8.15
8.05
8.30
0.215
O.H25
1.72
2.65
0.12
1 gram of acid did not completely dissolve in 100 ml.
b
Based on solubilities after 1 week, assuming all acid dissolved.
267
-------
The pH of the acid solutions varied from 1.75 to 3-80 (many
of the higher molecular weight acids did not completely dissolve before
the carbonate addition). The pH of the final solutions or slurries
ranged from 6.2 to 7.8 for CaC03 and from 7.5 to 8.5 for MgC03.
Most of the mixtures dissolved very little after the third day.
The Ca-Mg content of a few samples actually appeared to decrease between
3 and 7 days. Some of thase variations may be due to sampling and analytical
errors, while others may be real. Despite the apparent lack of equilibrium
in many of the mixtures, the final mole ratios (acidrmetal) shown in
Table 2 are indicative of the potential of the acid to dissolve the car-
bonate. For the monobasic acids of appreciable solr.bilizing power, this
ratio was approximately 2 for Ca and 1.5 for Mg; for the dibasic acids, the
corresponding ratios were 1 and 0.75« The results indicate that those acids
having a solubilization ratio of 2-COOH groups per atra of Ca should be
considered as potential limestone slurry additives as this is the stoichio-
metric value.
The data in Table IE also show that the low molecular weight acids
were more effective in solubilizing calcium or magnesium carbonate per gram
of acid than were the high molecular weight acids. This is due to (l)
more carboxy groups per gram and (2) increased solubility of the low molecular
weight acids in water. The substitution of chlorine, hydroxy, or phenyl
groups for hydrogen decreased the solubilizing power per gram as expected,
while acids whose calcium salt is insoluble were particularly ineffective
(e.g., oxalic, tartaric, and citric acids).
Batch Exploratory Tests; Laboratory tests in batch apparatus
were conducted to compare the effectiveness of various organic acids in
improving scrubbing efficiency and limestone utilization. The tests were
made by adding the acid to 1$ limestone slurry and using the mixture to scrub
a synthetic flue gas, with stirring, in a gas scrubber bottle. The S02
content of the gas, before and after scrubbing, was monitored by an infrared
analyzer. The S0£ removals and CaO utilizations were calculated by two
methods, one based on chemical analyses of the spent scrubber liquor and
the other on evaluation of the infrared recorder chart; the latter method
appeared to be more reliable. The acids tested were salicylic, citric,
formic, benzoic, and acetic. The results are presented in Table III.
Although the data are somewhat scattered, there is an indication
that salicylic and citric acids caused a decrease in the utilization of
limestone; citric acid, at very low concentration (acid: CaO mole ratio =
0.06), appeared to have no deleterious effect. Formic, benzoic, and acetic
acids increased the utilization of limestone, the overall S02 recovery, and
the time to 50$ removal. Reasons for the poor performance of salicylic and
citric acids may be the high value of the ionization constant, the low solu-
bility of salicylic acid, and the ineffectiveness of citric acid in dissolving
CaC03 ( Table II \ Formic and acetic acids are quite volatile and could result
in substantial losses in scrubbing a stack gas.
268
-------
TABLE III
Laboratory-Scale Batch Tests of the Use of Weak Organic Acids
ro
»
10
Test No.
Acid added
w« ft
lonizatlon constant •
Amount added
Grams
Males x ID3
Limestone fed
Grams
Moles CaO x 103
W«±fvr fed a
IICbwGJ. * CLLj g£ •
Scrubber conditions
Gas rate, l./min.
Solution temperature, °F.
Total S02 fed
Grams
Moles x ID3
Test results d
Maximum S02 removal, #
Time to 50? removal, ndn.
Scrubber solution pH x
Overall S02 removal, #
By chemical analysis f
By infrared analysis
CaO utilization based
ont S02 .removal, #g
By chemical analysis
By infrared . analysis
CaO utilization attributable
to acid addition, %
By chemical analysis
By infrared analysis
32o
None
—
-
1.0
9.2
100
4.375
121-138
0.675
10.5
84
20
3-1
82
65
94
74
-
3o4
Salicylic
1.06 x 10-3
1.0
7.2
1.0
9.2
99
3.826
120-125
0.620
9-7
81
21
4.2.
64
67
-7
330
}41
37)
Citric
2.1
10.9
1.0
9.2
100
4.273
110-119
0.509
7.9
88
15
3.8
56
75
48
64
-46
-ID
8.4 x. 10
1.0
5.2
1.0
9.2
99
4.253
122-124
0.443
6.9
83
13
3-9
83
74
62
55
-32
-19
-4
0.1
0.5
1.0
9.2
99-9
4.743
120-125
0.686
10.7
87
20
4.2
88
75
102
85
8
11
.w
Formic
1.76 x 10 "4
1.0
21-7
1.0
9-2
"•99
yyy
31**
y*j
Benzole
4.273 4.743
121-125 118-120
0.775
12.1
88
23
4.0
86
75
113
99
19
25
0.775
12.1
92
21
3.7
104
77
137
101
43
27
6.3 x
2.2
18.0
1.0
9-2
97-8
4.253
123-126
0.731
11.4
88
22
4.5
69
65
85
80
-9
6
ID'S
O.2
1.6
1.0
9-2
99.8
4.253
120-125
0.731
11.4
86
22
4.2
67
73
83
90
-11
16
.7fj
Acetic
1.75 x ID'
1.0
16.7
1.0
9-2
99
JJC-
4.253 4.743
120-125 116-125
0.797
12.4
88
24
5.9
58
76
78
102
-16
28
0.775
12.1
93
21
3-7
105
76
138
100
44
26
a First hydrogen ion at 25° C.; value for HsSOs is 1.7 x 10
,. PhyslcjB, 21st Edition. % Chemical Rubber Publishing Co.)
w 9
at 25° C., for
ysj, . % .
CruSSeg (72% -200 mesh) Spring Valley limestone; 51-5$ CaO, 1.3# MgO,
c Simulated flue gas; 0.3# S02, 3«5# Oa, 80.256 N, l6.0# C02 (dry basis).
d Maximum proportion removed as indicated by infrared analysis.
® By material balance.
1 By intagration of recorder chart.
8 100 (moles S02 removed)/ (moles CaO input).
n (CaO utilization with acid) - (CaO utilization without acid).
is 3-5 x ID"7 .at 18° C. (Handbook of Chemistry and
COa, 4# S102, 0.05$ S.
-------
Continuous Exploratory Tests; Exploratory tests with citric,
formic, benzoic and acetic acids were carried out in a small pilot plant
consisting of two unpacked spray towers in series with the recirculating
slurry impinging on a deflector plate to give a flat spray. Each- scrubber
was k inches in diameter by 2 feet long. The slurry of 1$ limestone was
recycled at 9 1/min to scrub about 5 cfm (L/G =—lj.00) of simulated flue
gas. The procedure in these tests was to start with fresh slurry in the
surge tanks and adjust system flows and conditions to give maximum removal
of S02. Operation was carried out without slurry addition until removal
started falling off, and then a flow of fresh slurry was started in an
open-loop manner; the slurry flow rate was varied as necessary to maintain
S02 removal at approximately the initial level. Each test represents
about 5 hours of steady-state operation. The amount of acid added was
10$ of the stoichiometric amount for reaction with the limestone feed.
The results and conditions are presented in Table IV.
Citric acid gave better removal at 10$ stoichiometric addition
than did the control test, but utilization of limestone was poor and a
high feed rate of makeup slurry was required to maintain S02 removal; in-
creasing the citric acid to 50$ of the stoichiometric limestone feed gave
even worse results. Formic, benzoic, and acetic acids were more effective;
the S02 removal and limestone utilization were higher at lower makeup feed
rate. When the supernatant liquid from a previous acetate run (No. 6k)
was used to slurry the limestone, the results were poorer, indicating a
possible loss of acetic acid due to volatilization.
The product solids contained a much higher proportion of sulfate
in those tests in which organic additives gave increased utilization of the
limestone and improvement in S02 removal. Since sulfate crystals are larger
than sulfite, this phenomenon may improve the settling rate of the slurry in
waste ponds; it may also reduce the possibility of sulfite pollution if there
is seepage into ground water or streams from the pond.
Physicochemical Properties of Four Selected Acids
Pour acids were chosen from the exploratory tests to determine
their properties (and those of their Ca and Mg salts) that are important to
use in limestone slurry scrubbing.
Aliphatic Aromaf-.ir.
Monobasic Glycolic, CH2OHCOOH Benzoic,
Dibasic Adipic, HDOC(CH2)4COOH Phthalic,
These were selected on the basis of potential stability and nonvolatility,
ability to solubilize CaC03 and MgC03 (TableII) and their commercial availa-
bility at a reasonable price.
270
-------
ro
TABLE IV
Effect of Weak Acid Type in Increasing Solubility of Limestone Used in Scrubbing Stack Gasa
A
Acid Additive0
Test No.f
Makeup slurry feed rate, 8 l./hr.
Stack gasn
Inlet temperature, °F.
SOa content, p. p.m.
To first-stage scrubber
From second-stage scrubber
Indicated SOa removal, $
Slurry from first-stage scrubber
pH at end of period
Composition, J % by wt.
CaO
Total S
Sulfite S
Indicated oxidation, $
Indicated CaO utilization, %
Chalk
None
66
16.5
300
3395
4io
88
5.6
33
16.1
11.2
30
85
Limestone13
None
63
20.6
305
3200
615
8l
4.3
32
0.2
_
- 1
(38)
Citric
67
19.8
305
3290
290
91
4.6
4o
2.6
1.3
50
n
Citrlcu
69
21.7
302
3255
730
78
4.7
26
0.1
, 0.04
_
•
Formic
61
14.0
300
3300
380
89
4.1
26
14.2
0.1
99
97
Benzole
70
8.0
304
3235
305
91
3.8
39
21.5
78
96
Acetic
6k
12.6
310
3450
395
89
3.9
32
17.2
0.5
97
95
Acetic6'
65
18.0
295
3^50
520
85
4.2
34
7-0
36
36
a Two spray scrubbers k- inches in diameter by 2 feet long in series, recycling 1$ limestone slurry at
9 l./min. to each; about 5 c.f.m. stack gas produced by burning a mixture of natural gas and fuel oil to which
carbon disulfide was added.
13 Spring Valley limestone from Ralph Rogers and Company, Inc., Tuscumbia, Alabama; 76$ -200 mesh.
0 Amount used vas 10$ of the stoichiometric amount for reaction with the feed limestone unless noted otherwise.
^ 50$ of stoichiometric to react with feed limestone.
e Supernatant liquor separated from test 6^4- was used to slurry the limestone.
f Tests were continued for about 5 br. at steady state.
S 1$ limestone content.
"• Values averaged over steady-state period.
Composite over steady-state period.
£ Solids separated by filtration.
? Overall oxidation (both scrubbers).
Estimated by similarity to other tests.
-------
Stability of Acids; Two series of tests were made to determine
the stability of the four acids under conditions considered to be more
drastic than those likely to be encountered in scrubbing operations.
These accelerated tests were made with both solutions and melts1 of the
acids as follows:
Solutions Melts
Conditions Benzoic Phthalic Adipic Glycolic Benzoic Phthalic Adipic.
Acid concen- a
tration, % 0.3 0.1 2 70 100 112.2 100
Temperature,
°C 75 75 75 55 136 212 162
Time, hr 5! 53" 59 28 18 0.5 10
Composition
of gas, $
S02 50 67 50 67 67 67 67
°2 25 33 50 33 33 33 33
N2 25 0 00 00 0
Gas flow rate,
ml/min 200 120 40 45 45 45 45
Phthalic anhydride was used rather than phthalic acid.
The ultraviolet absorption of the solutions and melts after treatment
was determined as a means of detecting any decomposition of the acids. The
solutions after treatment had a strong odor of S02. Since the S02 interfered
with the absorbance measurements, it was eliminated by boiling the solutions
for about 10 minutes. . Results of the measurements, shown in Figure 1 for
benzoic acid,2 indicate that there was no chemical attack on the acids. There
were no new absorption peaks observed in the treated samples as there should
have been if decomposition products were present. There was a possibility
that some glycolic acid was volatilized from the 70$ solution, but this could
have been due to our inability to attain the same concentration of the acid,
before and after S02 treatment, for the ultraviolet absorbance comparison.
It was concluded that all four acids were relatively stable.
Glycolic acid was available only as a 70$ aqueous solution; therefore it
was not included in the melt tests.
o
Only the curves for the melt and for benzoic acid volatilized from the melt
are shown. Similar results, however, were obtained for the solutions.
272
-------
A - Pure bensoic acid
B.C - Melt and condensed benzole acid
1.0
, 0.1
200
0.0
250
Wavelength, ran
FIGURE 1
Ultraviolet Absorption of Benzole Acid
273
-------
Solubility of Salts and Solubilizing Power of Acids: The
calcium saltsi of the four acids were solubilized in water at 25° and
50°C with the following results:
Solubility and pH of Calcium Salts of Organic Acids at 25° and 50°C
Solubility, % salt by wt. pH of saturated solution
Acid 25°C 5(TC *
Benzoic
Glycolic
Phthalic
Adipic
2.80 (4000)
1.52 (3200)
0.33 (650)
3.U* (6850)
3-83 (5^50)
3- 2k (6850)
0.25 (500)
1.98 (4300)
7.38
T.P8
6.36
7.57
6.83
6.60
7.U
7.55
o
Values in parentheses are ppm Ca in saturated solution.
The solubility of the calcium salts of the monobasic acids (ben-
zoic and glycolic) increased and the pH decreased with temperature. The
salts of the dibasic acids (phthalic and adipic) decreased in solubility
with temperature. Benzoic, glycolic, and adipic acids are especially
promising at 50°C, but phthalic acid gives only about 1/10 of the calcium
concentration as compared with the other three acids.
The solubilizing power of the four acids toward CaC03 and MgC03
was determined at three concentrations of acid (0.05, 0.1, and 0.2$) at
25°C and at one concentration (0.1$) at 50°C. The results are given in
Table V. In each case approximately: two carboxy groups were required to
dissolve an atom of calcium from CaC03, regardless of temperature or concen-
tration of acid; for MgC03, somewhat less than two carboxy groups were re-
quired. The pH' s of solutions saturated with CaC03 were about 8 at 25°C and
about 7 at 50°C; for the. solutions saturated with MgC03 the corresponding
pH values were 9 and 8.3 at 25° and 50°C, respectively.
Effect of Organic Acids on the Rate of Oxidation of Sulfite to
Sulfate: The rate of oxidation of sulfite in solution to sulfate by pure
oxygen was determined in the presence of the various organic acids. The
tests were made at pH = k by adding sufficient powdered CaC03 to maintain
the pH constant as oxidation proceeded. The comparison is made in Figure 2
for initial sulfite concentration of 0.0123 molal at an oxygen flow rate of
20 ml/min.
The magnesium salts were not tested because they could not be prepared in
sufficient purity for testing.
274
-------
Acid, 0.1$
O Benzole
A phtnalic
D Adipic
Glycolic
A None
10
Time, hours
FIGURE 2
Effect of Organic Acids on the Oxidation of Sulfite to
I I I II • I I I M .•< Mil • •• III I I I I • • I •• •' • • •*• ^ I I 1 I • • «-™ -™- I ,« Ml -III I I - - I —1- 1 - - - I _ _ "I
Sulfate in Calcium System at 50°C and pH i<-.0
275
-------
TABLE V
Solubllization of CaC03 and MgC03 by Organic Acids
Effect of Concentration of Acid at 25°C
ppm Ca when acid ppm Mg when acid
concn., "jo, is concn., $, is
Acid 0.05 0.1 0.2 0.05 0.1 0.2
Benzoic 97 180 340 140 210 330
Glycolic IJO 2TO 550 170 280 460
Phthalic l^O 240 480 160 250 420
Adipic 150 290 570 170 280 470
Effect of Temperature When Acid Concentration is 0.1%
pH ppm Ca pH ppm Mg
Acid 25°C 50°C 25° C 50°C 25°C 50°C 25"C 50°C
Benzoic
Glycolic
Phthalic
Adipic
••••••—^•M
T-95
T.98
8.00
7-92
^•••^^•i^
7.18
6.69
6.7T
6.80
^•MM^HB
180
270
240
290
(^•••W^
180
350
250
320
««•«••••
9-1
9.1
9.2
9.1
8.5
8.2
8.4
8.3
210
280
250
280
200
300
260
290
Benzoic, glycolic, and adipic acids, at 0.1% in solution, re-
tarded the oxidation of sulfite; less than 40$ of the original sulfite was
oxidized in 5 hours when these acids were present, whereas 42$ of the sulfite
was oxidized in 1 hour when no additive was present. Adipic acid has some
retarding effect on oxidation, but not as great as the other three acids.
The curves for adipic acid and no additive do not plot into 100$ at zero
time, indicating an initial quiescent period before the oxidation proceeds.
The first order rate constants for oxidation as determined by the curves in
Figure 2 are:
Acid K x 10s, sec"1
None 43.0
Benzoic 2.4
Glycolic 2,9
Phathalic 4-3
Adipic 32.0
A similar comparison of benzoic acid with no additive was made at
an initial pH of 4.0 but without holding the pH constant during the oxidation.
276
-------
Three replicas of the test without additive were made to determine the
reproducibility of the oxidation rates and also because of the diffi-
culty in obtaining identical starting solutions. The results are shown
in Figure 3.
When benzoic acid was present at 0.16$ the pH changed much less
rapidly as oxidation proceeded than when it was absent. Again, the amount
oxidized in 1 hour with no additive was greater than that oxidized in
5 hours with benzoic acid present—indicating a retardation effect of the
rate of oxidation by benzoic acid.
These results do not appear compatible with -those obtained in
exploratory tests on S02 absorption (Table 3V), in which the sulfate:sulfite
ratio in the product solids was higher when organic acids were used. This
may have been due to the fact that excess sulfite crystals were present in
the absorption tests. At the lower pH resulting from the acid addition
(on the order of 1.5)* this would produce a much higher dissolved sulfite
concentration in the solutions containing organic acid, and therefore might
promote the transition CaS03°0.5H20(c) -» CaS03'0.5H20(]_) -> CaS04-2H20(1) -*
CaS04»2H20(s) even if the oxidation rate in the solution phase were reduced
by presence of the acid.
Aqueous Species: The thermodynamic dissociation constants of the
acids1'2 at 25°C are as follows:
Acid log Kg log Kg.
Benzoic -k
Glycolic -3.831
Phthalic -3.14 -5.40
Adipic -4.42 -5.41
These constants denote the pH at which the activities of the undissociated
acid and the organic anion are equal; for example, the activity of benzoic
acid, HBz, equals the activity of the benzoate ion, 'Bz~, at pH 4.2, etc.
Analysis of the solubility and pH data of Table V indicate that
the benzoate ion forms a weak complex, CaBz"1", for which the following
equilibrium constant is obtained:
CaBz+ ?! Ca + Bz" K* = 1CT2-13 ± °-15 at 25°C
K* = 1CT1'0 ±0'5 at 50°C
This compares with the only reported literature value of ~10-°'2 by Bunting
and Thong3 at 3°°c and an ionic strength of 0.4; this latter value becomes
10~°'8 at I = 0, using the Davies activity coefficients-
I Stability Constants, Special Publication No. IT, The Chemical Society (1964).
Smolyakov and Primanchuk, Russ. J. Phys. Chem. 40 (3), 331-3 (1966).
Can. J. Chem. 48, 1654-6 (1969).
277
-------
100
90 -
Initial Solutions
pH S as
I O
II D
III A
IV -- k.05
Soln. IV contained 0.16$
benzole acid
Time, hours
FIGURE 3
Oxidation of Sulfite to Sulfate in the
System CaO-SOa-S03-iy) at 50"C With Og, Gas
278
-------
The calculation of stability constants from solubility data re-
quires very accurate data both with respect to the component concentrations
and the pH of the solution. Benzoic acid has very sharp ultraviolet ab-
sorption peaks, and its chemical determination is much more precise than
that of glycolic, phthalic, or adipic acids. Even so, the dissociation
constant calculated from the data in Table V was not confirmed by the
solubility of the salt, CaBz2—again due to the sensitivity of the data.
No consistent values of dissociation constants for complexes of Ca or Mg
with glycolic, phthalic, or adipic acids were obtained from either the
dissolution of carbonates (Table V) or the solubility of the salts; in
many cases the data gave negative values for some of the assumed species.
The significance of the value of the dissociation constant of the
complex CaBz , K*, is shown in the following tabulation of the amount of
calcium required to complex various amounts of the benzoic acid for two
values of K* when the benzoic acid concentration is approximately 0.1$:
Degree of complexing ppm Ca required if
as CaBx+, % K* = IQ-**-"3 K* = 1Q-J-*"
10 30 , 300
50 300 3,000
90 3,000 30,000
A further attempt to determine the presence of an aqueous complex
was made using both ultraviolet absorption and calcium ion specific electrode
with Job's method of continuous variation of the ratio Bz:Ca in solution.
The ultraviolet absorption data showed a strictly linear increase in absorbance
with total benzoate concentration with no breaks or irregularities due to a
significant amount of CaBz+ species. The specific ion electrode measurements
were not definitive but did indicate a small amount of calcium that was other
than ionic Ca"1"1"; this could have been CaOH+, CaHC03+, or other known species
in addition to CaBz . The conclusion is that if a complex between calcium
and benzoic acid is formed, it is a weak complex (highly ionized) that defies
accurate measurements by the methods attempted.
Small-Scale Tests with Benzoic Acid
Continuous scrubbing tests with benzoic acid were made in the 5-cfm
two-stage scrubber described earlier except that commercial spray nozzles
were used which permitted operation at a liquid rate of TOO ml/min (37 gal/Mcf)
in each tower. Limestone slurry (with filtrate recycled) was added to the
second tower to scrub the gas, and water was added to the first tower to
humidify and cool the gas before scrubbing.^ The flue gas was obtained by
burning natural gas, with S02 added from a cylinder to obtain about 3000 ppm
in the gas. This procedure simulated closed-loop operation under conditions
of about 50$ S02 removal so that the effects of an organic acid added to the
limestone slurry could be tested. The results are shown in Table VI.
279
-------
TABLE VI
Effect of Benzole Acid, on Sulfur Dioxide Absorption
by Limestone Slurry3
Test No.
Recycle slurry
Benzole acid added, %° 0 50 75
pH at end of period ^.5 3.7 4.3
Stack gas
Inlet temperature, °F 205 210 210
S02 content, ppm
To first-stage scrubber 3150 3150 3125
From second-stage scrubber 1565 915 7^0
Indicated S02 removal, $ 50 7! 76
n
First scrubber of water recycling at about 700 ml/
min; second of limestone slurry recycling at
, 700 ml/min.
Percent of that required to solubilize lime for
reaction with all the S02 entering the system.
Adding benzoic acid to the slurry in an amount equivalent to that
required to solubilize 50$ of the lime required for reacting with the S02
feed increased the S02 removal from 50 to 71$; 75$ of stoichiometric gave
a further increase to 76$ S02 removal. Other runs (not shown) with lesser
fractions of benzoic acid were made with intermediate results. The results
of the improvement in S02 removal with the percent of stoichiometric benzoic
acid for lime solubilization are shown in Figure Ij.. The results indicate
a rather steep increase in S02 removal for the first 25$ of stoichiometric
acid and a leveling of the response beyond this amount.
An extended closed-loop test was made using 20$ of the stoichiometric
amount of benzoic acid. No apparent loss in the efficiency of the slurry to
remove S02 was noticed in 18 hours, indicating that the losses of benzoic
acid were probably low. A similar test using 150$ of the stoichiometric
benzoic acid indicated, by ultraviolet absorption analysis of the liquid,
about 1% loss per hour.
280
-------
30
ro
oo
20
10
First scrubber: Water recycled at TOO ml./min.
Second scrubber: 2$ slurry of local limestone
(76# -200 mesh, 51-5# CaO) recycled at
760 ml./min.; makeup at 5 l./hr.; pH 3-7-5-0
Inlet gas: 3150 p.p.a. S02
I
I
10 20 30 40 50 60 70
Percent of stoichlometric benzole add to solubillze CaO for all sulfur in
80
FIGURE k
Effect of Proportion of Benzole Acid
Added to Limestone Slurry on Sulfur Dioxide Removal
-------
The vapor pressure of benzoic acid at 0.16$ was determined be-
tween kO° and 65°C (l04° and 149°F) and was expressed by least squares
by
log Vmz = 9-20 - lj-787/T
where T is the absolute temperature, °K. This implies an energy of
evaporation of benzoic acid of 22 kcal/mole from solution, compared with
21-9 kcal/mole for sublimation of the pure acid. The dependence of
Henry's law constant, h, for benzoic acid absorption becomes
log h = -10.27 + ^500/T
and the vapor pressure of 0.01 molar benzoic acid at various temperatures
and pH's are as follows:
Vapor pressure, ppb when temperature is
B§
2
k
5
6
These values indicate that losses from evaporation will be relatively low
in the range of scrubber conditions. However, a vapor pressure of 100 ppb
of benzoic acid would result in the loss of about 2.5 Ib of benzoic acid
per hour if used to scrub the gas from a 500-mw boiler. It is likely £hat
other losses, through adsorption on solids and faulty mist eliminators,
will be more important than evaporation per se.
Summary
The use of certain weak organic acids, particularly benzoic, appears
promising as a means of improving the effectiveness of limestone slurry in
removing S02 from stack gas. It has been determined that the stability,
volatility, and calcium salt solubility for benzoic, adipic, phthalic, and
glycolic acids are within ranges that appear acceptable for practical
operation. Only small amounts, on the order of 20 to kO% of that needed
to dissolve the stoichiometric quantity of limestone (based on amount of
S02 to be removed), are required to effect a major increase in S02 absorption
efficiency.
45°C U13°F) 5
136
129
85
19
2
OiC U22"FJ
217
206
137
31
k
55°c U31°F;
355
33T
225
52
6
282
-------
The method has the further promise of increasing limestone
utilization. If all the feed limestone were dissolved in the organic
acid, the limestone should remain in the scrubber loop until completely
reacted; without the organic acid, part of the limestone is lost with
the waste solids because complete conversion of CaC03 to CaSOQ'O.^T^O
is never obtained in one pass through the scrubber.
The potential advantages of organic acid addition are such
that further work is planned, both on a small scale and in pilot plant
tests.
283
-------
POTENTIAL WATER QUALITY PROBLEMS ASSOCIATED WITH LIMESTONE
WET SCRUBBING FOB SO? REMOVAL FROM STACK GAS
J. S. Jfcrris
Division of Environmental Research and Development
Tennessee Valley Authority
Chattanooga, Tennessee
Prepared for Presentation at
Second International Line/Limestone Wet Scrubbing Symposium
Sponsored by the Environmental Protection Agency
New Orleans, Louisiana
November 8-12, 1971
285
-------
TOTEHTIAL WATER QUALITY PROBLEMS ASSOCIATED WITH LIMESTONE
WET SCRUBBING FOR SOg REMOVAL FBCM STACK GAS
J. S. Morris
Division of Environmental Research and Development
Tennessee Valley Authority
Chattanooga, Tennessee
The present emphasis on rapid attainment of a higher quality total
environment has focused added attention on the interrelated effects on air,
water, and land resources that result from new technology for controlling
pollution. An inescapable byproduct of any separation process is a concentrated
waste. Without waste recovery it is rarely possible to develop practical con-
trol methods for specific wastes in any one of these three natural resource
areas without affecting at least one of the other two areas. In the limestone
scrubbing process for control of sulfur dioxide, land use is affected by the
necessity for disposing of large quantities of solid waste material, and dis-
charge of once-through scrubber water or bleed-off from recirculating systems
to receiving streams presents a potential problem of water quality degradation.
A scrubber system based on a lime process has a somewhat different
effect on water quality than one based on limestone. In evaluating the poten-
tial for each process, TVA has concluded that the use of line appears infeasible
because the system requires (l) dilution of the scrubber liquor, which necessi-
tates an "open loop" operation with an overflow to a watercourse, or (2) use
of a very expensive, high rate of recirculation. Since most of TVA's efforts
have thus been confined to limestone scrubbing, the following discussion is of
this basic process.
Itotenf.;i a,1 "Pollution Problems
Table 1 lists the soluble cation constituents of the effluent from
a typical limestone scrubber. This semiquantitative spectrographic analysis
indicates the wide range of cationic species found in such a system. These
constituents, together with the associated anions—mostly 803, SOl^, and NOj—
and the pH and suspended solids, represent the potential water pollutants. It
is assumed that operation of a scrubber system will provide adequate settling
for solids; thus, suspended material will not adversely affect water quality.
The minor and trace elements shown in table 1 require quantitative
analysis because of the low concentrations permissible under existing standards.
The maximum stream concentrations of the following elements are based on allowable
limits for potable water: manganese, 0.05 rag/1; zinc, 5-0 mg/1; silver, 0.05 mg/lj
copper, 1.0 mg/1; and lead, 0.05 mg/1. Stream standards generally apply to the
concentrations existing after the effluent is fully mixed with the receiving waters.
286
-------
Adverse effects from various concentrations of the remaining trace cations are
included in the Water Quality Criteria publication of the Resources Agency of
California referred to in the appendix. With the exception of zinc, manganese,
and copper, TVA has not quantitatively analyzed the minor and trace elements
listed. Chemical analysis of scrubber water during one preliminary pilot-plant
run resulted in the following maximum concentrations: manganese, 80 mg/1;
zinc, 9.0 mg/1; and copper, 5-0 mg/1. Each of these concentrations would likely
exceed limits for an effluent standard, and manganese would likely exceed appli-
cable stream standards.
TABLE 1
SPECTROGMPEEC ANALYSIS
Element Relative Distribution
Calcium Major
Iron Major
Aluminum Major
Silicon Major
Magnesium Major (-)*
Sodium Major (-)*
Potassium Major (-)*
Titanium Major '(-)*
Manganese Minor
Zinc Minor
Tin Trace
Copper Trace
Silver Trace
Vanadium Trace
Molybdenum Trace
Galium Trace
Mercury Trace*-*
Lead Trace**
Boron Trace
Nickel Trace
•*Less than other major components, but more
than minor quantities.
**Quantity is at the lower limit of detection
for emission spectrograph.
Quantitative analysis of the major soluble constituents from flyash
and limestone together with some preliminary data from pilot-plant operation
indicate that the characteristics of the scrubber water that are of most concern,
relative to potential for adverse effects on water quality, are those shown in
table 2. Table 2 also indicates an approximate allowable concentration of each
constituent based on various regulatory standards and on criteria recommended
ty TVA. Although limits generally apply to the concentrations after mixing,
strict adherence to a nondegradation standard or adoption of an effluent standard
287
-------
might require that these limits not be exceeded in the discharge itself.
Effluent standards, such as those recently adopted by the Ohio River Sani-
tation Commission for the Ohio River, are becoming more prevalent with
increased demand for more stringent controls.
A general discussion of the detrimental effects of the more important
of these characteristics is presented in the appendix.
TABLE 2
^Maximum Concentration
Milligrams "Per Liter
Calcium ) .
M . ) Hardness (as CaCOq) 200
Magnesium ) ^ J
Sulfate 150
Chloride 150
Total Dissolved Solids 500
PH 6.5 - 8.5
*Based on various regulatory standards and TVA
criteria for receiving streams.
Conceni
The constituents listed in tables 1 and 2 are likely to be present
in any scrubber solution; however, if one disregards the threat to the aqueous
environment from system leaks and pond seepage, the real potential for degrada-
tion of water quality is due to any scrubber solution discharged to the receiving
waters. Since operation of a scrubber system with complete recycle increases
the problems of scaling and may cause operational problems due to the buildup
of dissolved salts, many systems are being designed to discharge some of the
scrubber water. Figure 1 shows a simplified diagram of a typical limestone
scrubbing system. Item D, the discharge from the settling pond, may be any
portion of the total contents of the pond from no discharge to 100 percent dis-
charge. "No discharge would represent a fully closed-loop system with no potential
for pollution, and this system is discussed only in the conclusion. The following
discussion is confined to partial bleed-offs and to 100 percent discharge.
Table 3 shows the theoretical concentrations in scrubber water from a
limestone process utilizing a 3 percent WgO limestone, a 150 percent stoichio-
metric limestone feed for 3-5 percent sulfur coal, 100 percent NOg absorption,
100 percent MgO reaction, 25 percent formation of calcium sulfate, and 75 percent
formation of calcium sulfite. This theoretical analysis is based on the simul-
taneous solution of assumed solubility equations. The analysis represents the
buildup of concentrations for a system with 100 percent recycle; a system with
a partial bleed-off, with the necessary makeup of fresh water, would be expected
to contain somewhat lower concentrations. The reduction in concentrations is
dependent on the percentage of scrubber water discharged; however, for bleed-off s
288
-------
FIGURE I
TYPICAL SCRUBBER FLOW SYSTEM
00
GAS FROI^
BOILER
TO STACK
SCRUBBER
A
B
MAKE UP H20
CaC03
RECIRCULATION
TANK
POND
OVERFLOW
TO WATER-
COURSE
->-WET SOLIDS
-------
of 20 to 30 percent, the reduction is not expected to be .significant from the
standpoint of water quality. Table h compares the theoretical concentrations
with preliminary data from a pilot plant with a similar operating system. The
pilot-plant feed was approximately 100 percent stoichiometric, and the MgO
content was about 1 percent. The major differences between these calculated
and measured concentrations are in the concentrations of sulfites and nitrates.
The theoretical assumptions regarding N02 absorption are apparently erroneous,
since no pilot "plant data reviewed to date indicate more than 10 mg/1 NOo. The
higher sulfite concentration in the results for the pilot plant may indicate
a formation of MgS03, although no satisfactory balance of the complexes can be
made from the data. The most important information from the pilot-plant data
is not obtained from an analysis of the differences between them and the calcu-
lated values, but from the conclusion that the high concentrations of potential
pollutants theoretically predicted may actually occur in the system.
TABU;
THEORETICAL CONCENTRATIONS IN SCRUBBER HMD EFFLUENT
Calcium 950 mg/1
Magnesium 3*500 mg/1
Hardness (as CaC03) 16,500 mg/1
Sulfite 150 mg/1
Sulfate 9,500 mg/1
Nitrate 7,750 mg/1
Sodium 850 mg/1
Total Dissolved Solids 25,000 mg/1
pH 6-.0
TABLE U
THEORETICAL CONCENTRATIONS
KITH
PRELIMINARY PILOT
Theoretical ELlot ELant
Concentration Concentration
Milligrams Tier Liter Milligrams "Per Liter
Calcium 950 1,600
Magnesium 3,500 1,150
Hardness (as CaCOo) 16,500 7,100
Sulfite 150 5,1*00
Sulfate 9,500 6,500
Nitrate 7,750 1
Sodium 850 26
Total Dissolved Solids 25,000 13,000
PH 6.0 5.0
290
-------
The high quantities of dissolved salts in these scrubber systems.
represent a significant potential for creating a problem of water pollution.
Utilization of other limestones in different modes of operation may create
greater or lesser potential problems; however, for partially closed-loop
operations, table h should generally be representative. In situations where
stringent effluent standards are implemented or where strict interpretation
of a nondegradation standard is enforced, it is doubtful if sen er waters
can be .discharged without a very high degree of treatment. The potential
adverse effects from such discharges—regardless of standards—ate dependent
on the quantity of bleed-off and the flow of the receiving stream.
The effects on a receiving stream from the release of such concen-
trations were calculated for a hypothetical 1500-megawatt plant with an
assumed bleed-off of 30 percent, equivalent to about k,500 gpm, and an assumed
dilution ratio of the receiving stream to plant discharge at critical low
stream flow (? day-10 year minimum) of 135 to 1. The ratio is obtained by
assuming that the critical low flow is equivalent to the condenser cooling
water required for plant operation, 600,000 gpm or 13*4-0 cubic feet per second.
Although this may seem extreme, a random survey of some 20 existing plants
indicated that two plants are located on streams that have much less relative
stream flow. It is not actually necessary to assume any plant size, since the
ratio of discharge to available dilution would remain the same in all cases;
however, the assumptions indicate that the critical flow levels are represen-
tative of fairly large rivers. Figure 2 shows the resulting percentage of
total allowable concentration assuming typical background concentrations of
100 mg/1 hardness as CaCX^, 20 mg/1 S01+, 150 mg/1 total dissolved solids, and
0 mg/1 manganese, complete mixing in the stream, and a scrubber discharge
containing the theoretical concentrations shown in table k less the nitrate
concentrations. The manganese concentration was taken from the pilot-plant
data. Although only hardness and manganese exceed the assumed limits, the
percentage increase in the concentration is quite significant. Such increases
in total stream concentration are not generally acceptable. Refer to the
appendix for a discussion of actual adverse effects from the resulting concen-
trations.
One potential water-quality problem, that of the dissolved sulfite
concentration, appears to be much less severe than originally anticipated and
may, in fact, be quite insignificant. High concentrations of sulfite will
utilize large quantities of dissolved oxygen in the oxidation to sulfate:
0.2 milligrams of oxygen per milligram of sulfite. The presence of certain
metals in the flyash, such as iron, '.was- thought to be sufficient catalyst for
the oxidation of sulfite to proceed rapidly enough to significantly lower the
dissolved oxygen in the settling pond and ultimately in the receiving stream.
From pilot-plant results, it now appears that this reaction proceeds quite
slowly and, indeed, may be difficult to induce.
A complete theoretical computation of the chemical characteristics
of the effluent from the settling pond for a limestone scrubbing system operating
with a 100-percent overflow has not been made; however, some pilot-plant data
have been analyzed. Table 5 compares pilot-plant data obtained during open-loop
operation with those obtained during closed-loop operation. The open loop
operated with 1 percent MgO limestone and a limestone feed of approximately
150 percent stoichiometric for a unit burning coal with 3 percent sulfur. Although
the concentrations are less for the open loop, they are surprisingly high and of
291
-------
FIGURE 2
HARDNESS
STREAM CONCENTRATIONS AT LOW FLOW
50
13
SO,
ro
vo
ro
DISSOLVED
SOLIDS
MANGANESE
110
60
30
55
J L
1200
PERCENT OF MAXIMUM ALLOWABLE CONCENTRATION
BACKGROUND
:•:•:{ SCRUBBER CONTRIBUTION
-------
such magnitude as to be of concern. If the concentrations for closed-loop
operation are assumed to be representative of a 30~percent bleed-off, then
the total quantities of waste constituents discharged from each of the
operational modes are similar. The greater flow of the open-loop process
'(more than three times that of the closed-loop system) produces somewhat
greater amounts and, therefore, a greater potential for adverse effects on
water quality.
TABLE 5
CONCENTRATIONS IN OPEN-AMD CLOSED-SCRUBBER SYSTEMS
Closed-Loop Open-Loop
Mode Mode
, Milligrams per Liter Milligrams "Per Liter
Calcium 1,600 1,100
Magnesium 1,150 *
Hardness (as CaCOo) 7,100 **2,700
Sulfite 5,^00 1,700
Sulfate 6,500 2,600
Nitrate 1 6
Total Dissolved Solids 13,000 - **7,000
PH 5.0 5-3
*Not determined, MgO content of limestone less than 1%.
**Not determined, estimated minimum values.
Factors That Affect the Quality of Scrubber Water
Almost all input factors in a limestone scrubbing system will affect
the quality of scrubber water and its potential for water pollution. In partic-
ular, the coal, limestone, feed rate, type of boiler, type of scrubber, ratio
of liquid to gas, and pH of the scrubber water will have major effects on the
characteristics of the scrubber water.
The oxidation of sulfites to sulfates in the system will have a
major bearing on the quantity of dissolved salts in the system and, according
to TVA pilot plant results, will greatly affect the settling characteristics
of the reaction products. If the solids are mainly calcium sulfite rather
than calcium sulfate, as has been the case in much of TVA's work on the closed-
loop system, settling is very poor. The thin sulfite platelets form a gel-like
structure that settles very slowly. Although this settling problem has no
direct effect on water quality, the settling pond for the gel~like substance
would certainly require careful design to assure containment adequate to prevent
any threat to nearby watercourses. While formation of sulfate may be more desir-
able from the standpoint of settling, it may create a greater potential for
direct water pollution. The sulfate reaction products are much more soluble
than the sulfites, more than 50 times as soluble in the case of the calcium pro-
ducts; thus, higher concentrations of dissolved salts result from increased
oxidation.
293
-------
The chemical makeup of the limestone is quite important in defining
the potential effects on water quality. Among the major controlling charac-
teristics of the limestone is its magnesium content, which has an important
bearing on waste disposal because of the high solubility of magnesium sulfat.e.
There is some indication that small amounts of magnesium may appear as magnesium
hydroxide rather than sulfate in the products; this is desirable because the
hydroxide is extremely insoluble. In many areas of the country, however, calcitic
limestones are not readily available, and dolomite (CaC03'MgC(^) or dolomitic
-limestone would be the raw material. Figure. 3 shows the concentration of sulf ates
and hardness of the scrubber water as functions of the MgO content of the limestone,
Conclusion
A given concentration of a pollutant has the same detrimental effect
on water quality no matter what its source, but a minor detrimental effect from
the residual waste discharge from an efficient production process and treatment
facility might be considered acceptable while the same effect, if it results
from an attempt to provide control of wastes fr.om other sources sufficient to
meet an applicable standard, might not be considered acceptable. TVA believes
that a solution to a problem of air pollution that in turn creates a problem
of water pollution—even though possibly one of lesser extent—is not satis-
factory. Thus, TVA's full-size experimental unit being designed for installation
on unit 8 of the Widows Greek Steam ELant will be a closed-loop operation with
no degrading of water quality.
If an open loop is absolutely necessary for other systems, their
continued operation, in the immediate future, will be dependent on the charac-
teristics of the waste from the process, the quantity of discharge, and the
size of the receiving stream. It seems increasingly evident, however, that
more stringent controls, the implementation of effluent standards, and strict
enforcement of nondegradation policies will require a closed cycle or a high
degree of treatment. Although no thorough investigation of technically feasible
methods of treatment has been made, wastewater containing such high concentrations
of dissolved solids as are present in the scrubber flow generally requires ad-
vanced, sophisticated, and often expensive treatment. Treatment by distillation
is considered technically feasible, but since the estimated annual capital and
operating cost for treating a 30-percent bleed-off from a 1500-megawatt plant
is $1.5 to $2.0 million, this method is economically infeasible. Although other
treatment methods may be technically adequate and more economical, most processes
still require disposing of a concentrated waste product. Based on the present
state of the art of treatment of such wastes and environmental considerations,
complete recycle of the scrubber water, wherever technically feasible, should be
the operational mode for limestone systems.
Although it is quite apparent that limestone scrubbing systems can have
a potential for affecting water quality, the magnitude of this potential is not
adequately defined. TVA is continuing its efforts to gain additional information.
The Radian Corporation recently began a complete chemical analysis of TVA's pilot-
plant scrubbing process, and the resulting data will be used to refine the theo-
retical analysis of the system and to aid in evaluating effects on water quality.
The large-scale lime/limestone scrubbing tests scheduled to begin in March 1972
at. the TVA Shawnee Sfceam Hant will provide an opportunity to further define the
effects of various scrubbers and operating parameters.
294
-------
FIGURE 3
EFFECT OF MgO CONTENT OF LIMESTONE
ON
HARDNESS AND S04 OF SCRUBBER POND EFFLUENT
30
to
O
er
UJ
fe 20
or
uu
CO
CO
a:
o
CO
10
CO
o
cc
o
V)
0
o
HARDNESS
S04
PERCENT MgO IN LIMESTONE
295
-------
APPENDIX
The feasibility of releasing dissolved or suspended calcium salts
into watercourses depends on how much the content of these impurities in'
the watercourse is increased, which in turn depends on the amount of the
particular pollutant released in the effluent, the content in the watercourse
before influx of the effluent, and the water volume flow of the watercourse.
'Various regulatory agencies and advisory groups have set or recommended
limits above which the concentration of an impurity in a watercourse should
not be increased. Such limits for the dissolved salts released from a
limestone-wet scrubbing process are summarized herein.
Suspended Solids
Most regulatory agencies do not place quantitative limits on either
suspended solids or turbidity for raw water used as a source for domestic
water, but do generally specify that concentrations shall not be sufficiently
high as to be objectionable or interfere with normal treatment processes.
Both turbidity and suspended solids can be reduced to acceptable limits for
most purposes by conventional treatment methods; however, high concentrations
increase the cost of processing the water by increasing chemical requirements
and the volume of sludges to be disposed of. In some cases head losses
through the filters are increased with a resulting increase in the frequency
of filter backwashing.
Settleable suspended solids often blanket stream bottoms, killing
fish eggs and young and destroying much of the normal benthic biota which
serve as food organisms required for propagation and growth of fish. Turbid"
ity in streams also interferes with light penetration and limits or diminishes
the photo synthetic effects necessary for the primary productivity of fish-
food organisms. The Committee on Water Quality Criteria of the Federal Water
Pollution Control Administration! recommends a limit of 50 Jackson Turbidity
units in warm water streams.
issolved Solids
The following information is quoted from Water Quality Criteria.
issued by the Resources Agency of Calif ornia.^
(a) Domestic Water Supplies. The 1962 USPHS Drinking Water
Standards specify that the total dissolved solids should not exceed
500 mg/1 if more suitable supplies are, or can be made available.
This limit was set primarily on the basis of taste thresholds. The
1958 WHO International Standards set the '^permissible limit" at
Federal Water Pollution Control Administration. Water Quality Hriteria
(Report of the National Technical Advisory Committee to the Secretary of
the Interior). U.S. Government Printing Office, Washington, D.C. (April 1, 1968).
2
McKee, J. E. , and Wolf, H. W. Water Quality Criteria. The Resources Agency
of California, State Water Quality Control Board, Publication No. 3~A (1963).
296
-------
500 mg/1 and the "excessive limit " at 1500 mg/1, but no "maximum
allowable limit" is given. The 1961 WHO European Standards do not
include limits for total dissolved solids. It is generally agreed
that the salt concentration of good, palatable water should not
exceed 500 mg/1; however, higher concentrations may be consumed
without harmful physiological effects and may indeed even be more
beneficial. Each water with a total salt concentration over 1000
mg/1 should be judged on the basis of the local situation, alter-
native supplies, and the reaction of the local population.
(e) Fish and Aquatic Life. It has been reported that among
inland waters in the United States supporting a good mixed fish
fauna, about 5% have a dissolved-solids concentration under 72 mg/1;
about 50% under 169 mg/1; and about 95% under kOO mg/1.
Water quality objectives adopted by TVA for surface water in the
Tennessee River basin include a limit of 500 mg/1 of dissolved solids.
EH
TVA has proposed upper and lower pH limits for the waters of the
Tennessee Valley of 8.5 and 6.55 respectively. The pH values required for
industrial water supplies are covered in some detail in the FWPCA Water
Quality Criteria report. The proposed TVA criteria should be adequate for
a.11 of these but a few extreme cases.
Calcium
Calcium in water supplies has been suspected of producing undesir-
able physiological reactions but no definite relationship has yet been
established. The major concern is related more to its deleterious effects
in uses such as washing, bathing, and laundering and to the problem of
incrustations on cooking utensils and water heaters. Some industrial processes
require a very soft or low calcium water. Costs of conditioning water for
such industries is directly related to source concentration.
Thus the major detrimental effect of calcium hardness is economic.
TVA has proposed the following limitation on water hardness:
There shall be no substance added to the waters that will increase
the hardness to such an extent as to appreciably impair usefulness as
a source of water supply or interfere with other reasonable and necesso,ry
uses of the water.
Rivers iri the Tennessee Valley on which steam plants are located
generally range from 50 to 125 mg/1 hardness as calcium carbonate. In general,
a concentration of 100 mg/1 would not be objectionable for most uses if this
concentration were not significantly greater than the normal range for the area.
297
-------
Calcium Sulfate
No limits have been placed on calcium sulfate by regulatory agencies.
The following summary of beneficial uses and reported acceptable concentrations
is quoted from McKee and Wolf.
(a) Domestic Water Supplies. High calcium sulj.c,_~ concentrations
in water are disadvantageous for most household uses, but for drinking
purposes, 300 mg/1 or mere is not harmful. The taste threshold of calcium
sulfate has been reported to be 250 to 900 mg/1. Calcium sulfate signifi-
cantly increases plumbosolvency; raising the calcium sulfate concentration
from 25 to 250 mg/1 increased by 10 percent the amount of lead dissolved
from lead pipes.
(b) In general, calcium sulfate is beneficial in the brewing industry,
for it helps to maintain the acidity of the wort and therefore causes more
complete coagulation of albuminous matter. It also reduces the solubility
of the bitter substances of the hop.
(c) Irrigation. Gypsum in irrigation waters improves or restores
the permeability and tilth of soil having an unfavorable sodium ratio.
(d) Stock and Wildlife Watering. A saturated solution of•calcium
sulfate, used as the sole source of drinking water for rats, permitted
satisfactory growth. A concentration of 2^00 mg/1 of calcium sulfate
permitted normal growth and reproduction among rats.
(e) Fish and Other Aquatic Life. Trama reported that a saturated
solution of calcium sulfate at 20° C did not produce significant mortali-
ties among bluegills; but a later report from the same organization
indicated that this concentration was the 96-hour TI^ at 18-20° C in soft
water for the bluegill sunfish. It was also reported that 3200 mg/1
caused a 50 percent reduction in the rate of growth of the diatom,
Navicula seminulum. Using highly turbid water at 21-25° C and the mosquito-
fish (Garobusia affinis) as the test animal, Wallen et al. indicated that
the 96-hour Tim was greater than 56,000 mg/1, a concentration much larger
than the solubility.
TVA has proposed no specific limits for calcium sulfate; however,
the concentration is limited by the allowable sulfate concentration and the
permissible added calcium hardness.
Sulfates
McKee and Wolf discuss sulfate content as follows.
(a) Domestic Water Supplies. The 1962 Drinking Water Standards
of the USIHS recommend that sulf ates not exceed 250 ing/1, except where
a more suitable supply is not available. This limit does not appear to
be based on taste or physiological effects other than a laxative action
toward new users. Public water supplies with sulfate contents above this
limit are commonly and constantly used without adverse effects. The 1958
WHO International Standards established a "permissible limit" of 200 mg/1
298
-------
and an "excessive limit" of 1*00 mg/1, but set no maximum allowable limit.
The 1961 WHO European Standards include a recommended limit of 250 mg/1
for sulfates.
Korotchenok reported on heavily mineralized drinking waters in
western Turkmenia (USSR), noting that no outbreaks of disease have been
ascribed to waters in which the content of sulfate did not exceed 1295
mg/1. In British Somaliland well waters used for human consumption
contain very high sulfates, many having 2000-3000 mg/1. One village
is using water of 1^00 ing/1 sulfates. A survey in North Dakota indi-
cated that water containing less than 600 mg/1 of sulfates is usually
safe. In reviewing the literature, Moore quotes Sollman to the effect
that concentration of 1000 mg/1 of sulfates in water is harmless. A
cathartic dose is 1.0 to 2.0 grams, or a liter of water containing 1000
to 2000 mg/1 of sulfates.
Sulfates appear to have no detrimental effect on the corrosion of
brass fittings in domestic water systems, nor do concentrations less
than 200 mg/1 increase plumbosolvency.
Whipple is quoted by Moore to the effect that the taste thresholds
of sulfate salts were as follows:
Sodium sulfate 200-500 mg/1
Calcium sulfate 250-900 mg/1
Magnesium sulfate 1|-00-600 mg/1
(b) Industrial Water Supplies. Limiting or threshold concentrations,
or optimum ranges, for sulfates are assembled below:
Limits of recommended values. mg/1
Industrial process SO^ CaSOk Mgj30U. _Ea2JSQ'.i_
Brewing - 100-500 100-200 100
Garb beverages 250 - -
Concrete corrosion 25 - -
Ice making - 300 130-300 300
Milk industry 60 - -
Hioto processes - 100
Sugar malting 20 - -
Textiles 100 - - -
TVA has proposed a limit of 150 milligrams per liter for sulfates to
provide a reasonably satisfactory water for both domestic and industrial
use. Normal water treatment has little effect on sulfate concentration.
Sulfite,
The major concern in regard to high sulfite concentration is oxidation
to sulfate, thus lowering the dissolved, oxygen content of the receiving water.
299
-------
Magnesium
The effect of dissolved magnesium has been treated by McKee and Wolf.
(a) Domestic Water Supplies. Magnesium is an essential mineral
element for hviman beings; the daily requirement of magnesium is about
0.7 gram. Magnesium is considered relatively non-toxic to man and not
a public health hazard because, before toxic concentrations are reached
in water, the taste becomes quite unpleasant. At high concentrations,
magnesium salts have a laxative effect, particularly upon new users,
although the human body can develop a tolerance to magnesium over a
period of time.
The 19J46 USTHS Drinking Water Standards recommended a limit of
125 mg/1 but there is no limit in the 1962 standards. The 1958 WHO
International Standards have a "permissible Idjnit" of 50 mg/1 and an
"excessive limit" of 150 mg/1, but no maximum allowable concentration.
The 1961 WHO European Standards have a recommended limit of 125 rag/l>
but if the sulfate exceeds 250 mg/1, the magnesium is limited to 30 mg/1.
The taste threshold for magnesium (in MgSO^) has been reported as
100 mg/1 and for the average individual it is given as about 500 mg/1.
The negative correlation between hardness in water and cardiovascular
disease does not appear to hold for magnesium as it does for calcium; yet
one investigator reports the favorable use of magnesium gulf ate to treat
such cases and claims that magnesium, rather than calcium, is the beneficial
element in reducing cardiovascular attacks.
(e) Fish and Other Aquatic Life. The relative concentrations of
magnesium and calcium in water may be one factor controlling the distri-
bution of certain crustacean fishfood organisms, such as copepods, in
streams. Hart et al. cite a report that among U.S. waters supporting a
good fish fauna, ordinarily 5 percent have less than 3-5 nig/1 of magnesium;
50 percent have less than 7 ag/1; and 95 percent have less than Ik mg/1.
Magnesium chloride and nitrate can be toxic to fish in distilled
water or tap water at concentrations between 100 and UOO mg/1 as magnesium.
However, magnesium chloride, nitrate, and sulfate, at concentrations be-
tween 1000 and 3000 mg/1 as magnesium have been tolerated for 2-11 days.
Some fresh-water fish have been found in very saline lake water containing
over 1000 mg/1 of magnesium as well as additional sodium and calcium salts.
TVA has proposed no specific limit on magnesium.
300
-------
POTENTIAL UTILIZATION OF SOLID WASTE
FROM LIME/LIMESTONE WET-SCRUBBING OF
FLUE GASES*
by
Linda Z. Condry, Research Technologist
Richard B. Muter, Supervising Research Chemist
and
William F. Lawrence, Project Supervisor
Coal Research Bureau
West Virginia University
Morgantovn, West Virginia
Prepared for presentation before the Second International Lime/Limestone
Wet-Scrubbing Symposiun New Orleans, Louisiana, November. 1971
*Work supported by the Environmental Protection Agency under Contract No.
CPA 70-66.
301
-------
INTRODUCTION
The increasing waste disposal problem associated with coal combustion
may be compounded by the greater amount of waste generated if potential
alkaline earth sulfur dioxide abatement systems are implemented. Although
utilization schemes for normal flyash are available, "modified flyash," the solid
by-product resulting from lime/limes tone wet-scrubbing programs, represents a
new type of solid waste material having different chemical and physical properties
from regular flyash. In addition, simple disposal of this ash by land fill or
lagooning may cause serious water pollution problems especially if dolomite
is used as a'modifying agent. Previous research^ has shown that very little
potential exists for mineral separation and beneficiation of this new ash and
Indicates that the most economical utilization scheme would involve total or
whole utilization of the ash. For these reasons the Coal Research Bureau of
West Virginia University has been performing research under partial support from
(2)
EPA (Environmental Protection Agency) to develop and evaluate total
utilization processess. As a result of this research, several promising
utilization areas have emerged. These include production of autoclaved
materials such as calcium-silicate products, production of materials such as
mineral wool using high temperature processes and agricultural uses such as
soil amendments.
Sources of Modified Flyash
Several modified flyashes were investigated during the course of this
research; however, only two were of a wet-collected nature. They were
obtained.from the Kansas Power and Light Co., Lawrence, Kansas (designated
KPL), a limestone wet-scrubbing system, and from the Union Electric Company,
St. Louis, Missouri (designated SLO), a dolomite wet-scrubbing system. In
302
-------
addition, a sample from the Shawnee Steam Plant, Paducah, Kentucky (designated
TV A) was also examined. This ash originated from a limestone injection, dry-
collection system of the type which could possibly be employed before a wet-
scrubber .
Autoclaved Products
Because modified flyash liberates sorbed sulfur dioxide at elevated
temperatures, it was necessary to find a technique for the total utilization
of this ash which would either lock the sulfur into the product or require no
external application of heat. Preliminary research examining the natural
pozzolanic activity of limestone modified flyash indicated that some
cementitious setting occurred. Such setting is sufficient to obstruct power
plant process lines if water flow should stop but does not appear to impart
enough strength for structural products. For this reason and because
the technology has been well developed in Europe using normal flyash, auto-
claving was selected for further investigation. In addition to forming a stronger
material, this technique would also have cne advantage of locking sulfur into
the final product.
Various autoclaved products were investigated using modified ash as
the principal raw material. Those products exhibiting the greatest potential
for utilization of modified ash were:
a. brick or block,
b. aerated or foamed concrete, and
c. concrete materials.
A. Calcium-Silicate Brick
Calcium-silicate (CS) brick production was examined as a potential large
tonnage application of limestone modified flyash. Bricks produced by
autoclaving modified flyash have the added advantage of binding the sulfur
303
-------
components within a calcium-silicate matrix, and because they are full-
strength after curing, they are immediately available for marketing. In
addition, bricks made using this raw material surpass the standards for
conventional sand-lime brick.
Figure 1 illustrates a simplified process flowsheet for the production
of CS brick. As illustrated, this process involves dewatering of the ash,
addition of silica sand and lime, mixing of the resultant material, pressing
to the desired form, humidity curing and autoclaving. Brick made with, on a
dry basis, 40 percent KPL wet collected limestone modified flyash, 39 percent
silica sand (30 x 100 mesh) and 11 percent lime (95 percent pure CaO) exceed
ASTM (C73-67, C73-51) standards for grade SW (severe weathering) calcium-
silicate brick. The compress ive breaking strengths are in excess of 5000 psi
as compared to the ASTM specification of 4500 psi minimum.
Process parameters which were found to be of importance in the production
of CS brick include mixing and humidity curing. It was found that surface
cracking due to insufficient calcium-silicate bonding occurred if the various
components were not intimately mixed before pressing. This problem was solved
through the use of a muller-type mixer. Also, it was found .that the
compress ive strength of the final product could be increased significantly
if a humidity curing step was incorporated in the process prior to autoclaving.
The humidity curing was effected by storing the pressed green brick over water
at room temperatures for periods up to twelve hours.
Among the promising areas for utilization of calcium-Silicate brick are:
general low cost construction materials, interior walls, decorative walls,
patios, and thermal or acoustic insulatine walla or barriers.
B. Aerated Concrete
Originally patented in the U. S., aerated concrete has been widely
developed in Europe. Practically all new factory buildings in Sweden and
80 percent in West Germany are constructed, in part, using this material.'
304
-------
This lightweight structural material consists of small non-communicating gas
cells entrained in a calcium-silicate matrix.
A process flowsheet for the production of aerated or foamed cellular
concrete from modified flyash is included as Figure 2. Additions of
cement and a gas producing agent to the ash are required; however, for
this process, no pressing or humidity curing such as was required for brick
production is necessary. Also, if a dry-collected modified ash is used,
additional lime or sand is not required. On a dry basis the composition
of materials used for aerated concrete production was 90 percent dry-collected
modified ash, 10 percent portland cement (type I) and 0.16 percent aluminum
powder.
The resulting aerated concrete has a density which may be varied from 50
3 ' •
to 56 Ibs/ft and a compressive strength, which may be controlled, between
400 and 850 pal. Commercial aerated concrete varies from 20-50 Ibs/ft3 in
density and has compressive strengths on the order of 400-800 psi. This
material can be sawn, nailed, drilled, screwed or glued as with wood but has the
.thermal and fire resistant properties of concrete. A few potential areas
for utilization are in non-loadbearing walls, sandwich construction with
brick or concrete for insulating purposes and interior surfacing for exterior
walls. This material can be formed in desired shapes or cut from panels
and should be highly suitable for modular construction.
Research in this area was performed using both wet and dry-collected
modified ashes. Chemically very few elemental differences are apparent between
the two types of ash; however, the greater amount of reactive lime in the
dry-collected material made it more amenable than wet-collected material to
this process. As with foreign manufacture which uses normal flyash, aeration
and setting times are crucial in that setting must occur after aeratior is
completed and before the entrained bubbles collapse. It is felt that a major
advantage of modified flyash production may lie in the use of electrostatic
-------
precipitators in conjunction witn. wet scrubbers. A dry ash suitable for
aerated concrete production would be obtained in large quantities while the
overall efficiency of the sulfur dioxide abatement process would probably
not be affected. Such a process would greatly reduce the need for ash
dewatering and would .act to reduce the potential threat of increased water
pollution. In addition, a predominantly calcium sulface material would
be produced in the scrubber system.
C. Poured Concrete
The natural pozzolanic properties of limestone modified flyash indicated
that a potential use might be as a cementing agent in concrete materials.
Because research on CS brick production showed that autoc laving increased
calcium-silicate bond formation, further research was performed to investigate
the feasibility of using modified ash as the cementing agent for the production
of concrete block.
Included as Figure 3 is the process flowsheet employed in this phase
of research. It should be noted that the composition is the same as for CS
.brick with the exception that less dewatering is required. Also, It should be
pointed out that the material is poured into molds instead of pressed. As
with CS.brick, this composition consisted, on a dry basis, of 50% wet-collected
modified flyash, 39 percent silica sand (30 x 100 mesh) and 11 percent lime
(95 percent CaO). The resulting poured concrete which is similar to
3 .'•.-- o
concrete block had a bulk density of 90 Ib/ft as compared to 150 Ibs/ft for
conventional concrete block and a compressive strength of approximately 900
psi as compared to 1000 psi for conventional block. Although addition of
the necessary aggregate could raise the bulk density to 100 Ib/ft3, this would
still be only 2/3 the bulk density of standard block. Storage requirements
could also be reduced as this is a full strength, non-shrinking product
out of the autoclave and does not require 3 to 4 weeks curing time before
sale.
306
-------
An advantage of the use of modified ash is that no prior grinding or
crushing of the raw material is required. In addition, the possibility is
being examined that limestone modified flyash could be used for the production
of an autoclaved aggregate to be used as coarse material for this concrete.
The advantages of producing autoclaved calcium-silicate products
from limestone modified flyash would Include:
1. No sulfur dioxide capture or marketing problems would result as
processing occurs below the temperature for sulfur dioxide
regeneration;
2, Little or no storage or curing time is required before use
as the structural materials produced are durable, pre-shrunk
and pre-strengthened;
3. Mo exotic equipment is required for production as common
European technology is employed;
4. Ho pollution problems should occur at the plant as rejects
and process waters may be recycled; and
5. A potential large tonnage market exists in areas of low cost
housing and construction.
Fired Products
High temperature processing was also examined to develop other possible
(4)
uses for modified flyash, the most promising being mineral wool. The
alkaline earth constituents in limestone modified flyash reduce the
amount of heat necessary to make the ash sufficiently fluid for the production
of mineral wool. Comparative tests using normal flyash and bottom ash with
equivalent amounts of lime added as a fluxing agent gave a poor grade of
mineral wool and required more heat for processing. Conversely, mineral wool
fibers produced from modified ashes were uniform, quite resilient and were
superior to commercial fibers in protecting steel from rust corrosion.
AS mentioned previ< usly, the modified ash releases sorbed sulfur dioxide
upon heating to temperatures over 1700°F. Since this could constitute a new
307
-------
air pollution problem, it was felt that a study of sulfur dioxide evolution
from heated modified ash was in order. Accordingly, several modified flyashes,
both wet and dry-collected, were examined to determine their sulfur evolution
characteristics. In all instances, sulfur dioxide was evolved over a narrow
temperature range in a relatively pure form and in concentrations similar to
those considered suitable for sulfuric acid manufacture. For example, the
theoretical total yield of sulfur dioxide in sample KPL is 10.34 percent by
weight. At 2140eF, (melting temperature 2150-2160°F), 96 percent of its
total yield is given off in twenty minutes. Even with this long heating
period, concentrations of 9-13 percent sulfur dioxide in the off gas could be
maintained. The only impurities observed were carbon dioxide and water vapor,
both of which can be evolved at lower temperatures if desired.
• After evolution of the sulfur gases, the parent ash may be further
heated to a molten state for production of mineral wool or it may be cooled
and utilized for the production of fired ceramic materials.
Soil Amendment
Dried sample KPL (wet-collected,, limestone modified flyash) is being
examined as a soil amendment agent. This work is currently being performed under
sub-contract to Virginia Polytechnic Institute.* The results to date indicate
that modified flyash with its increased calcareous components may find use as
a soil amendment in the area of pH control. In addition, it was noted that
such uuage effects an increase in the boron supplying power of some soils
to which the ash was added. As with regular flyash, modified ash also has
the ability to improve the texture and drainage characteristics o? some soils.
A potential advantage for wet-collected modified flyash is that the slurry
may be sprayed directly on the soil area without the need for prior drying.
*A report of this research, performed by Dr. David Martens, is included as an
appendix in the final report of Contract No. CPA 70-66.
308
-------
Before such utilization is recommended, however, further research will be
necessary to determine whether a possible pollution problem due to water
leaching occurs.
SUMMARY
Wet-collected modified flyash has been found to be a suitable raw
material for the production of autoclaved structural products such as
bricks, aerated concrete and cement materials. Other areas of application
are in heat treating for mineral wool production and sulfur dioxide collection
as well as in soil amendment for agricultural purposes. All of the processes
discussed utilize the unique characteristics of modified flyash to eliminate
a potential solid waste disposal problem. A more detailed discussion of the
processes described in this paper may be found in the final report for Contract
No. CPA 70-66 which will be available from the Environmental Protection Agency
in the near future.
309
-------
REFERENCES
1. Anderson, Ronald E., Cockrell, Charles F., e£ ajL., "Study of Potential
for Recovering Unreacted Line From. Lines tone Modified Flyash by
Agglomerate Flotation," Final Report, Contract PH 22-68-18, National
Air Pollution .Control Administration, May, 1970.
2. Contract numbers CPA 70-66 and EHS D 71-11 between West Virginia
University and the Environmental Protection Agency.
3. ___, PFA Data Book, Central Electricity Generating Board,
England (1969).
4. Cockrell, C. F., Muter, R. B., and Leonard, J. W., "Study of the
Potential for Profitable Utilization of Pulverized Coal Flyash
Modified by the Addition of Limestone-Dolomite Sulfur Dioxide
Removal Additives," Final Report, Contract PH 86-67-122, National
Air Pollution Control Administration, April, 1969.
310
-------
FIGURE 1
Silica-§and
30 x 100 Mesh
Wet-Collected Limestone
Modified Flyash (1% Slurry)
Dewater Ash
to 58% H20
50% Ash + 11% CaO
+ 39% Sand
Paddle Mixer
(10 Minutes)
Pour into
Molds
T
24 Hours Thermal
Set (110°C)
Autoclave, 16 Hours
190 psig, 185°C
72 Hours Air
Drying
Lime (CaO)
96% Pure
FLOW SHEET: FORMED CONCRETE
311
-------
FIGURE 2
Mech. Precip. - Dry-Collected
Limestone Modified Flyash (927.)
8% Portland Cement
Type I, Normal
0.16% Aluminum
Powder
Paddle Mixer
(10 Minutes)
Pour into Molds
8 Hour
Air Set
Autoclave, 16 Hours
190 psig, 185°C
72 Hour Air
Drying
FLOW SHEET: AERATED CONCRETE
312
-------
FIGURE 3
Silica Sand
30 x 100 Mesh
Wet-Collected Limestone
Modified Flyash (1% Slurry)
Dewater Ash
to 34% H20
50% Flyash + 11% CaO +
39% Silica Sand
Paddle Mixer (10 Min.)
@ 20.4% H20
a.i.aKing Reactor
(1 Hour)
Forming Pressure
3000 psi, 17.7%
H20
'. » Hour Humidity Storage
95% Rel. Humidity @ STP
Autoclave, 8 Hours
190 psig, 185°C
72 Hours Air
Drying
Lime (CaO)
96% Pure
FLOW SHEET: CALCIUM-SILICATE BRICK
313
-------
PILOT SCALE RESEARCH AND DEVELOPMENT
F.T. Princiotta, Chairman
Participants:
B.N. Murthy, D.B. Harris, and J.L. Phillips
I.S. Shah
R.A. Person, C.R. Allenbach, I.S. Shah, S.J. Sawyer
R.J. Gleason
T.M. Kelso, P.C. Williamson, J.J. Schultz
N.D. Moore
A. Saleem, D. Harrison, N. Sekhar
J.L. Shapiro and W.L. Kuo
J.H. McCarthy and J.J. Roosen
A.L. Plumley and M.R. Gogineni
D.E. Reedy
John M. Craig, Burke Bell, J.M. Fayadh
Robert J. Phillips
Ivor E. Campbell and James E. Foard
315
-------
Pilot Scale Research and Development
(Parts I, H, and HI)
Second International Lime/Limestone Wet Scrubbing Symposium
New Orleans, Louisiana
November 8-12, 1971
F.T. Princiotta, Chairman
SUMMARY
In light of the recent trend toward preserving and improving
man's environment, legislation and genuine public concern has put
increasing pressure on electrical utilities to control the discharge
of gaseous and particulate pollutants. Since wet limestone processes
are generally considered closest to the state of the art in controlling
sulfur oxide emissions, research and development has been particularly
intense over the last several years in this area. The number of papers
in this session and the depth of investigation supports the conclusion
that more intensive and comprehensive investigation of wet limestone
processes has occurred relatively recently.
The Pilot Scale Research and Development Session of the Second
International Lime/Limestone Wet Scrubbing Symposium included papers
316
-------
which presented pilot plant results and pilot plant status for a
variety of scrubber types. Emphasis has been noted in selection of
scrubbers which are non-tortuous and not overly prone to scaling or
plugging when handling saturated or supersaturated limestone or lime
slurries. For the most part emphasis has also been on closed loop
systems, which although are the most prone to reliability problems,
are also the most acceptable since they minimize the extent of water
pollution problems.
The participants in the Pilot Scale Session were as follows:
Name Organization
B. N. Murthy Environmental Protection Agency
I. S. Shah Chemico
C. R. Allenbach Union Carbide
R. J. Gleason Cottrell Environmental Systems
T. M. Kelso Tennessee Valley Authority
N. D. Moore Tennessee Valley Authority
A. Saieem Ontario Hydro
J. L. Shapiro Bechtel
J. H. McCarthy Detroit Edison
A. L. Plumley Combustion Engineering
D. E. Reedy Universal Oil Products
B. Bell Zurn Environmental Engineers
R. J. Phillips General Motors Technical Center
I. E. Campbell Smelter Control Research Association
Tables 1 and 2 describe the pilot facilities and summarize
results of selected pilot plant studies presented in the Pilot Scale
Session. Please note, that only typical results were included in the
table, since it was infeasible to include all results from all the units
described. Also, the EPA Prototype Facility at the Shawnee Steam Plant
and the Bisehoff Prototype were included to show how these larger plants
compare to the pilot scale unit in terms of size and scrubber types.
317
-------
Also included were some selected pilot plant results from earlier
studies by Howden-ICI in the 1930's and 1940's in order to help put
the most recent results, as presented in the Pilot Scale Session, in
reasonable perspective.
Scrubber types which have received the most attention have been
the: venturi, Turbulent Contact Absorber, Hydrofilter (flooded marble
bed), spray tower and packed tower. Generally, pilot plant results
have indicated that under carefully selected operating conditions, all of
these scrubbers, with the probable exception of the packed tower due its
inherent plugging tendencies, can be operated with relatively high S0_
and particulate removal efficiencies, with acceptable reliability. Pilot
plant results have indicated that the desirable combination of good
removal efficiencies without excessive down-time can be achieved, under the
following conditions: (1) high liquid to gas flow rate ratios, (2) high
solids content in the scrubber slurry, (3) long residence times in a delay
tank following the scrubber and (4) proper selection of scrubber type.
318
-------
Table 1. Selected Wet Limestone Pilot Plants and Prototype Facilities
Status and Design Characteristics
Plant
Howden - ICI
(Fullham)
Howden - ICI
(Bankside)
EPA/Zurn at
Key West and
Shawnee Steam
Plant (Mobile
Unit)
Bischoff
Scrubber
Prototype at
Steag Power
Plant, Lunen,
W. Germany
Ontario-Hydro
Cottrell unit at
Ohio Power's
Tidd Plant
Status
Tested
in mid-
1930's
Tested
1948-49
Tested
1970-71
Started
Feb
1971
Tested
1971
Tested
1970-71
Design Characteristics
Size
acfm
3,150
11,700
1,500
110,000
4,000
1,000
Alkali
Lime and
limestone
Whiting
chalk
(limestone)
and lime
;
1
Various
Scrubber
Grid (wood) -
packed
Grid (wood) -
packed
Inspirating type;
limestone ! Zurn Dustraxtor
and lime
Lime and
calcium
hydroxide
Venturi in series
with spray tower
i
i
i
i
\
Limestone
Limestone
Spray tower
Venturi in series
with a wetted
film packed bed
Pt of Alkali
Addition
Scrubber
circuit
Scrubber
circuit
Fuel
Coal
1.7-2.0%S
Oil
/w
2.7%S
I
Scrubber
Oil (1-2%S)
circuit at Key Wes
! coal (1. 5-3^
at Shawnee
:
i
Gas inlet Coal
duct i 2-4. 5%S
;
Scrubber
circuit
Scrubber
circuit (alsc
Coal
1020-1600
ppm SO2 i
Coal
i 1100-2000
some sim- 1 ppm SO2 ir
ulated boil- \
*er injection)!
<*>
to
-------
Table 1 (Cont.). Selected Wet Limestone Pilot Plants and Prototype Facilities
Status and Design Characteristics
CO
ro
o
Plant
TVA unit at the
Colbert Steam
Plant
C.E. Prototype
Facility at
Windsor, Conn.
Mohave/Navajo
Pilot Facility
EPA Prototype
Facility at
Shawnee
Steam Plant
Status
Tested
1971
Started
up 4/70;
tests run lor
various
customers and
in-house tests
since; current-
ly EPA boiler
calcined lime-
stone tests
start 12^6/71
Started
up 7/71;
no data,
available due
to pilot
.plant problems
Testing
sched.
start
3/72
Desicrn Characteristics
Size
acfm
2,600
12,000
!
two
4,000
circuits
Alkali
Primarily
limestone
Na2C03 ,
lime,
limestone,
dolomite
Scrubber
1) Venturi-rod/
packed bed;
2) venturi-rod/
spray tower;
3) 3-stage TCA
(mobile bed)
scrubber
Hydrofilter
(flooded marble
bed) 1 or 2
beds
(
Limestone,
lime, soda
ash (w/ and
w/o regen.
by lime)
!
j
three !
30,000
circuits
|
Limestone
and
hydrated
lime
]
1) TCA ;
2) venturi;
3) polygrid
packed;
4) Lurgi
impingement
1) Venturi-spray
or packed bed;
2) versatile TCA;
3) hydrofilter
(marble flooded
bed)
Pt of Alkali
Addition
Scrubber
circuit
a) Scrubber
circuit;
b) boiler
injection
(simulated)
Scrubber
circuit
a) Scrubber
circuit:
y
b) boiler
injection
i
Fuel
Coal,
3000 ppm
SOo inlet,
4g/scf part
Oil
(S02, fly-
ash added)
,
Coal
0.3-0. 8%S
Coal
1. 5-3%S
-------
Table 2. Selected Wet Limestone Pilot Plants and Prototype Facilities
Typical Performance and Operating Parameters
Plant
Howden - ICI
(Fullham)
Howden - ICI
(Banks ide)
EPA/Zurn at
Key West and
Shawnee Steam
Plant (Mobile
Unit)
Bischoff
Scrubber
Prototype at
Steag Power
Plant, Lunen,
W. Germany ]
Typica
SO2
Removal, %
97-99
90-99
70
1 Performance
Particulate
Removal, %
97-99
approx 98
Not
measured
89 I 99.7
I
j
i
Stoich. %
Operating Parameters
L/G
p-pm/1000 cfm
100 lime; ? 97
130 lime-
stone
110 138
100 lime- Approx 600
stone (induced by
scrubber)
\
I
110 i
—
>
AP
in. HaP
1.2
5.7
12
10
(in venturi)
Comments
Scaling was a major
problem, but design
changes enabled fair
operating reliability;
L/G and scrubber size
excessive by modern
s tandar ds .
Reliable operation a
continuous problem;
Settling difficulties
encountered;
L/G and scrubber size
excessive by modern
standards.
Open-loop system;
Results shown for fresh
water—salt water slurry
SO2 removal approx 80%
Operating reliability good
Large prototype system;
Early results indicate
scaling and plugging
problems and mechanic;
difficulties;
System is down now; to b<
restarted in early 1972.
CO
ro
-------
Table 2 (Cont.). Selected Wet Limestone Pilot Plants and Prototype Facilities
Typical Performance and Operating Parameters
GO
INS
ro
Plant
Ontario Hydro
Cottrell Unit at
Ohio Power's
Tidd Plant
TVA Unit at
;he Colbert
Steam Plant3^
C.E. Prototype
Facility at
Windsor, Conn
Typical Performance
S02
Removal, %
72
55-98
1) Varied due
to scaling;
2) 77 (avg);
3) 92 (avg).
95+ (Na2CO3
1 bed);
80 (reliably
with CaCO3
and Ca(OH)«
2 beds)
!
Particulate
Removal, %
No part, in
flue gas
Not
measured
1) 99. 3;
2) 98.9-
99.3;
3) 98.3.
99
Stoich. %
130
100-120
1) 150;
2) 150;
3) 150.
100-110
(Na2CO3,
Ca(OH)2);
120-150
(CaC03)
Operating Parameters
L/G
gpm/1000 cfm
72
40
1)40;
2) 60 (to both);
3) 38.
15
(Na2C03);
20 per bed
(CaC03,
Ca(OH)2)
AP
in. H20
0.5-1
7-9 (total)
1) Approx 10
(total);
2) 16 (total);
3) 9.
Approx 6
(per bed)
1
Comments
Reliability of operation
has been good;
No serious scaling or
plugging problems.
Performance data based on
an 80-hour test; some
SO2 variation due to SO2
inlet variation;
Considerable plugging of
packed bed: 23 Ibs after
80 hours.
1) Major scaling in packed
bed- -deemed unacceptable;
2) A 354-hour test was run
with only 20 hours down
time — reliability good--
only minor scaling;
3) A 172 -hour test indicated
no scaling problems --
ball wear, grid erosion,
and particulate entrainment
were problems.
Unit designed to solve
field unit problems and
to predict performance
of new units.
-------
Table 2 (Cont.). Selected Wet Limestone Pilot Plants and Prototype Facilities
Plant
Mohave/Navaj o
Pilot Facility
SPA prototype
Facility at
Shawnee
Steam Plant
Typical Performance
so2
Removal, %
wm —
__
Particulate
Removal, %
— ^
«.—
Stoich. %
^ ^
--
Operating Parameters
L/G
;pm/1000 cfm
— —
--
AP
in. H20
— —
--
Comments
Plant aimed at evaluating
s crubbing of low SC>2
level flue gas ;
Plant versatile with
extensive instrumentation;
Plant to evaluate use of
soda ash scrubber
solutions with lime
regeneration;
Data should be available
within s e ver al months .
Most versatile wet
limestone facility;
Extensive instrumentation;
Tests are planned to
fully characterize wet
limestone scrubbing —
break- in, screening, and
long-term tests are
scheduled;
Unit to start up '3/72.
aNumbers listed opposite TVA Unit at the Colbert Steam Plant represent:
1) venturi-rod/packed bed;
2) venturi-rod/spray tower; and
3) three-stage TCA (mobile bed) scrubber.
-------
SULFUR DIOXIDE ABSORPTION STUDIES WITH EPA
IN-HOUSE PILOT SCALE VENTURI SCRUBBER
B. N. MURTHY and D. B. HARRIS
EPA, RESEARCH TRIANGLE PARK, NORTH CAROLINA 27711
J.L. PHILLIPS, RADIAN'CORPORATION, AUSTIN, TEXAS
FOR PRESENTATION AT THE
SECOND INTERNATIONAL LIME/LIMESTONE WET SCRUBBING SYMPOSIUM
NEW ORLEANS, LOUISIANA
NOVEMBER 8-12, 1971
325
-------
SULFUR PIOXIM ABSOHPTIOK 8TOH8 WITH EPA ft Pt-HOUSE PILOT-SCALE
VEHTURI SCRUBBER
B. K. Murthy 4 D. 1. Harris, EPA
and
J. L. Phillips, Kadlan Corporation
Tha in-house venturi pilot plant systems which was originally designed and
in use for pertlculate removal studios, was modified to Investigate the
lines tone vet-scrubbing process for sulfur dioxide removal. Changes included
replacement of ductwork with fiberglass-plastic piping, installation of
slurry mixing and hold tank* and clarifian and instrumentation for measuring
pi! values, liquid and gas flow rates, slurry sampling, and for gas analysis.
The venturi section was replaced by a stainless steel unit of similar dimen-
sions and modified flow nossle.
Vapor-liquid and solid-liquid mass transfer in the venturi scrubber ware
studied by treating flue gas containing sulfur dioxide with water, sodium
hydroxide solution and limestone slurry in separate ser&esof experiments.
The results indicated that the renturl scrubber alone is not an efficient
§02 absorption device. Liquid flew rate and liquid composition had an
appreciable effect°nthe overall mass transfer coefficient, indicating that
the liquid phase reaction was significant under the experimental conditions.
Experiments with limestone slurry indicated some solid-liquid mass transfer
in the venturi scrubber• Calcined limestone-flyash mixture removed more
sulfur dioxide in the wet-scrubber than lines tone.
The in-house venturl pilot plant studied also enabled the development of
suitable methods for slurry sampling and evaluation of laboratory analytical
methods for practical application in scrubber development.
326
-------
SULFUR DIOXIDE ABSORPTION STUDIES WITH EPA IN-HOUSE
PILOT SCALE VENTURI SCRUBBER
The .limestone wet scrubbing process for:sulfur dioxide absorption will
be studied on a large scale by EPA early in 1972 using three pilot plants on
a coal fired power generating station. The venturiscrubber is one of the three
systems planned for investigations.
To study the imnw+ant process variables in preparation for the large
scale test plans, an existing EPA in-house venturi pilot plant which was in use
for particulate removal studies, was modified. The main objectives of this
study were:
(1) To determine the extent of vapor-liquid and solid-liquid mass trans-
fer in the venturi scrubber.
(2) To investigate the factors influencing the important mass transfer
steps in the scrubber.
(3) To gather preliminary information on the absorption of NO .
3\
(4) To develop practical sampling and analytical methods for the proto-
type studies through trials under the pilot plant conditions.
EQUIPMENT
A flow diagram of the modified pilot plant system is shown in Figure 1.
The major changes in the equipment consisted of the following:
(1) The venturi scrubber, de-entrainer and the cyclone were replaced with
those of 304 stainless steel.
(2) The flow nozzle in the venturi was fitted with a 45° solid cone, to
handle slurry feed. The dimensions of the venturi scrubber are shown
in Figure 2.
i
(3) The ductwork downstream from the venturi up to the induced draft fan
was constructed of 8 in. diameter fiberglas reinforced plastic pipe.
(4) The I.D. fan was coated inside with a corrosion resistant paint.
(5) Two 55 gallon capacity sealed hold tanks fitted with stirrers and in-
let-outlet connections were installed in the slurry recycle line.
(6) A rectangular clarifier of 450 gallon- capacity was added to the
system to treat the slurry.
327
-------
OPERATION AND ANALYSIS
Flue gas, produced by burning natural gas in an incinerator, was diluted
with excess air to make up for the required volumetric flow rate through the
scrubber. The temperature and humidity of the gas could be adjusted by pre-
spraying with water in the incinerator. S02» NO and N02 were metered into
the flue gas stream before the mixer, which was a dry venturi. The limestone
or other dry solids were added to the flue gas at this point. The gas stream
then entered the venturi scrubber where it was mixed with the liquid reagent.
Downstream from the scrubber, the entrained liquid and particulates were
removed in a knock-out drum followed by a cyclone from where the gas was dis-
charged to the atmosphere through the I.D. fan.
Gas Sampling;
A Semi-automatic sampling system was constructed to monitor the flue gas com-
position at several points in the pilot plant. The sampling train included
a glass cyclone to remove entrained particulates and droplets.
Flue Gas Flow Measurement:
An ASME standard venturi gas flow meter was used to measure the gas flow rate
through the scrubber with an estimated accuracy within ±5%. The flow venturi
was located in a straight run of duct following the I.D. fan.
Liquid and Slurry Flow Controls:
Standard venturi flowmeters were connected to LSN differential pressure cells
and continuous recorders to measure liquid and slurry flow rates. The recorder-
chart readings were used to manually control the flow rates.
Liquid and Slurry Sampling;
Sampling points at several points in the slurry flow system were connected to
1/1 and 1/2 inch stainless steel tubings for sample draw out. Care was taken
during sampling to prevent oxidation or reaction with atmospheric carbon
dioxide.
Solid-Liquid Separation During Slurry Sampling;
Laboratory tests with a continuous centrifuge had, demonstrated that calcium
carbonate particles Gpuld-be_rapidly separated from a slurry of similar com-
position as expected in a scrubber, before analysis to determine the rate of
precipitation and dissolution in the system. However, under the pilot scale
scrubbing conditions in which the slurry contained fly ash and calcined lime-
stone particles, the performance of the centrifuge was not satisfactory.
The outlet liquid stream would become cloudy with suspended particles and the
flow lines in the centrifuge plugged with solids. Further work in developing
a satisfactory alternate method of solid-liquid separation has progressed else-
where.
328
-------
Gas Analysis:
S02 and NOx were analyzed by Whittaker Analyzers which were calibrated with
Span gases supplied in cylinders. The instrument output was connected to a
continuous recorder.
02 and C02 were analyzed by a laboratory gas chromatograph. Sampling was
done by the use of plastic bags.
Liquid £ Slurry Analysis:
Sulfite in the liquid sample w^ analyzed immediately by a modified iodine
method in which the pH of the solution was kept at 6.0-6.2 to prevent-inter-
ference from nitrites. The iodine solution was generated as needed for each
determination using standard iodate and excess iodide ion at low pH. The
excess iodine was back-titrated with arsenite solution using a dead-stop
technique for end point detection.
.Samples were shipped under stable conditions and analyzed at Radian Labor-
atories fos» the following species:
carbonate
total stilfinr-as sulfate-
total nitrogen
Ca
Mg
Na
K
Cl
The methods used for analysis are described in Reference 2.
RESULTS AND DISCUSSION;
Experiments we*»e conducted in Ifwo series. In the first series, the
vapor liquid mass transfer efficiency of the scrubber for sulfur dioxide
absorption was studied by using only water and sodium hydroxide solution
as reagents to circumvent problems of solids dissolving or precipitating
in the scrubber. The concentration of sodium hydroxide was in the range
of the dissolved reagent concentration in the wet-limestone scrubbing pro-
cess. The experimental conditions are shown in Table 1.
The effect of operating variables on the mass-transfer efficiency was
determined by calculating relative values of T%a'f the product of S02 mass
transfer coefficient and the interfacial area, from the familiar expression
for dilute gas concentration:
Y in p
z =^- • dy G „
Y outV y - y* = Kga *
329
-------
TABLE 1
COHDTIONS FOR SERIES 1.
FLITS GAS;
FLOW RATS
TEMFETIATUI32
NO
' NOg
COg
02
H20
(ACFM)
(MCZE
1000 TO 1300
250 TO 325
0.20
0 & O.cA-5
0 & 0.005
0.3 TO 1.3
20
0.6 TO 1*.2
BALANCE
KaOH IN WATER
(G MUUES/L)
(GFM)
0 TO 0.0^
10 TO 15
FKESStJKS PROP; (BJ.
7.8 TO 12.6
330.
-------
Kg = overall mass transfer coefficient for S0«
a = gas-liquid interfacial area per unit volume,
y = gas phase S02 concentration
y* = gas concentration in equilibrium with tb liquid
G = Average Molar gas flow rate through the scrubber
z = length of the scrubber
N = Number of transfer units
Relative value of Kga is then calculated as:
(K a) rel = Kga = Nt &
(KgtTref *
Equilibrium gas concentrations were calculated from liquid concentrations,
using a computer program. Experimental results (Table 2) indicated that
sulfur dioxide absorption did not reach equilibrium with the venturi scrubber.
Figure 3 shows the effect of liquid composition on Kga. The strong dependence
the rate of sulfur dioxide absorption on the liquid composition indicates that
liquid film resistance in the venturi scrubber is significant in the concentra-
tion range studied.
The effect of gas and liquid flow rates on Kga is shown in Figure 4. The
increase in mass transfer with liquid flow rate is expected to be due to
increase in surface area from the additional droplets. An increase in gas
flow rate did not appear to have an appreciable effect on mass transfer.
The pressure drop across the scrubber, however, was affected more by gas flow-
rate than the liquid flow rate (Figure 5). These results suggest that vapor-
liquid mass transfer in the venturi scrubber can be increased without appreciabJ
power input by increasing the liquid flow rate for the same gas flow rate.
Figure 6 shows the effect of flue gas temperature on the absorption rates for
water and alkali solution. Comparison of the results indicate that while a
decrease in temperature increased absorption with water, an increase in temper-
ature increased the absorption rate when the alkali solution was used. This
indicates that alkali absorption in the scrubber is controlled by the chemical
reaction mechanism in the liquid phase, since chemical reaction rate is a
strong function of temperature. However, under process conditions when the
equilibrium partial pressure of S02 is not zero, a lower temperature may favor
absorption. Further studies are needed before firm conclusions on the effect
of scrubber temperature on absorption rate can be made .
331
-------
TABLE 2
EXPERIMENTAL RESULTS - SERIES 1
CO
CO
ro
RUN NO.
1
1 A
2
3
3 A
3 B
4
5
6
6 A
7
8
8 A '
9
9 A
io
10 A
INLET FLUE GAS ,
FLOW RATE
(ACFM)
1040
1090
1020
1230.
1210
1210
1010
1010
1070
1120
1050
1270
1255
1060
1115
1090
1100
TEMP.
(oF)
300
320-
310
315
300
300
310
' 260
310
300
310'
325 .
3.00
260
250
• 315.
300
COMPOSITION, MOLE %.
S02 '
.198
.200
.211
.205
.199
.290
-.206
.197
.200
.196
.198
.246
-.202
.239
.200
.239
.207
NOX
0
0
.054
.056
.051
.05
.044
.039
.044
.05:
.051
,049
.05
.048
.05
.03
.05
C02
.85
.6
1.2
.9
.4
--
1..3
.6
1.1
—
.7
.5
.3
—
.10
o2
19.6
20.6
17.7
19.6
18.8
—
18.8
20.1
19.8
—
19.2
20.1
—
19.8
—
19.2
H20
2.4
2.3
4.0
2.5
2.3
--.
i.s
4.2
2.5
—
2.7
1.6
1.2
—
.6
.AQUEOUS
TtoOR
Moles/1
0
0
0
0.
0
0
0
0
y
.0282
.0416
.0365
.0370
.0349
.0393
.0280
' .0190
.0116
FEED
Feed
Rate
(SPM)
ip
10
10
10
10
10
15-
10
10
9.9
15
10
9.9
10
10
10
9.7
ous
s/1
282
416
365
370
349
393
280
190
1116
FEED
eed
Rate
(SPM)
ip
10
10
10
10
10
15-
10
10
9.9
15
10
9.9
10
10
10
9.7
SCRUBBER
(nH2o>
—
8.75
8.6
7.8
12.0
11.4
12.6
9.0
8.5
8.4
9.3
9.2
12.4
12.4
10.6
9.5
8.6
8.7
SD« OUT (MOLE %) -
EXPERIMENTAL
.1645
,1630
.1780
.1500
.1560
.2460
.1570
.1590
.0850
.0850
.0460
.1475
.0980
.1475
.1100
.158.0.
.1420
EQUILIBRIUM
.1210
..1260
.097
.111
.095
.228
.090
.086
0
0
0
0
0
0
0
0
.0775.
-------
TABLE 3
CONDITIONS FOR SERIES 2
FLUE GAS:
Flow Rate
Temperature
Composition (Mole %):
NO
NO,
1000 ACFM
300° F
H20
'2 WVJ "U2 w$ U2 "2W "2
0.20 0.045 0.005 0.8 20 1.0 BALANCE
SOLIDS:
Feed Rate: 150% Stoichiometrie
Type: Piqua Piqua
Limestone Limestone
(1)
(2)
TVA
Calcined
Limestone
Flyash
(3)
Composition: 86% CaCOg 86% CaCOs 35.2% as CaO
Particle Size:
(Microns)
Range:
Median:
2.0-50 2.0-32 2.0-40.3
15.5 12.7 6.5
LIQUID:
Flow Rate Trough Scrubber: 10 GPM
Composition: Saturated Liquid from Clarifier
Pressure Drop ("H?0); 7.0-8.5
333
-------
NOX Absorption: No absorption of NO was noticed from the gas phase analysis
since the change in NOX concentration was within the range of instrumental
error. Liquid . pjhaae analyses also did not indicate any significant absorption
of NOX.
Oxidation of absorbed S02 to sulfate was uniformly 10 percent in all the runs.
Since the absorption rate of S02 was considerably more (about U times) with
the alkali solution than with water, the rate of oxidation was proportional
to the concentration of the sulfite in the liquid phase, indicating a chemical
reaction or a liquid film mass transfer controlled mechanism for oxygen transfer
in the scrubber. However, the concentration of oxygen in a power plant flue gas
will be considerably less than that initfce!se_tests, and the same oxidation
conditions may not apply to the prototype scrubbing conditions.
The extent of solid-liquid mass transfer in the venturi scrubber was qualitatively
studied in the second series of experiments. Commercial grade limestone and
calcined:., limestone-flyash mixture from dry limestone injection studies at
Shawnee (TVA) were used to compare the effect of limestone type and composition
on flue gas sulfur dioxide removal. The experimental conditions are shown in
Table 3.
Three modes of operation were used to study the mass-transfer effects. In the
first mode, clear saturated liquor from the clarifier was fed to the scrubber
with solids injection, under steady state conditions. In the second mode,
solids injection was suddenly stopped and gas monitoring continued. The differ-
ence in sulfur dioxide removal between the first and the second mode indicated
whether any of the solids were hydrated in the scrubber. In the third mode
of operation, with the system under steady state conditions as in the first
mode, solids injection was suddenly stopped and the slurry from the hold tanks
was directly fed to the scrubber, bypassing the clarifier. The difference in
sulfur dioxide absorption between the first and the third mode of operation
indicated whether any hydrated solids were dissolved in the scrubber.
The results, shown in Figure 7, indicate that very little hydration and dissolution
of limestone or flyash-calcined limestone takes place in the venturi scrubber.
The calcined limestone-flyash mixture removed appreciably more sulfur dioxide
from the flue gas than limestone alone. The results also indicate that the
limestone part.icle size had a minor effect on the sulfur dioxide absorption
in the size range studied.
CONCLUSIONS;
The results of this investigation lead us to the following conclusions:
(1) The venturi scrubber aldne is not an effective device for sulfur-
dioxide removal from power plant flue gases.
(2) Liquid recirculation rate and composition have an appreciable effect
on the rate of sulfur dioxide absorption in the venturi scrubber.
334
-------
(3) Further work is needed to establish the effects of flue gas temper-
ature and composition on the effectiveness of the scrubber.
(4) Negligible hydration and dissolution of the solids take place in
the venturi scrubber.
(5) Limestone type, particle size and composition have considerable
effect on sulfur dioxide absorption in the venturi. Detailed
studies are needed to optimize the conditions.
33*
-------
REFERENCES:
.1. Technical Note 200-006-19, "Evaluation of Sampling Procedures for
Limestone-Wet Scrubbing Systems - Colbert Test Series", by Radian
Corporation, Austin, Texas
2. Technical Notes 200-004-03 to 09, Radian Corporation
3. Final Report for EPA Contract No. CPA 22-69-138, by Radian Corporation
336
-------
SOLIDS
S02,NO,N02
GO
CO
FURNACE
FLUE GAS
CYCLONE
VENTURI
SCRUBBER
MAKE UP
WATER
FAN
DE-ETRAINER
CONCENTRATED
No OH t
SOLUTION '
1 water
t>
-i
-------
10
00
MATERIAL
14 GA.
304 S.S.
ADJUSTABLE CONE
30° / 3/4" LIQUID FEED PIPE
FIGURE 2 VENTUR! DIMENSIONS
-------
INLET GAS TEMP: 3OO °F
GAS FLOW- HOO (ACFM)
LIQUID FLOW-' |Q (GPM)
"" SCRUBBER AP- 9" HoO
-2
I 2 3 4 5 X 10
NQOH CONC-. (G-MOLES/L)
FIG-3 EFFECT OF
LIQUID COMPOSITION ON SCRUBBING
339
-------
LU
_J
LU
ft
r
1000
-O
INLET GAS TEMP: 300 F
NQOH CONC-r
04 G-MOLES/L-
1500
GAS FLOW, ACFM
10
LIQUID FLOW, GPM
FIG-4-EFFECT OF GAS AND LIQUID FLOW
RATES ON KGQ
340
-------
15
10
X
•
z
cu
-------
LJ
GAS FLOW. IO50 ACFM
LIQUID FLOW; io GPM
_ AP: 8- 9-5" H~o
lMaOH(-03GMOLES/L.)
•© WATER
240 280 320
TEMPERATURE, *F
360
FIG. 6. EFFECT OF INLET GAS TEMPERATURE
342
-------
80
70
60
UJ
O
cr
UJ 50
a.
> 40
O
5
LU
cr
20
UJ
l-
V)
bJ
FLY ASH PLUS
CALCINED L.S.
6.5 MICRONS
(QUA L.S.
r MICRONS
; MICRONS
MODE OF OPERATION
FIG.7 EFFECT OF LIMESTONE TYPE AND
OPERATING CONDITIONS ON SULFUR
DIOXIDE ABSORPTION
343
-------
Chemical Construction Corporation
Pollution Control Division
SO;, REMOVAL USING CALCIUM BASED ALKALIES
PILOT PLANT EXPERIENCE
I. S. SHAH
Chief, Process Engineering and Development
Pollution Control Division
Chemical Construction Corporation
320, Park Ave., New York, N. Y. 10022
Presented at
2nd International lime - limestone
Wet Scrubbing Symposium
New Orleans, Louisiana
November 8 - 12, 1971
345
-------
Chemical Construction Corporation
Pollution Control Division
SQ2 REMOVAL USING CALCIUM BASED ALKALIES
PILOT PLANT EXPERIENCE
I. S. SHAH, CHIEF, PROCESS ENGINEERING AND DEVELOPMENT
Introduction
Since January 1970, Chemical Construction Corporation in cooperation with various
utility companies has undertaken extensive pilot plant work to obtain basic data for
the design of commercial plants to remove sulfur dioxide from boiler flue gases
using calcium based alkali namely limestone (CaCOs), lime (CaO), Dolomitic lime-
stone (CaCO3« MgCOs), Dolomitic lime (CaO« MgO), and carbide sludge. In addition
to obtaining basic design information, extensive useful information regarding scaling-
build up problems, effects of delay tank, particle size, and quality of limestone on
SO2 removal efficiency, were obtained.
In this presentation, we would like to present our experiences and important findings
from the pilot plant work done to remove sulfur dioxide from boiler flue gases, using
calcium based alkali.
Pilot Plant Work
Chemical Construction Corporation, has conducted several pilot plant test programs
for sulfur dioxide removal using calcium based alkali. Pilot Plants were installed at
eight different power generating stations, all of whom use coal as the fuel. Of the
total of eight, seven were balanced draft boilers, and one was a cyclone type boiler.
346
-------
Chemical Construction Corporation
Pollution Control Division
Pilot Plant Work (Cont'd)
Table I is the summary of boiler type, type of fuel and its sulfur and ash content,
SO2 and dust loadings in the flue gas, for the eight generating stations. A total of
fourteen limestones, five different types of lime, one dolomitic limestone, one
dolomitic lime, and one type of carbide sludge were tested. Table III summarizes
the composition of limestones tested, Table IV summarizes the composition of
various limes, and Table V summarizes the composition of dolomitic lime and lime-
stone, and carbide sludge. Both 32-5 mesh and 200 mesh limestones were tested.
Lime was always slaked in accordance with standard slaking procedure.
Pilot Plant Facility
A generalized pilot plant will include a two stage scrubber-absorber system, an
I.D. Fan, recycle pumps, agitated delay tank, slaking system, lime or limestone
slurry tank, a thickener, and a filter. Table II provides a summary of pilot plant
equipment installed at various power plants, the type of alkali used, and the mesh
size to which it was ground.
The flue gas from the generating unit was taken downstream of the Air heater. The
duct size and location were determined to obtain an isokinetic sample while the boiler
is operating at or near capacity, and the gas volume measured at the outlet of the
Scrubber system is 1500 cu.ft. /min (saturated). The scrubber-absorber system
consisted of two W-A venturi scrubbers arranged in series.
347
-------
Chemical Construction Corporation
Pollution Control Division
Pilot Plant Operation
Flue gas, isokinetically withdrawn downstream of the air heater, enters the first
stage scrubber, where SO9 and fly ash are simultaneously removed by intimate
contact with recycled slurry of CaSOg, CaSO., Ca(OH)2 or CaCOo and fly ash.
The recycled slurry is introduced into the scrubber tangentially through open pipes,
and it swirls down the converging section of the scrubber, thus completely wetting it.
This liquor forms a curtan of liquor across the throat, and the accelerating gas
shatters it into small droplets, which provide surface for dust collection and absorp-
tion of sulfur dioxide.
The cleaned flue gas and liquor enter the separator where liquor is separated by
gravity, and any liquor carryover is separated by cyclonic action.
The flue gas now enters the second stage-venturi absorber-where SO- is absorbed
by contacting with recycled slurry. The flue gas is then exhausted to atmosphere
after passing through an I.D. Fan.
The make up limestone or lime slurry is added to the suction side of the second stage
recycle pump, and the bleed from the second stage enters the suction side of the first
stage recycle pump. The bleed from first stage is sent to thickener, the overflow
348
-------
Chemical Construction Corporation
Pollution Control Division
Pilot Plant Operation (Cont'd)
returning to the scrubber-absorber system. The thickener underflow is then filtered,
the filterate returned to the system, and the cake is piled.
In the case where thickener and/or filter is not included, the bleed from first stage
or thickener underflow is sent to sewer or ash pond for disposal. In case a delay
tank is part of the pilot plant facility, the bleed from second stage, and first stage
product liquor are sent to delay tank. The delay tank liquor is then recycled to the
first stage scrubber.
Samples and Data
The flue gas samples were taken at the inlet of first and second stage, and outlet of
second stage, and SC»2, NOx and dust loadings were determined. Various liquor
samples were taken and analyzed-both liquid and solid phases. Temperatures and
pressures of various streams were measured.
Fly ash samples were collected and particle size determinations and mineral analysis
were made. Lime and/or limestone samples were analyzed and analysis compared
with suppliers' data. Coal samples were collected during pilot plant operation and
proximate and ultimate analysis were performed. Table VI summarizes the analy-
tical data collected during each pilot plant operation.
Boiler operating data, namely Air heater temp., excess oxygen, steam production, and
349
-------
Chemical Construction Corporation
Pollution Control Division
Samples and Data (Cont'd)
generating load, were also collected during the pilot plant operation.
SC>2 Removal Using Limestone
During the pilot plant work at various power plants, fourteen different limestones
were tested. The composition of these limestones is presented in Table m. The
CaCOs content ranged from a low of 86% to a high of 99. 3%. The effects of various
operating variables on SO2 removal efficiency were determined. Table Vn sum-
marizes the various variables and their ranges studied.
Effect of Limestone Quality on SO2 Absorption
The quality of limestone varies from mine to mine, and from region to region. Of
the limestones tested, limestones 2 through 10 were from one geographical region.
12 through 14 were from another region, and 1 and 11 were from two other regions.
Table VIE summarizes the data showing the effect of limestone quality on SO2 absorp-
tion.
For limestones 2 through 5, limestone 4 was found to be the least reactive, when
ground to 325 mesh, whereas limestone 4 was the most reactive when ground to 200
mesh.
For limestones 12, 13 and 14, the least reactive was 14 when ground to 325 mesh, whereas
it was the most reactive when ground to 200 mesh.
350
-------
Chemical Construction Corporation
Pollution Control Division
Effect of Limestone Quality on SO2 Absorption (Cont'd)
The SO2 removal efficiency ranged from 30 to 80 percent, and of the 14 limestones
tested, only .two were found to be reactive enough to provide 70%-fSC>2 removal
efficiency at a reasonable stoichiometry. This is not adequate to meet present
pollution codes for sulfur dioxide.
There is a wide variation in SC>2 removal efficiency with the variations in limestone
quality. From the data obtained so far, it is still not possible to accurately predict
SO2 removal efficiency, just from the knowledge of the composition of the limestones.
Effect of Particle Size on SC>2 Abosrption
Smaller the particle size, larger the surface area available for absorption of SO2-
The surface area for 325 mesh particles is approximately 70% more than that for
200 mesh particles. This clearly shows that limestone ground to 325 mesh should
provide higher SC>2 removal efficiency than that ground to 200 mesh. Table IX is a
summary of data showing effect of particle size on SO2 absorption. The data does
prove that 325 mesh limestones are more reactive than 200 mesh limestones, except
for limestone no. 14. The extra operating and investment cost to grind 325 mesh lime-
stDnes must be compared with the cost for a higher stoichiometric requirement and/or
for delay tank, to attain the same SC>2 removal efficiency.
Effect of Limestone Stoichiometry on SO2 Absorption
The CaCOs reacts with SO2 according to the following chemical reaction,
351
-------
Chemical Construction Corporation
Pollution Control Division
Effect of Limestone Stoichiometry on SO2 Absorption (Cont'd)
CaCO3 + SO2 + 2H2O > CaSO3- 2H2O + CO2
One mole of CaCO3 reacts with one mole of SO2, to form a mole of Calcium Sulfite
(CaSOs* 2H2O). Thus, a 100% stoichiometric limestone requirement means one mole
of CaCOs for each mole of SO2. The 150% stoichiometric requirement means 1. 5
mole of CaCO3 for each mole of SO2. Theoretically, higher the Stoichiometry,
higher the SO2 absorption. Table X summarizes the data showing the effect of
Stoichiometry on absorption of SO2. The data clearly shows that SO2 absorption
efficiency does increase with increase in Stoichiometry. The increase in SO2 absorp-
tion efficiency is not significant enough, compared with the increase in Stoichiometry.
Effect of Delay Tank on SC>2 Absorption Efficiency
As a result of SO2 absorption using limestone, calcium sulfite is form ed. This cal-
cium sulfite is presumed to form an inert coating on the limestone particle, thus mak-
ing limestone unavailable for further reaction. The purpose of the delay tank is to
remove this inert coating, and thus make more limestone available for absorption of
SO2. A delay tank has also been used in Howden - ICI process, to minimize and con-
trol scaling problems by holding the slurry leaving the scrubber for a period long
enough for the supersaturation to dissipate on the surface of suspended crystals.
During the pilot plant work, a study was undertaken to determine the effectiveness of
the delay tank in improving the SO2 absorption efficiency. The delay tank was included
either in the first stage recycle liquor loop, or second stage recycle liquor loop, or in
352
-------
Chemical Construction Corporation
Pollution Control Division
Effect of Delay Tank on SO? Absorption Efficiency (Cont'd)
both recycle liquor loops.
In the case of delay tank in the second recycle loop, the slurry leaving the scrub-
ber, and make up limestone slurry were held in the delay tank. The liquor from
the delay tank was sent to the scrubber for absorption of SO2. The bleed from the
second stage delay tank liquor, was sent to the suction side of first stage recycle
liquor pump. In the case of a delay tank in the first stage recycle loop, the bleed
from second stage, and the slurry leaving the scrubber, were held in the delay
tank.
Table XI summarizes the data showing the effect of delay time on SC>2 removal effi-
ciency increases with increase in delay time. In one case, the SC>2 removal efficiency
increased by as much as 19 points, whereas in the other cases, the increase was about
12 points.
The recycle liquor pH increased with the increase in delay time. In one case, the
pH increased from 4. 9 to 6. 5, and in the other cases it increased from 5. 2 to 5. 9.
The use of a delay tank was found to be more effective in the first stage recycle liquor
loop, as any percentage point increase in SC>2 removal efficiency is a full point increase
353
-------
Chemical Construction Corporation
Pollution Control Division
Effect of Delay Tank on SO9 Absorption Efficiency (Cont'd)
in efficiency, whereas in the case of a second stage loop, it is only a fraction of
a point. The use of delay tank in both the recycle liquor loops di d not produce
any significant increase in SO2 removal efficiency over the use of delay tank in
the first stage recycle liquor loop.
Overall Summary - SO2 Removal Using Limestone
The quality of limestone does have a marked influence on SC>2 removal efficiency,
with increase in stoichiometry, use.of a delay tank, use of 325 mesh ground lime-
stone, and with a high quality limestone, SC>2 removal efficiency as high as 80-85%
can be attained.
SC>2 Removal Using Dolomitic Limestone and Dolomitic Lime
The dolomitic limestone contains carbonates of both calcium and magnesium. The
composition of the dolomatic limestone is presented in Table V. A maximum SO2
removal efficiency of 30% was attained in a single stage venturi scrubber system.
The pH values for the feed slurry were about 8. 0 and for the recycle liquor between
4. 8 and 5. 5.
The composition of the dolomitic lime is given in Table V. With the use of dolomitic
lime in a single stage scrubber system, the maximum SO2 removal efficiency attained
was 92% (1800 ppm inlet) with 130% stoichiometric lime requirement. The presence
of magnesium oxide make dolomitic lime much more reactive.
354
-------
Chemical Construction Corporation
Pollution Control Division
SC>2 Removal Using Dolomitic Limestone and Dolomitic Lime (Cont'd)
It is our opinion, that dolomitic lime or limestone should not be used as an absorb-
ing reagent in the case of a throw away process, as the soluble magnesium-sulfur
compounds will result in water pollution problems.
SQr> Removal Using Lime
Five different types of limes have been evaluated to determine SO2 removal efficiency.
The compositions of these five limes are presented in Table IV. The summary of
the data showing the effect of stoichiometry, the effects of single stage and two stage
operations and effects of modified single and two stage operations on SC>2 removal
efficiency, is presented in Table XII. The data clearly shows:
(a) That increasing the stoichiometry, increases the SC>2 removal efficiency.
(b) That with the use of a two stage scrubber system a better utilization of
lime can be achieved. At 100% stoichiometry, the SC>2 removal efficiencies
for a single stage and two stage operations are respectively 60% and 83%.
(c) That the modified single or two stage systems provide still better utilization
of lime, as can be seen from the following:
% Stoichiometry % Efficiency
Single Stage 100 60.0
Modified Single Stage 100 72.0
Two Stage 100 83.0
Modified Two Stage 100 88.0
(d) That with the use of lime as an absorbing agent, adequate SC>2 removal effi-
ciency can be attained to reduce outlet SC>2 concentrations less than 150 ppm.
355
-------
Chemical Construction Corporation
Pollution Control Division
SO2 Removal Using Carbide Sludge
Carbide sludge is produced during the manufacture of acetylene using calcium car-
bide.
CaC2 + 2H2O • > C2H2 + Ca(OH)2
The composition of the carbide sludge tested in the pilot scale is presented in Table
V. The data for the single and two stage operations, showing the effects of stoichio-
metry, effects of modified one and two stage operations and effects of delay tank are
summarized in Table Xin. The data clearly shows:
(a) That increasing the stoichiometry, increases the SO2 removal efficiency.
(b) That use of two stage operation provides higher SO2 removal efficiency than
single stage operation.
(c) That modified system provides improved performance.
(d) That use of delay tank further improves SO2 removal efficiency, and
(e) That with the use of carbide sludge as an absorbing agent, adequate SO2
removal efficiency can be attained to reduce outlet SO2 concentrations below
150 ppm.
Thickener and Filter Tests
During every pilot plant operation, equipment vendors performed necessary tests to
determine thickener and/or filter performance. This was essential so that the vendors
can supply suitable equipment with guaranteed performance for the commercial plant.
Thickener underflow concentrations in the range of 40 - 60% solids can be attained
356
-------
Chemical Construction Corporation
Pollution Control Division
Pilot Plant Work (Cont'd)
with clean overflow. In some cases use of flocculants was found necessary. With
the use of a suitable vacuum filter, filter cake containing 60% - 80% solids can be
obtained. The cake characteristics were studied and were found suitable in most
cases for truck transportation.
Open Loop vs. Closed Loop Operations
In the open loop operation, make up lime or limestone slurry was added to the scrub
ber system, and the bleed from the scrubber system was discharged to the disposal
area. In the case of partially closed loop operation, the scrubber bleed was sent to
a thickener, the thickener overflow was returned to scrubber and the underflow was
sent to disposal. In the case of completely closed loop operation, the thickener
underflow was sent to a filter, the filtrate was returned to the scrubber, and filtered
cake was sent to disposal.
Essentially the same SO2 removal efficiency was obtained during open and closed loo
type operations. During closed loop operations, except for some scaling and build u
no noticeable difference in operation was observed.
Scaling and Build-Up Problems
The use of lime or limestone slurry as the absorbing agent results in a scrubbing li-
quor containing calcium sulfate, calcium sulfite and unreacted calcium carbonate or
calcium hydroxide. The recycle liquor being a slurry, the clear solution is always
357
-------
Chemical Construction Corporation
Pollution Control Division
Scaling and Build-Up Problems (Cont'd)
a saturated solution, containing CaSO3, CaSC>4 and CaCOg. The mechanism for
SO2 absorption and precipitation of sulfite or sulfate is believed to be as follows:
S02(g) + H20(D / H2S03(i)
CaC03som + H2SO3— '-CaSOgscHn + H2O + CO2
CaSO3 excess - '- CaSO3 saturated + CaSO3 ppt.
CaCOs solid - ' CaCO3soln
Sulfur dioxide from the flue gas dissolves in water to form sulfurous acid, which reacts
with calcium carbonate in solution to form calcium sulfite. This causes a super sat-
uration in terms of calcium sulfite, resulting in precipitation of calcium sulfite. Some
of the calcium carbonate from the solid phase dissolves in the solution thus causing
it to reach saturation. Any oxidation of sulfite to sulfate causes sulfate to precipitate
out. The inverse solubilities of both calcium sulfite and calcium sulfate, with in-
creasing temperature, does enhence the precipitation of both compounds. This pre-
cipitation occurs within the scrubber on various surfaces, causing the scaling and build-
up problems.
The basic scaling characteristics of solutions containing calcium sulfite and calcium
sulfate, are inherent in nature. Scaling and build-up will occur, regardless of the
type of scrubber-absorber selected, but how fast and how much will depend on type
of scrubber and method of operation. The proper point of introduction of lime or
limestone in the scrubber recycle liquor loop, the presence of delay tank, and selection
of scrubbing equipment with minimum internals, should minimize scaling and buildup
358
-------
Chemical Construction Corporation
Pollution Control Division
Scaling and Build-Up Problems (Cont'd)
problems. Based on pilot plant work done, it is our opinion that scaling and buildup
problems will be more pronounced during a closed loop operation than during par-
tially open or open loop operations. We are of the opinion that if and whenever poss-
ible, the slurry of calcium - sulfur compounds should be kept outside the path of gas
flow, to attain highest possible reliability for boiler operation, to generate electricity.
Summary
The limestone quality seems to vary from one region to another region, and has sig-
nificant effect on SC>2 absorption. Smaller particle size and higher stoichimetry help
increase SC>2 absorption efficiency. The presence of a delay tank increases effi-
ciency or help reduce limestone requirements for the same efficiency. With a suit-
able combination of limestone quality, particle size, reasonable stoichiometry and dela
time, SC>2 removal efficiency as high as 80-85% can be attained. With the use of lime
or carbide sludge as the absorbing agent, one can attain 90+% efficiency at 130% or less
stoichiometric lime requirement. Selection of alkali should be made after proper econ
omic evaluation which should include stoichiometric requirement, alkali cost, grinding
and preparation costs, disposal cost and installed equipment cost.
Scaling and build-up problems are inherent when using lime or limestone slurry as ab-
sorbing liquor. Proper methods of operation, and selection of scrubber-absorption
equipment, may help reduce the frequency and extent of these problems.
359
-------
Table I. Summary of Boiler Operating Data
(1) (2) (3) (4) (5) (6) (7) (8)
Boiler Type Balanced Balanced Balanced Cyclone Balanced Balanced Balanced Balanced
Draft Draft Draft Draft Draft Draft Draft
Fuel Coal Coal Coal Coal Coal Coal Coal Coal
% Ash 22-23 20-25 20-25 20-25 16.8-28.9 16.3-17.8 9-11 10.3-12.4
S
° % Sulfur 0. 68-0. 71 5. 5 - 7. 3 5. 5 - 7. 3 5. 5 - 7. 3 3. 2 - 5. 5 2.1 - 2. 3 2. 5 - 2. 8 3. 3 - 4.1
Inlet SO2, ppm 450-575 4000-5200 4000-5200 4000-5200 1500-3800 1100-1600 1500-2200 1700-2250
Inlet Dust, Gr/SCFD 3. 9 -11. 7 4. 3 - 8. 3 4. 3 - 8. 3 0. 85 - 2. 5 4. 0 - 10. 0 3. 0 - 6. 0 3. 0 - 4. 5 3. 0 - 5. 0
-------
Table II Summary of Pilot Plant Facility at Various Power Plants
(1) (2) (3) (4) (5) (6) (7) (8)
Pilot Plant Facility
I. D. Fan V V / \ V \S
Single Stage \X , I/- ^
Two Stage V * x' « x ^" ^^
Recycle Pumps ^- ,
Delay Tank . ,-
Thickener ^
Filter
Closed Loop ^
Open Loop \X' V-'" W" 'vX' ^
Slaking System v ' \^x
Lime Slurry Tank \x' vX" ^^
Limestone Slurry Tank
•
'Alkali Used
Limestone
Mesh Size
-325 vX" VX" ux v^-- v --' V
-200 v "' Yx" W-'" \ t•-•'' V-
Lime
Dolomite
Calcined Dolomite
Carbide Sludge
-------
TABLE III - SUMMARY OF LIMESTONE COMPOSITIONS
(1) (2)
Si02 4.26
Fe2°3 °'63
A1203 0.36
CaC03 95-98.25 91.84
MgC03 2.96
R2°3
CO C
JET O
ro
(3)
5.74
0.80
0.53
88.98
4.45
(4)
7.51
0.90
0.60
85.78
5.12
(5)
6.13
0.52
0.61
91.67
1.08
(6)
6.07
0.55
0.48
91.72
1.05
(7)
3.75
0.58
0.24
94.22
1.78
0.82
0.068
(8)
4.84
0.92
0.38
88.56
5.18
1.30
0.07
(9)
7.07
0.92
0.44
86.71
3.78
1.36
0.06
(10)
1.60
0.09
0.28
96.81
0.80
0.06
(11) (12)
0.
0.
0.
95
3.
0.
82
16
18
.21 99.3
51
34
(13) (14)
86.8 93.4
P 0.03-0-1
Na-Kpxides 0.2-0-22
LOI 41.50 40.56 40.20 40.83 40.84 41.98 41.73 40.56 42.93
Mesh-325 96 96 96 96 96 96 96 96 99+% 90+ 90+ 90+
-200 98 98 98 98 98 98 98 98 90+ 90+ 90+
-------
TABLE IV - SUMMARY OF LIME COMPOSITIONS
(1) (2) (3) (4) (5)
Cao
Mgo
Fe20 ]
A120|!
*2o3
LOI
77.3
0.2
21.5
1.0
85.58
0.52
0.77
0.11
0.15
0.26
0.014
12.75
79.82
0.23
1.09
0.07)
0.12)
0.19
0.01
18.95
95.0
1.08
1.3
0.95)
0.03
1.75
95.0
1.08
1.3
0.95
0.03
1.75
363
-------
TABLE V - COMPOSITIONS OF DOLOMATIC LIMESTONE & LIME & CARBIDE SLUDGE
Dolomatic Limestone Dolomatic Lime Carbide Sludge
Cao 60 57 Ca(OH)2 92.5
Mgo 40 43 CaC03 1.85
97% thru 83% thru Sio2 1.50
325 mesh 200 mesh MgO 0.70
1.60
364
-------
Chemical Construction Corporation
Pollution Control Division
Table VI Summary Of
Analytical Data
Flue Gas Fly Ash Coal Liquor Analysis
Dust loading, gr/SCFD Particle size Distribution Ultimate Analysis Slurry Analysis
SC>2 loading, ppm Mineral Analysis (as received &dry basis) % Solids
NOX, ppm P2°5 Moisture pH
Temp., °F SiC>2 Carbon
Pressure, " W. G. Fe2Os Hydrogen Filtrate Analysis
at Al2Os Nitrogen pH
1st stage inlet TiO2 Chlorine dissolved Solids
2nd stage inlet CaO Sulfur cl ppm
2nd stage outlet MgO Ash SO4 ppm
SOs Oxygen Ca ppm
K2O Mg ppm
Proximate Analysis SiO2 ppm
/ • j o j i. • \ SO2 ppm
(as received & dry basis) _, . A1, ,. ...
-,,_., J Phenol Alkalinity
% Moisture J
% Ash Total Alka inity
% Volatile Conductivity mmhos
% Fixed Carbon
BTU Dry Sollds
% Sulfur % S°2
% S04
365
-------
Chemical Construction Corporation
Pollution Control Division
Table VII variables and Their Ranges
Variables
Gas Veloctiy, ft/sec
Liquor to Gas Ratio, gallons/1000 ACFM
Limestone Stoichiometry
Limestone Mesh Size
Delay Time, Minutes
Inlet SO2 » ppm
Range
50 - 250
10 - 60
100 - 350
200 mesh and
325 mesh
10 - 60
450 - 5200
366
-------
TABLE VIII - EFFECT OF LIMESTONE QUALITY ON SO,, ABSORPTION
CO
CM
% CaCo3
91.84
88.98
85.78
91.67
91.84
88.98
85.78
91.84
88.98
85.78
99.3
86.8
93.4
99.3
86.8
93.4
99.3
86.8
93.4
99.3
86.8
93.4
Limestone
2
3
4
5
2
3
4
2
3
4
12
13
14
12
13
14
12
13
14
12
13
14
Mesh Size % Stoichiometry Inlet SO2 ppm
96% thru 150 4150-4250
- 325 mesh
96% thru 350 4050-4200
- 325 mesh
98% thru 300 4400
- 200 mesh
90% thru 150 1500
- 325 mesh
90% thru 250 1500
- 325 mesh
90% thru 150 1500
- 200 mesh
90% thru 250 1500
- 200 mesh
Efficiency
44.4
47.0
32.8
41.2
49.6
55.4
44.1
39.4
37.2
44.7
76.0
57.0
55.0
80.0
70.0
60.0
53.0
49.0
63.0
65.0
56.0
68.0
-------
Chemical Construction Corporation
Pollution Control Division
Table DC. Effect Of Particle Size On SO2 Absorption
Limestone
12
13
14
11
%CaCO3
99. 3
86. 8
93.4
95.21
91.84
88.98
85. 78
91.67
Mesh Size
90%+ thru 325
90%+ thru 200
90%+ thru 325
90%+ thru 200
90%+ thru 325
90%+ thru 200
90%+ thru 325
90%+ thru 200
96% thru 325
98% thru 200
96% thru 325
98% thru 200
96% thru 325
98% thru 200
96% thru 325
98% thru 200
Inlet SO2, ppm
1500
1500
1500
2900
4000
4200
4200
4200
%Efficiency
75.0
53.0
57. 0
49.0
55.0
62.0
62.0
54.0
44.4
26.1
47. 0
25.4
32. 8
30. 5
41. 5
34.0
368
-------
Chemical Construction Corporation
Pollution Control Division
TABLE X - EFFECT OF LIMESTONE STOICHIOMETRY ON SO,, ABSORPTION
% CaCo,
91.84
91.84
85.78
85.78
96.81
99.30
99.30
86.8
86.8
93.4
93.4
95.21
Limestone
2-325 mesh
2-200 mesh
4-325 mesh
4-200 mesh
10-325 mesh
12-200 mesh
12-325 mesh
13-200 mesh
13-325 mesh
14-200 mesh
14-325 mesh
11-325 mesh
% Stoichiometry
167
321
162
317
150
329
132
298
157
232
312
150
200
250
150
200
250
150
200
250
150
200
250
150
200
250
150
200
250
100
150
200
z
% SO? Efficiei
33.0
45.0
27.0
39.0
32.8
49.4
30.5
44.7
44.5
54.5
58.8
52.5
61.0
65.0
75.0
78.0
79.5
49.0
56.5
56.5
57.0
61.5
69.5
63.8
64.0
70.3
55.0
59.0
60.5
60.0
64.0
68.0
369
-------
TABLE XI - Effect of Delay Tank on SO? Removal
Delay Tank in l_st_Stage recycle liquor loop
Delay Time % SO? Removal Efficiency Recycle liquor pH Limestone
Tx = 0 min 46.0 4.9 3
T2 58.5 5.4
T3 65.4 6.5
T-L = 0 min 64.0 5.2 11
T2 74.0 5.6
T3 76.0 5.9
Effect of Delay Tank location on SO^ Removal Efficiency
Delay Tank Location Delay Time % SO? Removal Efficiency
1st stage T2 54.0
2nd stage T, 54.0
1st stage T3 65.0
2nd stage T~ 64.0
370
-------
TABLE XII - SO? Removal Using Lime
Lime
% Stoichiometry Inlet SO? ppm % Efficiency
100
140
220
100
100
130
145
225
100
130
100
130
150
100
140
1500
1500
1500
1500
1500
1500
1500
1500
1500
1500
4000-6400
4000-6400
4000-6400
4000-6400
4000-6400
83.0
94.0
95.0
88.0
60.0
64.0
71.0
77.0
72.0
88.0
66.0
75.0
80.0
80.0
89.0
Two Stage
scrubber
system
Modified Two
stage scrubber
system
Single stage
scrubber
system
Modified single
stage scrubber
system
Single stage
scrubber system
Two stage
scrubber system
371
-------
Chemical Construction Corporation
Pollution Control Division
Table Xin SC>2 Removal Using Carbide Sludge
% Stoichiometry
100
100
100
130
130
130
100
100
100
120
120
120
Single Stage
Modified Single Stage
Modified Single Stage
and Delay Tank
Single Stage
Modified Single Stage
Modified Single Stage
and Delay Tank
Two Stage
Modified Two Stage
Modified Two Stage
and Delay Tank
Two Stage
Modified Two Stage
Modified Two Stage
and Delay Tank
Inlet SO2 ppm
2500 -
2500 -
2500 -
2500 -
2500 -
2500 -
2500 -
2500 -
2500 -
2500 -
2500 -
2500 -
3000
3000
3000
3000
3000
3000
3000
3000
3000
3000
3000
3000
% Efficiency
62.0
69.0
72. 5
73
78
79. 5
74.0
78.0
92. 0
85. 0
92. 0
93. 0
372
-------
A PILOT PLANT TEST PROGRAM FOR
SULFUR DIOXIDE REMOVAL FROM
BOILER FLUE GASES USING LIMESTONE
AND HYDRATED LIME
AT
UNION CARBIDE CORPORATION
POWER STATION
MARIETTA, OHIO
Union Carbide Corporation
Dr. R. A. Person
Manager-Environmental Control
and Scope Engineering
Engineering Department
Ferroalloys Division
Niagara Falls, New York
Dr. C. R. Allenbach
Staff Engineer-
Environmental Control
Engineering Department
Ferroalloys Division
Niagara Falls, New York
Chemical Construction Corporation
I. S. Shah
Chief, Process Engineering and
Development
Pollution Control Division
New York, New York
S. J. Sawyer
Senior Process Engineer
Pollution- Control Division
New York, New York
Presented at
Second International Lime - Limestone
Wet Scrubbing Symposium
New Orleans, Louisiana
November 8 - 12,. 1971
373
-------
ABSTRACT
A pilot plant test program to determine the performance and
optimum operating conditions of a Chemico Venturi Scrubber System
for the simultaneous removal of fly ash and sulfur dioxide was
performed at Union Carbide Corporation's coal-fired power station
near Marietta, Ohio.
Hydrated lime and limestone were evaluated for sulfur dioxide
removal during the test program. Fly ash removal efficiency of the
venturi scrubber was determined simultaneously with sulfur dioxide
removal.
The test program was performed in three phases:
1. Fly ash scrubbing using water as the scrubbing media.
2. Sulfur dioxide absorption using a hydrated lime slurry.
3. Sulfur dioxide absorption using a limestone slurry.
This paper discusses the effect on sulfur dioxide absorption of
such operating variables as liquid to gas ratio, stoichiometry, number
of stages, and delay tank.
Based upon the results obtained from the pilot plant tests, it
was concluded that satisfactory technical performance using limestone
or hydrated lime in a two-stage venturi system could be achieved on
a full-scale plant. Estimated capital and operating costs for a full-
scale installation are presented.
374
-------
INTRODUCTION
Many coal-burning power generating plants are being required
to reduce particulate and sulfur dioxide emissions. Some approaches
being taken to control the particulate emission problem are wet
scrubbers, more efficient electrostatic precipitators and conversion
to oil. To decrease sulfur dioxide emissions, either the use of low
sulfur content fuels or various methods of desulfurizing the flue gases
can be utilized. Each approach poses many technical and economic
problems, and considerable effort is being expended to develop the
most acceptable solutions.
At U-nion Carbide's Marietta, Ohio power station, a captive high
ash and sulfur content coal has been burned since the start of opera-
tions in 1951. The station is equipped with mechanical collectors and
two-field electrostatic precipitators to control particulate emissions.
The Parkersburg, West Virginia/Marietta, Ohio Interstate
Abatement Conference issued specific recommendations in March and
April, 1970, calling for a 70% reduction of sulfur oxide emissions and
a decrease in fly ash emissions to a 0.14 Ib./mm Btu heat input level.
Compliance dates for these recommendations are April, 1972 for sulfur
oxides and March, 1973 for particulate. Proposed State of Ohio
regulations, now pending, are more stringent. .Particulate emissions
of 0.10 Ib./mm Btu heat input and sulfur dioxide emissions of 1.6 lb./
mm Btu heat input, equivalent to about a 1% sulfur coal (12,500 Btu/lb.),
are proposed for this size plant.
A survey of possible methods of reducing emissions pointed out
the desirability of evaluating the use of wet scrubbing, with a calcium
compound throw-away system, for control of both the particulate and
sulfur dioxide emissions. To evaluate such a system, a two-stage
venturi scrubber pilot plant supplied by Chemical Construction
Corporation was installed at the power station. Major aims of the
375
-------
program were to determine particulate removal capabilities of the
venturi scrubber and the sulfur dioxide removal efficiencies of
two alkaline materials—limestone and hydrated lime.
DESCRIPTION OF POWER STATION
The Union Carbide power station is a high-load factor station,
rated at 200 megawatts, located on the Ohio River near Marietta, Ohio.
This station supplies power to all units of Union Carbide's manufacturing
complex in Marietta. In addition to power, the station also provides
process steam to the complex. This additional requirement results in
a coal consumption equivalent to that of a 250 megawatt station.
The power station has four identical boilers, each currently
equipped with a cyclone-type mechanical collector and an electrostatic
precipitator. At full capacity, the flue gas quantity from each boiler
is approximately 280,000 ACFM at 320° F. entering the collectors.
The coal burned at the plant has been, until recently, Meigs #9
Seam from Noble County, Ohio, with a typical ash and sulfur content
of about 20% and 5%, respectively. Tests were made with this coal,
although currently higher quality blend coal is being used to comply
with interim emission reductions recommended by the Interstate
Abatement Conference.
OBJECTIVE OF TEST PROGRAM
The objective of the pilot plant test program was to confirm
engineering design parameters and develop design data for a commercial
installation. The key elements of the program were:
376
-------
To correlate particulate and sulfur dioxide removal
efficiencies with various scrubber operating conditions.
To establish a relationship between outlet dust loadings
and pressure drop across the venturi throat.
To determine sulfur dioxide absorption efficiencies using
limestone and hydrated lime, and to evaluate the effect
of the following variables:
Limestone
(a) particle size - nominal 200 mesh by Down and
325 mesh by Down
*(b) stoichiometric feed rate
(c) liquid to gas ratio
(d) delay tank
s
Hydrated Lime
(a) single-stage versus two-stage scrubbing
(b) stoichiometric feed rate
(c) liquid to gas ratio
(d) delay tank
4. To confirm the optimum design conditions using limestone
and observe the degree of scaling or depositing when
operating for an extended test period.
PILOT PLANT
A slip stream for the pilot plant was withdrawn after the air
preheater and before the mechanical collectors. Location and size
of the gas withdrawal duct were determined by traverses to assure
that an isokinetic flue gas sample was taken for the pilot plant.
The isokinetic condition corresponded to capacity operation of the
boiler and 1500 cfm saturated gas flow from the pilot plant.
377
-------
A flow sheet of the 1500 cfm two-stage pilot plant is shown
in Figure 1. Figures 2 and 3 illustrate the installation.
The alkali feed, for a two-stage configuration, is fed to the
second stage recycle loop with the second stage bleed going to the
first stage recycle loop. This gives an alkali flow countercurrent
to the flue gas producing the most efficient absorption system.
The delay tank is part of the first stage recycle loop. Its
purpose is to allow for additional reaction time and increased utilization
of the alkali by providing time for removal of reaction products that coat
the unreacted alkali. During two-stage testing, the second stage bleed
is also fed to the delay tank.
The main components of the pilot plant are two wet approach
venturi scrubber-separator units. The flue gas is drawn into the scrubber
at the top and the recycle liquid is fed through tangential nozzles. The
liquid creates a curtain across the venturi throat and the accelerated
flue gas breaks up the liquid into fine droplets which effectively interact
with the fly ash and sulfur dioxide.
The flue gases are disengaged from the slurry, in both separator
stages, by centrifugal and gravitational forces.
The recycle slurry is circulated by stainless steel pumps and all
recycle piping is of stainless steel construction. Magnetic rotameters
are used to measure the flow.
An induced draft fan is used to pull the stack gas through the
system. It is of stainless steel construction and is capable of
pulling 1800 cfm with a maximum pressure drop of 56 inches w.g.
378
-------
The other accessory equipment consists of:
(a) Two feed tanks of 5000 and 7000 gallon capacity
(b) A delay tank of 4500 gallon capacity
(c) Feed pumps with 20 gpm capacity
(d) Feed rotameters
(e) Stainless steel orifice plate and manometer, located
upstream of the fan, for flue gas flow measurement
(f) Delay tank pump identical to the recycle pumps
(g) Stack gas reheater
RESULTS
Fly Ash Scrubbing
Phase I of the test program was to determine particulate removal
when scrubbing with water. The throat velocity and liquid to gas ratios
were adjusted to vary the pressure drop across the throat over a wide
pressure range. The particulate sampling procedures employed were
the methods recommended by the Industrial Gas Cleaning Institute.
The results were as follows:
(a) At inlet dust loadings ranging from 5.2 to 9.6 grains/SCFD
and averaging 7-8 grains/SCFD, the outlet dust loadings
ranged from about 0.08 to 0.02 grains/SCFD over the range
of 1.5 to 18.5 inches w.g. pressure drop.
(b) Particulate sampling was continued during the sulfur dioxide
absorption program. The results showed that the outlet dust
loadings were comparable to those obtained when scrubbing
with water. During the two-stage scrubbing test work,
additional particulate removal occurred in the second stage.
379
-------
Sulfur Dioxide Absorption with Hydrated Lime
Phase II of the test program was to determif.j the sulfur dioxide
absorption capabilities of hydrated lime. The hydrated lime, a by-
product of the generation of acetylene from calcium carbide, was
obtained as a nominal 50% solids content sludge. This sludge, if
used on a commercial basis, would be 90-95% calcium hydroxide
on a dry basis. However, the sludge fed to the pilot plant was
only 80-85% calcium hydroxide, with 5-10% being calcium carbonate.
The tests performed were to determine the optimum operating
conditions, to remove 70% and 90% of the inlet sulfur dioxide, which
ranged during the program from 2000-3200 ppm sulfur dioxide on a
dry basis. Both single-stage and two-stage tests were performed.
The results of the single-stage tests were as follows:
(a) At stoichiometries ranging from 90% to 150%, based upon
inlet sulfur dioxide concentration, absorption efficiency,
or removal of sulfur dioxide, ranged from 60-80%.
(b) Use of a delay tank improved the absorption efficiency
at the lower stoichiometries but had less effect at the
higher hydrated lime feed rates.
(c) For the delay tank tests no improvements were obtained
for residence times beyond 25 minutes.
(d) Increasing the liquid to gas ratio from 20 to 60 gpm/1000
ACFM, significantly improved the sulfur dioxide removal
efficiency.
After completion of the single-stage tests, two-stage tests were performed,
In general, the effects of the various operating parameters were similar
to that of single-stage tests with all values of sulfur dioxide removal
about 10-20% higher. Values above 90% were obtained.
380
-------
The use of a delay tank was found more beneficial at the
lower stoichiometries just as it was for single-stage operation.
Higher reagent utilization was found to occur when the delay tank
was used, and therefore, a commercial design would be significantly
benefited by its use.
Sulfur Dioxide Absorption with Limestone
Phase III of the test program was to determine the absorption
capabilities of an available limestone analyzing 95% CaCO and
O
3.5% MgCO with the remainder being mineral inerts. This limestone,
O
from Carntown, Kentucky, is a material currently utilized in other
operations at Marietta and is normally received by barge. Only two-
stage tests were performed since past experience indicated that 70%
sulfur dioxide removal could not be obtained in a single-stage unit.
The results of the limestone tests were as follows:
(a) Increasing the stoichiometric feed rate from 130% to 230%
increased the removal efficiency about 10%.
(b) The effect of liquid to gas ratio on sulfur dioxide removal
efficiency was significant. Increasing the liquid to gas
ratio by 30% improved the sulfur dioxide removal efficiency
by better than 5%.
(c) The gain in sulfur dioxide removal efficiency was about
10% higher when using 325 mesh limestone than when using
200 mesh limestone.
(d) The use of a delay tank increased sulfur dioxide removal
by as much as 10%.
(e) Values in excess of 75% sulfur dioxide removal were
obtained using a delay tank and 325 mesh limestone.
381
-------
Extended Limestone Test
The final portion of the program was to operate the pilot plant
continuously for three days using 325 mesh limestone. The system
was run in an open loop configuration with no attempt made to recover
the bleed stream liquid. The results of the tests were as follows:
(a) A removal efficiency of better than 70% for the optimum
design conditions selected was confirmed.
(b) The unit operated smoothly and continuously during the
entire test period. No changes in the venturi or separator
pressure drops were observed.
(c) Inspection of the piping, rotameters, pumps and tanks
showed that scale formation or depositing was insignificant.
SUMMARY OF RESULTS
The results of the test program for the Meigs Seam coal and the
specific removal reagents are summarized below:
1. Fly ash levels down to 0.02 gr/SCFD were obtained.
This value is well below the Interstate Abatement Conference
Recommendations and the proposed State of Ohio Regulations.
2. A sulfur dioxide removal efficiency of 70%, the requirement of
the Interstate Abatement Conference, was readily attainable
with a single-stage scrubber using hvdrated lime or a two-
stage unit using limestone.
3. A sulfur dioxide removal efficiency of 90%, which would
exceed the proposed State of Ohio Regulations, was attained
using hydrated lime in a two-stage scrubber system.
4. A continuously operated three-day test using limestone
developed no scaling or depositing problems.
382
-------
PROJECTED ECONOMICS OF FULL-SCALE PARTICULATE
AND SULFUR DIOXIDE SCRUBBING SYSTEM
It was concluded from the pilot plant tests that satisfactory
technical performance using limestone or hydrated lime in a two-stage
venturi system could be achieved on a full-scale plant. Studies for
the projected full-scale installation were based upon an individual
scrubbing train for each boiler in order to assure maximum reliability
and on-time performance. Predicted removal efficiencies for the
full-scale system were 99%"for particulate, 90% for sulfur dioxide
with hydrated lime, and 70% for Carntown limestone at a system
pressure drop of 20 inches w.g.
Design and engineering work by Chemical Construction Corporation
under way since January, 1971 and updated for information obtained
from the pilot plant program, yielded estimated costs for a complete
installation utilizing a two-stage hydrated lime system as follows:
Capital Costs
Initial Scrubbing System on One Boiler $ 3,900,000
Settling Basin for Scrubber Effluent 900,000
Hydrated Lime Materials Handling System 800,000
Subtotal $ 5,600,000
Three Additional Scrubbing Systems 9,000,000
Subtotal $14,600,000
Development Costs
First Scrubbing System $ 1,500,000
Remaining Scrubbing Systems 1,500,000
Subtotal $ 3,000,000
Capital Plus Development Costs Total $17,600,000
383
-------
Annual Operating Costs
One Scrubbing System
Energy $ 80,000
Maintenance 150,000
Absorbent (Hydrated Lime) for
90% Sulfur Dioxide Removal 250,000
Total $ 480,000
Four Scrubbing Systems Total $1,920,000
The preceding cost estimates included the following specific
items for each scrubbing system:
1. Chemico integral venturi scrubbers-mist eliminator units,
with each stage a separate vessel.
2. -Delay tank.
3. Flue gas reheater using process steam.
4. Steam turbine-driven fan to supply system pressure drop.
5. Recycle of pond water to system.
The estimates include connection into an exhaust stack, but not the
stack itself.
-------
Absorbent Cost Comparison
The following table shows a material cost comparison of hydrated
lime vs. limestone for a scrubbing system on one boiler burning 250,000
tons of 5% sulfur content-coal per year.
Basis: Total Sulfur - 12,500 tons/year
Total SO
- 25,000 tons/year
70% Removal of Total SO - 17,500 tons/year
90% Removal of Total SO - 22,500 tons/year
70% Sulfur Dioxide Removal
Hydrated Lime
(50% solids)
Limestone
Tons
Required
58,000
58,000
53,360
53,360
Cost Aon, $
3.40 (1)
3.90 (2)
4.10 (3)
4.30 (4)
90%
Total Cost/Pound of
Cost Cost/Ton of Sulfur Dioxide
$ Coal Burned, $ Absorbed, £
197,
226,
218,
229,
200 0
200 0
776 0
448 0
Sulfur Dioxide
.79
.91
.88
.92
Removal
0.56
0.65
0.63
0.66
Hydrated Lime 70,900
3.40
241,060
0.96
0.53
Cost Assumptions:
(1) Hydrated Lime @ $1.00/ton
(2) Hydrated Lime @ $1.50/ton
(3) Limestone @ $1.70/ton and grinding @ $0.80/ton
(4) Limestone @ $1.70/ton and grinding @ $1.00/ton
At the most favorable material costs for 70% sulfur dioxide removal, hydrated
lime has about a 10% lower cost advantage.
A salient feature of hydrated lime favoring its selection over limestone is
that about 55% as much absorbent solids and 65% as much absorbent -
sulfur dioxide reaction product solids are handled in the system. Less
waste product storage capacity would be required.
385
-------
ACTION TAKEN. BY FERROALLOYS DIVISION OF UNION CARBIDE
Concurrent with the activity discussed in this presentation,
changing business conditions indicated that a high quality (10% ash,
0.8% sulfur) captive coal supply could be made available for the
Marietta power station. Utilization of this coal, together with
precipitator upgrading, has been determined to be the minimum capital
solution for reducing Marietta power station emissions to meet existing
recommendations and proposed regulations.
386
-------
Flue Gas
f
Sample Point
T7
Venturi
I
Meter
00
00
XXX
Bleed
Sample! Point
Separator
Pump
Tank
Delay
Tank
HXl
4=iventuti
T7
I
Meter
Sepa-
rator
—txj-
Pump
Tank
3 MX
Reagent —.
Water
IMeter
Feed Tank
Cleaned
Flue Gas
Sample Point
+-O
Orifice Reheater
Plate
Fan
Reagent
Water
I
Meter
Feed Tank
-IX]
-CXJ
FIGURE 1
TWO-STAGE VENTURI SCRUBBER SYSTEM
SCHEMATIC FLOW SHEET
-------
Figure 2
ENTRIES TO VENTURI SECTIONS - MARIETTA PILOT INSTALLATION
-------
FIGURE 3
TWO-STAGE VENTURI ':'\rSTEM - MARIETTA PILOT INSTALLATION
389
-------
LIMESTONE SCRUBBING EFFICIENCY OF
SULFUR DIOXIDE IN A WETTED FILM
PACKED TOWER IN SERIES WITH A
VENTURI SCRUBBER
By
R. J. Gleason
COTTRELL ENVIRONMENTAL SYSTEMS, INC.
A Division of Research-Cottrell
Bound Brook, New Jersey
Prepared For:
Presentation at Second International
Lime/Limestone-Wet Scrubbing Symposium
New Orleans, Louisiana
November 8-12, 1971
391
-------
ACKNOWLEDGEMENT S
The work upon which this publication is based was
performed pursuant to Contract EHS-D-71-24 with the
Environmental Protection Agency.
Experiments were carried out at the Tidd Power Station
of Ohio Power, a subsidiary of American Electric Power. The
cooperation of AEP and the Tidd Power Station personnel
played a key part in executing this study.
Part of the work reported here was conceived and per-
formed by the Tennessee Valley Authority personnel. The
progress resulting from the limestone tests is due, in part,
to TVA participation.
392
-------
LIMESTONE SCRUBBING EFFICIENCY OF
SULFUR DIOXIDE IN A WETTED FILM
PACKED TOWER IN SERIES WITH A
VENTURI SCRUBBER
Sulfur dioxide removal from a power plant flue gas
was studied in a 1,000 CFM pilot plant containing a
packed tower and venturi scrubber combination.
Absorption efficiencies were determined for a range
of SO~ concentrations in the gas-phase and limestone
concentrations in the liquid slurry. Results of the
absorption tests indicate removal efficiencies to
be strongly affected by the inlet SO- concentration
and the limestone stoichiometry. Reaction products
deposition in the absorption tower were determined
for various operating modes and/or conditions.
Scaling of the tower absorber was controlled by
maintaining a high liquid-to-gas ratio and high
slurry concentration in the circulating liquor.
393
-------
I. INTRODUCTION
Control of power plant sulfur dioxide emission has
been under study at Research-Cottrell since the early 1960's.
Initially, R-C investigators were concerned with the absorp-
tion characteristics of the Flooded Disc Scrubber (FDS) , a
venturi-type wet scrubber. Results of this experiment work
demonstrated a low degree of SO2 removal (40 to 60%) when
lime was employed as a scrubbing agent. Scaling within the
scrubber and auxiliary equipment was difficult to control
during maximum absorption conditions. To achieve greater
sulfur dioxide removal and controlled scaling/ other conven-
tional tower absorbers were screened. Bench-scale studies of
select packing types showed that a particular wetted film
packing with exceptionally low pressure drop characteristics
and high specific surface was well suited for SO- scrubbing
systems. In 1969, a 1,000 CFM pilot plant consisting of a
FDS in series with a wetted-film packed tower was installed
in Ohio Power's Tidd Plant. The results of earlier bench
work and the status of the subsequent pilot plant program
were reported in March, 1970 at the First International
1»2
Symposium on Wet Limestone Scrubbing.
Since the First Limestone-Wet Scrubbing Symposium,
Cottrell Environmental Systems was contracted by EPA to
execute a pilot plant study at the Tidd Power Station in-
volving both lime and limestone. One objective of this pro-
gram was the evaluation of the wetted-film packed tower
absorption capabilities with limestone wet scrubbing.
The process conditions for the limestone tests were
based upon bench scale studies performed by the Chemical
394
-------
Development Division of the Tennessee Valley Authority,
Muscle Shoals, Alabama. The absorption system simulated,
to some respect, the Howden-ICI actual plant operation
of the Fulham Power Station, London.
Test results reported in this paper were previously
presented at the 69th National Meeting of the American
Institute of Chemical Engineers, Cincinnati, Ohio, May
16-19, 1971.
395
-------
II. EQUIPMENT
A layout of the pilot plant system is given in Figure
II-l. The absorption section of the pilot plant contained a
Flooded Disc Scrubber (venturi scrubber) in series with a
packed tower. The flue gas, containing both particulates and
sulfur dioxide, passed first through the venturi where the
entering gas quickly cooled to its dew point (#120°F). Flue
gas was removed isokinetically from the suction of the power
station I.D. Fan.* Scrubbing solution or slurry entering the
absorber tangentially above the throat flowed cocurrently
through the venturi and into a cyclonic demister. The fly-ash-
stripped gas then passed vertically through a conical hat
gas/liquid splitter before entering the packed tower.
Slurry flow rates to the scrubber and other process units
were measured by venturi-type flow meters. In-line pH probes
were installed at the discharge of each scrubber. Also,
immersion-type pH elements were placed in the clarifier and the
hold tank.
The venturi scrubber is a cocurrent absorber with a
variable throat orifice. Pressure drop across the scrubber
can be varied by adjusting the disc position within a venturi
throat. Figure II-2.
The packed tower is a countercurrent absorption device
containing a packing with low pressure drop characteristics
and high specific surface (68 sq.ft./cu.ft.). The packing,
a rigid material fabricated from corrugated sheets of asbestos
* Flue sampling contained particulate concentration of
approximately 4.0 gr/scf from a pulverized coal combustion,
396
-------
V-VENTURI SCRUBBER
D-DEMISTER
A-ABSORBER
C-CLARIFIER
H-HOLDING TANK
M-MIXING TANK
B-BULK FEEDER
GIV-GAS INLET VENTURI
GOV-GAS OUTLET VENTURI
GOO-GAS OUTLET ORIFICE
VV -VENTURI VENTURI
AV -ABSORBER VENTURI
CV -CLARIFIER VENTURI
VBV-VENTURI BY-PASS VENTURI
ABV-ABSORBER BY-PASS VENTURI
VP -VENTURI PUMP
AP -ABSORBER PUMP
-CLARIFIER PUMP
-HOLDING TANK PUMP
-MIXING TANK PUMP
-HAND VALVE
CP
HP
MP
FIG.U-1 PILOT PLANT LAYOUT
-------
GAS FLOW
HEIGHT
13
VARIABLE
LIQUID
INLET
r CONDUIT
/LIQUID DISENGAGES
/ FROM WALL
FIG.E-2 DIMENSIONS FOR FLOODED DISC SCRUBBER
398
-------
coated with neoprene, was five feet in height and sixteen
inches in diameter. Pressure drop across the packing at gas
velocities between 8 and 10 feet per second is approximately
one inch of water with well-irrigated conditions.
Flooding characteristics of this packing were approxi-
mated by using the Eckerts flooding line correlation and a
i*
packing factor of 12.0. The calculated flooding profile of
the wetted film packing are compared to other high capacity
packing in Figure II-3. As illustrated graphically, the wetted-
film packing allows significantly greater liquid rates than
either 3 inch plastic Pall rings or 3h inch Intalox saddles.
PROCESS EQUIPMENT
Process Units
Size
Material Of
Construction
FDS Scrubber
Packed Tower
Tower Slurry
Tank
Mixing Tank
Venturi Slurry
Tank
Cyclonic
Demister
Clarif ier
Blower
6" to 8" diameter
See Figure 2.
16" diameter x 5'
1,500 gallons agitated
1,500 gallons agitated
55 gallons
4 ' diameter x 4 '
1,200 gallons
1,000 cfm at Ap = 40
inches H2O
316 SS
316 SS
CS
CS
CS
SS
CS
CS
399
-------
CALCULATED FLOODING MASS VELOCITIES
FOR THE WETTED-FIU1 AND OTHER HIGH CAPACITY PACKING
ji I
102
103
GAS MASS VELOCITY, LB/(HR)(SQ,FT.)
10*
400
-------
III. RESULTS
The limestone programs were carried out in two separate
test series: several types of limestone were first processed
in an open-loop system and absorption efficiencies were com-
pared for each alkali type. In the second program, scale
accumulation within the tower absorber was investigated
using a limestone selected from the first test result. For
the scaling runs, a closed-loop operating mode was used to
minimize the process makeup water.
A. EFFICIENCY TESTS - OPEN-LOOP
Calcium carbonate materials (chalk, cement kiln dust,
and two limestone grinds) were processed under similar opera-
ting conditions. Chalk (89%-200 mesh) and finely ground lime-
stone (89%-325 mesh) allowed approximately the same absorption
i.e. 96% removal. Limestone ground to 75%-200 mesh gave an
efficiency between 81 to 88% while the cement dust showed the
poorest results with 73% absorption. A chemical analysis for
each of these materials is given in Table III-l.
Process modes selected for the pilot tests utilized the
packed tower as the main contacting device. The venturi
scrubber removed fly ash and a small portion of the inlet SO2
by operating at a moderate pressure drop (6 to 7 inches of
water). Fresh limestone slurry was fed first to the tower
circulating tank where the bulk of the absorption took place.
For maximum calcium carbonate utilization, the reaction
products and the residual limestone were bled from the tower
slurry tank and passed through a venturi circulation tank.
The input limestone and the bleed were balanced so that a
constant calcium concentration remained in the tower hold tank
401
-------
TABLE III-l
CALCIUM CARBONATE MATERIALS RECEIVED FROM TVA
Type
Tiftona
Limestone
Tiftona
Limestone
Cement
Kiln Dust
Selma
Chalk
Particle
Size
75%-200M
89%-325M
90%-200M
89%-200M
Amount
Received
Ibs.
4700
350
1000
1000
CaO
50.5
50.5
41.5
43.1
MgO
1.5
1.5
2.4
0.59
K00
0.4
0.4
3.1
0.49
Na.,0
0.3
0.3
0.19
0.18
Ign.
Loss
41.5
41.4
22.8
35.1
402
-------
Block diagrams illustrating the operating modes are
shown in Figure III-l. A limestone slurry containing 2% by
weight calcium carbonate was pumped to the Tower Slurry
Tank from the Mixing Tank at a flow equivalent to the SO2
stoichiometry desired.
For six of the seven efficiency tests, the stoichio-
metric input was between 100 and 110%.* One run was performed
at 160% of the equivalent S02 to determine crudely the lime-
stone stoichiometry effects. The sulfur dioxide concentration
to the FDS was between 1,135 and 1,980 ppm. Although no
particulate measurements were made, the normal fly-ash loading
was from 3 to 4 grains per SCF.
The tower gas velocity for these tests was between 9.3
and 10.7 ft/sec. Liquid flows for most of the tests were
maintained at 40 gallons per 1000 cu.ft.. A single test
operated at L/G = 20 gallons per 1000 cu.ft.. Pressure drop
on the tower was around 1.0 inches of water for most tests.
Tests with chalk and cement were operated at slightly higher
tower pressure differentials.
Operating results for these experiments are given in
Table III-2. Each efficiency measurement listed is an average
of four readings over a three-hour period. Liquid-to-gas
ratio in the tower had considerable influence on SO2 reduction.
At L/G of 40 gallons per 1000 cf and a coarse grind limestone,
the efficiency across the tower was 81.6%; however, at L/G of
20 for the same operating mode, the SO2 removal dropped to
58.2%.
* Based on total SO- feed to system.
403
-------
GAS IN
^•SLURRY
RECYCLE
VENTURI
SLURRY
TANK
"r
GAS OUT
PACKED
TOWER
DEMISTER
1
^-SLURRY
RECYCLE
CAC03
WATER
TOWER
SLURRY
TANK
Z
MIXING
TANK
SLURRY FEED
COUNTER-CURRENT SLURRY FLOW
WATER
GAS IN
4-
GAS OUT
PACKED
TOWER
DEMISTER
t ,r
TOWER
SLURRY
TANK
1
^SLURRY
'RECYCLE
WATER
MIXING
TANK
WATER FEED TO FDS
FIG, IBM OPERATING MOOES USED FOR LIMESTONE EFFICIENCY TESTS
404
-------
TABLE III-2
O
01
Material Used
Test No.
Tower Liquid Rate, GPM
FDS Liquid Rate, GPM
Gas Velocity, Tower, Ft/Sec.
Gas Velocity, FDS, Ft/Sec.
Tower Pressure Drop, inches H-0
FDS Pressure Drop, inches H«O
CaO/SO2 Ratio
SO- Concentrations, PPM
FDS in
FDS out
Tower out
Fraction of SO2 Removed, %
FDS
Tower
Overall
Gas Temperature, °F
FDS in
FDS out
Tower out
Liquid Temperature, °F
FDS in
FDS out
Tower in
Tower out
L/G ratio, FDS, Gal. per 1000 cf
L/G ratio, Tower, Gal. per 1000 cf
1
iUMMARY DATA SHEET FOR THE OPEN-LOOP EFFICIENCY TESTS
Limestone
Selma
Chalk
A2
35
9.4
10.7
99
3.9
12.7
1.11
1550
1405
48
9.4
96.6
96.9
371
122
114
119
126
112
122
10.4
39
Cement
Dust
A3
36
9
10.7
105
6.5
6.4
1.06
1135
1025
268
9.7
73.5
76.4
377
122
116
121
124
114
122
10
40
Fine
A4
24
6
9.3
132
1.1
7.3
1.16
1650
1288
62
21.9
95.1
.96.1
361
118
110
118
125
110
117
10
40
Coarse
AS
24
6
9.3
132
0.9
7.3
1.00
1980
1795
209
9.4
88.4
89.3
361
116
109
116
124
110
116
10
40
Coarse
A6
3i
8
9.6
121
1.2
7.1
1.00
1467
1155
212
21.2 2
81.6
85.5
348
92
91
40
75
89
124
10
40
Coarse
A7
18
9
10.7
98
1.3
7.9
1.10
1637
1415
592
13. 62
58.2
63.8
391
102
99
40
97
95
106
10
20
Coarse
AS
32
8
9.6
87
1.3
9.4
1.62
1485
1402
199
5.6
85.8
86.6
369
101
98
41
95
96
101
10
40
1. Operating conditions for each task shown are an average of four readingsmeasured over a
three hour period.
2. Water was fed to the FDS for this test.
-------
B. SCALING EXPERIMENTS
During the limestone efficiency tests, some scaling
did occur in the packed tower after the chalk/ cement dust,
and limestone were used as sorbate. Limestone alone did not
show appreciable scale accumulation after 40 hours of contact.
To determine definitively the scaling characteristics of the
limestone/SO- absorption, a second test program was undertaken.
The operating conditions for minimum scaling were studied
in four continuous 40-hour tests. Process parameters such as
liquid-to-gas ratio, tower slurry tank residence time, and
tower slurry tank temperature were varied. Scale deposition
was measured by weighing the packing before and after each
run. In most cases, new packing was installed for the
subsequent test. Following these four preliminary experiments,
an 80-hour continuous operation was executed and the scale
accumulation measured. The closed-loop mode for the major
portion of this test series is shown in Figure III-2.
1. Scaling
The conditions for each run are summarized in Table A-l
of Appendix A. Although some operational difficulties were
encountered, and several process parameters were not consis-
tent throughout any run, a general trend in scale accumulation
could be seen. With high tower liquid-to-gas ratio, lower
solids buildup was measured than with low L/G. Residence
time for the slurry in the tower hold tank showed no effect
as the hold time was varied from 30 to 10 minutes.
Profile of solids buildup on the packing going from the
top to the bottom section did show a pattern of scale. Very
little scale deposited in the top section while a consistent
406
-------
CLARIFIERJ
VENTURI
SLURRY
TANK
TOWER
SLURRY
TANK
FIG.TH-2 TWO STAGE CALCIUM CARBONATE SCRUBBER
407
-------
quantity precipitated in the bottom four elements. This
profile of solids buildup suggested an absorption super-
saturation taking place within the tower with an induced
encrustation after one or two feet in the packing. From the
profile of solids deposited and the reduced deposition at high
L/G, one could conclude that incoming solution from the hold
tank was at a lower supersaturation, but once supersaturation
did develop within the tower, the rate of encrustation was
constant.
Based on the aforementioned reasoning, the operating
conditions selected for the long-term demonstration tests
combined a low level tower slurry tank volume with a high
tower liquid-to-gas ratio, i.e. a ten minute residence time in
the hold tank and a L/G of 40 gallons per 1,000 cubic feet.
Stoichiometry at the start of the run was near 100% for the
first 40 hours and 120% for the last 40 hours. Two limestone
grinds were used throughout the run; for the first 40 hours,
a limestone having 75%-200 mesh was employed, while during
the second 40 hours the same material having 61%-200 mesh was
fed. Solids concentration in the slurry was held between 4.4
and 8.9%. Chemical and particle analysis for these limestone
materials are given in Table III-3.
2. Absorption
The absorption efficiency varied from a low of 55% to a
high of 98%. Near the end of the run, the absorption was
higher. A profile plot of the Stoichiometry, slurry concen-
tration, and tower efficiency is shown in Figure III-3.
During the continuous scaling run, a blower motor failure
interrupted the test half way through the experiment. The
408
-------
TABLE III-3
Limestone Materials Supplied By
TVA For The Scaling Test Series*
Designation
Particle Size
Amount Received, Ibs
Chemical Analysis
Shipment
A
61%-200 Mesh
2000
Shipment
B
75%-200 Mesh
4700
CaO %
MgO %
K2O %
Na20 %
Ign. Loss
50.8
—
—
—
41.5
50.5
1.5
0.4
0.3
41.5
* Limestone supplied by TVA called "Tiftona
Limestone". Shipment "A" was a special
supply of CaCO3 for the scaling test series;
Material "B" was received from TVA for the
first test series.
409
-------
FIGURE 111-3
TASK C-6 EFFICIENCY PRQEHE, SLURRY CQWBURATIQN AMD STDlCHIQMEtfiy
DURING THE RUN, SEE TABLE M
-pi
o
\i
<* 04
0-
l I ' •
' ! i
„__. L::J.-"n,_-jTJ | .' r '
"^V"
J.
\
• •'
::••..
^, ^x"
S ii ilppV r
•~vL.l|l\IM *^
: -i
- -gS^
.„.: i
QNCi
- 4-
D STOICHldMETRY
!
;
! - . . '
/
.s '
!
:r
10
20
30 40 50
HOURS IN OPERATION
60
70
-1,2
-1-.-0-
80
-------
packing was removed from the tower and each section was
examined and weighed. Most of the encrustation deposited
during the 40 hours was on the packing periphery. The en-
crustation had a mud-like consistency, not the hard scale
observed with hydrated lime systems. Scale was not evident
on the well-irrigated surfaces. For the second 40 hours, the
same packing was used; encrustation that did develop was again
predominantly at the periphery. Measured weight gain after
40 and 80 hours is given in Table III-4.
Within the first 40 hours, 10.5 pounds of solids had built
up and during the next 40 hours an additional 12.5 pounds were
deposited, weight gain determined on dried basis. No pressure
increase could be measured throughout the 80-hour test. The
amount of solids clinging to the packing was a small fraction
of the packing void volume. Pressure measurements on the
tower were approximately 1 inch of water at the start and
finish of the tests, as given in Table A-l in Appendix A.
411
-------
TABLE III-4
Measured Scaling Accumulation On The Packing
PO
Test Number
Tower L/G
gals/1000 cf
Residence Time
For Slurry
Run Time, hrs.
Packing Number
1 weight gain, Ibs.
2 weight gain, Ibs.
3 weight gain, Ibs.
4 weight gain, Ibs.
5 weight gain, Ibs.
TOTAL :
C-2
30
30
40
7
14
13
16
14
64
C-3
40-60
30
40
2
9
10
10
12
43
C-4 *
40
10
40
4
13
15
14
20
66
C-5 **
40
10
40
2
5
8
8
8
31 .
C-6
45
1C
40
.5
1.5
1.5
3.5
3.5
10.5
**
..i
)
80
3
3
7
6
4
23
* Temperature on the tower circulating slurry not controlled
at the inlet gas dew point. Heat losses on the slurry—10
to 15°F.
** Packing was weighed slightly wet.
Packing
Position
-------
IV. DISCUSSION
A. LIMESTONE EFFICIENCY TESTS - OPEN-LOOP
The venturi scrubber removed only a small fraction of
the inlet S02 with each type of limestone. Between 6 to 22%
of the SO2 was sorbed for an inlet stoichiometry from 1 to
1.6, a liquid-to-gas ratio of 10 gallons per 1,000 cu.ft.,
and a pressure drop across the throat of 6.5 to 12.7 in. of
H20. Such low scrubbing efficiency would not be practical for
commercial application.
The packed tower performed outstandingly good with the
limestone and chalk. The finely ground limestone (89%-325 mesh)
removed nearly 95% of the tower inlet SO2 (1288 ppm). Chalk
with a through 200 mesh fraction of 89% absorbed 97% of the
S02 at 1405 ppm inlet. Presumably, the high surface area
characteristics of the chalk improved the absorption con-
siderably. Cement kiln dust absorption of SO2 was the lowest
for the three limestone types tested. Although the cement dusts
mesh size was very close to the chalk (90%-200 mesh) only 73%
of the inlet SO2 was removed. The cement dust may have been
"dead burned" by the calcination conditions in the cement kiln,
hence an appreciable portion could have been slow dissolving.
Liquid-to-gas ratio in the tower had a dramatic affect
on the SO2 efficiency. For a L/G of 40 gallons per 1,000
cubic feet and an inlet SO2 at 1402 ppm, the SO2 absorption
was 85.8% while at L/G of 20 and an inlet S02 of 1415 ppm,
the removal dropped to 58.2%. Such sensitivity to liquid
flow implies a significant liquid-phase absorption resistance.
The flow rates through the tower at the liquid-to-gas
ratio of 40 gals/1000 cf WPT-» well below the calculated
413
-------
flooding conditions indicated in Figure II-3. For example,
at a L/G =40 and a gas velocity of 9.6 ft/sec (Test No. A8
of Table III-2), the liquid and gas rates were ll,700(lbs)/
(hr)(sq.ft.) and 2,160(Ibs)/(hr)(sq.ft.), respectively. These
values fall well below the flood condition. In fact, the
liquid rate could have been doubled to 23,400(Ibs)/(hr)(sq.ft.)
or an L/G = 80 and still remain below 70% of the flooding
liquid rate.
B. SCALING EXPERIMENTS
1. Scaling
The experimental technique used in measuring the scale
accumulation on the packing was designed to qualitatively
indicate the operating conditions affecting the deposition.
Chemical composition and physical properties of the solid
deposits were not analyzed. As previously mentioned, solids
that did precipitate on the packing surface had a mud-like
consistency, not the hard encrustation normally observed with
lime.
Because of the packing design, a small diameter tower
such as the pilot unit has a significant fraction of surface
area that is not well-irrigated. When the corrugated sheets
are cut in circular shapes, several peripheral layers are not
exposed to the main irrigated pattern. It is in this peripheral
region that much of the solids accumulated.
The scaling experiments, Tests C2 to C5, were originally
designed to screen several tower process variables, i.e.
liquid-to-gas ratio and tower slurry tank residence time.
Tower slurry concentration and limestone stoichiometry were
414
-------
planned at 5% by weight and 100 to 110%, respectively. How-
ever, throughout the preliminary tests these variables were
poorly controlled. Slurry concentration for tests C2 to C5
fluctuated considerably in some cases and stoichiometry was
sporadic because of poor limestone feed control. No
quantitative conclusions could be made with the measured
scaling for the 40-hour tests, C2 to C5. Yet, two general
trends could be seen: 1) the high liquid-to-gas ratio showed
lower scaling, and 2) residence time in the tower slurry tank
did not affect the solids accumulation. The operation and
problems experienced in tests C2 and C5 were either eliminated
or closely scrutinized during the final continuous test.
For the 80-hour run, solids buildup on the packing was
reduced considerably as shown in Table III-4. The 40-hour
preliminary experiment showed 30 to 66 pounds of solids
accumulation in the five feet of packing while the final 80
hour run had 23 pounds. More important, however, the deposi-
tion that did occur was not a hard encrustation but a soft
sludge precipitate and a considerable portion of this sludge
collected on the packing periphery.
2. Absorption Efficiencies
The absorption efficiency for the preliminary scaling
runs varied considerably because of the fluxuations in lime-
stone stoichiometry, slurry concentration, inlet gas composi-
tion, and liquid-to-gas ratio. To avoid process variations in
the long-term run, a strict control of the operation was
maintained for the planned 80-hour tests. However, even with
close scrutiny of these process variables, the absorption
efficiency fluxuated considerably. To explain the efficiency
415
-------
variation, the operating conditions for the 80-hour run were
examined carefully. Slurry chemical analysis performed by
TVA and Radian Corporation were combined with the absorption
efficiency, inlet SO2 concentration, and slurry concentrations
* 5,6
measured in the field.
As a first step in the data analysis, the calcium car-
bonate concentrations were determined for the absorbate over
the entire 80-hour run by digital computer simulation of the
absorption process. With a known CaCO3 analysis as a start-
ing point, the CaC03 slurry concentration was calculated for
each point in time that an efficiency measurement was made.
The predicted carbonate slurry concentration fit well with the
chemical analyses. The computer simulation did not take into
consideration process changes such as spills, leaks, or
uncontrolled water addition. A comparison of the computed
and analyzed carbonate concentration is given in Table IV-1.
TABLE IV-1
LIMESTONE CONCENTRATION IN THE HOLD
TANK DURING TASK C6
Calcium Carbonate Concentration, %
Date
1/25
1/26
1/26
1/26
1/29
1/29
1/30
1/30
Time
2230
1300
1500
2100
1400
2100
1230
0103
Computer Predicted
1.89(Start)
1.19
1.19
1.46
0.36 (Start)
1.44
1.32
1.78
TVA
1.89
0.54
0.94
0.36
2.16
1.65
1.60
Radian
1.13
—
416
-------
Using the computer-estimated value of limestone concen-
tration, the analyzed sulfur dioxide absorption efficiency in
the tower, and the measured tower hold tank slurry concentra-
tion, a linear correlation was developed which predicts the
absorption efficiency for each of the limestone materials
employed.
For the first 40 hours, where limestone ground to 75%
-200 mesh was used, an outstandingly good correlation was
realized. The absorption efficiency, predicted to within
+ 1.9%, showed sensitivity to inlet SO2 concentration and lime-
stone concentration as seen below:
Y = 165.05 - 0.0463(ppm) + 30.48(% CaC03) - 9.126(SL)
where Y = SO- absorption efficiency, %,
ppm = tower inlet SO- concentration in ppm,
% CaCO- = concentration of limestone slurry in
the hold tank, %,
SL = concentration of all solid in hold
tank, %.
Statistical parameters and the variable range for this correla-
tion are listed in Table IV-2.
For the second half of the run, a similar linear correla-
tion having a precision of + 3% efficiency was developed for
limestone with 61%-200 mesh. Here the last 26 hours of
operation were studied so that a mixture of the two limestone
types could be avoided. The predicted efficiency showed less
417
-------
TABLE IV-2
STATISTICAL PARAMETERS FOR
EFFICIENCY CORRELATION
(First 40 Hours Run*)
Number of Data Points
Correlation Coefficient
Standard Error For Estimate
Significance of Regression (F)
% Efficiency Range, Y
Sulfur Dioxide Inlet Range
% CaCO, Range
% Total Solids Range, SL
30
0.987
1.9
349
53% to 97%
1160 to 1900 ppm
1.131% to 1.89%
5.4% to 9%
* Limestone Used - Tiftona Limestone 50.5% CaO, 75% - 200
Mesh.
418
-------
sensitivity to inlet SO2 concentration and qreater sensitivity
to the limestone concentration, i.e.
(IV-2)
Y = 56.273 - 0.0178{ppm) + 50.313(% CaCO) - 4.15(SL)
See Table IV-3 for statistical limitations.
To make use of these efficiency correlations, the lime-
stone slurry concentrations must be known. Three factors in-
fluence the residual limestone concentration in the tower
slurry liquor: 1) the actual absorption efficiency for the
process, 2) the stoichiometric feed ratio of CaCO^/SO^/ and
3) the overall slurry concentration. For a system with 6%
total slurry, the limestone concentration can be predicted by:
% CaC03 = 1.23 - 0.033(% Y) + 2.236 R (IV-3)
where R = stoichiometric feed ratio, moles of CaCO^/moles of SO^
Table IV-4 presents the condition for equation (IV-3) .
Using this expression and equation (IV-1) or (IV-2) ,
the absorption efficiency can be predicted for either lime-
stone material for a liquid-to-gas ratio of 45 gallons per
1,000 cf., and a total slurry of 6% by weight.
Clearly, if absorption efficiency is dependent upon
the limestone slurry concentration and the inlet S02 concen-
tration, then one or both of these conditions must be con-
trolled for a desired SO2 removal. To illustrate this point,
equations (IV-1) and (IV-3) were combined and the resulting
relationship was plotted in Figure IV-1.
419
-------
TABLE IV-3
STATISTICAL PARAMETERS FOR
EFFICIENCY CORRELATION
(Last 26 Hours Run*)
Number of Data Points
Correlation Coefficient
Standard Error For Estimate
Significance of Regression (F)
% Efficiency Range, Y
Sulfur Dioxide Inlet Range
% CaCO-, Range
% Total Solids Range, SL
26
0.95
3.08
69
75.5% to 97.9%
960 to 1380 ppm.
1.317% to 1.847%
5.5% to 8%
* Limestone Used - Tiftona Limestone 50.8% CaO, 61% - 200
Mesh.
420
-------
TABLE IV-4
STATISTICAL PARAMETERS FOR
STEADY STATE WT.% LIMESTONE CORRELATION
Number of Data Points = 16
Correlation Coefficient = 0.995
Standard Error For Estimate = 0.054
Significance of Regression (F) = 741.8
% CaCO3 Range = 0.1616% to 2.07%
% Efficiency Range, Y = 53% to 92%
Range For Stoichiometric Ratio - 1.0 to 1.3
421
-------
ND. 341 -2D DIETZGEN GRAPH
2O X 2C »EO ISCH
EUGENE DIETZGEN CD.
OF CA€03/ltoL£ OF -S0;?
-------
As the graph illustrates, the efficiency is very sensitive
to the stoichiometry and the inlet SO- concentration. For
a stoichiometric addition of limestone, SO- removal can vary
between 85 and 62% as the inlet SO, concentration changes
from 1000 to 2000 ppm. This effect on inlet concentration
indicates a significant liquid-phase resistance. It is
evident in equation (IV-1) and (IV-2) that the total solid
concentration has an adverse effect on the absorption
efficiency. This is logical in that the viscosity and flow
characteristics would change with increasing solid concentra-
tion. A similar slurry influence was observed by Chertkov
with hydrated lime slurries in a board packing. He observed
decreasing mass-transfer coefficients as a slurry concentration
increased.
Simultaneous to this negative coefficient for slurry
concentration (SL), the residual calcium carbonate in slurry
shows a positive coefficient. Hence, with increasing calcium
carbonate concentration, the absorption efficiency improves.
The net result of increasing or decreasing the slurry concen-
tration must be analyzed and the effects of total slurry and
calcium carbonate concentration should be taken into considera-
tion.
The absorption efficiency for the two limestone grinds,
61% and 75%-200 mesh, showed a considerable difference in
absorption efficiency. For a limestone feed of 120% of
stoichiometric and a SO2 inlet concentration of approximately
1300 parts per million, the absorption efficiency expected
with the 61%-200 mesh would be approximately 77%. On the other
hand, the 75%-200 mesh material for the same conditions would
allow near 85% removal. These differences in efficiencies
are demonstrated graphically in Figure IV-2. The curves shown
423
-------
NO. 341-2C DIETZQCN GRAPH
20 X 2- = E9 ISCM
EUGENE DiETZGEN CO.
.«»C€ N j 5. *
ro
i-
•fafiURE
ABseimeNWiTt
-twesw
SLURRY IN A -HETTEfr-Fttff mm T0WER
•IV-2
11
12 1}3
.4.. 1 5
i
TOICHIOWE'RY, MOLES
OF CA^/ftJLE OF
-------
represent the expected absorption for a gas containing 1288 ppm
S02 inlet concentration and a total slurry concentration of 6%.
To illustrate further the influence of the limestone
grind, a single point measurement for 89%-325 mesh was also
plotted in Figure IV-2. The finer material allowed approxi-
mately 11 to 12% more SO2 removal for the same operating con-
ditions as the 75%-200 mesh. One could conclude from these
results that the surface area available for limestone
dissolution within the slurry tank and tower and absorber played
an important role in the overall mass-transfer resistance.
425
-------
V. SUMMARY AND CONCLUSIONS
1. A limestone slurry, circulating through a
high specific-surface packed tower can
absorb greater than 90% of the flue gas SO-.
Absorption efficiency is adversely affected
by increasing S0~ concentration and by high
slurry concentration. SO- absorption can
be improved by increasing the calcium car-
bonate slurry concentration in the absorbing
liquor.
2. A finely ground limestone (90%-325 mesh)
increases the SO- absorption by 11 or 12% over
a material with (75%-200 mesh). Absorption
increases with higher liquor-to-gas ratio
in the tower.
3. Scale formation in a limestone/SO- scrubbing
system can be controlled by maintaining a
reaction product slurry in the absorbing
liquor and by circulating a high liquid rate.
4. A significant liquid phase resistance exists
for the limestone/S02 system while absorbing
the S02 with inlet concentration up to 1,000
ppm.
5. Sulfur dioxide absorption in a venturi type
scrubber using limestone absorbate is limited
to an efficiency between 10 and 20% for Ap = 7"
of water.
426
-------
VI. REFERENCES
1. Walker, A. B., "Mass-Transfer Characteristics of Variable
Annular Throat Venturi Scrubbers", Paper presented to the
National Air Pollution Control Administration International
Symposium of Wet Limestone Scrubbing, March 16-20, 1970.
2. McKenna, J. D., "Evaluation of a Two-Stage Particulate
Scrubber and Gas Absorber Applied to Power Plant Flue
Gas", Paper presented to the National Air Pollution
Control Administration International Symposium of Wet
Limestone Scrubbing, March 16-20, 1970.
3. Lessing, R., "The Development of a Process of Flue Gas
Washing Without Effluent", Journal of the Society of
Chemical Industry, November, 1938.
4. Frisch, N. W., "Calculated Flooding Velocities For A
Wetted Film Packing and Several Commercial Packings",
Cottrell Environmental Systems, Inc., Technical Memorandum
TM70-19, December 15, 1970.
5. Letter from Potts, J. M. of TVA to Gleason, R. J. of
CES, January 19, 1971.
6. Barkley, J., (TVA), Schwitzgebel, K., (Radian), et. al.,
"Chemical and X-ray Analysis of Samples Taken During The
Runs: C5(ll: p.m. 1/21/71) and C6(3:00 p.m. 1/6/71) at
the Tidd Plant in Brilliant, Ohio", Technical Note
200-006-12, February 26, 1971.
7. Chertkov, B. A., "Coefficients of Mass-Transfer During
Absorption of SO, from Gases by Lime Suspension", Khim.
Prom. No. 7, pp 533-36 (1962).
427
-------
APPENDIX
428
-------
TABLE A-l
OCSCITZOTtS TOR TASK. C-2
1/13
2130
0
700
10
30
110
8.35
1.2
6.9
1
1350
1525
695
54. 5
329
118
111
110
115
115
115
4.8
6.3
5.3
4.7
1/14
0100
700
10
30
115
8.35
2.2
7.0
0.6
1580
1600
1020
36.. 3
342
122
110
122
122
122
120
4.6
6.2
5.0
4.6
1/14
0300
B
700
i.0
30
130
8.35
2.2
7.0
1
1500
1550
880
43.3
350
122
112
122
120
120
118
4.8
6.4
5.2
4.6
1/14
0600
B
700
10
30
137
8.35
3.6
7.0
1
1550
1230
850
30.9
~
350
120
112
122
120
115
120
4.6
6.5
5.2
4.6
1/14
1000
B
700
10
30
108
8.35
4.3
7.0
0.9
1930
1050
__
4r.6
349
122
110
122
119
122
4.8
6.4
5.9
4.8
1/14
1200
B
700
10
30
8.35
6.0
6.8
1500
1560
660
•»•»
57.7
~
—
—
__
—
--
--
1/14
1420
B
700
10
30
110
8.35
7.0
8.0
1
1325
1500
336
— —
77.6
—
345
122
115
105
121
118
120
5.2
5.3
5.5
4.7
1/14
2030
B
700
10
30
115
8.35
8
7.0
1.7**
1032
1225
520
<••*
57.6
— ~
340
120
116
116
120
118
118
5.6
6.6
5.2
4.7
1/14
2300
B
700
10
30
116
8.35
6.8
6.9
1.0
1063
1250
570
••
54.4
*"•
335
122
118
115
120
120
118
5.2
6.6
5.5
4.7
1/15
0100
B
700
10
30
98
8.35
6.0
7.0
1.1
1030
875
320
15.1
63.5
69.9
338
120
110
112
120
115
118
4.9
6.5
5.6
4.7
1/15
0330
700
10
30
88
8.35
13.2
7.0
1.1
1030
950
460
8.8
SI. 6
55.4
335
110
10.2
110
115
108
110
4.2
6.4
5.4
4.6
Date
Time
Limestone Used*
Gas Flow, cfm
FDS L/C ratio/ gal/mcf
Tower L/C ratio, gal/mcf
Gas velocity FDS, ft/sec.***
Gas velocity Tower, ft/sec.
Tower pressure drop, inches B.O
PDS pressure drop, inches n*20
CaO/SO2 ratio (inlet analysis)
SO? Concentration, ppm
FDS in **••
FDS out
Tower out
Fraction of SO* removed, %
FDS
Tower
Overall
Gas temperatures, TT
FDS in
ro FOS out
10 Tower out
Liquid temperatures, °F
FDS in
FDS out
Tower in
Tower out
PH measurements
Tower outlet
FDS outlet
Hold tank
Clarifier tank
•Limestone "received from TVA for the "around-the-clock* tests designated as A}Old Tiftona from TVA shown as Type B.
••Stoichiometry was 1.0 at 2015
•••FDS gas velocity based on outlet gas volume.
*••• Some difficulties in SO, analyses were experienced during this test
High efficiencies and negative absorption on FDS should not be con-
sidered valid.
-------
TABLE A-l cont'd
OPERATING OONPITIOK8 FOR TASK C-3
CO
o
Date
TilM
Limestone Used*
Gas Flow, cfra
FDS L/G ratio, gal/mcf
Tower L/G ratio, gal/mcf
Gas velocity FDS, ft/sec.***
Gas velocity tower, ft/sec.
Tower pressure drop, inches H20
FDS pressure drop, inches H20
CaO/SGj ratio (inlet analysis)
SO2 concentration, ppa
FDS in****
FDS out
Tower out
Fraction of SO* removed. %
FDS *
Tower
Overall
Gas temperatures. °F
FDS in
FDS out
Tower out
Liquid temperatures. °F
FDS in
FDS out
Tower in
Tower out
pH nasurements
Tower outlet
FDS outlet
Hold tank
Clarifier tank
Hold tank volume
1/11
1600
A
700
10
50
76
8.35
1.0
_.
1
_—
__
— •
•»••
__
—
——
._
-.
— «.
._
_•
—
— —
._
--
4»^
600
1/12
0930
A
700
10.0
45. 5
76
8.35
1.0
5
1
1350
500
150
63.0
60.0
88.9
330
125
127
130
122
12
125
6.0
4.0
6.3
-_
600
1/12
1730
A
700
10.3
61.0
76
8.35
1.0
4.9
1
182S
17*5
625
2.2
65.0
65.8
34?
122
118
122
122
118
121
5.2
4.4
5.6
5.2
760
1/12
2230
A
700
11
40
76
8.35
1.2
4.4
1
1920
733
429
61.9
41.5
77.7
164
122
121
122
122
121
122
5.6
4.1
5.7
5.1
700
1/13
0100
A
700
—
..
76
8.35
1.7
4.4
1
2000
—
500
__
-—
75.0
242
122
118
122
122
120
122
5.3
3.8
5.4
5.0
700
'Limestone received from TVA for the "around-the-clock" tests designated as A Old Tiftona from TVA shown as Type B
***FDS gas velocity based on outlet gas volume.
••** Some difficulties in SO, analyses were experienced during this test.
High efficiencies and negative absorption on FDS should not be con-
sidered valid.
-------
TABLE A-l cont'd
OPERATING CONOITIOHS FOR TASK C-4
Ci)
Date
Tine
Limestone Used
Gas Flow, cfra
FDS L/G ratio, gal/mcf
Tower r./G ratio, gal/mcf
Gas velocity FDS, ft/sec.***
Gas velocity Tower, ft/sec.
Tower pressure drop, inches H,0
FOS pressure drop, inches H.O
CaO/S02 ratio
SO, Concentration, ppm
2 FDS in •*•*
FDS out
Tower out
Fraction of SO, removed, %
FDS
Tower
Overall
Gas temperatures, *P
FDS in
FDS out
Tower out
Liquid temperatures, *F
FDS in
FDS out
Tower in
Tower out
pU measurements
Tower outlet
FDS outlet
Hold tank
Clarifier tank
1/18
1330
B
700
10
40
138
8.35
• «
—
__
•»•
—
_ —
—
^ —
—
— ^
_—
__
—
__
— —
__
—
1/18
1700
B
700
10
40
138
8.35
0.8
7.8
1.6
1150
1390
400
71.3
335
110
105
112
125
108
111
4.9
6.6
5.6
7.1
1/18
190C
B
700
10
40
138
8.35
0.5
6.4
1.0
1750
1900
150
92.2
—
335
90
80
100
120
83
100
5.7
6.9
6.3
7.1
1/18
ilJO
B
700
10
40
138
8.35
0.7
7.0
.9
1700
1900
1125
40.8
—
330
105
100
105
115
100
110
5.5
7.0
5.7
7.1
1/18
2330
B
700
10
40
138
8.35
0.7
7.0
1.0
1325
1500
320
78.7
—
330
105
100
105
110
100
-
5.2
5.4
5.7
7.1
1/19
0130
B
700
10
40
138
8.35
0.7
7.0
0.9
1230
1250
350
72.0
—
330
105
100
105
110
100
110
5.3
7.2
5.9
7.2
1/19
0530
B
700
10
40
138
8.35
0.7
7.0
1.2
950
1025
488
52.4
—
330
105
102
105
112
100
110
4.2
7.4
5.8
—
1/19
1030
B
700
10
40
138
8.35
0.8
7.8
0.9
1250
1500
—
—
-—
335
115
105
112
120
105
-
4.7
6.7
5.7
—
1/19
1440
B
700
10
40
138
8.35
1.4
7.9
0.9
--
—
700
— ^
~
— -
330
110
105
110
118
105
-
4.9
6.3
5.5
-
1/19
1600
B
700
10
40
138
8.35
1.2
7.2
1.4
1350
1350
680
**
49.7
~~
333
112
105
112
120
105
112
4.9
6.7
5.5
~
1/19
2230
B
700
10
40
138
8.35
0.9
7.0
• MM
1475
1400
800
5.1
42.9
45.8
338
112
105
108
120
105
~
4.8
6.5
5.5
™
•••FDS gas velocity based on outlet gas volume.
•••« some difficulties in SO. analyses were experienced during this test.
High efficiencies and negative absorption on FDS should not be con-
« * j *. « 1 * «3
sidered valid.
-------
TABLE A-l cont'd
COIPITTOIIS FOR TASK C-4
CO
ro
Date
Time
Limestone Used
Gas Flow, cfm
FDS L/G ratio, gal/racf
Tower L/C ratio, gal/mcf
Gas velocity FDS, ft/sec. **•
Gas velocity Tower, ft/see.
Tower pressure drop, inches H-O
FOS pressure drop, inches H-O*
CaO/SO, ratio
SO. Concentration, ppa
2 FDS in ••••
FDS out
Tower out
Fraction of SO. removed* %
FDS
Tower
Overall
Gas temperatures, *F
FDS in
FDS out
Tower out
Liquid temperatures, *r
FDS in
FDS out
Tower in
Tower out
pR measurements
Tower . outlet
FDS outlet
Hold tank
Clarifier tank
1/19
2230
B
700
10
40
138
8.35
1.6
•7.0
—
16SO
1625
1040
1.6
36.0
37.0
325
110
103
105
120.
105
-
4.7
7.1
5.3
1/20
0330
B
700
10
40
138
8.35
1.6
7.0
1.0
850
702
400
17.4
42.4
52.9
332
105
104
102
112
102
102
4.6
6.4
5.4
1/20
0330
a
700
10
40
138
8.35
1.6
7.0
• •
500
425
100
15.0
76.5
80.0
325
100
98
102
112
102
—
-
—
—
*.*FDS gas velocity based on outlet gas volume
as
sidered valid.
'
-------
TABLE A-l cont'd
OPEH?TTT«; CONDITIONS FOR TASK C-5
u>
D.-ito
Time
Limestone Used
Gas Flow, cfm
FDS L/G ratio, gal/mcf
Tower L/G ratio, gal/mcf
Gas velocity FDS, ft/sec.***
Gas velocity Tower, ft/sec.
Tower pressure drop, inches H
FDS pressure drop, inches H,O
CaO/S02 ratio
SO, Concentration, ppa
2 FDS in •••*
FDS out
Tower out
Fraction of SO- removed, %
FDS
Tower
Overall
Gas temperatures, *F
FDS in
FDS out
Tower out
Liquid temperature*, *P
FDS in
FDS out
Tower in
Tower out
pH measurements
Tower outlet
FDS outlet
Hold tank
Clarifier tank
1/20
1130
•*
700
10
40
141
8.35
OR
• o
6£
• w
1250
800
17
36.0
97.8
98.6
335
85
70
100
110
75
80
5.6
50
• V
6.3
1/20
1230
g
700
10
40
141
8.35
1.0
6.4
1075
600
200
44.2
66.7
81.4
330
108
102
105
115
102
105
5.0
6.0
7.1
1/20
1400
B
700
10
40
141
8.35
1.0
7.2
1.1
1050
425
235
59.5
44.7
77.6
330
110
105
105
115
105
110
5.2
4.9
6.2
1/20
1700
B
700
10
40
141
8.35
0.8
6.6
1.1
950
350
100
63.1
71.4
89.4
330
105
102
105
115
100
105
4.8
5.2
6.4
1/20
1900
B
700
10
40
141
8.35
1.1
7.2
2.1
850
300
185
64.7
38.3
78.3
333
115
112
108
118
112
115
5.1
4.9
6.6
6.4
1/20
2100
B
700
10
40
141
8.35
1.1
6.7
1.4
750
250
155
6».6
38.0
79.3
332
115
112
111
118
112
115
4.5
4.9
6.4
5.9
1/20
2300
B
700
10
40
141
8.35
1.1
7.0
1.1
850
300
285
64.7
S.OO
66.5
328
117
113
112
118
115
118
5.3
4.8
6.4
5.7
1/21
0100
B
700
10
40
141
8.35
1.1
7.0
2.0
775
378
330
SI. 3
12.7
57.4
330
118
112
110
118
115
118
4.6
4.8
6.2
5.6
1/21
0300
B
700
10
40
115
8.35
1.9
7.0
1.7
600
250
160
58.3
36.0
73.4
332
112
112
108
118
112
118
5.0
4.8
6.4
5.5
1/21
0500
700
10
40
115
8.35
1.9
7.0
1.45
500
234
125
53.2
53.4
75.0
332
110
105
102
126
106
™
4.8
4.7
6.0
5.5
1/21
0730
700
10
40
138
8.35
3.1
7.0
— —
600
300
ISO
50.0
50.0
75.0
--
MO*
—
^^
—
~
•—
~~
1/21
0915
700
10
40
108
8.35
5.4
5.8
1.45
750
275
120
63.3
54.6
84.0
335
115
110
85
110
""
4.8
4.3
5.7
4.3
•**FDS gas velocity based on outlet gas volume.
•••• some difficulties in SO, analyses were experienced during this test.
High efficiencies and negative absorption on FDS should not be con-
sidered valid.
-------
TABLE A-l cont'd
Date
Tine
Limestone Used
Gas Flow, efm
FDS L/C ratio, gal/mcC
Tower L/G ratio, gal/mcf
Gas velocity FDS, ft/sec.***
Gas velocity Tower, ft/sec.
Tower pr ensure drop,' inches H,O
FDS pressure drop, inches H.O
CaO/SO, ratio
SO, Concentration, ppm
2 FDS in ••••
FDS out
Tower out
Fraction of SO, removed, %
FDS *
Tower
Overall
Gas temperature*, *F
FDS in
FDS out
Tower out
Liquid temperatures, *F
FDS in
FDS out
Tower in
Tower out
pB measurements
Tower outlet
FDS outlet
Hold tank
Clarifier tank
1/21
1100
B
700
10
40
111
8.35
5.0
5.8
1.27
825
225
210
72.8
6.7
74.6
335
120
115
110
*•
120
4.9
4.5
5.7
5.2
1/21 1/21
1300 1500
B
700
10
40
132
8.35
5.0
6.0
1.3
890
225
240
74.8
73.1
340
115
70
105
115
115
4.8
3.7
5.4
5.1
B
700
10
40
115
8.35
3.2
5.2
0.8
1125
1050
330
7.7
68.6
70.7
335
120
120
118
120
120
3.0
3.V
5.7
5.0
IJTTOHS FOR TASK
1/21
1700
•*
700
10
40
115
8.35
3.3
4.6
1.0
1300
1250
370
5.7
70.4
71.6
335
118
119
114
118
119
3.8
4.1
5.8
4.8
1/21
1900
g
700
10
40
115
8.35
3.6
4.2
0.9
1500
1375
355
8.4
74.2
76.4
335
119
119
115
118
119
5.9
4.6
6.2
4.2
C-5
1/21
2100
B
700
10
40
115
8.35
3.9
4.6
C.9
1600
1550
425
3.1
72.6
73.5
335
121
120
118
120
121
5.8
4.9
6.0
4.2
1/21
2300
B
700
10
40
115
8.35
4.3
4.0
0.9
1650
1525
360
7.6
76.4
78.2
333
120
120
118
119
121
5.3
5.1
5.9
4.2
••• FDS g*s velocity based on outlet gas volume.
sidered valid.
-------
TABLE A-l cont'd
OPERATING CONDITIONS FOR TASK C-6
en
Date
Time
Limestone Used
Gas Flow, cfm
FDS t/G ratio, gal/mcf
Tower L/G ratio, gal/mcf
Gas.velocity FDS, ft/sec.***
Gas velocity Tower, ft/sec.
Tower pressure drop, inches l
FDS pressure drop, inches H2
CaO/SO- ratio
SO. Concentration, ppm
FDS in
FDS out
Tower out
Fraction of S03 removed, %
L-ne *
Tower
Overall
Gas temperatures, *F
FDS in
FDS out
Tower out
Liquid temperatures,
FDS in
FDS out
Tower in
Tower out
pH measurements
Tower outlet
FDS outlet
Hold tank
Clarificr tank
1/25
1830
B
700
10
45
127
8.35
1.0
6.4
0.9
1700
1580
540
7.1
65.9
68.3
335
105
108
40
85
108
108
5.4
4.6
5.3
3.2
1/25
2130
B
700
10
45
141
8.35
1.0
7.0
1.0
1700
1635
620
3.9
62.1
63.5
328
120
120
114
118
118
118
5.4
4.4
5.5
3.4
1/25
2400
B
700
10
45
141
8.35
1.1
6.5
0.9
1920
1900
880
1.0
53.7
54.2
335
120
120
112
120
121
121
4.9
5.2
5.3
3.4
1/26
0200
B
700
10
45
141
8.35
o.«»
6.8
1.0
1930
1760
560
8.8
68.2
-1.0
330
118
118
115
118
118
120
5.2
5.0
5.7
4.7
1/26
0400
B
700
10
45
141
8.35
1.0
7.3
1.0
1890
1735
460
8.2
73.5
75.7
330
120
120
118
,120
'120
122
5.4
4.8
5.4
4.6
1/26
1115
B
700
10
45
141
8.35
1 0
6.0
1.0
1680
1560
320
7.2
79.5
81.0
118
118
112
120
118
118
5.5
5.4
6.2
4.3
1/26
1230
*»
700
10
45
135
8.35
l.C
6.2
1.2
1360
1220
100
10.3
91.8
92.7
118
118
112
120
118
120
5.9
6.8
6.4
4.5
1/26
1500
B
700
10
45
135
8.35
1.0
5.8
1.2
1260
1160
34
8.0
97.1
97.4
330
118
118
112
120
118
120
6.6
6.8
6.1
4.9
1/26
1715
B
700
10
45
135
8.35
0.9
6.2
0.9
1380
—
335
118
115
110
118
115
118
4.3
6.6
6.0
4.9
1/26
2230
B
700
10
45
141
8.35
0.9
7.0
1.1
1400
1320
240
5.8
81.9
82.9
333
118
115
114
118
115
118
5.7
4.8
6.4
4.6
1/27
0300
B
700
10
45
141
8.35
0.9
7.0
0.9
1565
1440
280
8.0
80.6
82.2
335
115
116
114
118
114
118
4.8
4.8
6.3
4.5
•••FDS gas velocity based on outlet gas volume.
-------
TABLE A-l cont'd
Dat«
Tim*
Limestone Used
Gas Flow, cfm
FDS L/C ratio, gal/mcf
Tower L/C ratio, gal/mcf
Ga« valocity FDS, ft/sac.***
Gas velocity Tower, ft/sec.
Tower pressure drop, inches H.O
FDS pressurw drop, inches H.O
CaO/SO, ratio
SO, Concentration, pp»
* FDS in.
FDS out
Tower out
Fraction of SO- removed, %
FDS *
Tower
Overall
Gas temperature, *F
FDS in
FDS out
Tower out
Liquid temperatures, *F
FDS in
FDS out
lower in
Tower out
pH Bcasurements
Tower outlet
FOS outlet
Hold tank
Clarifiex tank
1/29
1405
A
700
10
45
127
8.35
0.8
6.4
1.2
1440
1320
360
8.4
72.8
75.0
342
122
122
119
126
122
125
5 8
** • W
5 4
W * ^
7.6
3.6
1/29
1620
A
700
10
45
127
8.35
0.8
7.0
1.2
1440
1340
220
8.0
83.5
84.8
340
112
110
100
118
112
115
3.5
4.2
5.4
5.0
1/29
2100
A
700
10
45
127
6.35
0.8
7.0
1.2
1360
1280
210
5.9
83.6
84.6
342
122
120
118
125
121
123
5.8
5.4
S.4
5.0
i/30
0200
A
700
10
45
141
S.35
0.7
7.0
1.2
1280
1140
280
11.0
75.5
78.2
335
120
118
118
122
118
120
5.6
5.4
6.2
5.2
1/30
0600
A
700
10
45
141
6.3i
0.6
7.0
1.1
1480
1380
280
6.8
79.8
81.1
330
120
118
116
122
118
120
5.4
4.7
6.2
4.9
1/30
1230
A
700
10
45
151
8.35
0.6
6.4
1.3
1450
1340
400
7.6
70.2
72.5
325
120
120
118
122
122
5.4
5.5
5.8
5 A
.2
1/30
1330
A
700
10
45
141
8.35
1<9
• «
1420
1300
320
8.5
75.4
77.5
—
*"*
— **
1/30
1415
A
700
10
45
141
£.35
~~.
1 2
J» • •»
1240
120
90.4
•••»
•—
zz
1/30
1505
A
700
10
45
141
8.35
0.6
6.0
1.2
1260
1120
140
11.2
87.5
88.9
330
120
118
115
122
118
122
5.5
6.0
6 2
Q • •
5.2
1/30
1830
A
700
10
45
141
8.35
1.1
7.2
1.2
1160
1100
84
5.2
92.4
92.8
330
116
112
115
122
113
118
5.1
4.4
6.4
5.1
1/30
2100
A
700
10
45
141
8.35
1.1
7.6
1.2
1000
960
20
4 .0
97.9
98.0
320
120
118
114
120
120
120
5.3
4.4
6.3
4.9
1/30
2330
A
700
10
45
141
8.35
0.9
6.4
1M
.2
1000
979
20
4n
. V
97.9
98.0
325
118
112
115
121
112
118
4.4
4.0
6.2
5.0
1/31
0100
A
700
10
45
141
8.35
0.9
•7 O
/ . u
i 3
x . *
1080
1000
72
2.9
92.8
93.3
330
118
112
116
120
114
120
4 A
.8
4.9
6.0
5.0
•••FDS gas velocity based on outlet gas volume.
co
-------
REMOVAL OF SULFUR DIOXIDE FROM STACK GASES
BY SCRUBBING WITH LIMESTONE SLURRY:
TVA PILOT PLANT TESTS
I. SCRUBBER-TYPE COMPARISON
By
T. M. Kelso, P. C. Williamson, and J. J. Schultz
Division of Chemical Development
Tennessee Valley Authority
Muscle Shoals, Alabama
II. EXPERIMENTAL DESIGN AND DATA ANALYSIS
FOR SPRAY AND MOBILE-BED SCRUBBERS
By
N. D. Moore
Division of Power Resource Planning
Tennessee Valley Authority
Chattanooga, Tennessee
Prepared for Presentation at
Second International Lime/Limestone Wet Scrubbing Symposium
Sponsored by the Environmental Protection Agency
New Orleans, Louisiana
November 8-12, 19J1
437
-------
REMOVAL OF SULFUR DIOXIDE FROM STACK GASES
BY SCRUBBING WITH LIMESTONE SLURRY:
TVA PILOT PLANT TESTS
ABSTRACT
I. SCRUBBER-TYPE COMPARISON
By
T. M. Kelso, P. C. Williamson, and J. J. Schultz
Division of Chemical Development
Tennessee Valley Authority
Muscle Shoals, Alabama
TVA plans to install limestone slurry scrubbing on a 500-mw
boiler (No. 8 unit at the Widows Creek station) as a demonstration project.
Since adequate design data for limestone scrubbing are not available, a
pilot plant (2000-cfm capacity) has been constructed at the Colbert Steam
Plant and operated for about 10 months. The limestone, pulverized to about
70$ minus 200 mesh, is a high-calcium type from the Widows Creek area. Most
of the tests have been at a CaO:S02 mole ratio of 1.5, an L/G (gal slurry/
Mcf of gas) of 50 to 60, and a slurry solids content of 10-15$; the stack gas
contains about 5000 ppm of S02. All long-term tests have been with a closed-
liquor loop.
Since others have shown that packed scrubbers are effective con-
tactors but subject to scaling, three scrubber types were tested—ranging
from fully packed (with fixed packing) to an empty tower. These systems
were:
1. Ventri-Rod (a type of venturi) followed by a packed crossflow
scrubber.
2. Three-stage mobile-bed scrubber (bouncing bed of hollow plastic
balls).
J. Ventri-Rod followed by open spray tower.
With the Ventri-Rod - packed crossflow scrubber configuration,
about 80$ of the S02 was removed from the gas stream. This system could not
be operated for more than about 120 hours because of plugging of the packing
either by deposition of slurry solids or by scaling.
438
-------
With the Ventri-Rod - open-spray tower configuration, about
77$ of the S02 was absorbed. This system was operated continuously
without difficulty for over 350 hours and routinely shut down.
Tests are in progress with the mobile-bed scrubber. In the
tests that have been completed, 90-92$ of the S02 in the flue gas has
been removed.
Several problems other than scaling have been encountered, in-
cluding (l) slurry solids deposition on surfaces, (2) poor settling of
solids, (3) unstable operation (irreversible decline in pH) under some con-
ditions, (4) corrosion and erosion of surfaces, and (5) mist carryover.
All of these have been fairly well resolved (although further improvement
would be desirable) except the last two. Test programs aimed specifically
at erosion and mist carryover are under way.
The following conclusions can be drawn from these studies:
1. Good absorption (75-90$) can be obtained without scaling
if the proper scrubber type is used. Longer test runs are
desirable, however, to confirm the nonscaling operation.
2. At the high liquor rates required for S02 removal in
limestone slurry scrubbing, and for fly ash of the type
involved in these tests, excellent collection of the dust
in the incoming gas is achieved with no special provisions.
Solids carryover in mist is a major problem, however, and
must be solved if limestone slurry scrubbing is to be a
usable process.
3. Care must be taken to maintain an adequate rate of limestone
dissolution in order to avoid unstable operation. Other
variables may also be effective in accomplishing this but
have not been tested.
439
-------
II. EXPERIMENTAL DESIGN AND DATA. ANALYSIS
FOR SPRAY AND MOBILE-BED SCRUBBERS
By
N. D. Moore
Division of Power Resource Planning
Tennessee Valley Authority
Chattanooga, Tennessee
The experimental designs and statistical analyses of the data
from several pilot plant runs on limestone scrubbing are discussed.
Scrubber types involved are a mobile bed (three beds of hollow plastic
balls) and three configurations of an assembly consisting of a Ventri-Rod
followed by sprays in an open tower. The significant variables for each
type are identified and linear regression equations derived.
A mathematical model for fly ash removal is developed from data
obtained in a mobile-bed scrubber operated at a power plant with water as
the scrubbing medium. The fly ash model is compared with experimental
data obtained in scrubbing with limestone slurry.
440
-------
REMOVAL OF SULFUR DIOXIDE FROM STACK GASES
BY SCRUBBING WITH LIMESTONE SLURRY:
TVA PILOT PLANT TESTS
I. SCRUBBER-TYPE COMPARISON
By
T. M. Kelso, P. C. Williamson, and J. J. Schultz
Division of Chemical Development
Tennessee Valley Authority
Muscle Shoals, Alabama
In addition to the small-scale work described earlier in this
symposium ("Removal of Sulfur Dioxide from Stack Gases by Scrubbing with
Limestone Slurry: Small-Scale Studies at TVA" by J. M. Potts, A. V. Slack,
and J. D. Hatfield), pilot plant studies on limestone slurry scrubbing have
been carried out. The purpose of both studies is to supply information
needed for design of the full-scale demonstration scrubbing facility that
TVA plans to install on the No. 8 generating unit (550-mw) at the Widows
Creek Steam Plant (near Chattanooga, Tennessee). The pilot plant is located
at the Colbert Steam Plant (on the No. 3 unit), which is near TVA1s chemical
research facilities at Muscle Shoals, Alabama.
The main purpose of the pilot plant work has been to identify the
problems that affect operational reliability and to find ways of solving
them. (Scaling, for example, is known to be a major problem of this type.)
Part I of this paper deals mainly with this aspect of the test program. Some
effort was also made toward optimizing the operating parameters; Part II covers
the experimental design and data analysis associated with this phase of the
work.
The pilot plant, which is still in operation, receives a sidestream
of gas of about 2000 cfm (measured at scrubber outlet conditions--saturated
at 120°F) from the outlet duct of the boiler (Babcock and Wilcox, front-fired,
pulverized coal), in which coal containing about k^> S and 15$ ash is burned.
The stack gas entering the pilot plant contains about 3000 ppm S02 and 3 to
k grains per scf (dry basis) of fly ash.
In the pilot plant the gas is scrubbed with a circulating limestone
slurry containing 12 to 15$ solids. This level of solids was selected because
a higher content increases pumping difficulty and increases settling tendencies
in the scrubber; a lower one reduces the beneficial effect of high solids
content on scaling and S02 removal.
441
-------
Pulverized limestone is slurried and fed into the liquor circu-
lation loop continuously. Typical analyses of the three limestones used
are given in Table I. Preliminary tests in the pilot plant did not indicate
any significant difference in the reactivity of the three; consequently,
most of the tests were made with limestone from Tiftonia, Tennessee, which
is near the Widows Creek Steam Plant.
TABLE I
Typical Limestone Analyses
Screen analyses,
Quarry Chemical analyses, % cumulative % retained on
location CaO MgO AlgQ3 SiQ2 100 mesh 200 mesh 325 mesh
Birmingham,
Alabama 53.8 1.2 0.07 1.^ 14 28 38
Scottsboro,
Alabama 50.0 k.O 0-52 2.2 8 22 37
Tiftonia,
Tennessee 52.3 1.2 0.13 3.4 10 27 UO
Three different types of scrubbers have been tested; fixed packing,
mobile packing, and spray tower. The objective of the tests was to determine
which of the three would give the best combination of reliability and S02
removal efficiency.
Packed Scrubber (Fixed Packing)
The packed scrubber selected for testing was the crossflow type
(Fig l); the horizontal configuration would have the advantage of fitting
better into the duct system from boiler to stack. The scrubber, manufactured
by the Ceilcote Company, consisted of a horizontal housing (glass-reinforced
plastic) filled with spiral-type polypropylene packing (Tellerettes) irrigated
across the top with nozzles set at right angles to the gas flow. A section
of the packing at the gas outlet was left unirrigated to serve as an integral
entrainment separator.
The crossflow was preceded by a venturi scrubber, the primary function
of which was to remove fly ash. The type used was the Ventri-Rod (manufactured
by Environeering, Inc.), essentially a horizontal housing containing a "window"
section of irrigated parallel rods (3/^-in. diameter; stainless steel) spaced
to develop the pressure drop necessary for particulate removal (Fig 2). The
pressure drop across the rod section could be adjusted by two guillotine-type
movable gates. The rod section was followed by a chevron-type entrainment
separator. The entire scrubber, except for the rods, was made of mild steel
coated with bitumastic material.
442
-------
LttUOR
OUTLET
FIGURE 1
Packed Crossflow Scrubber
-------
INLET
VENTRI-ROO
ELEMENT,
OUTLET
GAS
ENTRAPMENT
SEPARATOR
SLURRY1
INLET
DRAIN
FIGURE 2
Horizontal Ventri-Rod
-------
The overall pilot plant (Fig 3) consisted of the scrubbers
connected in series, recirculation tanks equipped with agitators and
pumps, settling tanks, and a slurry preparation tank. Separate slurry
circulation loops were provided for the two scrubbers. Pulverized lime-
stone was slurried in the slurry preparation tank, with liquor from the
recirculation tank in the crossflow scrubber loop, and pumped continuously
at a controlled rate back into the tank. From there it was pumped to the
scrubber and returned to the tank by gravity. The recirculation tank
had a working capacity of about 500 gallons. In most of the tests, the
L/G ratio (gal slurry/1000 scf of gas) was maintained at about 40; conse-
quently slurry residence time in the tank was about 6 minutes. This was
about twice the retention time used in the ICI-Howden experiments.
Slurry from the crossflow recirculation tank flowed by gravity
to the Ventri-Rod recirculation tank. The working capacity of the latter
was 500 gallons; at the usual L/G of 15, residence time in the tank was
about 17 minutes.
Slurry from the Ventri-Rod tank flowed by gravity to the settling
tanks. Supernatant liquor from the settling tanks was returned to the
crossflow recirculation tank, at a withdrawal rate regulated to maintain
the solids content in the recirculated slurry at 12 to 15%. The product
solids and fly ash were periodically pumped from the settling tanks and
discarded.
Unstable pH: In preliminary tests with this scrubber configuration,
two major problems developed that prevented sustained operation: (l) decline
in the pH of the scrubbing slurry finally resulted in an unresponsive slurry
and (2) uncontrollable solids buildup in the circulating slurry.
In the initial attempts to start the pilot plant on a continuous
multishift operation, the crossflow circulation tank was charged with lime-
stone and water proportioned to form a slurry containing 5$ solids. The
flow of gas was started through the crossflow scrubber and the slurry circu-
lated from the recirculation tank through the scrubber and back to the tank.
No fresh limestone was added to the system until the initial batch of limestone
in the slurry was exhausted. A forward flow of slurry to the Ventri-Rod
scrubber was then established, the scrubbing loop closed, and the addition of
fresh limestone started (stoichiometry, 1.0). The pH of the slurry immediately
began to drop, despite the continual addition of fresh limestone, to such an
extent that the slurries became completely unresponsive. The pH declined
finally to 3.7 and the S02 removal decreased to about 30$. Addition of lime-
stone in large quantities to samples of the slurry did not appreciably increase
the pH. Electron microscopic analyses indicated that the limestone particles
in the slurry were covered with a thin crystalline layer of material which
appeared to be calcium sulfite. The layer of sulfite apparently prevented
the limestone substrate from further dissolution and reaction.
445
-------
INLET
FLUE-
GAS
1
VENTRI-ROD
SCRUBBER
•4
I
V
or
01
*»*
RECIRCULATION
TANK
f
SOLIDS
DISCARDED
CROSS-FLOW
SCRUBBER
\
RECIRCULATION
TANK
SETTLING
TANKS
FLUE GAS
TO
STACK
PULVERIZED
LIMESTONE
SLURRY
PREPARATION
TANK
MAKE-UP
WATER
FIGURE 3
Pilot Plant with Ventri-Rod and Crossflow Scrubber Configuration
-------
The mechanism of the limestone blinding is not yet known (see
paper in this symposium, "Removal of Sulfur Dioxide from Stack Gases by
Scrubbing with Limestone Slurry: Small-Scale Studies at TVA" by J. M. Potts,
A. V. Slack, and J. D. Hatfield); however, it is obvious that low pH triggers
it. The exact pH at which the slurry becomes irreversibly unresponsive has
not been clearly defined but there is some indication from the pilot plant
work that if a level no lower than 4-5 can be maintained, the slurry will
return to full responsiveness.
If the pH falls below about 4.0, the pH cannot be restored by
any procedure practical in large plant operation, and the slurry must be
discarded.
The procedure developed to maintain pH at a high level during
startup involved filling the circulation tanks with fresh limestone slurry
and then starting the forward feed flow, slurry circulation, and limestone
feed simultaneously with the flow of gas. Consequently, during the initial
startup period before plant operation was stabilized the circulating slurries
contained a large excess of limestone. Moreover, the stoichiometry was
increased to 1.5 in an effort to maintain a stable pH.
Even with a large excess of limestone in the scrubbing circuits,
the pH of the slurries and the removal of S02 tended to decline during the
first 20 to 30 hours of operation. After this initial period, however,
there was a rapid inc-rease and then stabilization. This phenomenon may be
a function of the ionic strength of the circulating slurry. A typical de-
cline and recovery is shown in Figure 4.
Unstable pH also occurred in two other tests in which the plant
had to be shut down for a period of about 1 hour during each test for emergency
maintenance. The flow of slurry to the crossflow scrubber was maintained
during the shutdown to prevent drying out of the papking. During both periods,
a small amount of flue gas (estimated to be JOO cfm ) leaked through the
scrubber. Even this small amount of gas depleted the alkalinity of the slurry
and made it necessary to abort one of the tests because the slurry had become
completely unresponsive. In the second test, lime was added to the scrubber
slurry to raise the pH to about 5-0 in the crossflow scrubber loop and 4-8
in the Ventri-Rod loop. The pH rapidly dropped back to about 4.0 and remained
in the range of 4.0 to 4.5 for several hours. After approximately 7 hours of
operation, the pH of the makeup limestone slurry (made with circulation tank
liquor) increased from 4-5 to 6.1 within an hour. After 3 hours of additional
operation, the pH of the slurry to the crossflow scrubber increased from 4.2
to 5.0 within an hour and immediately the removal of S02 increased from 49 to
(Table II).
447
-------
-p.
-pi
CO
100
90
5 80
o
2
UJ
* 70
M
O
CO
60
5 10 15 20 25 30 35
OPERATING HOURS SINCE START-UP
40
FIGURE
SOP Removal and yK in Crossflow Recirculation Liquor
-------
TABLE II
Increase in
Sulfur Dioxide
Removal
pH of liquor, from
March Time
17 5 p.m.
o p.m.
7 p.m.
8 p.m.
9 p.m.
10 p.m.
11 p.m.
12 p.m.
18 1 a.m.
2 a.m.
3 a.m.
4 a.m.
5 a.m.
6 a.m.
7 a.m.
8 a.m.
9 a.m.
10 a.m.
11 a.m.
12 m.
1 p.m.
2 p.m.
3 p.m.
S02
removal, f
54
59
64
49
50
48
47
45
46
47
50
_42.
80
86
87
84
83
83
82
77
77
76
78
F-ll
a Mixing tank
4-7
4.6
6.6
6.1
5-5
6.8
4.7
4.6
4-.L
r~ 6.1 "I
! 6.0 i
J 6.1 [
6-5
6.8
6.8
6.5
6.8
6.7
6.7
6.9
6.7
6.6
6.8
F-12
Crossf low
4.1
4-5
4.4
4.3
4.2
4.2
4.3
4.0
4.0
4.1
4.2
4.2
5-0
5-1
5-2
5.8
5-6
5-4
5-5
5-4
5-4
5-6
5.4
F-13
Ventri-Rod
4.2
4.6
4.5
4-3
4-3
4-3
4.3
4.1
4.2
4.2
4.3
4.4
1 4.3
1 4.6
[_ 41£__
"~4-9
4.9
4.9
4.9
4.9
4.9
5-0
4.9
a Lime added: 40 Ib to circulation tank of crossflow scrubber
and 20 Ib to circulation tank of Ventri-Rod.
449
-------
The combination of high initial limestone content in the slurry
and a continuing makeup at 1.5 times stoichiometric made it possible to
operate without trouble from unstable pH. Lower stoichiometry may be
feasible but has not yet been adequately demonstrated.
Solids Settling Rate; The second problem that prevented sustained
operation in the preliminary tests was the uncontrollable solids buildup
in the circulating slurry. In the original pilot plant configuration, the
settling tanks had a total capacity of about 800 gallons but due to the
slow settling characteristics of the slurry, the resulting 3-1/2 hours'
retention time was inadequate. The solids content in the slurry would in-
crease linearly from about 5$ in the starting slurry to approximately 25
to yyf> after about 30 hours of operation, a level that caused plugging of
the packing in the cross flow scrubber.
The settling tank was replaced with three tanks having a usable
capacity of about 4,000 gallons, equivalent to about 20 hours' retention
time. In subsequent tests with the additional settling capacity, the solids
content of the slurry could easily be controlled at the desired 12 to l$%
level.
Although the additional settling time made it possible to operate
the pilot plant satisfactorily in these and the later tests, settling remains
a problem. Even after several weeks, the settled solids layer contains only
about k-0% solids and the volume occupied in a waste pond therefore would be
relatively large. Tests on methods for getting a more compact settled layer
are planned.
Scaling; Because of the uncontrolled pH decline and the buildup
of solids, the initial pilot plant tests were usually of only 30 to kO hours'
duration. After these problems were corrected, attempts were made to operate
the pilot plant for longer periods. The longest period of sustained operation
(about 125 hr) was terminated because of scaling of the scrubber packing.
Before this test was made, the packing was removed from the crossflow
scrubber and cleaned. At the outset of the test, the pressure drop across
the scrubber was 4 in. t^O. There was a steady increase during the test until
after 125 nr of operation (Fig 5) the pressure drop exceeded 17 in., which
caused the seals in the scrubber sump to blow. Many of the Tellerettes were
cemented together into a large mass by a hard, tenacious scale. Also, much
of the open area in the individual Tellerettes was blocked or obstructed by
the scale formation.
Several attempts were made to wash the packing in place with high-
velocity streams of water. Practically none of the tightly adhering scale
could be dislodged. The packing was removed, new packing installed, and the
test resumed. The results were essentially the same as in the previous test.
450
-------
a:
UJ
o ii -
CO
UJ
12 24 36 48 60 72 84 96 108 120
OPERATING HOURS
FIGURE 3
Ventri-Rod Crossflow Scrubber Pressure Differential
and Slurry Solids Content
451
-------
Fetrographic examination of the scale formed on the packing in-
dicated a bulk phase (70-95$) of CaS04-2H20, a minor phase (10-20$) of
CaS03-0.5H20, and very minor phases of CaC03 and CaMg(C03)2.
These tests indicated that a packed scrubber of the type tested
is very prone to plugging, either by solids deposition or by chemical scaling.
Although the S02 removal efficiency was probably acceptable (about 80$ under
the conditions tested), it was concluded that such a scrubber should not be
selected for the Widows Creek project because operation would likely be un-
reliable.
Spray Tower
In the spray tower design used (Fig 6), the venturi scrubber was
integrated into the scrubber by placing a Ventri-Rod element near the bottom
of the tower with the spray pointed upward. The objective was to get more
S02 removal in the venturi (which was quite low in the venturi - packed scrubber
arrangement), by inducing a fountain effect that would give more slurry holdup
in the scrubber. The auxiliary pilot plant equipment, i.e., tanks, pumps, and
blowers, were essentially unaltered.
The spray tower was 32 in. square and 20 ft tall, with a superficial
gas velocity of 5 ft/sec at a total gas flow of 2000 scfm. The Ventri-Rod
element, mounted in the inlet end of the 32-in. housing, was irrigated from
the bottom side cocurrent with the upward gas flow. Two banks of spray nozzles
were located in the open tower at 'y-f.t intervals above the Ventri-Rod element.
A chevron-type entrainment separator was mounted in a horizontal outlet section
and equipped with a spray on the upstream side to remove accumulated solids.
(A simplified sketch of the overall spray tower assembly is shown in Fig j).
A set of screening tests (open loop) was run first to check the
effects of major operating variables (see Part II of this paper). An ex-
tended test run then was made, at conditions indicated to be optimum, to
identify any problems associated with extended plant operation. The L/G
was 60 gal/Mcf—20 each to the spray banks and the Ventri-Rod unit. Pressure
drop, at the gas throughput of 2000 acfm, was about 15 in. H20 across the
venturi and 1 in. through the rest of the scrubber. Stoichiometry was 1.5.
The plant was operated for 35^ hr with only a total of 20 hr of
downtime for maintenance. S02 removal averaged 77$ over the entire test
(Fig 8). The pronounced decrease in S02 removal occurring about 270 hr after
startup and lasting about two shifts was caused by poor distribution of slurry
in the tower. A large hole was eroded in the bottom spray header, resulting
in poor slurry distribution and reduced S02 removal.
452
-------
ENTRAINMENT
SEPARATOR
VENTRI-ROD.
ELEMENT/
ADJUSTABLE
SLIDE GATE
TO
MECIRCULATION
TANK
FIGURE 6
Ventri-Rod - Spray Tower
453
-------
MIST ELIMINATOR
01
INLET
FLUE
GAS
f
MAKE-UP
WATER
1
, , FLUE GAS
TO
STACK
VENTRI-ROD ELEMENT
HUMIDIFICATION Tuff"*
SPRAY
SOLIDS
DISCARDED
SETTLING
TANKS
RECIRCULATION
TANK
PULVERIZED
LIMESTONE
RECIRCULATION
TANK
SLURRY
PREPARATION
TANK
FIGURE 7
Pilot Plant with Ventri-Rod - Spray Tower Scrubber Configuration
-------
o
UJ
(E
N
O
CO
en
en
o|
3 <
-IK
CO Q.
01
>-
o
pi
rffeo
V)
80
70
60
6.8
6.4
6.0
5.6
5.2
16
12
8
4
DOWNTIME
24 48 72 96 120 144 168 192 216 240 264 288 312 336 360
CUMULATIVE OPERATING HOURS SINCE START-UP
FIGURE 8
S05 Removal, pH, and Solids Concentration in Spray Tower Reclrculation Slurry
-------
After the test was routinely terminated, the spray tower was
inspected. No significant scaling had occurred but there were heavy
deposits of limestone and reaction products in the tower above the upper
spray bank. These soft deposits, which were several inches thick, occurred
only in areas not wetted by the sprays; more than 50$ of the mist eliminator
inlet was sealed off. Although the deposits were heavy, the total pressure
drop across the scrubber system, including the mist eliminator, had increased
by only 1 in. H20--from 16 to 17 in. H20. The accumulation could be removed
easily with high-velocity jets of water.
There was some indication in this run that better operation would
be obtained if the spray banks were raised to a higher level, to give more
space between the Ventri-Rod and the upper sprays. Therefore the sprays were
raised 5 ft and another set of open-loop screening tests was carried out.
(A third spray was also installed in the top of the scrubber to prevent solids
accumulation.) The results indicated that the changes improved operation.
It was also indicated (see Part II) that higher gas velocities through the
scrubber could be used to reduce the scrubber cost and the area requirements
by making the cross-sectional area smaller. The previous screening tests
had been made at superficial gas velocities of k.6 to 6.2 ft/sec; the new
test series indicated that 7.8 ft/sec might be used without significantly
reducing S02 removal.
An extended closed-loop run (l68 hr) was made, therefore, at (l)
7-8 ft/sec, (2) L/G of 55 gal/Mcf (2O to Ventri-Rod, 5 to first spray bank,
15 to second, and 15 to the new top spray), and (3) 1-5 stoichiometry. The
S02 removal was only 66%, however, as compared with the 76% predicted from
the screening tests and the j8% obtained in the earlier run at 4.6 ft/sec.
Possible reasons for the lower removal are discussed in Part II.
These tests indicate that the spray tower is a reliable scrubber
for S02 removal, but is not as good a contactor as a fixed-bed scrubber.
Also, the limitation in gas velocity may be an important cost factor.
Mobile-Bed Scrubber
The mobile-bed scrubber type tested was the Turbulent Contact
Absorber (TCA), manufactured by Universal Oil Products Company. It consists
of three beds of mobile packing spaced vertically (Fig 9) and is designed to
operate at 8-13 ft/sec. Each packing unit consists of a 1-ft static depth of
1.5-in. hollow plastic spheres, retained by wire mesh grids k ft apart. The
flow of gas through the bed causes the spheres to move rapidly in a random
pattern between the two retaining grids. The scrubbing liquor is fed above
the top retaining grid through an open-pipe distributor and flows counter-
currently to the gas flow.
456
-------
•*•
' LIQUOR OUTLET
FIGURE 9
Mobile-Bed (TCA) Scrubber
457
-------
A chevron-type mist eliminator manufactured by the Heil Process
Equipment Company was installed in the horizontal housing at the outlet
of the scrubber. The pilot plant configuration with the TCA scrubber is
shown in Figure 10.
After the usual series of screening tests (Part II), a 172-hr
closed-loop run was made at a gas velocity of 12.5 ft/sec, a total L/G of
*4-8 gal/Mcf, and 1.5 stoichiometry. The total pressure drop through the
scrubber was 9 in. H20 during the entire test period.
After routine termination of the run, inspection of the scrubber
showed a small accumulation of solids immediately below the bottom retaining
grid of the first stage. Otherwise, the scrubber elements were clean with
no significant evidence of accumulated solids or scale.
Removal of S02 was excellent, averaging about 92^. The main problems
were ball wear, erosion of grids, and particulate entrainment.
Further tests of the mobile-bed scrubber are being made. If the
indicated resistance to scaling can be verified, and the problems of erosion
and particulate carryover solved, the scrubber should be a good candidate
for large-scale use.
Erosion
Severe and rapid erosion by the circulating slurry has been a major
oroblem with all three of the scrubber types. For each of three, frequent
replacement of pipefittings, valves, and pump components, has been necessary.
Other problems have been specific to the particular equipment.
Ventri-Rod; During preliminary tests with the crossflow scrubber
assembly, a hole was eroded in the Ventri-Rod housing. A 1/2-in.-thick mild
steel wear plate welded over the hole was cut through after about 200 hr of
operation. A test panel coated with Urecal (a urethane-based coating) was
installed in the housing in the area prone to highest erosion. After 300 hr
of operation, the welds supporting the panel were eroded through, but the
coating on the panel was undamaged.
Spray Tower; During the tests with the spray tower system, the
303 stainless steel spray nozzles and the mild steel lower spray header had
to be replaced at intervals of about 300 hr.
Mobile-Bed Scrubber; In the tests made with the TCA scrubber, the
weight loss of the plastic spheres has averaged about 10$ in 300 hr of operation.
During the same period, the weight loss of the bottom retaining grid was about
30$. In tests now being made, finer limestone (88$ through 325 mesh) is being
used in an effort to reduce erosion. A small test loop is also being assembled
in which test specimens can be studied under conditions more closely controlled
than is feasible in the scrubber.
458
-------
en
MIST
ELIMINATOR
m
i'V
l^/
....
P.
•
^
4
HUMIDIFICATION
INLET ^R
PI UF ^ ^
GAS w
PULVtKIZbD ' """'
LIMESTONE
U 9
if " H
SPRAY
r
J
MAKE-UP
WATER — ^
> & T '
r
f
y^pj j->
i ^
t
^
1 ,,
/
/
4
f
i
-WTP-^
i
i
FLUE GAS
TO
i ' STACK
i
.c5
^id
j
/•"<^ •
==-^>^ J r%. - ~ ^~ - j .a^TP-*'
^^S^^t'^dLl^- 2!1
SLURRY'^ RECIRCULATION RECIRCULATION'0" 1 SETTLING
PREPARATION TANK TANK ' TANKS
TANK
SOLIDS
DISCARDED
FIGURE 10
Pilot Plant with TCA Scrubber Configuration
-------
Particulate Loading in Exit Gas
The objective in the test program has been to remove fly ash from
the gas to the degree that a "clear stack" is obtained. The degree of dust
removal required to give a clear stack varies with several conditions, but
it seems generally accepted that a level of 0.02 gr/ft3 (dry) or less is a
reasonable objective.
In the TVA pilot plant tests, the particulate loading in the exit
gas from the scrubbers tested has averaged about 0.0} gr/scfd from the cross-
flow, 0.027 and 0.05 gr/scfd from the spray tower (for k.6 and 7.8 ft/sec gas
velocity, respectively) and 0.07 gr from the TCA scrubber (12-5 ft/sec velocity).
The particulate loading has been a linear function of the apparent gas velocity
in the barrel of the scrubber.
Petrographic analyses of the particles indicate that the material
in the exit gas is calcium carbonate or calcium sulfite - sulfate rather than
fly ash. Apparently, the fly ash is readily captured and removed in the first
stage of each scrubber configuration, even at the low pressure drop in the
mobile-bed scrubber (see Fart II). The particulate in the exit gas comes from
the mist of scrubber slurry that passes the chevron mist eliminator. Tests
are currently in progress to develop a more effective mist elimination system.
Tests of scrubber configurations in the pilot plant are continuing.
Evaluation of the spray tower and TCA scrubber have not been completed.
Summary
Of the three scrubber types tested, all should give adequate S02
removal if properly designed. As to reliability, the fixed bed (crossflow
type) is unacceptable because of scaling and plugging. The spray tower and
mobile-bed types avoid this problem, but each has other problems that must be
solved. The advantages and disadvantages of each, as indicated by the pilot
plant work, are listed in Table III.
Further tests are under way to continue evaluation of the spray tower
and mobile-bed types. Other scrubber configurations may also be tested. The
main remaining problems, on which special studies are planned, are erosion,
mist carryover, and slow settling of solids. Ways for avoiding unstable pH
at low limestone stoichiometry will also be explored.
460
-------
TABLE III
Evaluation of Scrubber Types
Spray Tower
Advantages
1. No internals for build up of mud
or scale
2. Infinite turndown
Spray nozzle and distributor
replacement only routine mainte-
nance item
Low pressure drop through tower
-AP-2.0 in. H20 or less
Mobile-Bed Scrubber (TCA)
Advantages
1. Gas velocity of 12-5 ft/sec
possible with good S0e recovery
2. Fly ash removal possible in
same unit
3- Slurry distribution may require
only open pipes with splash
plates
4. Excellent gas contacting
efficiency
Disadvantages
1. Gas velocity limited (however,
preliminary indications are that
the velocity may be increased
significantly without loss of
efficiency)
2. Separate fly ash removal system
required
3- High velocity (atomizing nozzles)
required for slurry distribution,
requring high-pressure pumps
4. Gas contacting efficiency rela-
tively low
Disadvantages
1. Some internals (support grids,
etc.) required that could promote
solids accumulation
Turndown limited
Bouncing balls and supporting grids
must be replaced at regular inter-
vals because of erosion
Pressure drop across scrubber of
9 in. or more required
Mist elimination problem aggravated
by high gas velocity
461
-------
II. EXPERIMENTAL DESIGN AND DATA ANALYSIS
FOR SPRAY AND MOBILE-BED SCRUBBERS
By
N. D. Moore
Division of Power Resource Planning
Tennessee Valley Authority
Chattanooga, Tennessee
As described in Part I of this paper, three scrubber types have
been tested in the TVA pilot plant studies on limestone slurry scrubbing of
S02« The fixed-packing type was abandoned because of scaling, leaving the
spray tower and mobile bed as the candidates for full-scale application.
In the tests of each type, short-term tests (open loop) were made to evaluate
effects of the major operating parameters and to determine the best combination
for extended closed-loop runs aimed at evaluating operational reliability.
N The basic experimental plan used for all the screening tests was
the 2 (N = 1, 2, 3, . . . ) factorial or fractional factorial. At the com-
pletion of each test series, the data were analyzed by analysis of variance
and regression analysis.
For both scrubber types, preliminary tests indicated that of the
several operating parameters, four had the most significant effect:
1. Liquid to gas ratio (L/G)
2. Gas velocity (v)
3. Pressure drop (AP)
4. Distribution of slurry between scrubber units
Spray Tower
The first test series consisted of 20 tests with the spray tower
as shown in Figure 6 in Part I. This was a 1/2 replicate of a 2 factorial
experiment, with four center points added. The five variables studied
and the range of each variable were as follows:
462
-------
Variable
Range
1. L/G to Ventri-Rod (L/GVR), gal/Mcf
2. Pressure drop across Ventri-Rod
3- L/G to first spray above venturi
4. L/G to second spray above venturi
5. Gas velocity, cfm
ft/sec
in. H20
20-40
7-15
0-10
0-10
2000-2700
4.6-6.2
The other operating conditions are given in Part I. Detailed test data
are listed in Table IV.
TABLE IV
Test Conditions and Results from Screening Tests
of Spray Tower Scrubbing (First Series)
Gas AP
Test flow, APvR, overall, L/G, gal/Mcf $ S02 removal,
No.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
cfm
270O
2000
2000
2000
2350
2700
2700
2000
2000
2350
2700
2000
2700
2000
2350
2700
2000
2700
2700
2350
in. H20
6.5
6.0
tt.5
15.0
10.0
13.2
13.9
15.6
7-2
11.1
7-4
15.2
15.0
7-4
11.3
7-0
7.0
15.0
7-2
11.6
in. H20
10.4
-
15.0
15-6
10.6
15.9
16.9
17.0
8.5
12.4
8.3
16.7
17.0
8.7
12.7
8.9
8.4
17.0
8.4
14.1
VR
4o
20
20
20
30
20
40
4o
4o
30
20
4o
20
4o
30
20
20
4o
40
30
SI
0
10
0
10
5
0
10
10
10
5
10
0
10
0
5
0
0
0
10
5
S2
10
0
0
10
5
10
10
0
10
5
10
10
0
0
5
0
10
0
0
5
solids
12
25
16
16
16
13
12
16
15
13
16
17
16
14
17
13
16
20
18
% average
64
55
58
68
58
64
71
60
59
59
60
61
60
53
56
50
59
66
57
60
Range
59-69
53-58
57-62
54-63
70.5-71.4
60.0-60.3
57-60
58-61
60-63
59-62
51-55
56-57
49-51
59
66
56-59
58-6!
463
-------
Although each variable had a statistically significant effect
on percent S02 removal, the two largest effects were the pressure drop
across the venturi and the L/G to the second spray above the venturi.
The regression equation derived from the data is as follows:
Predicted $ S02 removal =
59-9 + IT xx + 51 x2 + 15 X3 + 47 x4 + 19 X5 + 31 xs
T£ T5T T& TS 15" ib
where
X2 =
X3 = (L/Gsl-5)/5
X4 = (L/GS2-5)/5
X5 = (Gas velocity-2350)/350
X6 = Xi-Xs
This equation accounts for 90$ of the total variation of the data.
The remaining 10$ is composed of 3.0$ due to curvature (i.e., x^2) and J.0$
error. The analysis of variance is tabulated in Table V and the predicted vs
observed values are given in Figure 11. Based upon the analysis of the data,
95$ of the actual or observed values will be within 3$ of the predicted percent
S02 removal.
Two of the results from this test series were encouraging. First,
the contribution to S02 removal from the lower spray (si) was small compared
with other variables. Since this spray was located in a position just above
the venturi and was subject to severe erosion, removing it would simplify
operation without any great loss of S02 removal efficiency. Secondly, in-
crease in gas velocity, within the range studied, improved S02 removal slightly
rather than reducing it.
The S02 removal in the extended run made under closed- loop con-
ditions in the venturi - spray tower (Part l) checked with the predicted
removal from the open- loop screening tests within experimental error.
The second spray tower test series was run after the spray banks had
been raised to a higher level and a wash spray (53) installed at the top of the
scrubber. The series consisted of two blocks of tests composed of a full factorial
for two variables (22) plus one center point in the first block (Table Vl).
464
-------
TABLE V
Analysis of Variance for Screening Tests
a
of Spray Tower Scrubbing (First Series)
Source
Mean
L/GVR
L/G
S1
Velocity
; velocity
Residual
Lack of fit
Curvature
Interaction
Error
Total
Sum of squares
71,760.2
18.06
162.56
14.06
138.06
22.56
60.06
48.42
39.6750
13.6
26.0750
8-75
72,224
Degrees of
freedom
1
1
1
1
1
1
1
10
1
9
Mean
square
18.0625
162.5625
14.0625
138.0625
22.5625
60.0625
2.897) 2.9 (pooled estimate)
2.92 )
20
Correlation coefficient = 0.96; index of determination
b = °'92
All values significant at a = 0.01, where a indicates
degree of confidence (i.e., 0.01 indicates 1% possibility
of variable not being significant).
465
-------
S02 REMOVAL, %
55 60 65
OBSERVED S02 REMOVAL, %
FIGURE 11
Comparison of Predicted and Observed Results in Screening Tests of Spray Tower Scrubbing (First Series)
-------
TABLE VI
Test Conditions and Results from Screening Tests
a
of Spray Tower Scrubbing (Second Series)
Test
No.
1
2
3
5
6
7
8
9
APV
in.
7
7
15
15
11
15
15
15
15
b
R>
HgO
•3
.4
.1
.2
.2
.1
.0
.1
.0
AP
overall,
in.
10
10
17
18
14
18
17
18
Hao
•
.
»
,
*
*
*
*
18.
3
6
9
3
5
0
8
0
2
VRb
20
38
20
38
30
20
20
20
20
L/G
SI
10
10
10
10
10
0
20
0
, 20
S2
10
10
10
10
10
18
0
0
17
1°
S02
S3 removal
10
10
10
10
10
9
10
10
10
67
71
75
75
70
64
62
58
72
.1
.6
•9
•3
.1
•3
• 7
.8
.8
a Gas flow, 2700 cfm (6.2 ft/sec); L/G to humidi-
, fication spray, 5-5 gal/Mcf.
Ventri-Rod scrubber.
The variables studied and the range of each were as follows:
Variable
1. L/G to Ventri-Rod, gal/Mcf 20-38
2. L/G to SI, gal/Mcf 0-20
3. L/G to S2, gal/Mcf 0-l8
4. AP across Ventri-Rod, in- H20 7-15
All tests were conducted at the maximum system velocity of 2700 ft /min
(6.2 ft/sec). In the first block of five tests, the AP across the venturi
and the L/G to the venturi were varied.
the following equation:
Analysis of these data resulted in
S02 removal = 72 + 3AP + L/G - 1-5(L/G)(AP)
(2)
where AP = (APVR-ll)/4
L/G = (L/GVR-29)/9
Table VII gives the analysis of variance. The L/G was not significant over
the range tested.
467
-------
TABLE VII
Analysis of Variance for Block 1
of Second Test Series on Spray Tower
Degrees of Sum of Mean
Source freedom scuares square F value
Mean
Ap
L/G
(AP)(L/G)
Residual
Error0
Total
1
1
1
1
1
(12)
5
25,920
36
k
9
5
25,97^
p
36 12
4 i.y
9 ?
5 1.6?
3
Significant at a = 0.01.
Not significant or significant at ci> 0.10.
0 Error variance assumed to be 3.0 with 12 degrees
of freedom.
The next block of tests was run with L/G to sprays SI and S2
varied but with venturi AP and L/G held constant. The resulting equation
was
% S02 removal = 65 + 3-25L/G1 + 3.75L/G2 + l.asCL/G^L/G.,) (3)
where L/G^ = (L/G to Sl-lo)/10
L/G2 = (L/G to s2-9)/9
The analysis of variance is given in Table VIII.1
This block indicates that the two sprays are both important in
S02 removal but are independent and the greater the L/G to each the greater
the SO removal. Thus raising the sprays in the scrubber did increase the
effectiveness of the lower spray (Si).
From previous error estimates, the error variance is estimated to be 3.0
with 12 degrees of freedom (see Table v). This estimate is used as the
error throughout the Ventri-Rod - spray tower test work.
468
-------
TABLE VIII
Analysis of Variance for Block 2
a
of Second Test Series on Spray Tower
Degrees of Sum of Mean
Source freedom squares square F value
Mean
L/G!
L/G2
L/G-^L/GS
Total
1
1
1
1
4
16,770.25
42.25
56.25
6.25
16,875.00
42.25
56.25
6.25
b
14.08
18.75C
2.08
a Error variance assumed to be 3-0 with 12 degrees
, freedom.
Significant at a = 0.01.
Not significant or significant at a>0.10
A third series of tests (Table IX) were run to test the system
at the maximum velocity attainable, 3400 ft3/min (7.8 ft/sec). This group
consisted of a 1/2 replicate of a 25 factorial experiment with 4 center
points added for a total of 20 tests. Again, open- loop conditions were used.
Analysis of the test results gave the equation:
% S02 removal = 68-9 + O-QZL/Gy^ + 0.^6^/G^ + 2.J58L/G2 +
- i. 55(1/6^) (L/G) + o.95(VGvR)(ApvR) - i-93(L/G XL/G )
o J. &
- 0.85(L/G2) (L/G5) + 2.2(L/G2)(APVR) + l^L/GgXAP^) (4)
where
L/GVR = (L/G to Ventri-Rod-28)/7
L/GI = (L/G to si-io)/5
L/G2 = (L/G to S2-io)/5
L/G3 = (L/G to S3-io)/5
(AP to Ventri-Rod-12)/3
469
-------
Each of the variables in the above equation is significant at a = 0.2 or
lower. The analysis of variance is given in Table X, along with the
t - statistic for each variable.
TABLE IX
Test Conditions and Results from Screening Tests
of Spray Tower Scrubbing (Third
Test
No.
1
2
3
4
5
6
T
8
9
10
11
12
13
Ik
15
16
IT
18
19
20
AP,
VR
14.5
9
15
15
12
8.9
15.1
15.1
13-3
13.5
9.0
15.0
14.9
9-2
12.6
11.8
11.0
15.0
13.0
12.1
in. H20
Overall
19.8
19.1
21.6
2J.1
21-5
22.0
23.0
23.1
22.6
23.2
24.5.
l.8b
*'ll
2.0b
20.5
24.5
21.5
20.5
25.9
20.6
a
Series)
L/G
VR
20
35
20
35
28
35
35
35
20
28
20
20
35
35
28
20
35
20
20
28
SI
5
15
15
5
10
5
15
5
15
10
15
15
15
15
10
5
5
5
5
10
S2
15
5
5
15
10
5
5
5
15
10
5
15
15
15
10
5
15
5
15
10
!1
15
15
15
5
10
5
5
15
15
10
5
5
15
5
10
15
15
5
5
10
S02
removal
76.8
68.9
70.4
77-6
77-3
65.8
68.2
64.3
71-3
68.7
68.9
68.3
75-6
70.7
70.4
65.7
67-1
57-0
70.1
69.5
a Gas flow, 3400 cfm (6.8 ft/sec); L/G to humidi-
b fication spray, 4.4 gal/Mcf.
Inch Hg.
470
-------
TABLE X
Analysis of Variance for Third Test Series on Spray Tower*
Source
Regression
Error
Total
Degrees of
freedom
10
19
Sum of
squares
367-73
22.59
390.32
Mean
square
36.67
2.51
F ratio
Variable
t - statistic
Variable
t - statistic
L/G to VR
L/G to SI
L/G to S2
AP to VR
(L/GvR)(L/G3)
(L/GVR)(AP)
1-73
2.40
5.30
1.78
-3.69
1.71
(L/GX)(L/G2)
(L/6j)(L/Gg)
(L/G2)(AP)
(L/G3)(AP)
-lf.58
-2.08
Ml
3-21
Correlation coefficient = 0-97; index of determination
= °-^'
Significant at a = 0.01.
The most significant variables in the third series are pressure
drop across the venturi and the L/G to the upper spray, S2. These tests
indicate that although increasing the velocity above 2700 ft3/min (6.2 ft/sec)
does reduce the S02 removal, this effect can be offset by judicious choice
of slurry distribution. The following tabulation illustrates the effect of
velocity on S02 removal when the slurry distribution to the sprays is held
constant.
Velocity, APyR> L/G, gal/Mcf S02
ft/sec in. 1^0 VR Sll S2 S3_ removal, %
k.6
6.2
7-8
k.6
6.2
7-8
k.6
6.2
7-8
15
15
15
15
15
15
15
15
15
20
20
20
30
30
30
38
38
38
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
72
76
68
72
75
71
72
75
72
471
-------
Under these conditions, 2700 cfm gives the maximum removal. By
varying the distribution to the sprays, however, good results can be ob-
tained at 5^00 cfm--as shown in the following tabulation.
Velocity,
£t/seca
7-8
7-8
7-8
7-8
7-8
APVR,
in. H2(
9
9
15
15
15
L/G, gal/Mcf
) VR
35
20
20
20
35
SI
5
15
5
c;
15
s2
15
5
15
15
15
V i
tt
s
10
15
10
15
S02
removal, %
73
TO
76
74
78
3400 cfm.
In view of these results, an extended closed- loop run (165 hr) was
made at ^400 cfm, 15&VR, 20L/GVR> 5VGS1> 15L/GS2> and 15L/GSy The
average removal, however, was only 66$ as compared with the predicted 76$.
The cause of this is unknown; the removal was about J0% at the end of the
run (stopped to install a different type of scrubber), and perhaps would
have improved further if time had been available for extending the run.
Mobile-Bed Scrubber
In screening tests with the mobile-bed (TCA) scrubber, the variables
were gas velocity, L/G to the main slurry inlet (Si) over the mobile beds,
and L/G to the wash spray (S2) in the top of the scrubber. Test conditions
and results for the 16 tests are given in Table XI. In developing the equation,
test 5 was not included and tests 11 and 16 were averaged to give a total of
14 tests. The equation derived is as follows:1
(5)
% S02 removal = 64-58 + 5.04L/G2 + 0.855V- L/G! - 0.94l(L/Gi
where V = velocity (cfm) /100, L/G! = (L/G to Sl)/10,
and L/G2 = (L/G to S2)/10
The analysis of variance and t - statistics are given in Table XII. The
t - statistics for each variable are significant at ey = 0.05 or lower.
It should be remembered that all regression equations strictly apply only
within the experimental region and that extrapolation can be hazardous.
This is particularly evident in the TCA system as it is very easy to flood
the beds at certain conditions and in fact too large an L/G may be detri-
mental.
472
-------
TABLE XI
Test Conditions and Results for Screening Tests
on Mobile-Bed Scrubber
Test
No.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Gas
flow,
ft/sec
10.5
16.2
16.2
10.5
16.2
10.5
16.2
10.5
16.2
14.4
13.3
8.3
8-3
8.3
8-3
13-3
AP through
scrubber,
in. H20
7-1
16.5
19.6
5-6
24.0
9-0
16.7
9-0
18.6
22.0
7-7
6.1
5-4
8-5
7.7
7-2
L/G,
gal/Mcf
SI
12
4
12
12
32
52
4
52
12
35
20
15
15
65
65
20
S2
29
10
16
17
15
29
16
17
10
16
17
36
21
36
21
17
% S02
removal
81
77
82
71
74
95
76
95
79
94
84
86
79
89
86
77
L/G to humidification, 15 gal/Mcf.
It is interesting to note that velocity was not a significant in-
dependent variable in these tests. It will appear if the level of significance
is selected at a = 0.25, but the coefficient is positive indicating that
increase in velocity will increase S02 removal.
At the conclusion of this series, an extended closed-loop run
(172 hr) was carried out at conditions selected on the basis of the screening
tests (12.5 ft/sec, j4 L/G to main slurry inlet, and 14 L/G to the upper
wash spray). Average removal was 92$, close to that predicted.
473
-------
TABLE XII
Analysis of Variance for Screening Tests
a
on Mobile-Bed Scrubber
Degrees of Sum of Mean
Source freedom squares square F ratio
Regression 3 621.18 207-06 19-5b
Error (residual) 10 106.0^ 10.6
Total 13 727-22
Variable t - statistic
L/Gj, 2.9103
(velocity)(L/GS) 5-teiu
-2.6095
Correlation coefficient = 0-92
Significant at a = 0.01.
Particulate Removal
Since removal of particulate matter as well as S02 is required
in wet scrubbing with limestone slurry, small-scale tests were made to
determine the pressure drop necessary for good removal. This study was
conducted at the Widows Creek plant in order to have the same type of dust
as in the later full-scale scrubbers; no limestone was involved.
The equipment used to determine required pressure drop was the
"Dust Difficulty Determinator," a small orifice-type test scrubber (with
a following mobile bed) manufactured by Environeering, Inc. Composition
of the coal burned during the test period is given in Table XIII, particle
size of the ash in Table XIV, and results of the tests in Table XV.
Data analysis was performed on tests 16 through 22, 2k through 26,
29 through 42, and ^5. The first 15 were preliminary, obtained during the
shakedown period; the others were rejected because equipment difficulties
were believed to have made the results atypical.
474
-------
TABLE XIII
Composition of Coal Burned During Dust Removal Tests at Widows Creek
Total Weight percent, dry basis
apple
1
2
3
4
5
6
moisture, $
6.6
5-7
5-6
7.4
7-0
7-7
Volatile matter
28.4
31-5
38.2
31.0
31-2
32.2
Ash
20.7
21.0
17.4
18.3
18.1
19-1
Fixed carbon
50.9
47-5
44.4
50.7
50.7
48.7
Sulfur
1.1
2-9
U.I
2.5
3A
3.1
The analysis indicated that both L/G and AP affect the degree
of dust removal. The following equation was developed:
Y = 0.12(-AP)-°-54 (6)
where
Q
Y = outlet grain loading, gr/ft
L = liquid pumping rate, gal/min
G = gas rate, acfm x 10"
AP = pressure drop, in. H20
If the pressure drop across a venturi is defined as
AP = L . V1'87 (7)
G
where V is gas velocity (ft/min),
then substituting equation 7 in 6 gives
Y =
0.12G
_ _O.54
0. 54
L0.54
(L) .
475
-------
TABLE XIV
Bahco Particle Size Analysis of Fly Ash During Widows Creek Tests
Test No. i 2 3 31 32 33 3^ 35 36 1
Micron Inlet (6-28-71) Inlet (6-29-71) Inlet (6-50-71) Outlet (6-28-71)
less than % leas than % less than
0 J»C*«
3-3
6.5
9.6
1J.2
18.0
23.9
^ *
29.5
J2.4
/" *
3-1
13.0
*.j • **
27-6
48.2
69-5
85.0
89.0
90.3
«V 0 ** *•* A
4.6
20.7
r*w i
42.5
57-1
64-3
69-5
72.7
74.3
**»*
2-7
14.0
31.8
50.6
65-7
79-4
83.8
85.4
/ —
3.8
17-0
36.0
59-6
78.6
89.6
92-3
93-1
5-4
19.5
37.1
47-7
54.0
63.6
70.3
73-5
6.2
31.8
49.7
62.9
70.2
74.5
77.1
78.6
4-3
20.7
42.9
60.0
69.2
77-7
81.3
82.8
I'1?
18.4
41-3
63-7
79-9
87.7
89.7
90.4
4.2
20.6
40.8
51.7
56.1
5M
66.0
68.6
i.
9-
22.
35-
48.
65-
71.
76.
4
5
2
0
4
6
3
74 94.6 86.7 92.4 94.6 86.1 87.5 88.9 91-6 84.5 91-6
(Filter No. 702)
-------
TABLE XV
Dust Removal from Power Plant Stack Gas (Widows Creek Tests)
&PLS APHS
L/G, gal/Mcf
Inlet Outlet
loading, loading,
Run No.
1
2
3
4
5
6
7
8
9
10
11
12
13
tt
15
16
17
18
19
20
21
22
23
A
25
26
27
28
29
30
31
32
33
3^
35
36
37
38
39
to
41
42
43
1*
45
in. H20
_
-
-
3-6
3-9
4.0
6-5
3-1
0.8
1.6
3-0
3-o
4.0
4.9
4.0
2.2
1.8
2.1
2.1
2.1
-
-
_
-
-
_
0.2
0.2
0.2
1.2
0.8
0.4
1.7
3-2
3-2
2
2
2
_
_
_
_
_
-
in. H20
33
11
20
10
28
11
11
21
46
10.4
21.4
32
5
10.3
15
-
-
-
-
-
-
5-5
10.9
4.9
3-4
^3
4.4
-
-
-
_
-
-
-
-
_
_
-
_
4.9
4-5
4.2
8.8
10. 1
10.1
in. H20
33
11
20
15.6
31-9
15
17-5
24.1
46
11.2
23
35
8
14.5
19-9
4.0
2.2
1.8
2.1
2.1
2.1
5-5
10.9
4-9
3-4
M
4.4
0.2
0.2
0.2
1.2
0.8
0.4
1-7
3-2
3-2
2
2
2
4.9
4-5
4.2
8.8
10.1
10.1
LS
20
20
24
17
17
19
7.4
18
18
20
19
18
20
24
13
27
27.7
28.,3
20.2
17.5
12.8
-
-
-
-
-
-
2.6
3.9
5.2
5-2
3.9
2.6
2.6
5-2
7.8
4.7
9-4
13
-
-
-
-
-
—
HS
34
19
17
22
22
14
15
15
15
15.1
14.2
14.0
16
14
14
-
-
-
-
-
-
15.2
10.0
7.2
5-0
7-2
10.0
-
-
-
-
-
-
-
-
-
-
-
-
7.6
5.4
2.1
5.1
4.0
3.0
T
54
39
41
39
39
33
22.4
33
33
35-1
33.2
32
36
38
27
27
27-7
28.3
20.2
17.5
12.8
15.2
10.0
7.2
5-0
7-2
10.0
2.6
3.9
5-2
5-2
3.9
2.6
2.6
5.2
7.8
4.7
9-4
13
7-6
5-4
2.1
5-1
4.0
3.0
gr/scfd
1.20
2.00
5-40
2.70
0.62
2.30
2.00
2.00
0.34
1.68
1.40
1.57
4.0
3.0
2.8
5-9
3-2
2-9
2.7
2-9
3.0
1-9
4-5
5-9
6.01
4-53
3-59
6.2
2.2
3-9
6.6
4.0
3-7
3-4
3-5
2.7
2.4
3-7
3-3
2.2
-
5-2
5-5
5-2
4.4
gr/scfd
0.0068
O.0097
0.027
0.12
0.014
0.10
0,16
0.027
0.0032
0.006
0.010
0.010
0.007
0.006
0.006
0.011
0.009
0.009
0.025
0.007
0.031
0.006
0.068
0.016
0.013
0.018
o. 031
2.7
0.11
0.09
o.o4
0.044
0.13
0.075
0.028
O.o4£
o.o64
0.038
0.026
0.02
0.0246
0.0445
0.0743
0.0483
0.03
HS = orifice section of scrubber.
LS = mobile-bed section of scrubber.
T = total.
477
-------
Now let G = pAV, where A is the cross-sectional area.
Substituting G = pAV in equation 8 and approximating coefficients,
Y = 0.12[~Al (9)
Thus the outlet grain loading is a function of the cross-sectional area
and the resistance (i.e., L where L = liquid pumping rate) across the
area. The analysis of variance, for equation 6 as the regression model,
is given in Table XVI.
TABLE XVI
Analysis of Variance for Dust Removal Tests
Sum of Degrees Mean
Source squares of freedom square F_
Mean 362.52 1 a
Regression 191-97 1 191-97 75
Residual 62.29 2£ 2.56
Total 616.78 25
(R2= 0.755; R = 0.87)
Significant at a = 0.01.
To confirm the model (equation 6), data supplied by the Western
Precipitation Corporation from tests on various venturi scrubbers in the
western United States were compared with the Widows Creek data (Fig 12).
Good agreement was obtained.
In sampling at the Colbert pilot plant, the particulate in the
gas from the scrubber has been found to be mainly calcium salts picked up
in the scrubber (see Part l). The grain loading attributable to fly ash
(based mainly on petrographic examination of the total particulate and
estimation of the fly ash percentage) has been less than 0.02 gr/scfd.
Correlation with the Widows Creek model is shown in Table XVII. Agreement
was fairly good, considering the observed values for fly ash were estimated.
478
-------
.13
.12
.11
.10
.09
.08
r
.07
.06
.05
.04
.03
.02
.01
I i
• PREDICTED BY MODEL:
Y (DUST IN OUTLET GAS, GR /SCFD) • 0.12 X ~°'54
WHERE X - A P•L/G
* OBSERVED DUST IN OUTLET GAS
* STANDARD DEVIATION
•O
-o
X
I
10
15 20 25
60 65 70 75 80 85 90 95 100
30 35 40 45 50 55
X
FIGURE 12
Correlation of Observed and Predicted Dust Removal Data from Widows Creek Tests
Data from the Western Precipitation Corporation have been added (denoted by No. 1-12)
-------
TABLE XVII
Correlation with Widows Creek Model
APyRa Total L/G,b
in. H20
6.5
6.0
11*. 5
15.0
X
10.0
13.2
13.9
^ *
15.6
— ^x • v
7.2
11.1
7.1*
15.2
15.0
7-1*
11.3
7.0
7-0
15.0
7-2
11.6
15.1
gal/Mcf
56
38
28
1*6
36
66
58
68
1*6
1*6
58
36
1*8
1*6
26
38
1*6
56
1*6
68
L/G- Ap
361*
228
1*06
720
1*60
1*75-2
917.^
901*. 8
1*89.6
510.6
31*0.1*
881.6
51*0.0
355.2
519.8
182
266
690
1*03.2
533-6
1026.8
Fly ash in
outlet gas,
gr/scfd
Observed
0.0101*
0.0093
0.0063
0.001*0
0.001*7
0.0033
0.0037
0.0138
0.001*8
0.0031
0.0041
0.0059
0.0030
0.0027
0.001*2
0.0028
0.0032
0.0023
0.0027
0.0012
0.0016
Predicted^
0.0050
0.0063
0.001*7
0.0031*
0.001*1*
0.001*3
0.0030
0.0031
0.0042
0.001*1
0.0051
0.0031
0.001*0
0.0050
0.001*1
0.0072
0.0059
0.0035
0.001*7
0.001*1
0.0028
a Pressure drop in Ventri-Rod scrubber.
b Total slurry flow to venturi and spray tower.
C Based on Widows Creek model:
Y (outlet loading, gr/scfm) = 0.12 X'0-54
where X = L/G • AP
480
-------
SULPHUR DIOXIDE REMOVAL BY LIMESTONE SLURRY
IN A SPRAY TOWER
A. Saleem
D. Harrison
N.. Sekhar
The Hydro-Electric Power Commission
of Ontario
T or ont o, Canada
Presented At
Second International Lime/Limestone
Wet Scrubbing Symposium
New Orleans, USA
November 8 - 12, 1971
481
-------
SULPHUR DIOXIDE REMOVAL
BY LIMESTONE SLURRY IN A SPRAY TOWER
A. Saleem
D. Harrison
N. Sekhar
The Hydro-Electric Power Commission of Ontario
Toronto, Canada
ABSTRACT
A MOOO-cfm pilot plant for removing sulphur dioxide
from power plant flue gases has been operated for one
year. The sulphur dioxide was removed by contacting
flue gas from a coal-fired boiler with a limestone
slurry containing 12 per cent by weight of solids in
water in a spray tower. Automatic control was used
to maintain the pH of the slurry between 5*6 and 6.0.
A spray tower was selected because of its simplicity
of operation, and because it has a minimum of solid
surfaces for scale accumulation. The variables
studied included shell velocity, liquid to gas ratio,
pressure at the spray nozzles and the slurry reten-
tion time. The main variables affecting the sulphur
dioxide removal efficiency were gas velocity and
liquid to gas ratio. During a continuous 1000-hour
run, deposits that formed on the gas inlet duct and
on the demister were removed by water jets. The
results indicate that a spray tower can be operated
with a high degree of reliability and that a sulphur
dioxide removal efficiency between 70 to 80 per cent
can be achieved using from 20 to 30 per cent above
the stoichiometric amount of limestone.
482
-------
INTRODUCTION
Ontario Hydro, like most North American utilities,
is faced with the problem of finding ways to control pollutant
emissions from fossil-fuel power plants. Considerable research
and development work is being directed toward finding a process
to control sulphur oxides emissions and the limestone slurry
process has been selected as the most promising for immediate
development. This process is being studied in a ^OO-cfm
pilot plant which draws flue gases from a 300-MW coal~fired
boiler, after the electrostatic precipitator. This paper des-
cribes the results of recent pilot plant studies.
The sorption of sulphur dioxide by aqueous limestone
slurry constitutes a three-phase system, namely, gas, liquid
and solid. When sulphur dioxide is transferred into the liquid
phase, hydrogen and sulphite ions are formed. Calcium carbonate
is also transferred to the liquid phase where it is partially
ionized to form calcium and carbonate ions. The calcium and
sulphite ions combine chemically to form sparingly soluble
calcium sulphite which crystallizes as a separate solid phase.
Oxygen in the flue gas is transferred into the liquid together
with the sulphur dioxide and partially oxidizes the sulphite
resulting in the formation of calcium sulphate, which also
crystallizes because of its low solubility. In the presence
of a high concentration of sulphite ions in solution, the
sparingly soluble calcium sulphite is converted to the more
soluble calcium bisulphite. The equilibrium between sulphite
and bisulphite ions is reversible and pH controlled.
483
-------
The transfer of sulphur dioxide is believed to be
controlled by the slow rate at which calciuia carbonate dis-
solves. To overcome this limitation, the slurry contacted
with a unit volume of gas must contain calciuia carbonate sub-
stantially in excess of the stoichiometric requirement. With
a given per cent solids in the slurry, this can be accomplished
by having either a large liquid hold-up in the scrubber or by
having a large liquid to gas ratio. The former approach is
less desirable because it not only increases the pressure drop
across the scrubber but can also lead to the formation of
highly supersaturated solutions of calcium sulphate and sul-
phite with the consequent formation of hard scale over the
scrubber surfaces. It therefore appears more desirable to
achieve the desired rate of mass transfer by using a high
liquid to gas ratio and a spray tower is particularly suitable
for this.
Reliability is the key factor affected when a
scrubber plugs due to the deposition of solids. Plugging by
solids can occur in two ways, namely, by settling of solids in
parts of the equipment where the slurry is relatively stagnant,
or by the crystallization of dissolved salts from supersatu-
rated solutions on internal surfaces. Our experience indicates
that deposit formation due to settled solids can be controlled
by irrigation of the affected surfaces with water.
The crystallization of salts from supersaturated sol-
utions can be a much more difficult problem to control. Both
calcium sulphate and sulphite have strong tendencies to form
supersaturated solutions from which the salts crystallize in an
484
-------
unpredictable way. If crystals form on the internal surfaces
of the scrubber, a very hard, cenacious film is formed which
tends to promote further crystallization. To control this type
of scale formation, it is necessary to keep the concentrations
of dissolved calcium sulphite and sulphate as low as is practi-
cal which is probably at or near the saturation level. In
practice, this can be accomplished by maintaining a very high
liquid to gas ratio.
Scrubbers that are designed for high gas sorption
efficiency at low liquid rates, such as packed towers and
flooded bed and tray types, are therefore considered unsuitable
for use with the limestone slurry process, unless very high
liquid rates are maintained, in which case the main advantage
of these devices is lost. In the spray tower where very high
liquid rates are needed to increase gas sorption efficiency,
there is very little possibility of forming highly supersatu-
rated solutions with attendant scaling problems. This has been
verified in pilot plant testing where no hard scale formation
has occurred during a year of operation, which included a 1000-
hour continuous run. On the other hand, hard scale formation
has frequently been observed in flooded bed and tray-type
scrubbers.
When large liquid to gas ratios are used, coupled
with low sulphur dioxide concentrations, the equilibrium back
pressure of sulphur dioxide over the slurry surface is negli-
gible. For this reason, a true countercurrent-type scrubber
is not essential and a single equilibrium stage is adequate to
obtain a high mass transfer rate.
485
-------
A spray tower offers further advantages in handling
large volumes of flue gas such as those from power plants. It
has the highest turn down ratio and the lowest pressure drop
of the alternatives considered. Further, it has a minimum of
internal surfaces and so reduces the possibility of plugging
even if scale formation does take place. Although its mass
transfer characteristics are not as good as those of the other
scrubbers, it is considered adequate in this respect.
DESCRIPTION OF PILOT PLANT
Figure 1 is a diagram of the pilot plant, a detailed
description of which is contained in Appendix I. Four thousand
cfm of flue gases at 260 F from a coal-fired boiler were drawn
off after the electrostatic precipitator and passed through a
spray tower and then to the stack. The spray tower was 32 inches
square, 16 feet high, and contained 6 banks of sprays, each
fitted with 9 full-cone spray nozzles. The slurry, containing
about 12 per cent by weight of solids, was recirculated by cen-
trifugal pumps from two interconnected 500-gallon tanks contain-
ing a total of about 900 gallons of slurry. A pH sensing probe,
located near the pump feed on the reclrculating tank, switched
on the dry limestone feeder and the clarified liquor puinp so that
a mixture of fresh limestone and clarified liquor from the
settling tank was added to the slurry recirculating tank each
time the pH fell below the set point of 5.8. This fresh feed
caused an overflow of spent slurry from the recirculating tank
to the settling tank. Fresh make-up water was used to wash the
demister and to flush the pump seals. There was no liquid waste
stream.
486
-------
The sulphur dioxide concentrations were measured by
an infrared analyzer. The operating variables studied included
the gas velocity, liquid to gas ratio, nozzle pressure and the
slurry retention time.
RESULTS AND DISCUSSION
The results are expressed in terms of the overall
mass transfer coefficient K , the number of transfer units N
ga7 og
and the overall efficiency of sulphur dioxide removal. In the
calculations, it was assumed that the equilibrium back pressure
of sulphur dioxide over the slurry surface was negligible. The
following formulae were used to calculate K and N values:
ga og
Nog = ln Vy2
Kga = (GVV) ln Vy2
where: G* is the molar flow rate of flue gas
V is the effective volume of the spray tower,
y1 and y~ are the inlet and outlet concentrations of
sulphur dioxide.
Effect of Gas Velocity; The performance of a single bank of
sprays was studied under various gas and liquid flow rates.
The liquid flow rate was varied by changing the nozzle pressure.
Figure 2 shows the effect of gas velocity on the mass transfer
coefficient K under various liquid flow rates when using the
ga
spray bank having nine 1/M—inch orifice nozzles. It can be
seen from Figure 2 that K increases with increasing gas velo-
ga
city. The strong dependence of K on gas velocity indicates a
ga
487
-------
significant gas phase resistance to mass transfer. Figure 3
shows similar data for another spray bank containing nine
1/2-inch spray nozzles. Because of the larger orifices, more
slurry could be sprayed at a given nozzle pressure than with
the lA-inch nozzles. The marked increase in K_Q with gas
6<*
velocity can also be seen in Figure 3. It is interesting to
note that when gas velocity is increased at constant liquid
flow rate, the liquid to gas ratio decreases, but in spite of
this decrease, the mass transfer coefficient shows an increase.
Increasing the gas velocity not only decreases the gas phase
resistance, but probably decreases the liquid film resistance
as well by inducing some surface regeneration of the liquid
droplets.
Effect of Nozzle Pressure and Size: The effect of nozzle
pressure was studied by comparing the performance of a bank
with lA-inch orifice nozzles with that of a bank with
1/2-inch orifice nozzles. The results, given in Table I, show
that under similar gas and liquid flow rates the mass transfer
coefficient K is practically unchanged in the nozzle pressure
ga
range of 10 to 20 psig. Further, the two-fold change in orifice
size did not affect the mass transfer coefficient.
488
-------
TABLE I
EFFECT OF NOZZLE SIZE AND PRESSURE OH
MASS TRANSFER COEFFICIENT AT CONSTANT LIQUID FLOW RATES
Run
No
1
2
3
*f
5
Nozzle
Orifice
Size (in)
1/2
1A
1/2
1A
1/2
6 1A
Effect of Liauid
PN
(psig)
10
20
10
20
10
20
Flow
L
(IGPM)
^7.5
1*6.5
*f 6.0
W.5
*f 6.0
*f 6. 5
hate: The
VQ
(fDS)
^.79
^.79
7.19
7.19
9A7
8.97
liquid
L/G
( I gal/
1000 cu ft)
23.65
23.15
15.23
15. MO
11.98
12.3^
flow rate
V
(Ib moles/hr
cu ft atra)
2.09
2.06
2.26
2.30
2.75
2.83
at a given
gas velocity was varied by operating various spray banks simul-
taneously. Figure If shows the effect of liquid flow rate on
mass transfer coefficient K _ at a constant gas velocity of
ga
8.6 ft/sec. It can be seen from Figure *f that the liquid flow
rate has a marked effect on K . The increase in 1C is due to
ga ga
increased gas-liquid interfacial area and the greater volume of
slurry present to dissolve and react with the gas. Figure 5
shows the number of transfer units as a function liquid flow
rate. Extrapolation of the linear part of the plot gives the
order of magnitude of the wetted wall effect. Figure 6 shows
the overall efficiency of sulphur dioxide removal as a function
of liquid to gas ratio at a gas velocity of 8.6 ft/sec.
489
-------
Slurry Retention Time: Retention time is defined as the time
the slurry is retained in the recirculation tank before being
returned to the scrubber. Some retention time is necessary to
eliminate any supersaturation which may be present in the liquid
at the scrubber outlet. The recirculation of supersaturated
solutions increases the probability of scaling. The length of
retention time also affects the pH recovery of the slurry. A
plot of typical pH recovery of a sample of slurry taken at the
scrubber outlet is shown in Figure 7. The exact shape of this
curve is dependent upon the nature of the limestone, its parti-
cle size, the solids concentration in the slurry and the amount
of sulphur dioxide absorbed per unit volume of slurry. Some
experiments were also conducted to study the effect of pH
recovery on the mass transfer rate. The results are shown in
Table II. A decrease in mass transfer coefficient with
decreasing retention time was noted, probably due to the
lowering of the pH of recirculating slurry.
TABLE II
EFFECT OF RETENTION TIME ON MASS TRANSFER RATE
Run
1
2
£
6
7
8
Cmins )
17.2
8.6
|J'5
5.7
2.9
2*2
L
(IGPM)
*iO rt
M-Q f\
99.0
99.0
Hf6*5
196.5
196.5
(fSs)
7.19
7.19
7-19
7.19
7.19
7.19
7.19
7.19
K
Clb moles/hr cu ft atm)
1.^-2
1.20
1.98
1.75
2.98
2.59
3.72
3-18
490
-------
General Operating Experience
Operation of the pilot plant was automatic except for
filling the limestone feeder hopper and discharging the spent
sludge from the settling tank. The process was controlled by
monitoring the pH of the slurry. The limestone feed was auto-
matically turned on when the pH in the slurry recirculation
tank dropped below 5.8 and was turned off when the pH was 5.8
or more. The limestone consumption was from 1,2 to 1.3 times
the stoichiometric amount. A typical composition for the
slurry is given in Table III.
During the pilot plant operation, which included a
1000-hour continuous run, deposits accumulated in the demister
and in the flue gas inlet zone. These deposits were soft and
were easily removed by water jets or mechanical scraping. The
chemical compositions of the deposits are given in Table III.
The deposit in the demister zone contained mainly calcium sul-
phate along with calcium carbonate and fly ash. The deposit in
the gas entry zone contained about equal amounts of fly ash and
calcium salts. The scrubber shell was found to be remarkably
free of scale and deposits. The spray banks were also free of
deposits except for the top two banks which were used infre-
quently. Deposits usually occurred on surfaces which were not
adequately irrigated with slurry or water.
The bulk settling rate of the spent slurry was
approximately three inches per hour and gravity settling gave
a sludge containing about 35-l+0 per cent solids by weight.
This could be dewatered to approximately 80 per cent solids by
491
-------
vacuum filtration. The sludge could be discharged directly
into a settling pond but, if the sludge has to be transported
for disposal, some conditioning might be necessary to improve
its handling properties. Even after vacuum filtration, the
sludge is serai-plastic and when jarred or vibrated forms a
uniform levelled mass.
The average pressure drop through the spray tov/er
including the deraister was about 0.5 to 1.0 inch of water.
TABLE III
CHEMICAL COMPOSITION OF SOLIDS IN
THE SLURRY AND THE DEPOSIT SAMPLES
Typical
Recirculatj
Demister Deposit
Gas Inlet Deposit
; Slurry
sit
lOSit
Calcium
Sulphate
30
7^
30
Calcium
Sulphite
Per Cent
56
—
if
Calcium
Carbonate
bv Weight
12
3
15
Balance
ash, etc
2
23
51
COirJLUSIONS
• A spray tower is suitable for scrubbing large volumes of flue
gases with limestone slurry. From 70 to 80 per cent of the
sulphur dioxide can be removed from the flue gases and,
although considerable energy is needed for pumping the slurry,
the gas phase pressure drop across the tower is very low.
• The mass transfer coefficient is a direct function of both
gas velocity and the liquid flow rate. The increase in mass
492
-------
transfer coefficient with increasing gas velocity is of
particular interest and may lead to reducing the size of the
spray tower.
• Deposit formation in the scrubber can be prevented by irri-
gating surfaces with slurry or make-up water and by limiting
the degree of supersaturation in the liquid phase.
• The consumption of limestone can be kept within 1.2 to 1.3
times the stoichiometric requirement by maintaining the pH
of the recirculating slurry between 5.6 and 5.8.
493
-------
LIST OK SYMBOLS
a - gas-liquid interface, sq ft/cu ft of active
scrubber volume
G - gas flow rate, cu ft/hr at 120 F
G* - molar gas flow rate, Ib moles/hr
K - mass transfer coefficient, Ib moles of S0p/hr cu ft
E of scrubber volume/atm of driving force
L/G - liquid to gas ratio, I gal/1000 cu ft of flue gas
at 120 F
N - number of transfer units
Pfl - nozzle pressure of the slurry, psig
Rt - retention time, defined as the time the slurry is
retained in the recirculation tank before being
returned to the scrubber, minutes
VG - gas velocity through the scrubber shell, ft/sec
y« ,y2 - concentration of sulphur dioxide in the gas at the
scrubber inlet and outlet, ppm by volume
494
-------
APPENDIX I
PILOT PLANT DESCRIPTION
Figure 1 is a diagram of the pilot plant. Flue gas
was drawn from the main duct, after the electrostatic precipi-
tator, and through the scrubber shell by an induced draft fan
(F1). The flue gas entered the scrubber at about 260 F. After
scrubbing it was cooled to about 120 F and discharged to the
atmosphere from a 25-foot-high chimney. The gas flow rate was
regulated with a damper (D1) and was monitored with a pitot
tube. The 32-inch-square scrubber was constructed from 1/16-
inch-thick 316 stainless steel sheet. The shell was assembled
in 7 flanged sections with a total height of 16 feet. The
bottom section was conical and a chevron-type demister was
installed in the top section. The slurry was sprayed into
the tower from 6 spray banks (S1-6). The top M- banks sprayed
counter-currently to the gas flow and the bottom 2 sprayed
cocurrently. Each bank was equipped with 9 full-cone spray
nozzles arranged to completely cover the tower cross section
with spray. The top bank had 1/2-inch orifice nozzles and the
remaining banks had 1/V-inch orifice nozzles. These nozzles
when operated under 10-20 psig pressure gave a spray angle of
about 80 degrees with an average drop size of 2500 microns.
The slurry flowed out of the tower through the bottom conical
section and a 6-inch discharge pipe which was immersed in the
slurry tank (T1) to provide a liquid seal.
495
-------
The slurry, containing about 12 per cent solids,
was recirculated from 2 interconnected 500-gallon tanks (T1 &
T2) containing about 900 gallons of slurry. This capacity was
chosen to provide a minimum of 3 minutes delay time in the tank
for the recirculating slurry. The slurry in the tanks, kept in
suspension by paddle-type agitators (A1 & A2), was recirculated
by three centrifugal pumps (P1-3). The slurry flow through the
banks was regulated with diaphragm valves and metered through
orifice plates.
Limestone was added to the recirculating slurry auto-
matically using a pH control system. A pH sensing probe, placed
near the pump suction in tank (T2), operated a screw-type lime-
stone feeder (F1) and the clarified liquor recirculating pump
(1%). The dry limestone feed was mixed with the clarified
liquor in a mixing cone (C1) and introduced near the bottom of
tank (T1). Spent slurry was discharged from an overflow in
tank (T1) into the settling tank (T3). The settled sludge
containing 35-l*O per cent solids was pumped into the station's
settling pond.
The sulphur dioxide in the flue gas before and after
the scrubber was monitored with an infrared analyzer. The
moisture content of the incoming gas was determined by wet and
dry bulb temperature readings. The make-up water was used for
deraister washing and flushing pump seals.
496
-------
APPENDIX II
TABLE IV
SUMMARY OF DATA OBTAINED WITH BANK NO
(NINE 1 A-INCH ORIFICE FULL-CONE SPRAY NOZZLES. COUNTERCURRENT
Ron
No_
1
2
3
it
5
6
7
8
9
10
11
12
PN
(psie)
25
25
25
20
20
20
15
15
15
10
10
10
L
(IGPM)
52.0
52.0
52.0
if6.5
W.5
»f6.5
M-1.5
M.5
M-1.5
33.5
33.5
33.5
VG
+7.7
38.7
36.5
Uo.8
32.7
29.95
V
(Ib moles/hr
cu ft atm)
2.20
2.5^
2.98
2.06
2.30
2.8M-
1.86
2.11
2.*f5
1.51
1.72
1.92
497
-------
APPENDIX II (Cont'd)
TABLE V
SUMMARY OF DATA OBTAINED WITH BAM NO 6
(NINE 1/2-INCH ORIFICE FULL-CONE SPRAY NOZZLES. COUNTERGURRENT
Run
PN
L
JSfi_ (psie) (IGPM)
1
2
3
4
5
6
7
8
9
10
11
12
25
25
25
20
20
20
15
15
15
10
10
10
86.0
86.0
89.5
78.0
7^.5
78.0
60.0
61.5
64.0
47.5
46.0
46.0
T /r S00 Cone
v VG 2
VG (I eal/ (ppm)
(fps) 1000 cu
4.8
7.2
9.5
if. 8
7.2
9.5
4.8
7.2
9.5
if. 8
7-2
9.5
42.8
28.5
22.4
38.9
24.9
20.3
29.9
20.4
16.7
23.7
15.2
12.0
% so2
ft) In Gas Exit Gas Removed
1500
1600
1525
1500
1560
1620
1500
1560
1620
1450
1560
1600
TABLE VI
SUMMARY OP DATA OBTAINED OPERATING VARIOUS
Run
1
2
3
if
5
6
L
(IGPM)
50
100
136
186
232
282
VG
(fps)
8.6
8.6
8.6
8.6
8.6
8.6
VG
(I gal/
1000 cu ft) ]
13.9
27.7
37A
51.5
64.2
78.1
S02 Cone
(ppm)
In Gas Exit G^g
1020 600
1020 410
1020 340
1030 290
1030 210
1040 170
550
780
725
560
780
810
650
840
900
700
925
970
63.4
51.2
52.5
62.6
50.0
50.0
56.6
46.1
i.i. cr
fT. J
51.7
li^\ *7
39.3
V
(Ib moles/hr
cu ft atm)
2.87
3.10
4.09
2.83
2.99
3.80
2.40
2.66
3.23
2.09
2.26
2.75
BANKS SIMULTANEOUSLY
% so2
Removed
59.8
66.7
71.9
79.6
83.7
N
0.53
0.91
1.10
1.27
1.59
1.81
V
(Ib moles/hr
2.75
4.71
5.68
6.54
8.71
9.36
498
-------
<£>
FIGURE 1
ANT FOR FLUE GAS
-------
3.5
I- 3.0
t
8
111
O
2
3
2.5
2.0
1.5
PN L
O 25PSIG 52 IGPM
A 20 PSIG 46.5 IGPM
o 15 PSIG 41.5 IGPM
• 10 PSIG 33.5 IGPM
678
VG-FPS
10
FIGURE 2
EFFECT OF GAS VELOCITY ON MASS TRANSFER COEFFICIENT
AT DIFFERENT LIQUID FLOW RATES AND NOZZLE PRESSURES
(BANK S-4. COUNTERCURRENT)
(ORIFICE DIA OF NOZZLE —1/4")
500
125098-RD
-------
<
u
o:
LJ
o
m
j
4.0
3.5
3.0
2.5
2.0
1.5
PN
678
VG-FPS
10
FIGURE 3
'EFFECT OF GAS VELOCITY ON MASS TRANSFER COEFFICIENT
AT DIFFERENT LIQUID FLOW RATES AND NOZZLE PRESSURES
(BANK S-6, COUNTERCURRENT)
(ORIFICE DIA OF NOZZLE = 15/32")
501
125103-RD
-------
VG = 8.6 FPS
PN ~ 15 PSIG
ISO 200
L - IGPM
250
300
FIGURE 4
EFFECT OF LIQUID FLOW RATE ON MASS TRANSFER COEFFICIENT
AT CONSTANT GAS VELOCITY AND NOZZLE PRESSURE
502
125097-RD
-------
1.5
1.0
O)
o
0.5
VG - 8.6 FPS
100 200
L- IGPM
300
FIGURE 5
EFFECT OF LIQUID FLOW RATE ON THE NUMBER OF TRANSFER UNITS
AT CONSTANT GAS VELOCITY AND NOZZLE PRESSURE
503
125096-RD
-------
75
3
UJ
a:
50
25
VG = £.6 FPS
^ ISPSIG
25 50 75
L/G - I. GAL/1000 CU FT (I20°F)
100
FIGURES
EFFECT OF L/G ON THE EFFICIENCY OF SULPHUR DIOXIDE REMOVAL
AT CONSTANT GAS VELOCITY AND NOZZLE PRESSURE
504
I25IOI-RD
-------
I
a
SOLIDS - 14.0% BY WT OF SLURRY
CaC03 - 1.68% BY WT OF SLURRY
20 30
TIME- MINS
FIGURE 7
pH RECOVERY OF SPENT SLURRY
505
I25IOO-RD
-------
THE MOHAVE/NAVAJO PILOT FACILITY
FOR
SULFUR DIOXIDE REMOVAL
BY
J. L. SHAPIRO AND W. L. KUO
Bechtel Corporation
Vernon, California
Presented At
Second International
Lime/Limestone Wet Scrubbing Symposium
New Orleans, Louisiana
November 8-12, 1971
507
-------
I. INTRODUCTION
The Mohave/Navajo S02 Removal Research Program is a project for testing
alkali absorption processes on a pilot-size basis. The major thrust of
the work is to determine the characteristics of these processes as applied
to boilers fired with low-sulfur coal.
The program was initiated in June 1970. The owners of the Navajo Generat-
ing Station, which is under construction near Page, Arizona, asked the
Bechtel Corporation (engineers-constructors of the Station) to survey
the status of S02 removal technology for application to the Station.
Schedules call for initial operation of each of three 770 Mw units in
1974, 1975, 1976, respectively.
Coal dedicated to Navajo will come from the Black Mesa mine. Surveys
have indicated that over the life of the plant, this coal will have
sulfur content varying from about 0.3% S to 0.8% S with an average of
about 0.5%. The heating value of this sub-bituminous coal is about
11,000 BTU/lb.
We need not, for this audience, describe the state of technology that
we observed in the Summer of 1970. However, it is worthy of note that
there was considerable surprise evidenced in the industry that utilities
with such very low sulfur coal were seriously considering sulfur dioxide
removal. The concerns of the industry had been directed towards reducing
flue-gas concentrations of sulfur dioxide to levels approaching the
508
-------
400 ppm that are the expected values for Navajo at the boiler outlet.
Therefore, relatively little was known concerning the performance that
could be achieved, the operational problems that might be expected
when attempting to reduce concentrations of SO,, from 400 ppm on down,
or the economic trade-offs involved.
Incidentally, events during the past year have shown that while Western
utilities have access to low-sulfur coals, they may still be required by
state emission regulations to apply flue-gas absorption equipment.
Another feature distinguishes the approach adopted for the Navajo Station
from many others. High efficiency particulate removal equipment will be
located upstream of the S02 absorption equipment, so that the designs
and constraints of the system will be based almost completely on S0?
absorption.
These facts shaped the basic nature of the research program that we
describe here.
The characteristics of the Navajo Station are very similar to those of
the Mohave Generating Station. The latter consists of two 770 Mw
units which also burn Black Mesa coal. Both units of the Mohave Station
have been completed. Therefore, the owners of the two stations have
joined in sponsoring this experimental program located at Mohave.
Southern California Edison Company was selected by the participating
owners as Project Manager of the program.
509
-------
A flexible, dual-loop system was designed and installed at Mohave.
Construction was completed in July, 1971. As it was desired to test
as many different types of absorbers as possible in a short time,
two identical testing loops were built, each with individual controls
and instrumentation. In each loop provision was made for installation
of two absorbers in parallel with ducting that could pass the flue
gas through either one. At the present time, the four absorbers
installed contain a single stage venturi, a turbulent contact absorber,
a combined Lurgi-impingement tray, and a low pressure-drop egg-crate
packing, respectively.
The Mohave units have electrostatic precipitators that are operating
<
at very high efficiency. Ducting inlets to the test loops were
placed both upstream and downstream of the precipitators so that the
effects of particulate loading can be tested. As we shall show, the
process loops are designed to simulate on a small scale, the operations
that a full-scale unit may be required to undergo.
Downstream of each absorber is an electrically powered flue-gas reheater,
installed mainly to enable measurement of the demister efficiencies.
The present program plan provides for testing three absorbents -
limestone slurry, lime slurry, and soda-ash solution. Parametric
variations of L/G, gas flow rate, and pH are made, while measuring
SC>2 absorption efficiency and particulate removal. In addition
the possibility of regenerating the soda-ash absorbent (using lime)
is being investigated.
510
-------
II. PILOT PLANT DESCRIPTION
The pilot plant facility is constructed on three skid-mounted platforms
as shown in Figures 1 and 2. It accommodates the four previously men-
tioned scrubbers and the two identical loops which take a flue gas
sidestream of 500 acfm to 4000 acfm from the Mohave Station (Unit 1)
exit duct. A 10' x 50' trailer equipped with control panels is at
one quadrant of the pilot plant. The trailer also contains a wet chemistry
laboratory with chemical analysis equipments and an atomic absorption
spectrometer, permitting analyses on-site.
Figure 3 shows the flow diagram. The sample flue gas is withdrawn, by
the suction generated by a 20" fan, from either upstream or downstream
of the electrostatic precipitator. Special care was taken in designing
the sampling valve for the gas intake at the electrostatic precipitator
upstream line in an attempt to obtain representative fly ash concentra-
tion at all flow rates. At the downstream of the precipitator, since
the fly ash distribution consists mainly of small particles, the gas
is withdrawn through a simple 14" collecting elbow parallel to the main
gas flow. The gas flow rate up to 4000 acfm automatically is controlled
by an orifice flow controller through a butterfly valve in the gas line
to the blower.
Upon entering the gas system of the pilot plant, the flue gas is auto-
matically monitored for gas temperature, pressure at many key points,
and S02, NOX and Oo concentrations at locations in and out of the scrubber.
Dust loadings of the feed stream and the effluent stream of the scrubber
are periodically measured by manually collecting gas samples.
511
-------
The flue gas from the scrubber passes through an electric preheater
which is controlled by a temperature controller. Heat input of the
reheater is accurately monitored so that by heat balance, the quantity
of entrained liquid can be calculated. The gas is then discharged
through a short stack.
For the measurement of S(>2, NOX and 0£ concentrations, the on-line
monitoring system has its own gas circuit shown in Figure 4. The gas
is sampled through a microporous dust filter in the collecting point
and a condensate trap prior to entering into the electrochemical cells.
Shutoff valves and air-back-purge are provided for cleaning the filter
if necessary. Electric time and solenoid valves are used to switch
alternately from one gas to the -other so that one set of analyzers can
be used to monitor both feed and effluent streams. Standardized gas
mixture is used to calibrate the electrochemical cell periodically.
For the measurement of dust loading in the gas in and out of scrubber,
fiber glass filters for total mass and inpaction-type particle classi-
fiers for size distribution are employed.
The four types of absorbers to be tested are as follows:
A. Turbulent Contact (UOP)
A 19-inch ID by 27-foot height absorber, with three trays of
mobile packing (each tray containing 850 of pinch ping-pong
balls) and a 24-inch-square demister section. It is rated at
1500 cfm and requires up to a 60-gpm solution circulation rate
(stainless steel construction).
512
-------
B. Venturi (Chemico)
A 30-inch ID Venturi (6-1/16 inch to 7-1/8 inch throat) scrubber,
with a 30-inch ID centrifugal separating chamber. It is rated
at 1500 cfm and requires 30- to 60-gpm solution circulation
(stainless steel construction).
C. Polygrid Packed (Heil)
A 24-inch ID by 15-foot-high Heil absorber, which will be tested
with a 7-foot depth of polygrid (Ecodyne Corporation, formerly
Fluor Products Company) "eggcrate" packing. It is rated at 2000
to 3000 cfm and required 25- to 50-gpm solution circulation
(fiberglass scrubber, polypropylene packing).
D. Lurgi Impingement (Peabody)
A 30-inch ID by 14-foot-high scrubber, combining a Lurgi Venturi
and three Peabody impingement trays. It is rated at 1500 to 2000
cfm and requires 40- to 50-gpm solution circulation (stainless steel
construction).
The liquor flow rate to each scrubber is controlled by a magnetic flow
recorder-controller in the 3" main PVC line from the hold tank. A liquid
level controller is provided in the hold tank to adjust the tank volume
to allow sufficient hold time for the supersaturated calcium sulfite
slurry from the scrubber to desupersaturate. A hold tank of 3-15 minutes
can be achieved with the equipment sizes selected. 'To control solids
content and to maintain a constant alkalinity and pH control in the liquor
to the scrubber, a continuous withdrawal of liquor is necessary for solids
513
-------
removal or for regeneration. For this purpose, a centrifuge is provided
for loop A while a settling tank with a long winding channel is in use
for loop B.
The discharge stream from each loop is measured and analyzed before
respectively flowing to the centrifuge or settling tank with the treated
solution being returned continuously to maintain close solids and pH
controls. The solids in loop A are processed continuously during the
test, while the solids in loop B are accumulated in the bottom of the
settling tank and measured before being sent to the centrifuge at the
end of the test.
For the testing of both calcium and sodium absorbents in the removal of
S02 from flue gases somewhat different process systems are required due
to different solubility of these absorbents in water. In the calcium
system, the scrubbing liquor contains fine particles of absorbents and
is circulated through the scrubbers as a slurry. On the other hand,
the sodium absorbent is passed through the system in the dissolved
state, with only a small percentage of fly ash as solids.
In the System I mode of operation, as shown in Figure 5, the S02 is
absorbed from the flue gas by either a lime or limestone slurry that is
circulated to the scrubbers from the hold tank, fed from one of the two
lime-limestone feed tanks. The rate of product withdrawal is controlled
by a specific gravity controller in the hold tank which activates a
valve to the centrifuge. The liquid level in the hold tank is adjusted
by a level controller connected to the makeup water line. This level,
in combination with the liquid circulation rate, is used to maintain
hold time up to 10 minutes or more.
514
-------
In the System II mode of operation, as shown in Figure 6, the SC>2 is
absorbed from the flue gas by a sodium solution circulated to the
scrubbers from the hold tank. The rate of product withdrawal from
the hold tank is controlled by a pH recorder controller located in the
liquid line from the hold tank. The regeneration in the reaction tank
is controlled by a pH recorder-controller. Liquid in the reaction tank
is pumped with level control into the centrifuge. The liquid from
the centrifuge (which constitutes the regenerated sodium sulfite absorp-
tion solution) flows by pressure back into the hold tank.
515
-------
III. ANALYSIS PROCEDURES
Test procedures of some of the more complicated data measurements are
as follows:
A. Analyses of Gas Streams
1. SOo Concentration
862 in the inlet and outlet gases is measured continuously by
a Dynascience monitoring instrument that utilizes an electro-
chemical transducing cell. This unit is routinely calibrated
with a standard gas mixture. Periodically, the accuracy of
SC-2 analyzer is checked either by the iodometric method or by
the barium sulfate turbidimetric method.
2. NOX Concentration
in the inlet and outlet gases is measured continuously by
another Dynascience monitoring instrument. This unit is cali-
brated routinely with a gas mixture, and its accuracy is checked
by a modified phenol disulfonic acid method.
3. Dust Loading
Particulates in the inlet and outlet gases are measured in dup-
licate for each test Isokinetic sampling is used. Particulate
mass is composed of the solids (dried to constant weight at HOC)
caught on the filter thimbles or filter pads, including particulates
deposited in the filter probe and sample lines. The filter pads
used in the outlet gas have a reported efficiency of 99.7 percent
for 0.3 particulates.
516
-------
4. Trace Metals
Small amounts of metallic elements, such as mercury and chromium
may be present in the inlet and outlet gases and are measured
by a Varian Techtron Atomic Absorption Spectrophotometer. These
elements are measured in both the particulate filtered from the
gases and in the impinger liquids through which the gases pass
after flowing through the filters.
5. SO, Concentration
A gravimetric method of determining 803 in the flue gas is used.
The method involves collecting gas samples in two fractions:
the condensate and the isopropy alcohol scrubber. The absorbed
SO-J is precipitated as barium sulfate using barium chloride.
6. Fluorine Concentration
Part of the fluorine contained in the coal, present in the flue
gas, is condensed or absorbed in the traps of the 803 apparatus.
The contents in these traps is combined and analyzed by a SPADNS--
zirconium spectrophotometric method.
B. Analyses of Liquor Streams
The pH is measured continuously by an industrial pH meter, which is
routinely calibrated with standard buffer solution; density is recorded
continuously by a differential pressure transmitter located in the
hold tank. This unit is checked periodically by gravimetric methods.
Na, Ca, Cr, and other cations in the solutions are measured by the
atomic absorption unit. SC>2 concentration (sulfite + bisulfite) is
measured by iodometric titration. Spectrophotometric methods are used
for 804, NC-3, etc. Sulfate determination is also confirmed by barium
gravimetric method.
517
-------
IV. TEST PLAN
The test plan is divided into four parts
A. Startup of pilot plant includes check-out of facility and
equipment and calibration of instrumentation systems, both manual
and automatic controls.
B. Initial reference experiments using blowdown water from cooling
towers (8 experiments)
C. Parametric Experiments (228 experiments)
D. System Performance Tests (2 tests)
•
The test plans are scheduled for testing for two levels of pH (7, 9),
three levels of liquid to gas ratio (3.3 to 100) and gas flow rate
(500 acfm - 3000 acfm). The range over which the gas, liquid, and
L/G vary differs with each type of scrubber. For each scrubber,
maximum allowable ranges of gas, liquid, and L/G are tested.
The system performance tests involve two extended runs, each lasting
two weeks in duration, for testing system performance from the viewpoints
of erosion, corrosion, and scaling.
The parametric experiments and system performance data are to provide
essential comparative data for each of the absorbers and absorbents
tested.
518
-------
In order to acquire meaningful experimental data for subsequent analysis,
operating variables and their responses throughout the scrubbing system
are measured, during each run. The relevant experimental data collected
include: (1) gas flow rate; temperature; absolute pressure; pressure
differential across the scrubber; S02, 803, and NO concentrations;
and dust loadings of the feed and effluent gases; (2) absorbent liquid
circulation rate; pH; temperature; and chemical composition; (3) liquid
composition of feed and effluent of the regeneration tank; (4) slurry
density in solids and desupersaturation time in the hold tank; (5) water
losses and water droplet entrainment; and (6) quantity and composition
of makeup water and absorbent feed.
The experimental data obtained from each scrubber are analyzed to assess
the validity of the assumptions adopted for the mathematic model describing
the behavior in each scrubber. Further, scrubber performance data such
as S02 and particulates removal efficiencies and pressure drop across
the scrubber are correlated as a function of operating parameters such
as gas flow rate, L/G ratio, pH, etc., for each scrubber absorbent
combination.
519
-------
V. CONCLUSION
The utilities sponsoring this research feel a sense of urgency which
has made the program highly specific and compressed in time. The
testing, initiated in July, 1971, will be completed by March, 1972.
We believe that the results will add significantly to the body of
information on alkaline wet absorption and will be of particular
value to those using low-sulfur fuel.
520
-------
(71
[\5
FIGURE 1
GENERAL PLANT ARRANGEMENT
S02 REMOVAL PILOT PLANT
EQUIPWIENTSKID
W/PLATFORWl ABOVE
D
480V POWER
WATER
110 LB.AIR
-B-M.4V-Q
4" ORA»N
-------
FIGURE 2
cn
ro
ro
-------
FIGURE 3
GENERAL FLOW DIAGRAM SO2 REMOVAL PILOT PLANT
cn
ro
CO
LEGEND
FRC15
FRC53
AE 4. 5.6
DP
FLUE GAS
PRECIPITATOR
\\\\\\\\\\\V
T-3
MAKEUP
TANK
SCRUBBER
ABSORBER
GAS FLOW RECORDER CONTROL
LIQUID FLOW RECORDER CONTROL
pH CONTROL ANALYSIS ELEMENT
DIFFERENTIAL PRESSURE
PRECIPITATE
-------
FIGURE 4
PILOT PLANT GAS MONITORING SYSTEM
ABSORBER OUTLET GAS
ABSORBER
INLET GAS
PRESSURE
REGULATOR
AND FILTER
POROUS STAINLESS STEEL FILTERS
/-40'LONG
/ HEATED 1/4"
SS SAMPLE LINES-•
MANUAL OR
AUTOMATIC
SWITCH
VALVES
CYLINDERS OF GAS
FOR CALIBRATION
GLYCOL-H20
VACUUM
GAGE
A COOLING SYSTEM
DIAPHRAGM
VACUUM PUMP
CONDENSATE
AND EXCESS
GAS VENT
MANUAL
SELECTOR
SWITCH
524
-------
FIGURE 5
en
t>o
en
SYSTEM I -CALCIUM ABSORPTION
170°F,750MMHG
39 PPM S02
L ,
o
H>h
B-1
ID BLOWER
2500 CFM, 30"
1 ^
S-l 1
SCRUBBER
1 y\
— < —
Ca(OH)z,
0.075 Ib/min
r-H20,
1 1.8lbs/min
PRECIPITATOR INLET GAS
250°F,750MMHG.
390 PPM S02
1500 CFM
2.8 gr/cf FLY ASH
H20. 0.9 LB/MIN
T-3
LIME TANK
200 GAL
0.2 LB/MIN CALCIUM
SULFITE + SULFATE
0.6 LB/MIN FLY ASH
0.4 LB/MIN H20
P-5
PUMP
100GPW1.20'
P-1 T-1
PUMP HOLD TANK
100GPM.70' 1200 GAL
F-1
CENTRIFUGE
-------
FIGURE 6
SYSTEM II-SODIUM ABSORPTION
170°F,750MMHG.
39 PPM S02
130°F
i- 0.09 LB/MIN LIME
I—- H20 3.5 Ib/min.
01
ro
o>
- PRECIPITATOR OUTLET GAS
250°F,750MMHG.
390 PPM S02
2500 CFM
0.65 LB/MIN
ORIGINAL CHARGE
400 GAL H20
71.5 LB Na2 C03
tl
T-3
LIME TANK
200 gal.
H-1
HEATER
30KW
B-1
ID BLOWER
4000 CFM, 20"
riuu urm, &u
S4 *
CRUBBER
p— ^
HI
t=»^r
^
NOTE:
IF 5% OF S02 OXIDATION
IS ASSUMED AND 3000 PPM
S04 IN MAKEUP H20 IS
ASSUMED ONLY
0.012 Ib/min NazCOs IS REQ'D
0.118 Ib/min Ca(OH)2 IS REQ'D
- 0.04 LB/MIN Na2C03
P-1
PUMP
100 GPM, 70'
T-1
HOLD TANK
1200 GAL
T-2 P-3
REACTION PUMP
TANK 5 GPM, 20'
200 GAL
CALCIUM SULFITE
0.22 Ib/min.
H20 0.22 Ib/min.
F-1
CENTRIFUGE
-------
DETROIT EDISON PILOT PLANT AND FULL-SCALE DEVELOPMENT PROGRAM
FOR ALKALI SCRUBBING SYSTEMS - A PROGRESS REPORT
J. H. McCarthy
Project Manager, Air
Quality Control Projects
J. J. Roosen, Director
Environmental Studies
Division, Engineering
Research
The Detroit Edison Company
2000 Second Avenue
Detroit, Michigan 48226
For presentation at the Second International
Symposium on Lime and Limestone Wet Scrubbing
Sponsored by Environmental Protection Agency,
Office of Air Programs, Division of ControT
Systems
527
-------
Detroit Edison Pilot Plant and Full-Scale Development
Program for Alkali Scrubbing Systems—
A Progress Report
This paper includes the material which was originally to
be presented in two separate papers—4c, "Detroit Edison Pilot
Plant for Alkali Scrubbing Systems" and 7c, "Detroit Edison
Full-Scale Development Program for Alkali Scrubbing Systems."
The text of Paper No. 7c is not included in this proceedings.
528
-------
DETROIT EDISON PILOT PLANT AND FULL-SCALE DEVELOPMENT PROGRAM
FOR ALKALI SCRUBBING SYSTEMS - A PROGRESS REPORT
ABSTRACT
In September, 1970, The Detroit Edison Company announced its
plans to construct a full-scale limestone scrubbing system for control of
sulfur oxides on its 280 megawatt River Rouge Unit No. 1. At that time,
and of this date, there is no commercially proven system utilizing limestone
for scrubbing in operation. Proven commercial sulfur oxide control systems
for the utility industry are ones that: 1) can be ordered from a manufacturer
as we do with a boiler, 2) can be installed in a coal burning plant with a
reasonable construction time and cost, 3) can be started up and operated
with a high degree of service reliability so that sulfur oxide is controlled
the majority of the time that the generating unit is running and, 4) incor-
porate a practical method of handling and disposing of large quantities of
waste materials. The River Rouge project is considered a research and
development program. The Detroit Edison Company developed a pilot plant
program which would provide for the detailed examination of the sensitivity
of a sulfur oxide removal system to variations in the different process
parameters. This paper describes the basis for formulation of the pilot
plant program, how the pilot plant results must relate to the design and
selection of equipment for the full-scale installation, and preliminary
results of several investigations carried out in the pilot plant to date.
The specific objective of the pilot plant program is to study,
evaluate, and demonstrate the system behavior with respect to: 1) level of
sulfur dioxide removal from the flue gas, 2) level of entrainment and outlet
particulates in the flue gas leaving the scrubbing system, 3) sensitivity of the
system to turn down in the flue gas rates, 4) short-term scaling and solids
deposition problems, 5) mechanical problems related to equipment and operation,
6) the effectiveness of lime, limestone, and other available alkalis.
INTRODUCTION
The Detroit Edison Company has announced plans to construct a
full-scale limestone scrubbing system for control of sulfur dioxide on its
280 megawatt, River Rouge Unit No. 1. The steam generator of this unit has a
front-fired pulverized fuel furnace that is equipped with mechanical cyclone
separators followed by electrostatic precipitators. When constructed
in 1956,, these devices represented the latest in particulate emission
control technology. However, more restrictive air pollution codes
indicated the need for upgrading the Collection equipment. Various
529
-------
alternatives for upgrading the performance of equipment at the plant were
explored. Among the alternatives was the addition of a wet scrubbing, device
after existing equipment.
Impending regulations on emission of sulfur oxides indicated
that fly ash control systems should not be considered independent of sulfur
oxide control, even though the latter had not been commercially proven.
Therefore, it was decided that scrubbers should be installed since they
would also provide for S02 control. Proven commercial sulfur oxide control
systems for the utility industry are ones that: 1) can be ordered from a
manufacturer as we do with a boiler, 2) can be installed in a coal burning
plant with a reasonable construction time and cost, 3) can be started up and
operated with a high degree of service reliability so that sulfur oxide is
controlled the majority of the time that the generating unit is running and,
4) incorporate a practical method of handling and disposing of large quantities
of waste materials. Examination of available information on sulfur oxide
control and work that had been done from the 1930's to now, indicated that
the most promising concept for add-on systems to existing units was the
limestone wet scrubbing system. The major factors favoring the use of
limestone in such a system, at this time, are the low cost and wide
availability of limestone, apparent simplicity relative to sulfur recovery
systems, and minimum potential for water pollution problems.
RIVER ROUGE SCRUBBER PROJECT
The River Rouge project is considered a research and development
program. Because of the lack of demonstrated performance of full-scale
systems and the variety of scrubber designs, it was decided that this
project would include a pilot scrubber installation as well as a full-scale
alkali scrubbing system. This parallel approach was needed for the following
reasons.
530
-------
1. Specific pilot plant data concerning performance of various
scrubber designs and the process chemistry of these systems
were not available.
2. System reliability in terms of equipment dependability and
operating efficiencies were not known.
3. Criteria regarding the effectiveness of different limestones
had not been adequately established.
4. Actual costs as quoted by equipment suppliers bidding to
specifications were not available.
5. Physical and chemical characteristics of waste products,
with regard to water pollution potential and solid waste
disposal methods, required further identification.
6. Operating experience with this type of equipment had to
be acquired.
Bechtel Corporation is designing the full-scale limestone scrubbing
system for River Rouge Unit No. 1, and participating in Detroit Edison's
pilot scrubber program.
PILOT PLANT PROGRAM
In order to accomplish the objectives outlined above, a pilot
scrubbing installation capable of handling up to 2500 ACFM of flue gas from
an operating power boiler was constructed at the River Rouge Plant. The
pilot program was designed to examine the following:
1. Level of sulfur dioxide removal from flue gas.
2. Levels of entrainment and outlet particulrtes in flue gas
leaving the scrubbing system.
3. The sensitivity of the system to turn down in flue gas rates
4. Short-term scaling and solids deposition problems.
531
-------
5. Mechanical problems related to equipment and operation.
6. The relative effectiveness of various alkalis such ag
limestone, lime, and others.
In order to minimize the number of limestones that would have to
be evaluated in the pilot plant, laboratory screening tests were carried out.
The reactivity of available stones was compared with that of reagent grade
calcium carbonate from two different sources. Results indicated that Presque
Isle and Rogers City limestones were comparable with regard to reactivity.
Arrangements were made to have a supply of Fresque Isle limestone ground to
various degrees of fineness, bagged, and shipped to the pilot plant. The
chemical and physical characteristics of the limestone are given in Table I.
Included also are typical coal and ash analyses for the unit supplying flue
gas to the pilot plant.
The physical arrangement provides the flexibility to study several
different types of scrubbers with minimum system modifications. The scrubber
types to be evaluated were: 1) a series of two Venturis, 2) a venturi followed
by a sieve tray absorber and, ^3) a venturi followed by a mobile bed absorber.
Series of Two Venturis
A schematic of the pilot scrubber system employing two Venturis
is shown in Figure 1. The piping system was arranged so that gas could be
taken from either before or after the precipitators of the No. 3 Unit at
the River Rouge Plant. After the two scrubbers, the gas flows through a
metering orifice and an induced draft fan to heaters which reheat the clean
gas to approximately 200 F before it is emitted to the atmosphere. As can
be seen in Figure 1, the slurry system is the most complex part of the pilot
plant. The slurry is prepared by mixing limestone with clarified liquid
from the sludge thickener in a slurry mixing tank. Bagged limestone in various
532
-------
degrees of fineness are available. From the mixing tank, the slurry of
alkali scrubbing material is transferred to a large agitated holding tank
from which it is metered to the second-stage recycle tank. Slurry is
pumped from this tank to the second-stage scrubber at a rate to satisfy a
specific liquid to gas (L/G) ratio. A portion of the solution from the
second-stage recycle tank is bled to the first-stage recycle tank. Solution
from this tank is circulated through the first-stage venturi to satisfy a
specific L/G. A portion of the slurry from the first-stage recycle tank
is bled to the sludge thickener. Clarified liquid is pumped to the slurry
mixing tank. Underflow from the thickener, spent slurry and fly ash, is
removed to the plant ash handling system.
A typical operation of this pilot plant arrangement consisted
of:
1. Preparing a mixture of slurry in the large mixing tank and
filling the recycle tanks to a predetermined level.
2. Establishing gas flow and monitoring inlet temperatures until
a constant value is obtained.
3. Starting the recycle pumps at a high rate.
4. Monitoring ph of the slurry until it reaches a level of
approximately 6.0.
5. Adjusting slurry recycle to satisfy preselected L/Gs.
Pressure readings, temperature, and sulfur dioxide concentrations
are taken at frequent intervals. The instrumentation to make these measure-
ments is installed in a small laboratory that has been built adjacent to the
pilot plant. Samples of limestone slurry are taken from the slurry mixing
tank and analyzed for calcium carbonate content. With these measurements,
the amount of limestone slurry needed to meet a specific stoichiometric
533
-------
value is determined. The slurry feed and discharge rate are then adjusted
as needed. Similarly, the L/G is adjusted to meet a predetermined value by
holding the gas rate constant and monitoring the liquid flow using magnetic
flow meters. The ph of the recirculating solutions is monitored in both
the recycle tanks to control scaling in the scrubber and to determine the
correlation between ph and SO- removal. Suspended solids are measured at
the inlet and outlet of the thickener to determine its efficiency.
The results of investigations conducted employing a system of
two venturi scrubbers indicate that outlet dust loadings of less than 0.02
grains per standard cubic foot of gas can be achieved. This performance,
however, has yet to be adequately demonstrated on a full scale unit. This
system has been operated for up to 120 hours without serious plugging or
scaling problems. During this series of tests, the system removed from
55 to 65 percent of inlet S02, independent of the ratio of excess limestone
to sulfur dioxide. The optimum removal efficiencies during these investiga-
tions were obtained with liquid to gas ratios between 30 and 40 gallons per
minute per 1000 cfm. When the same system was operated with 20 percent excess
hydrated lime, greater than 90 percent of sulfur dioxide was removed. These
results indicated that, in the case of a venturi scrubber, there is little
internal circulation within the drops; consequently, relatively insoluble
alkalis such as limestone do not readily dissolve and react with sulfur
dioxide which comes in contact with them for short periods of time. On the
other hand, the large number of small droplets and large amounts of surface
area produced in these units provide for excellent particulate removal.
Venturi Followed by Sieve Tray Absorber
A schematic of the pilot system employing a venturi followed by
a sieve tray absorber is shown in Figure 2. in this system, the alkali
scrubbing material is transferred from the slurry mixing tank to the large
534
-------
agitated holding tank from «hich it is metered to the tray absorber feed
tank. The slurry is then metered to the absorber section of the scrubber.
The limestone slurry then cascades over several stages of trays and flows
back to a recycle tank. Solution from the recycle tank is metered to the
venturi section of the scrubber. A portion of the solution in the recycle
tank is pumped to the sludge thickener. As with the previous system, the
clarified liquid is returned to the slurry mixing tank for preparation of
new scrubbing solution. Sludge from the bottom of the thickener is removed
to the plant's ash handling equipment as with the previous system.
To date, typical operation of this installation has consisted of:
1. Preparing a batch of scrubbing slurry in the mixing tank.
2. Filling the slurry holding tank, absorber feed tank, and
recycle tank to a predetermined level.
3. Establishing a flow through the venturi section at a
predetermined L/G.
4. Establishing gas flow through the unit, and immediately
thereafter, a flow of scrubbing liquid to the absorber
section.
5. Adjusting the rate of flow to the absorber section at a
predetermined L/G.
Limited experience with this system to date indicates that outlet
dust loadings of less than 0.015 grains per standard cubic foot of gas can
be achieved. It should be noted, however, that these results represent the
performance of a l/200th scale of the River Rouge unit. Difficulties have
been encountered to date in establishing a ph level high enough to control
scaling due to precipitation of dissolved solids in the absorber section of
the scrubber system. In order to provide more precise ph control, modifica-
tions are presently being made to the recycle and alkali feed systems of the
pilot plant. As of this writing, significant sulfur oxide removal data have
not been obtained. 535
-------
Operating experience to date has indicated that it is extremely
important to have an on-site capability for performing chemical analyses.
This has been found necessary because certain parameters must be monitored
and the results of that monitoring fed back into the operation of the
scrubber in order to examine such characteristics as scaling and deposition
potential. This is particularly true in establishing the degree of super-
saturation in the supernatant from the sludge thickener.
For each test series, the pilot plant is manned for 24-hour per
day operation. On each 12-hour shift, a crew of three men is employed.
This crew is composed of an operator from the power plant, a chemical
engineer, and an engineering technician who has been trained in chemical
analysis. Recently, an analytical chemist has been added to the staff during
the day time part of the operation.
Future Plans
Beyond the present test series, it is planned to study the effective-
ness of other alkalis including certain chemical process waste products available
locally. Of equal importance to the effectiveness of the systems in removing
S02 are solutions to the waste disposal problems. Therefore, a concurrent
program, utilizing waste materials generated by the pilot plant, is underway.
This effort includes the development of practical means of handling, trans-
porting, and ultimately disposing of the sludge.
As mentioned earlier, engineering of a full-scale limestone
scrubbing system for River Rouge No. 1 is proceeding in parallel with the
pilot program. The ultimate configuration of this unit and its operational
schedule is dependent upon the outcome of the present pilot program. The
current schedule for completion of the full-scale scrubbing system is
January 1973.
536
-------
TABLE I
CHEMICAL AND PHYSICAL PROPERTIES OF LIMESTONE, COAL, AND ASH
DETROIT EDISON PILOT SCRUBBER PROGRAM
Presque Isle
Percent
by Weight
Silica
Si02
0.60
Limestone
Iron Oxide Calcium
Fe203 Oxide, CaO
0.35 54.5
Magnesium Oxide
MgO
0.60
Sulfur
(S03)
0.085
Loss on Ignition
43.82
Calcium Carbonate, percent 97.5
Magnesium Carbonate, percent 2.5
Coal
01
-1
Percent
by Weight
Volatile
Matter
36
Fusion
Percent
by Weight
Loss on
Ignition
3.4
Ash
Silica
Si02
40
Iron
Moisture
As Fired
6.3
temp (fluid)
Oxide Alumina
Fe70
30
3 A1203
17
Total
Ash
12
- 2350
Sulfur Carbon
2.8
F Heating
Calcium Magnesia
Ca 0
2.6
MgO
0.6
72
Value -
Hydrogen
4.8
12000 BTU/lb
Titanium Sodium
Ti02
0.9
N320
0.3
Nitrogen
1.7
•
Potassium
K20
1.8
Oxygen
5.5
Carbon
C
2.9
Sulfur
S03
1.1
-------
GAS INLET
BEFORE
PRECIPITATOR
EXHAUST
FIRST
VENTURI
SECOND
VENTURI
GAS INLET
AFTER
PRECIPITATOR
01
to
00
FIRST
STAGE
RECYCLE
TANK
LIME OR LIMESTONE
WATER
SECOND STAGE
RECYCLE TANK
SLURRY
HOLDING
TANK
SLURRY
MIXING
TANK
FIGURE I
DETROIT EDISON PILOT SCRUBBER
NlSAJLTtNVBLVa. VIA- <
SLUDGE
THICKENER
PUMP
•ASH
HANDLING
SYSTEM
-------
HEATER
EXHAUST
GAS INLET
BEFORE
PRECIPITATOR
DRAIN
SIEVE
TRAY
ABSORBER
SLUDGE
THICKENER
VENTURI
GAS INLET
AFTER
PRECIPITATOR
ABSORBER
FEED TANK
LIME OR
LIMESTONE
RECYCLE\/
TANK I
SLURRY
MIXING
TANK
ASH
SETTLING
BASIN
SLURRY
HOLDING
TANK
FIGURE 2
DETROIT EDISON PILOT SCRUBBER
« tm
stifi/f TO AW An Ort DO CD l?nHICIAI IDATiriM
-------
RESEARCH & DEVELOPMENT IN WET SCRUBBER SYSTEMS
A. L. PLUMLEY
M. R. GOGINENI
C-E Combustion Division
Windsor, Connecticut
Presented at
SECOND INTERNATIONAL LIME/LIMESTONE WET SCRUBBING SYMPOSIUM
November 8-12, 1971
Sheraton-Charles Hotel, New Orleans, Louisiana
541
-------
RESEARCH & DEVELOPMENT IN WET SCRUBBER SYSTEMS
A. L. Plumley
and
M. R. Gogineni
The subject of this paper is the research and development sup-
porting Combustion Engineering's product line of air pollution
control and by-product recovery systems to meet Federal standards
for air and water pollution. Major emphasis of this program
centers on providing design criteria and engineering support
for commercial systems.
R&PD PROGRAM OBJECTIVES
The following have been established as both short and long range
objectives of C-E Research and Development Program for Air Pol-
lution Control Systems:
A. Complete development of limestone furnace injection
demonstration systems to meet contract obligations for
SO and particulate removal.
B. Establish design criteria for future systems at lower
cost using the present APCS principles.
C. Develop non-furnace injection systems for both utility
and industrial boiler applications.
D. Develop improved air pollution control systems to reduce
other air pollutants and to allow expansion of a company
product line.
E. Develop by-product utilization processes to avoid poten-
tial water and land contamination and to offset operating
costs of current and future air pollution systems.
The R&D program is divided into two principle areas of effort:
1. Air Pollution Control Systems Development
2. Chemical Process Research
AIR POLLUTION CONTROL SYSTEMS DEVELOPMENT
Air Pollution Control Systems Development is comprised of performance
work on full scale systems in the field and experimental work done
on our test facilities in the laboratory. Design criteria are con-
firmed in this area of development.
542
-------
The first area of development involves the APCS contracts which
C-E has obtained. These contracts are listed and described in
another paper at this conference, entitled "Contract Applications
of the C-E Air Pollution Control Systems".
The second area of Air Pollution Control Systems Development is
comprised of all the experimental work done in the laboratory on
our scrubber test facilities. There are two operating APCS test
facilities available in the Kreisinger Development Laboratory
for evaluation of wet scrubber technology: A 12,000 CFM Proto-
type system and a 1,200 Pilot system (Figures 1 to U).
These APCS test facilities are comprised of all system components
utilized by the existing C-E installations and are assembled in a
modular arrangement to allow rapid system modifications. In this
way, a number of flow arrangements can be studied.
CHEMICAL PROCESS RESEARCH
The Chemical Process Research section is concerned with two major
areas. The first is the fundamental understanding of reaction
mechanisms associated with our wet scrubber work and the second
is the disposal of waste materials from APCS operations. The more
fundamental work is carried out in various types of laboratory
bench reactors and is designed primarily to provide supplementary
information to the larger scale test facilities and to screen
variables to be evaluated during pilot and prototype tests.
As the overall program develops, co-ordination of various technolo-
gies related to environmental control, both solid and liquid wastes,
as well as air pollutants, will be carried out.
TEST WORK
Test work is currently underway which is directed at providing
sufficient design information to resolve problems remaining on
the demonstration field installations thus allowing them to meet
contract guarantees as well as federal and local regulations on
S00 and particulate removal. Field development is expensive and
time consuming since generation of power takes precedence over
the check-out of field modifications and other work necessary in
evaluation of new processes and new process development. We have
experienced considerable time, months of delay, between modifi-
cations and their evaluation in these field systems. As a result,
most of our development work is being carried out on the two
operating APCS test facilities in the laboratory.
Prototype (Figures 1 and 2)
The APCS prototype is approximately 1/U the linear scale of C-E's
field units and utilizes a 25 sq. ft. marble bed contactor to scrub
the flue gases from the oil fired packaged boiler. This boiler has
a flue gas output of 12,500 to 15,000 ACFM measured at lU.J psia in
543
-------
125°F. The inlet gas temperature to the scrubber is variable; the
600°F flue gas from the boiler can be reduced to 250°F by means of
a finned tube heat extractor prior to the scrubber.
Major components of the system are the scrubber, heat extractor,
reheater, demister, clarifier or retention tank, and the slurry
or hold tank vith provisions for additive and fly ash injection.
The piping and pumps can be readily modified to allow study of
several different processing arrangements.
Pilot (Figures 3 and 1*)
The APCS pilot system features multiple three sq. ft. marble beds,
as well as 1 sq. ft. rod type contacting surfaces which may be
operated in various combinations to determine the most effective
use of additives for SO removal from flue gas. Provision has
been made for evaluation of other types of scrubbing surfaces.
Separate collection of particulate and SO is also possible in
this system. The major system components in addition to the
scrubber with its multiple beds are the demister, clarifier,
slurry and delay tanks additive and fly ash injection systems.
Here, too, piping and pumps can be modified to allow several
different processing arrangements. The system is currently being
used to carry out preliminary screening of variables so that only
optimum modes will be demonstrated on the prototype proper.
Bench Scale Facilities
There are three principal bench scale facilities that are utilized
in the APCS studies. In general, the purpose is not to perfectly
scale down APCS either from field or laboratory, but to provide
small flexible laboratory tools in which to study various phenomenon
including packed-bed mass transfer rates, scaling phenomena and rates
of reaction to provide a basis for correcting certain problem areas
observed in operation of either the larger scale laboratory or field
facility. The first of these facilities, the bench scale scrubber,
is a tube which has been flanged and divided into three sections (See
Figures 5 and 6). The lower section contains a gas inlet in the
liquid drain, the middle section contains the bed, the nozzle and
the bed drain, and the upper section contains the mist eliminating
section and the gas outlet. These sections are removable for modi-
fications and repairs. The rest of the apparatus is concerned with
controlling concentration, temperature and flowrates of reactants.
Instrumentation includes temperature controllers, SO monitors, tem-
perature monitors, and the pH meters.
The next type of facilities are stirred reactors. We have a CFSTR
or Continuous Flow Stirred Tank Reactor. This equipment is used to
understand scrubber reaction kinetics by direct determination of dif-
ferential rate data for the various reactions taking place in the
scrubber. This is important in design of system controls as well as
in sizing of system components. In this reactor the contents are
well stirred and uniform in composition throughout. Thus, the exit
stream from this reactor has the same composition as the fluid within
544
-------
the reactor and analytical results on this stream represent the
behavior of the system.
The second type of stirred reactors are batch type reactors.
These consist of a pair of temperature controlled flasks. System
reactions, specifically those relating to potential deposit for-
mation, and for preparation of scrubbing solutions, delav tanks,
holding tanks, etc. are being studied. (Figure 7)
The holding effects of ponds to allow for completion of reactions
including precipitation of solids and dissolution of potential
reacting materials which are being recirculated in the overall
system are being simulated in bench tanks.
CHEMICAL PROCESS DEVELOPMENT
WASTE DISPOSAL
A comprehensive program for utilization of APCS by-products is
currently underway. This six phase program covers:
1. Nationwide environmental survey of present waste
disposal procedures and ecological data relating
to the needs of potential APCS customers.
2. Determination of physical and chemical properties
of waste products.
3- Studies on direct disposal of both solid and liquid
wastes.
U. Beneficiation and utilization of liquid and solid
waste products.
5- Pilot plant studies of promising utilization of
procedures.
6. Economic and technical summary and recommendations.
A significant part of the understanding of physical and chemical
properties is being carried out with the assistance of C-E's
modern X-ray diffraction and X-ray spectrographic equipment and
recently acquired thermal analysis equipment. This instrument
(Figure 8) is capable of performing simultaneously the three
most widely used thermoanalytical techniques: Differential Thermal
Analysis (DTA), Thermal Gravimetric Analysis (TGA) and Differential
Thermogravimetric Analysis (DTGA).
The instrument has been utilized so far in determination of melting
point, decomposition temperature and sinterability of waste sludge
from APCS operations (Figure 9). These are all properties needed
for determination of appropriate utilization applications of the
material.
545
-------
Characterization
Work is underway on development of a sludge "library" by cataloging
of ten different chemical and physical properties of sludges from
various APCS operations in both field and laboratory studies. This
information is essential since the utilization or disposal process
in each instance will depend on both the additive used and the mode
of APCS operation. (Figure 9).
A number of areas for utilization of modified and unmodified fly ash
are shown in Figures 10 and 11. These schemes range from production
of light weight aggregate developed some years ago to compression
and sintering of drain pipes which is under laboratory development.
Large volume usage will be necessary since 1-2 tons/day/MW of sludge
would be produced from a coal-fired boiler firing 3% S coal and
ash with an APCS for emission control.
Beneficiation
A number of small brick and pipes have been successfully produced
from APCS sludge by Industrial Material Technology, Inc. Samples
formed by high pressure without sintering appeared most promising.
Efforts are continuing to improve durability and optimize fabrica
tion methods for large quantity production.
ACRES, Inc. has submitted a proposal to define the geotechnical
properties of the APCS sludge. Properties involved are important
in designing structural fill - dams, dikes, pavement and other
load bearing land-fill usage. ACRES, Inc. has been retained by
Hiagra Mohawk to handle ash disposal problems.
The evaluation of APCS sludge for hydrothermal bricks is contin-
uing at Michigan Technology University.
Environmental Survey
A report summarizing the results of an environmental survey has
been completed. Compilation of current Federal, State and Local
pollution codes is continuing.
Direct Disposal
A meeting of the National Ash Association in Washington discussed
plans for transportation of fly ash by rail from the utility to
the disposal site. It was decided to also include modified fly
ash (from lime/limestone wet scrubbing). Representatives of
leading coal producers were present to consider use of abondoned
mines as disposal sites. No firm decisions were reached.
The EPA Office of Solid Wastes is continuing to evaluate an
unsolicited C-E proposal for study of direct disposal procedures
for APCS sludge.
546
-------
BENCH STUDIES
An extensive laboratory investigation designed to reduce the liquid
blow-down required to control calcium sulfate (gypsum) deposit for-
mation during APCS operation is underway at KDL. This work is essen-
tial to avoid replacing an air pollution problem with water pollution,
The three-point bench s+ -1y is directed at development of a "closed
cycle" system in which th^ only liquid discharge from the APCS is
that conveying solids which can be recycled, thus requiring no blow-
down to the river.
The most promising technique resulting from the bench study will be
followed by a demonstration run on the 1200 cfm APCS pilot facility.
Other bench scale support studies are underway to provide aid in
understanding of scrubber chemistry and mass transfer within the
APCS and similar SO removal system. Included are:
A. Reliability of extrapolation of bench units to larger
systems.
B. Confirmation of the marble bed as a completely back
mixed system as part of fundamental studies on mass
transfer within the bed.
C. Basic studies on oxidation of sulfites and the sub-
sequent relationship to both SO absorption in the
scrubber and control of both sulfite and sulfate
deposit formation.
D. Use of a Continuous Stirred Tank Reactor to aid in
equipment sizing by determination of rate controlling
steps in various phases of the APCS operation. Dis-
solution of alkali (CaO and CaCO ) as well as preci-
pitation rate of sulfates and sulfites are under study.
TEST RESULTS
Experimental work since 1970 dealt with specific problems in the
field demonstration system. Control drain line scaling and opti-
mization of slurry recycle (both clarifier and scrubber underflow)
have been studied.
The under bed addition of slurry and dry injection of additive in
Se SB stream have given comparable results in both SO removal
and chemical efficiency of additive. Both procedures resuit in
passage of both additive and SO through the bed and it may be
conceded that the intimate mix?ng thus achieved is required for
maximum efficiency of operation.
SO, removal of 90 - 95 percent have been achieved vith both dry
injection and slurry addition of hydrated dolomite About 10,
more hydrated lime is required than hydrated dolomite to give
comparable SO removal.
547
-------
Experimental work in 1970 on the APCS prototype test facility
emphasized optimization of slurry additive or tail-end systems
utilizing calcium carbonate as the additive.
In a series of single marble bed tests at 1000 ppm SO , 50 to
300$ stoichiometric limestone slurry was fed through underbed
spray nozzles with no recycle. SO removal ranged from UO to
75$. The formation of calcium sulrate (CaSO.) deposits in the
scrubber, is being controlled by blowdown tests.
Experimental work in 1971 on the Air Pollution Control System
(APCS) test facilities has emphasized optimization of tail-end
slurry additive systems which can-be installed in Unit No. 1
at the Meramec station of Union Electric and Unit 6 at the
Paddy's Run Station of Louisville Gas & Electric. Both single
and dual marble bed studies have covered use of both calcium
carbonate (for Union Electric) and calcium hydroxide (for
Louisville) with inlet SO concentrations of 1000 and 2000 ppm.
Following a series of successful pilot and prototype tests
extended runs to demonstrate reliability were carried out with
dual bed operation and recycle using lime in a scale up of
previous pilot studies.
During an 80 hour run with an appropriate recycle of scrubber
effluent sulfur removal in excess of 80$ was maintained without
scaling. Dust loadings at the outlet were about 0.03 gr/SCF.
Louisville Gas & Electric approved construction of a full scale
APCS at Paddy's Run Station on the basis of these successful tests.
An extended run to demonstrate reliability was carried out utilizing
dual bed operation with maximum recycle of scrubber effluent through
the slurry tank using calcium carbonate. This scale up of previous
studies extended for 125 hours without deposit formation and resulted
in sulfur removal efficiency in excess of 75$-
Laboratory and pilot studies are continuing to understand the overall
chemical balance as related to blowdown and waste disposal. Approval
of full scale APCS construction is expected in the near future.
EPA CONTRACT FOR RESEARCH IN SCRUBBER TECHNOLOGY
C-E has received its first prime contract from the Office of Air
Programs - Environmental Protection Agency (EPA). This $250,000
contract is for the "Optimization of a Lime/Limes tone Wet Scrubbing
Process for SO and Particulate Removal in a Marble Bed Scrubber.
A general outline of the program is as follows:
548
-------
Impending national performance standards and standards "being set
by state and local air pollution control office authorities re-
quires the rapid development of as many SO emission control
processes as possible. Several processes are now being developed,
including the Combustion Engineering (C-E) Lime/Limestone Wet
Scrubbing process, which .has been under development for about
three years. Two full scale units of this process have been
installed on separate 125 MW power plants.
The purpose of this contract is to conduct research and development
on small pilot scale, large pilot scale and full plant versions of
the c-E SO scrubbing process in order to accelerate its commercial
development. Principle capabilities to be developed and demonstrated
include adequate SO and particulate removal, proven operational
reliability, adequate equipment lifetime, adequate capability for
scale-up, and sufficient estimates of process costs.
The following specific work steps are anticipated:
1. Engineering Analysis - Existing C-E data and all related
data from other EPA contractors to develop the final test
program and calibrate mathematical process model.
2. Soluble System Studies - KDL Prototype determination of
the approach to vapor-liquid equilibrium and the mass
transfer capabilities of marble bed scrubber and valida-
tion of input to mathematical process model using Na CO
solution.
3. Prototype Test Program - KDL - To provide optimization of
the present C-E APCS based on input from 1, 2, and k.
U. Prototype tests - KDL - Verification of prototype as model
for prediction of field unit operation.
5. Prototype Tests - KDL - Final calibration to determine
optimum operation for field units.
6. Field tests - to demonstrate on the field units, under a
separate contract, the optimum operating conditions deter-
mined from the KDL Prototype tests and to arrange for the
final confirmation of mathematical process r.odel.
549
-------
ft
c
X
m
in
en —
o
-------
O
c
ro
TO STACK
A
MAKE-UP
REHEATER
DEM!ST€R
MARBLE/
BEDS \.
>^
FLUE GAS
""r^/^*" *
r KUUI
V.P. BOILER
SLURRY
TANK
LIQUID
SLOWDOWN
TO
VACUUM
FILTER
A PCS PROTOTYPE-TAIL-END SYSTEM
-------
o
c
^o
m
' 11
< n
i
I
CM
-------
TO STACK
J A I
DEMISTER
VENTRI
SECTION
CJl
01
CO
T,
o
c
;a
iM
GARBLE
BEDS
V V V V
DRAIN
TAWK
•>3
LJ
0-*
V
TO
SEWER
WELL
V-'ATER
V,
ftrt'tt
[R.1
I
V V V
MIX
TANK
IFfJ
V
CLARIFIER
TO
DISPOSAL
ARCS PILOT PLANT
CYCLONE
FEEDER
«
I
SI
K
v
'.N
i.
t
TO
SE17ER
-------
I
30
ni
A,
-------
01 rn
01
tn
cn
TO STACK
-M
GAS OUTLET SAMPLE
SPRAY NOZZLE
ROTOMETER
LIQUID INLET TEMP.-
ADDITIVE
-^HXH-1
IMMERSION
HEATER
^DEIVISSTER
B
m GARBLE BED
-DOWNCOMER
13
r
GAS IWLET TE^'P.
?
FURNACES
GAS INLET SAMPLE
JRNACES
CAfl'IDI ST 1
j
ROTOMETERS
LIQUID OUTLET SAMPLE, pH
-:>TO DRAIN
SCHEMATIC OF BENCH SCALE SCRUBBER SYSTEIV1
-------
o
c
a
m
en
en
o>
-------
c>
c
-43
tjl
-------
o
c
33
m
n
•
r
u.
;
3
- >
c
:
••:
2
„ , , V tjj
0
O OK
Z 2?
o o x
P H-
-0
x-J
L'J
-------
C
P5
01
WATER TO
CLARIFIER
WET
VACUUM FILTER
^=>q
!
PELLETIZER
CEMENT K
ILN
DRYING KILN
u
V
MIXER
AUTOCLAVE
=CH
SINTERING
MACHINE
LIGHT
WEIGHT
AGGREGATE
SULFURIC ACID
OR SULFUR
^C> MINERAL WOOL
o
ADDITIVE FOP,
ASPHALT
LAND FILL
SLUDGE DISPOSAL SCHEMES FOR A SINGLE SCRUBBER SYSTEM
-------
Q
C
70
m
PELLETIZER
SINTERING
MACHINE
LIGHT
WEIGHT
AGGREGATE
1
HJ P
RESL IFLYASHJJ
SCRUBBER JFLYASHC"
o
V
S02
REMOVAL 0
CALCIUM S
A A CALCIUM
\} if CALCIUM C/S
! t
! Lr=3 WATER err:
1 .. -,,
VACUUM
FILTER
SLUICED
' TO
POND
ULFATE
OXIDE
^RBONATE
VACUUM
FILTER
POZZOLAN
- crtn
^zm?."rr r OK
CONCRETE
1 0
I
C02
.]
Kl
700 °C
1292 °F
MINE
LAND-FILL DRA'NA^,nM
NEUTRALIZATION
S02
PROCESSING
FACILITY
11
SO2
n
LN
1200 °C
2192 °F
aO BACK TO SCRUBBERLL-^— — ru J
SLUDGE DISPOSAL SCHEMES FOR A DUAL SCRUBBER SYSTEM
-------
LIME SCRUBBING
Of
SIMULATED ROASTER OFF-GAS
by
D. E. REEDY
AIR CORRECTION DIVISION
Universal Oil Products Company
Darien, Connecticut
For
Presentation at the Second International
Lime/Limestone Wet Scrubbing Symposium
New Orleans, Louisiana
November 8-12, 1971
561
-------
Introduction
In recent years, the lime/limestone absorption system
has gained considerable impetus as an economically
attractive flue gas SO2 control scheme, over such
alternates as lower sulfur fuels. There are other
possible uses for the technology being developed for
power plant flue gases, such as metallurgical roast-
ing operations, but the level of understanding of the
lime/limestone/S02 chemistry does not lend itself to
reliable extrapolation to the higher SO2 levels and
S02/CO2 ratios often encountered in industrial
emission sources.
In this study, the absorption of SC>2 from an ambient
air stream containing 1% S02 and 1% CO2 utilizing a
lime slurry was accomplished using a pilot plant
(R)
Turbulent Contact Absorber (TCA) . The main objective
was to demonstrate the capability of high SO2 removal
efficiencies and to determine the chemical efficiency
for lime usage under a specific set of conditions.
Description of Equipment
A schematic diagram of the pilot plant equipment set-
up is shown in Figure 1.
562
-------
Air Correction Division
Universal Oil Products Company
Tokeneke Road • Darien. Connecticut 06820 203-655-8711
uop
TO FAN
3 STAGE
TCA®
TOWER
ROOM
AIR
SO2
C02
.ENTRAINMENT
SEPARATOR
Y
*sx
II ~V^
TANK
1
I
V"
Co (OH)2
MAKE-UP
TANK
MAKE-UP
STREAM
t i
METERING
TANK
i
I
RECIRCULATION
PUMP
PURGE
STREAM
MAKE-UP
PUMP
FIGURE 1- EQUIPMENT FLOWSHEET
563
-------
Vaporized SC>2 and C(>2 were injected into the ambient
air inlet duct approximately four duct diameters
ahead of the absorber. The absorber itself was one
of our standard, one square foot Turbulent Contact
Absorbers*!* Three contacting stages, each containing
approximately one cubic foot of static packing,
were utilized in this study. Recirculation liquor
was introduced countercurrent to the gas flow, above
the top contacting stage, through a full cone nozzle.
Figure 2 depicts the unrestrained contacting motion
©
achieved with the spherical packing in a TCAT The
entrainment separator used in this study is similar
to that shown in the illustration.
Liquor draining from the lower stage was collected
in the scrubber sump connected in parallel with an
agitated retention tank that provided 2-3 minutes
holdup. The slurry purge from the system was with-
drawn from the discharge of the recirculation pump.
Slurried, hydrated lime, with a particle size distri-
bution similar to that shown in Figure 3, was added
to the liquor recirculation line downstream of the
purge withdrawal point.
Liquor samples referred to in subsequent sections were
taken directly from the retention tank, while gas
samples were withdrawn immediately after the entrain-
ment separator.
564
-------
Air Correction Division
Universal Oil Products Company
Tokeneke Road • Darien, Connecticut 06820 203-655-8711
uop
SCRUBBING
LIQUID
RETAINING
GRIDS
CLEAN AIR OUTLET
MIST
ELIMINATOR
MOBILE
PACKING
SPHERES
HOT GAS
INLET
SLURRY
DISCHARGE
FIGURE 2
TURBULENT CONTACT ABSORBER
565
-------
ULJ
M
z
X
1-
oe
UJ
i-
Ul
ae,
O
100
90
80
70
60
50
40
30
20
10
0.1
0.2
Air Correction Division
Universal Oil Products Company
Tokenska Road • Darlen, Connecticut 06820 203-655-8711
uop
FIGURE 3
LIME PARTICLE SIZE DISTRIBUTION
0.5 1 2345 10 20
SIZE .MICRONS
50
100 200
-------
Discussion of Results
Table I is a summary of the operating data from
pilot plant operations. A quick analysis of these
data yields some interesting observations.
Gas velocity variations, illustrated by a compara-
sion of Test 2 and Test 5, seemed to have a relatively
small effect on S02 removal (the magnitude and
direction of change in SO2 removal in these two
tests must be attributed to "data scatter").
Although the liquor residence time has been shown
to be essentially a function of recirculation rate
/B\
only, in aTCA4* a comparison of results from Tests 1
and 4 would indicate that holdup time, or L/G, are
not significant factors in SO2 removal.
The controlling variable in this study is the pH of
the scrubber outlet liquor. The relationship is
shown in Figure 4. This is nothing new, of course,
for many studies of limestone scrubbing of power
plant flue gases have shown the importance of
limestone "quality" when translated into terms of
scrubber effluent pH.
567
-------
TABLE I
SUMMARY OF PILOT PLANT DATA
Test
Number
1
2
3
4
Gas Velocity, L/G, Scrubber AP, Outlet Liquor
Ft./Min. Gal/1000 CFM in W.G. pH % Solids
Inlet S02, S02 Removal,
Vol % %
en
CTI
oo
900
630
800
800
800
820
820
58
82
69
82
85
73
73
9.5
7.5
8.5
8.5
9.2
8.9
8.9
8.0
5.8
13.0
8.0
5.8
11.2
12.7
17.8
17.8
7.0
0
1
1
1
1
1
1
.92
.32
.01
.01
.01
.04
.04
82
65
98
78
73
89
98
DER 11/71
-------
Air Correction Division
Universal Oil Products Company
Tokttwke Road • D«ri«n. Connecticut 06820 203-655-8711
uop
100
90
Q
ui
CD
8 80
CO
o
to
70
60
FIGURE 4
SO2 ABSORPTION VS. pH
* 8 10
pH IN SLURRY
12
569
-------
Slurry samples from several tests were analyzed
for calcium, carbonate, sulfur and sulfite as
shown in Table II, and then the solids were
filtered from the slurry and analyzed for the
same components. Assuming that the only sulfur
species present are the sulfite and sulfate these
data show that only a few percent of the hydrated
lime input is being lost with the slurry purge
stream. A somewhat higher percentage of the
hydrated lime may be lost as carbonate, as shown
by the following mol ratios:
Molar Ratios
Test CO3=/Ca++ S/Ca++ CO3=/S SO3=/S
3
5
7
0.16
0.09
0.11
0.76
0.97
0.90
0.22
0.09
0.12
0.96
0.93
0.91
These ratios would indicate that 10 to 15 percent
of the calcium values are being lost as carbonate,
but it should be pointed out that spot checks of
various batches of the hydrated lime indicated
that there was some carbonate present prior to its
use in the absorption tower, probably as a result
of exposure to the atmosphere.
570
-------
TABLE II
RETENTION TANK ANALYSES
Test No. 11!
Total Stream,Wt.%
Calcium 5.5 5.3 2.2
Carbonate 1.1 0.7 0.2
Sulfur 3.3 4.2 1.7
Sulfite 4.2 - 2.4
Solids,Wt.%
Calcium
Carbonate
Sulfur
Sulfite
34.1
8.3
20.6
49.8
29.8
3.7
23.1
54.0
28.7
4.7
20.7
47.3
DER 11/71 571
-------
The data are consistent in that operation at a
lower pHf Test 5, resulted in an increased re-
jection of CC>2 and higher utilization of calcium.
In Test 3, the makeup lime slurry concentration
was about 10% while in Test 7 the makeup rate
was approximately doubled and the input lime con-
centration cut in half. A comparison of the above
mol ratios for these two tests would indicate
that either the makeup slurry concentration or
the recycle slurry concentration is affecting the
loss of calcium values as carbonates.
The sulfite/sulfur ratio would indicate that very
little sulfate is being formed in this operation,
and that sulfate formation is relatively insentitive
to pH. These test runs were of relatively short
duration, however, and these numbers may not represent
fully equilibrated operation.
Figure 5 shows that, at pH's above the neutral point,
the liquid in the retention tank was not completely
in equilibrium with the solids and that a long time
was required for this equilibrium to be attained
as the residual lime went into solution. Again,
however, the amount of undissolved lime remaining in
the slurry was very small.
572
-------
Air Correction Division
Universal Oil Products Company
Tokeneke Road • Darien, Connecticut 06820 203-655-8711
uop
12r
FIGURE 5
SLURRY pH VS. TIME
10
01
•vj
10
I
a
8
1
TIME, HOURS
-------
Conclusions
I. Lime scrubbing can achieve high SO2 removal
efficiencies from gases containing moderately
"high.502 levels.
II. Recirculation liquor pH is a dominant factor
in this operation and must be maintained at
a high level to achieve high scrubbing
efficiencies.
III. In spite of the high pH required to achieve
the desired SC>2 removal, chemical efficiencies
for lime utilization in excess of 80% can be
realized.
References
1. Chen, B.H. and Douglas, J.M., Canadian Journal of
Chem. Engr. 46, 245 (1968).
2. Pollock, W.A., Tomany, J.P. and Frieling, G.,
ASME 66-WA/CD-4 (1966) .
it U. S. GOVERNMENT PRINTING OFFICE. IB7X 746761/I4OI
574
-------