Escolar Documentos
Profissional Documentos
Cultura Documentos
265
,-_-. CLEAN GAS TO
STACK OR ESP
FLUE GAS
FROM FURNACE
PACKED
TOWER
1'It:-----::�
QUENCH
.------.. -,.- l' �
I \ I \ , \ TANK
VENTURI
SCRUBBER
PURGE TANK
PURGE
CAUSTIC
pH is allowed to get above 8.5, an undesirable reaction ($66/t) [4], including slaking. Since lime is a weaker
with the carbon dioxide in the products of combustion base than NaOH, it does not pose the same problems
will occur. This results in unwanted usage of caustic with pH control, as is the case with caustic. The neu
by the following reaction: tralizing reaction is as follows:
CO2 + 2NaOH ..... Na,COJ + H20 (3) 2HCl + Ca(OH)2 ..... CaC12 + 2H20 (4)
Another disadvantage with using caustic is its rel Another reagent that has been used for removal and
ative cost. At an average cost of about $200/ton neutralization of acid gases is limestone which is cal
($200/t) [4], the cost of neutralization with caustic cium carbonate. It is relatively inexpensive at a cost
can be significant. of approximately $25/ton ($28/t) including shipping
Lime is used as a neutralizing reagent in some cases. and grinding. Since it is a weaker base than lime and
Lime is slaked to form calcium hydroxide, which is caustic it does not pose pH control problems. Addi
less soluble than caustic, and must be slurried before tionally, since it exists in the carbonate form, it will
use. Lime is less expensive at a moderate $60/ton not react with the carbon dioxide in the flue gas, re-
266
suIting in consuming the reagent with undesired by maximum pH at which CaC03 dissolution could occur
products. The reaction between HCI and CaC03 is as was determined as a function of the CaCl2 concentra
follows: tion.
The reactivity apparatus consisted of a 2.0 L reactor
2HCl + CaC03 --> CaCl2 + H20 + CO2 (5) equipped with a pH controller. The controller regu
lated the addition of a mixture of N2/C02/HCI gas
This reaction can occur in the pH range of 6-7 with bubbled into the reactor. The reactor was initially filled
no reaction with the CO2 or inherent problems in con with a known amount of reagent in water. As the
trolling pH. reagent dissolved, pH was controlled by the addition
Even with these inherent advantages, limestone has of the acid gas. The neutralized product, CaCI2, is very
not been widely used for HCI removal and neutrali soluble in water, whereas CaC03 is slightly soluble.
2
zation. The physical characteristics of limestone's low The concentration of Ca + ion in solution as a function
solubility in water has greatly limited its use. To use of time was used as an indication of the reactivity of
this material, it must be mined and crushed to be CaC03• The reagents tested were reagent-grade CaC03,
slurried with water for scrubbing and neutralization. crushed limestone (200-325 mesh particle size), and
The feeding of a slurry poses potential problems for process precipitated CaC03 sludge. The reactivity tests
scaling and plugging in the packing and the recircu represented an initial screening of the CaC03 reagents
lation piping. In addition, most crushed and slurried to evaluate relative performance potentials.
limestone solutions have poor reactivity rates, even The bench-scale HCI scrubber apparatus is illus
with a strong acid such as HCI, due to low specific trated in Fig. 2. The synthetic flue gas was prepared
surface area. by mixing CO2, N2, and HCl. The flowrates were con
The purpose of this investigation was to determine trolled with calibrated flowmeters. The gas was
if a form of calcium carbonate that is precipitated sparged through the scrubbing solution in the contac
directly from solution in a water softening process tor, where HCI was absorbed, before exiting through
would offer an enhanced reactivity with hydrochloric a KOH impinger to a vent.
acid compared to crushed limestone due to higher spe The scrubber solution was held at constant pH in
cific surface area. Since this material is a waste product, the reaction tank by adding CaC03 reagent under pH
the costs will be primarily associated with handling feedback control. The solution was circulated through
and transportation expenses. In addition, the param the contactor and kept at constant level by placing an
eters of pH, CI and CO2 partial pressure, and CaC03/ additional effluent line at the desired level. A liquid
CaCl2 solubility, which affect reactivity, were studied. blowdown stream was used to maintain constant sys
tem volume. The system operating conditions are listed
below:
EXPERIMENTAL APPROACH
Three methods were used to examine the feasibility Total gas flowrate 0.10 ftl /min (2.83 L/min)
of using a process precipitated CaC03 sludge as a re Liquid circulation rate 100 mll min
agent for HCI removal and neutralization. Contactor liquid 50 ml
The first employed a computer simulation program volume
which estimated component solubilities and equilib Hold tank liquid 250 ml
rium conditions. The second approach involved the volume
use of a laboratory-scale reactivity apparatus for mea Contactor diameter 2.4 cm
suring the reaction rate of an alkaline reagent with Temperature 25° C
absorbed HCI gas. The third method used a bench CaC03 sludge wt% 10.0 wt%
scale HCI scrubber to determine actual operating char solids
acteristics and to compare the CaC03 sludge with lime
stone and NaOH as neutralizing reagents. The tests included operating at pH 4 and 6, HCI
The computer simulation program [5] uses theoret gas concentrations of 0.5% and 2.0%, and CO2 gas
ical equilibrium correlations to estimate solid/liquid/ concentrations of 0% and 10%. A finer crushed lime
gas equilibrium concentrations. Given partial pressure stone (95% < 400 mesh) and 1.0 N NaOH were used
of CO2 in the gas, concentration of CaCl2 in the liquid, as neutralizing reagents for three of the tests. The
and pH, the program predicted the equilibrium Caco3 operating conditions were chosen to approximate con
concentration in solution. Using this simulation, the ditions at full-scale incinerator scrubbers [6, 7, 8].
267
HCL ANALYSIS
KOH
IMPINGER
PERISTALTIC
PUMP
BLOWDOWN
LEGEND:
, ---- ----
I
, FIC = FLOW INDICATOR I CONTROLLER
, '
I PR = PRESSURE REGULATOR
l __
SP = SAMPLE POINT
REAGENT MAKEUP
EXPERIMENTAL RESULTS
Operating data collected during each 3-hr test in
cluded run time, CaC03 makeup tank weight, hold The computer simulation program, "LIQEQ," pre-
tank pH, and temperature. Table 1 shows the sampling . dicted the maximum operating pH at which CaC03
and analysis matrix with sample points indicated in dissolution could occur for a given CaCl2 concentra
Fig. 2. Samples collected during the 3-hr tests included tion. These results are shown in Fig. 3. A CO2 partial
one gas inlet and outlet sample, three liquid hold tank pressure of 0.1 atm was used for these cases. From
samples, and one final hold tank sample for weight Eq. (5), the steady state CaCl2 concentration can be
percent solids and specific gravity analysis. predicted for any system based on the gas HCI con
The inlet and outlet gas samples for HCI determi centration and removal efficiency. For systems oper
nation were sparged through three H20 impingers in ating near the pH on this curve, the CaC03 dissolution
series and the impinger liquid was analyzed for CI and reactivity will be more dependent on the CO2
by ion chromatography. The reaction tank samples partial pressure.
2
were filtered and analyzed for C03- by nondispersive The reactivity tests compared the reaction rate of
2
infrared analysis, Ca + by atomic absorption spectro HCI with CaC03 sludge, crushed limestone (200-325
photometry, and CI- by ion chromatography. mesh particle size), and reagent grade CaC03• At pH
268
TABLE 1 SAMPLING AND ANALYSIS MATRIX
Reaction Tank
Wt% Solids 1 3 Filter. Weight
Calcillm (Ca++) 3 3 AA
Chloride (Cl-) 3 3 IC
Carbonate (C0 =) 3 3 IR
3
CaC0 Slurry 18 -
R eagent
3
Feed Rate Tank Weight
5.0 the CaCOJ sludge was as reactive as the reagent limestone is evident from the low reagent utilization
grade CaCOJ and significantly more reactive than the (11.5%). The limestone used for this test was very fine,
crushed limestone. At pH 6.2 the CaCOJ sludge was with a mean particle diameter of 12 X 10-6 m (95%
more reactive than the reagent grade CaCOJ• The reac < 400 mesh). A pH of 6.0 could not be maintained
tivity differences are most likely due to differences in with the limestone.
specific surface area, which affects dissolution rates
significantly. The reactivity test results indicated a po
tential for CaCOJ sludge as a neutralizing reagent.
DISCUSSIONS AND CONCLUSIONS
The bench-scale HCI scrubber test results are shown
in Table 2. Tests CS- l through CS-8 used CaCOJ The parameters for evaluation of the CaCOJ sludge
sludge as the neutralizing agent. Tests CS-9 and CS- were pH, CO2 partial pressure, and HCI partial pres
10 were run under the same conditions as test CS-3 sure. From the bench-scale scrubber results, the CaCOJ
with finely crushed limestone (95% < 400 mesh) as sludge demonstrated more reactivity than the crushed
the reagent in CS-9, and 1.0 N NaOH in CS-lO. Test limestone. This was evident from the higher reagent
CS- l l used the fine limestone at a higher pH of 5.9. utilization for CaCOJ sludge for a given set of condi
For all of the bench-scale scrubber tests the HCI tions.
removal efficiencies approached 99.9%. The key op The system operating pH has a large effect on re
erating variable becomes, then, the reagent utilization. agent reactivity. All of the tests at pH 4 had higher
Reagent utilization is defined as the ratio of reacted utilizations due to the increased solubility of CaCOJ
reagent to reagent fed to the system. Utilization is a at low pH. The computer simulation illustrated in Fig.
direct indication of reagent reactivity. Figure 4 shows 3 supports these results. From the bench-scale scrubber
percent CaCOJ utilization as a function of pH for the tests it is evident that the operating pH must be limited
different test conditions. to 6.0 or less. Even at this sub-neutral pH, HCI removal
From test CS-12, which was run with 20 wt % efficiencies will approach 100%. At pH 4 reagent uti
limestone reagent at pH 5.9, the low reactivity of the lizations are high (> 88%), whereas at the higher pH
269
7 �-------,
6.5
5.5
I
Cl. 5
4.5
3.5
3 �----��----�--r-
o 10 20 30 40
of 6 the utilization drops to 35% for the CaC03 sludge. From a cost analysis of HCl neutralizing reagents,
This will allow reagent addition with no possibility of CaC03 sludge is more cost effective than caustic, lime
pH control overshoot. It will require, however, that a [Ca(OH2»), and limestone. Figure 5 shows the annu
set point below 6 be maintained to be controllable. alized reagent costs as a function of HCl load (kg/h)
System pH had little effect on HCI removal effi for a scrubber with 99.9% removal efficiency at pH 5.
ciency. The results in Table 1 show that removal ef Reagent utilizations used for comparison were 100%
ficiencies of 99.9% are attainable even at a reaction for caustic, 95% for lime, 65% for CaC03 sludge, and
tank pH of 4.0. This is well within the RCRA regu 50% for the limestone. Reagent costs used were $200/
lation of 99.0% minimum removal efficiency. ton for caustic, $60/ton for lime, $25/ton for lime
Partial pressure of CO2 in the gas had a more pro stone, and $1O/ton for the CaC03 sludge. The CaC03
nounced effect on reagent utilization for the pH 6 tests sludge cost was taken as the limestone cost less proc
with the lower HCI concentration of 0.5%. A minimal essing costs of $15/ton. At lower operating pH the
effect was seen for the 2.0% HCl tests and for all of cost effectiveness increases for CaC03 sludge due to
the pH 4 tests. The CO2 effect can be attributed to the increased reagent utilization.
low driving force for dissolution of CaC03 when high CaC03 sludge is an acceptable neutralizing reagent
2
liquid C03 - concentrations are present. Most incin for the scrubbing of HCI incinerator gas for several
eration flue gas compositions are in the 5-10% range reasons. It is much more cost effective than caustic as
for carbon dioxide concentrations and should not have a reagent and provides identical removal efficiencies
an adverse effect on reactivity. with better pH control capabilities. CaC03 sludge pro-
270
TABLE 2 LABORATORY-SCALE HCL SCRUBBER RESULTS
CaC0 HCl
3
HCl CO Solids CO = Ca++ Cl- Util. Removal
Test pH (%) (% t (Wt %) (mg/t ) (mg/L) (mg/L) (%) (%)
271
100
90
80
c
0 60
ij
�
:5
(")
50
0
U
<ll 2% HCI, 0% C02
U
40
�
0
20
2% HCI, 0% C02
10
Limestone
0
3 4 5 6 7
pH
272
900
800
700
600
1ii
0
()
-Vi
c 500
"
Q) C
0>",
'" If)
Q) :::>
0: 0
_.r:::.
"'I-
:::>�
400
c
c
«
300
200
100
o
100 200 300 400
vides more reactIvlty and higher reagent utilization EPA-600/2-84-052. u.s. Environmental Protection Agency, Cin
cinnati, Ohio, February 1984.
than the crushed limestone and there is no additional
[4] Chemical Marketing Reporter. 231 no. 27, (July 6, 1987).
processing involved. The CaC03 sludge can be slurried [5] "LIQEQ" computer equilibrium simulation program for
and used with no pretreatment. estimating flue gas desulfurization system liquid species concentra
tions. Developed by Radian Corporation, Austin, Texas.
[6] Bonner, T., et al. "Hazardous Waste Incineration Engi
neering." Pollution Technology Review (Noyez Data Corporation)
REFERENCES (no. 88, 1981): 29-40, 136-194.
[7] Trenholm, A., Gorman, P., and Jungclaus, G. Performance
[I] Bureau of National Affairs, Inc. "Environmental Reports." Evaluation of Full-Scale Hazardous Waste Incinerators. Volume II.
Current Developments (April 20, 1979): 2273. Incinerator Performance Results. EPA Contract No. 68-02-3177,
[2] Environmental Protection Agency. Resource Conservation November 1984.
and Recovery Act. Federal Register, January 23, 1981. [8] Gorman, P., et al. "Particulate and HCI Emissions From
[3] Keitz, E., et al. "A Profile to Existing Hazardous Waste Hazardous Waste Incinerators." In Incineration Treatment and Haz
Incineration Facilities and Manufacturers in the United States." ardous Waste, EPA-600/9-84-015 (1984), 151-159.
273