Kinetics of the reaction of Ca(OH)
2
=$y ash sorbent with SO
2
at low temperatures
Chiung-Fang Liu, Shin-Min Shih
∗, Ren-Bin Lin
Department of Chemical Engineering, National Taiwan University, Taipei 106, Taiwan Received 31 May 2001; received in revised form 20 August 2001; accepted 28 August 2001
Abstract
A di4erential 5xed-bed reactor was employed to study the reaction betweenCa(OH)2=$y ash sorbent andSO2under the conditions
similar to those in the bag 5lters of the spray-drying $ue gas desulfurization system. TheCa(OH)2=$y ash sorbent was prepared at
weight ratio of 70=30, slurrying time of 16h, and slurrying temperature of 65◦C. Foil-like calcium silicate hydrates were found in the sorbent. TheCa(OH)2=$y ash sorbent was highly reactive towardsSO2 and its initial reaction rate and maximum conversion
were much higher than pure Ca(OH)2. Increasing the relative humidity of the gas signi5cantly increased the initial reaction
rate and the maximum Ca utilization of the sorbent. Temperature andSO2concentration had slight e4ects on the initial reaction rate
and negligible e4ects on the maximum conversion. The reaction kinetics ofCa(OH)2=$y ash sorbent withSO2was well described
by the surface coverage model which assumes the sulfation rate being controlled by chemical reaction on sorbent grain surface and takes into account the surface coverage by products. The results of this study are useful to the design and operation of the dry or semidry processes usingCa(OH)2=$y ash sorbent to removeSO2from $ue gas.?2002 Elsevier Science Ltd. All rights reserved. Keywords: Flue gas desulfurization;Ca(OH)2; Kinetics; Mathematical modelling; Multiphase reaction; Pollution
1. Introduction
Reducing SO2 emission from power plants is a main
issue for the environmental protection. Many $ue gas desulfurization (FGD) processes are available for the
re-duction ofSO2emission. The dry and semi-dry FGD
pro-cesses have the advantage of lower capital cost than the wet processes commonly adopted in power plants; how-ever, the conversion of the sorbent, which is mostly hy-drated lime, in the dry and semi-dry processes is low. How to increase the utilization of hydrated lime has been an important subject for the application of the dry or semi-dry processes.
Many researchers have shown that sorbents prepared
from $y ash and hydrated lime have higher SO2
cap-ture and Ca utilization than hydrated lime (Jozewicz & Rochelle, 1986; Ho, 1987; Jozewicz, Chang, Sedman, & Brna, 1988a; Jozewicz, Jorgensen, Chang, Sedman, & Brna, 1988b; Martinez, Izquierdo, Cunill, Tejero,
∗Corresponding author. Tel.: 23633974; fax:
+886-2-23623040.
E-mail address: smshih@ccms.ntu.edu.tw (S.-M. Shih).
& Querol, 1991; Ho & Shih, 1992, 1993a; Davini, 1995; Sanders, Keener, & Wang, 1995; Garea et al., 1997a; Garea, Viguri, & Irabein, 1997b; Ishizuka, Tsuchiai, Mu-rayama, Tanaka, & Hattori, 2000). The sorbents were prepared by slurrying hydrated lime and $y ash in water for a certain period of time and drying the slurry subse-quently. Fly ash is a pozzolanic material, its main
compo-nents areSiO2; Al2O3, andFe2O3 (Taylor, 1964). In the
presence of water, amorphous silica in $y ash would re-act with hydrated lime to form calcium silicate hydrates (xCaO · SiO2·yH2O), which may lead to sorbents having
higher reactivities towardsSO2than hydrated lime.
The literature on the kinetic model of the reaction of
theCa(OH)2=$y ash sorbent withSO2, however, is scarce.
Recently, Garea et al. (1997b) studied the sulfation of
$y ash=Ca(OH)2(3=1wt. ratio) sorbent at low
tempera-tures and proposed a kinetic model based on the nonideal surface adsorption model (Irabien, Cortabitarte, & Ortiz, 1992) to describe the reaction of their sorbent.
In this work, the reaction of Ca(OH)2=$y ash
(70=30wt. ratio) sorbent withSO2was studied under the
conditions similar to those of the bag 5lters in a semi-dry scrubbing system by using a di4erential 5xed-bed 0009-2509/02/$ - see front matter?2002 Elsevier Science Ltd. All rights reserved.
reactor, and a kinetic model describing the reaction of the sorbent was derived.
2. Experimental section
The hydrated lime used was reagent grade Ca(OH)2
(purity¿ 95%; Hayashi Pure Chemical Industries, Ltd). Fly ash was from Boiler 3 of Shin-Da pulverized-coal power plant of Taiwan Power Company. The
chemi-cal compositions of the $y ash are: 59.0% SiO2; 26.7%
Al2O3, 5.5% K2O, 1.6% CaO, 1.3% TiO2, 1.2% Na2O,
0.9%MgO, 0.47% SO3, 0.05%V2O5, and 2.7% ignition
loss (Ho, 1987). The volume mean particle diameters of
Ca(OH)2 and $y ash were 7.4 and 8:1m, respectively.
The hydrated lime and $y ash, together with 80g
deionized water, were placed into a 250ml
polypropy-lene conical beaker at aCa(OH)2=$y ash weight ratio of
70=30 and a water=solid weight ratio of 10. The beaker was then sealed with a rubber stopper at the mouth and
inserted into a water bath at 65◦C
. The slurry was stirred
with a magnetic stirrer for 16h. After slurrying, the
wa-ter in the slurry was evaporated in a vacuum oven, the
solid phase left was further vacuum-dried at 105◦C
. The dried cake obtained was crashed into powder and sealed in a bottle before use. The particle size distribution of the
sorbent was 98wt% in the range of 0.6–60:6mwith a
volume mean particle diameter of 10:3m. The speci5c
(BET) surface area of the sorbent was 38:0m2=g. The
speci5c surface areas of sorbents prepared at di4erent slurrying times were found to increase with slurrying time, as shown in Fig. 1. Except those used for studying the e4ect of speci5c surface area, all the samples used
were prepared at a slurrying time of 16h.
Experiments for the reaction of the sorbent with SO2
were carried out by using a di4erential 5xed-bed reactor. The details of the experimental setup and procedure were described in Ho and Shih (1992). In this study, about
30mg sample was used for each run. The sorbent
pow-der was dispersed into quartz wool; the wool was then set into the quartz sample pan. The sample pan had
dimen-sions of 10mmo.d. and 15mmheight and was perforated
at the bottom to facilitate the passage of the sweep gas. The sweep gas entered the bottom of the reactor, passed
through the 25mmi.d. and 365mmlength outer tube, and
went downward through the sample pan and the 10mm
i.d. and 315mm length inner tube. The sweep gas was
comprised ofSO2; H2O, andN2. The SO2 andN2 gases
were supplied from cylinders and H2O vapor was
pro-vided by a water evaporator. Prior to each run, the
sam-ple bed was humidi5ed for 30min by humid N2 with a
relative humidity at which the experiment was to be per-formed, which had been proved to be long enough for the sample to equilibrate with the gas stream. After humidi-5cation, the reactive gas was admitted into the sweep gas
to start the run. The total gas $ow rate was 4l=min(STP).
The di4erential condition of the reactor with respect to
interparticleSO2 levels at the selected total gas $owrate
and sample weight was con5rmed by the results of ex-periment using di4erent sample weights (Liu, 1999). The results showed that no e4ect of sample weight on the
re-action when the weight was smaller than 40mg.
The utilization or conversion ofCa(OH)2for a reacted
sample was determined from itsSO2−
3 =Ca2+ molar ratio.
TheSO2−
3 content in a sample was determined by
iodo-metric titration, and theCa2+content by EDTA titration.
3. Results and discussion 3.1. XRD analysis
The XRD patterns of unreacted sorbent and
sam-ples reacted with SO2 for 10 and 60min are shown in
Fig. 2. The main peaks of the unreacted sorbent are those
ofCa(OH)2. The presence ofCaCO3in the sample is due
to the CO2 contamination during the sample preparing
process. The quartz is contained in the $y ash. The peaks of calcium silicate hydrates (C–S–H) formed during the preparing process are not found in this 5gure, indicating that the calcium silicate hydrates are amorphous. The
reaction product is CaSO3 ·0:5H2O. The intensities of
the characteristic peaks ofCaSO3·0:5H2Oincrease with
reaction time. 3.2. Structure
Fig. 3(a) is the SEM micrograph of a typical
Ca(OH)2=$y ash sorbent particle. It was evident that the
particle is highly porous and constituted by foil-like sub-stances which are calcium silicate hydrates. The SEM micrograph of a particle of the sorbent which had
re-acted withSO2to a conversion of 0.59, Fig. 3(b), shows
that the reacted particle is less porous and the foil-like substances become thicker.
The nitrogen adsorption and desorption isotherms for the sorbent show that the sorbent exhibits a type IV isotherm with a type H3 hysteresis (IUPAC, 1985). Type H3 hysteresis is associated with slit-shaped pores or the space between plate-like particles. Observation by SEM con5rms this pore shape (Fig. 3(a)).
The pore volume distributions of sulfated samples are shown in Fig. 4. The volumes of micropore, mesopore, and total pores smaller than 3000 NA at di4erent reaction extent are listed in Table 1. The pores of the sorbent are mainly the mesopores and macropores. Pore volumes of sulfated samples are observed to decrease with reaction extent, indicating that reaction takes place inside the pores and the reaction products formed on the pore walls have larger molar volumes than the solid reactants.
The speci5c surface areas of unreacted and sulfated samples are listed in Table 1. One can see that the
0 2 4 6 8 10 12 14 16 18 0 5 10 15 20 25 30 35 40
BET surface area, m
2 /g
Slurrying time, h
Fig. 1. E4ect of slurrying time on the speci5c surface area ofCa(OH)2=$y ash(70=30wt. ratio) sorbent. Slurrying conditions: 65◦C, L=S = 10=1.
10 20 30 40 50 60 70 80 250 500 750 1000 q h h h h c s h h s 60 min 2θ 250 500 750 1000 q q s h h h h c s h h s 10 min Intensity(CPS) 250 500 750 1000 h: Ca(OH)2 c: CaCO3 s: CaSO3 .0.5H 2O q: quartz h h h h c h h 0 min
Fig. 2. XRD patterns ofCa(OH)2=$y ash sorbent(70=30wt. ratio) reacted at 60◦C, 70% RH, and 1000ppm SO2 for di4erent reaction times.
speci5c surface areas of samples decrease with reaction extent and the reduction of surface area is much greater in the initial period. This indicates that the formation of product crystals would reduce the pore volume and make the rough reactant surface to become more smooth. This explanation can be con5rmed by the SEM observation (Fig. 3(b)).
3.3. E2ects of reaction variables
The experimental results for the reaction ofCa(OH)2=$y
ash sorbent with SO2 are shown in Fig. 5 in terms of
conversion X versus time t. As can be seen from these plots, the reaction is rapid in the initial period, but the
conversion levels o4 after about 10min, and the sorbent
is incompletely converted for a reaction time as long
as 1h. This reaction behavior is similar to that of pure
Ca(OH)2 (Ho & Shih, 1993b). However, the initial
re-action rate and the maximum conversion (for 1h) of
theCa(OH)2=$y ash sorbent are remarkably higher than
those of pure Ca(OH)2 reacted at the same
experimen-tal conditions. For example, at the conditions of 60◦C,
70% RH, and 1000ppmv SO2, the 1h conversion of the
Ca(OH)2=$y ash sorbent was 0.60, which is much higher
Fig. 3. SEM micrographs of (a) Ca(OH)2=$y ash(70=30wt. ratio) sorbent and (b)Ca(OH)2=$y ash sorbent(70=30wt. ratio) reacted at 60◦C,
70% RH, and 1000ppm SO2for 60min.
10 100 1000 0.00 0.05 0.10 0.15 0.20 0.25 0.30 o 0min 5min 40min dV/d(logD), cm 3 /g Pore Width, A
Fig. 4. Pore volume distribution ofCa(OH)2=$y ash(70=30wt. ratio) sorbent reacted at 60◦C, 70% RH, and 1000ppm SO2for di4erent reaction
times.
The e4ect of temperature (60–80◦C) on the reaction
was weak; after 10minreaction, the di4erences between
the conversions at 60◦C and 80◦C were within the
experimental error (±0:02). Increasing temperature could
raise the chemical reaction rate constant, but reduce the adsorption amounts of water vapor and reaction gas, the
Table 1
BET surface areas (m2=g) and pore volumes (cm3=g)aofCa(OH)2=$y ash(70=30wt: ratio) sorbents reacted at 60◦
C, 70% RH, and 1000ppm SO2
t (min) X Surface area Micropore vol. Mesopore vol. Total pore vol.
(¡ 20 NA) (20–500 NA) (17–3000 NA)
0 0 38.0 0.0024 0.1078 0.2318
5 0.39 19.8 0.0026 0.0701 0.1670
(23.8) (0.0031) (0.0843) (0.2009)
40 0.59 15.1 0.0016 0.0402 0.1026
(18.9) (0.0021) (0.0525) (0.1341)
aThe value in parentheses is based on the weight of the unreacted sorbent.
e4ects caused by the temperature change might o4set each other and resulted in the weak overall e4ect of tem-perature on the reaction.
The major factor a4ecting the reaction was the relative humidity of the gas phase. The initial reaction rate and the maximum conversion of the sorbent increased signif-icantly with increasing relative humidity.
The reaction was also a4ected by SO2 concentration
slightly. The di4erences between the conversions at 5000
and 1000ppm SO2 in initial stage (before 7:5min) were
higher than the experimental error, but the di4erences
after 7:5minwere within the experimental error.
The e4ect of the speci5c surface area of sorbent on the reaction was studied by using sorbents prepared at di4er-ent slurrying times. As can be seen Fig. 6, both the ini-tial reaction rate and the maximum conversion increased with increasing sorbent speci5c surface area.
3.4. Analysis of reaction kinetics
The reaction rate ofCa(OH)2=$y ash sorbent withSO2
is fast in the initial stage and decreases abruptly, leaving the sorbent incompletely converted. Relative humidity is the major factor a4ecting the reaction, whereas the
in$u-ences of temperature and SO2 concentration on the
re-action are small. This rere-action behavior of Ca(OH)2=$y
ash sorbent is similar to the reaction behaviors of pure
Ca(OH)2 with SO2 (Ho, 1987; Ho & Shih, 1993b; Ho,
Shih, & Lee, 1996) and with CO2 (Shih, HO, Song,
& Lin, 1999). From previous kinetic studies by Ho et al. (1996) and Shih et al. (1999), the surface coverage model was found to be the most suitable model to describe the
reaction between Ca(OH)2 and SO2 or CO2. The same
model was employed in this study to describe the
reac-tion kinetics ofCa(OH)2=$y ash sorbent withSO2.
The hypothesis of the surface coverage model is that the sorbent is made up of plate grains and that the reaction is controlled by chemical reaction on the surface of a grain and the reacting surface area of the grain decreases with the deposition of solid product. According to this model, the reaction of a sorbent reaches a maximum conversion when its reacting surface is fully covered by the product. Thus the reaction rate of sorbent per unit initial surface
area, rs, can be expressed as
rs= ks; (1)
where ks is a function of temperature, concentrations of
reacting species, and relative humidity, is the fraction of surface area which is not covered by product. The rate of conversion of the sorbent is
dX=dt = Sg0Mrs= Sg0Mks; (2)
where Sg0is the initial speci5c surface area of the sorbent
and M is the sorbent weight per mole Ca.
How changes with reaction time depends on the reaction rate and the way by which the product deposits on the surface; one may assume that
−d=dt = kprs= kpks; (3)
where kp is a proportional constant, which is a function
of temperature, concentrations of reacting species, and relative humidity. Eq. (3) can be integrated to get as a function of time
=exp(−k1k2t); (4)
where
k1= ksSg0M; (5)
k2= kp=(Sg0M): (6)
By substitution of Eq. (4) into Eq. (2), one can inte-grate Eq. (2) to obtain the relation between conversion and time
X = [1− exp(−k1k2t)]=k2: (7)
From Eqs. (2), (4), and (7), the rate of conversion can be expressed as a function of conversion
dX=dt = k1(1−k2X ): (8)
Equation similar to Eq. (7), named as asymptotic equa-tion, has been used to describe the 5lm growth for the oxidation of metals at low temperatures (Scully, 1975; Evans, 1981). Also similar equation which is obtained by the pore-plugging model has been applied to the sulfa-tion of limestone at high temperatures (Lee & Georgakis, 1981).
0 10 20 30 40 50 60 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 70oC 30%RH 50%RH 70%RH 80%RH 1000ppmSO2 3000ppmSO2 5000ppmSO2 X t, min (b) 0 10 20 30 40 50 60 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 80oC 30%RH 50%RH 70%RH 80%RH 1000ppmSO2 3000ppmSO2 5000ppmSO2 X t, min (c) 0 10 20 30 40 50 60 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 60oC 30%RH 50%RH 70%RH 80%RH 1000ppmSO2 3000ppmSO 2 5000ppmSO2 X t, min (a)
Fig. 5. Plot of conversion versus time for the sulfation ofCa(OH)2=$y
ash(70=30wt. ratio) sorbent: (a) 60◦C, (b) 70◦Cand (c) 80◦C.
The two parameters, k1 and k2, in Eq. (7) can be
ob-tained by least-squares 5tting of Eq. (7) to the
experi-mental data. One can easily see from Eq. (7) that k2 is
the reciprocal of the maximum conversion. Figs. 6 and 7 show that the experimental data are described by Eq. (7)
very well. According to the de5nitions of k1 and k2, they
are functions of initial speci5c surface area, temperature,
SO2 concentration, and relative humidity. These
func-tions can be derived by analyzing the values of k1 and k2
obtained at di4erent experimental conditions.
The values of k1 and k−1
2 obtained for samples
of di4erent initial speci5c surface areas are plotted in Figs. 8 and 9, respectively. One can see that
both k1 and k−1
2 follow the linear relationship with
Sg0M (M = 108:8g=mol Ca) in accordance with
Eqs. (5) and (6), respectively. This indicates that the present model can adequately describe the reaction
ki-netics and that ks and kp for the sorbent did not change
with the speci5c surface area. The values of k1 and k2
for pure Ca(OH)2 (Sg0= 9:5m2=g; M = 74g=mol Ca)
reacted under the same conditions were also plotted in
the above 5gures; one can see that ks of Ca(OH)2 is
about the same as that of Ca(OH)2=$y ash sorbent, but
kpofCa(OH)2 is smaller.
The following analyses were based on the experimental results obtained for samples having a speci5c surface area
of 38:0m2=g.
As shown in Fig. 10, plots of ln k1versus relative
hu-midity, RH (%), are linear and the 5tting lines have nearly
the same slopes for the temperatures andSO2
concentra-tions considered, indicating that the relaconcentra-tionship between
k1 and relative humidity is
k1=e0:0102RHf(y; T); (9)
where f(y; T) is a function of SO2 concentration,
y (ppm), and temperature, T.
Values of ln(k1e−0:0102RH) were plotted againstlny in
Fig. 11. The slopes of the straight 5tting lines seem to decrease as temperature increases; however, the di4er-ences between the slopes are within the range of varia-tion estimated from the error of the data. Thus the slope was considered to be independent of temperature. Using the average value of the slopes of the straight lines in
Fig. 11, the relationship between k1e−0:0102RH and y can
be written as
k1=e0:0102RHy0:17g(T); (10) where g(T) is a function of temperature.
Values of g(T) for di4erent temperatures were
calcu-lated from k1e−0:0102RHy−0:17 and the average value for
each temperature is shown in Fig. 12. The plot shows that g(T) can be expressed by the Arrhenius law. The linear least-squares 5tting of the data gives
k1= 0:83e0:0102RHy0:17e−9977=RT: (11)
The small value of the apparent activation energy,
9:977kJ=mol, indicates that the e4ect of temperature on
the reaction rate is weak.
As can be seen from Fig. 13, relative humidity is the only variable a4ecting the value of k2, and the logarithmic
0 10 20 30 40 50 60 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
BET surface area Symbol (m2/g) 27.2 29.8 32.0 35.0 38.0 Model X t, min
Fig. 6. E4ect of Sg0 on the reaction ofCa(OH)2=$y ash sorbent withSO2. Reaction conditions: 60◦C, 70% RH, and 1000ppm SO2.
0 10 20 30 40 50 60 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 Model RH,% Exp. 30 50 70 80 X t, min
Fig. 7. Comparison of model predictions and experimental data for the reaction ofCa(OH)2=$y ash at 60◦C, 1000ppm SO2, and various relative
humidities.
plot of k2 versus RH is linear. The relationship between
k2 and relative humidity obtained from Fig. 13 is
k2= 66:0RH−0:864: (12)
Since the constants in Eqs. (11) and (12) contain the initial speci5c surface area, Sg0, of the sample, they are
only applicable to the sample which has a Sg0 value of
38:0m2=g. Thus for sorbents with other values of Sg0,
Eqs. (11) and (12) can be rewritten as
k1= 0:0218 Sg0e0:0102RHy0:17e−9977=RT; (13)
k2= 2508:0 S−1
g0 RH−0:864: (14)
Eq. (7) together with Eqs. (13) and (14)
consti-tutes the kinetic model for the sulfation ofCa(OH)2=$y
ash(70=30wt. ratio) sorbent. Conversions calculated by
the kinetic model are compared with the experimental results in Fig. 14, which shows that the calculated re-sults are in good agreement with experimental data with
a standard deviation (N−1) of 0.02 in conversion.
3.5. Discussion on reaction kinetics
From the above kinetic analysis, one can see that rel-ative humidity is the major factor a4ecting the reaction.
0 1000 2000 3000 4000 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 0.18
Ca(OH)2/fly ash Ca(OH)2 k1 , m in -1 Sg0M, m2/mol Ca
Fig. 8. Plot of k1 versus Sg0M. Reaction conditions: 60◦C, 70% RH,
and 1000ppm SO2. 0 1000 2000 3000 4000 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
Ca(OH)2/fly ash Ca(OH)2
1/k
2
Sg0M, m2/mol Ca
Fig. 9. Plot of k−1
2 versus Sg0M. Reaction conditions: 60◦C, 70%
RH, and 1000ppm SO2.
This implies that the water adsorbed on the sorbent sur-face plays an important role in the reaction. Fig. 15 shows
the water adsorption isotherms on the sorbent at 60–80◦C
; one can see that the amount of water adsorbed increased with increasing relative humidity and decreased with in-creasing temperature. The monolayer water adsorption capacity at each temperature was estimated by the BET equation, and the water layer thickness equivalent to the
30 40 50 60 70 80 0.1 0.3 0.2 60oC 1000ppm 3000ppm 5000ppm k1 , m in -1 RH 30 40 50 60 70 80 0.1 0.3 0.2 70oC 1000ppm 3000ppm 5000ppm k1 , m in -1 RH 30 40 50 60 70 80 0.1 0.3 0.2 80oC 1000ppm 3000ppm 5000ppm k1 , m in -1 RH (a) (b) (c)
Fig. 10. Relationship between k1 and relative humidity: (a) 60◦C,
(b) 70◦Cand (c) 80◦C.
amount of water adsorbed on the sorbent surface at each relative humidity was calculated. The thickness is about
1000 0.1 5000 3000 500 0.06 0.07 0.08 0.09 60oC 70oC 80oC k1 /e 0.0102 R H , m in -1 y, ppm
Fig. 11. Relationship between k1 andSO2 concentration.
Fig. 12. Arrhenius plot of k1.
80% RH, being nearly independent of temperature (Liu, 1999).
The fundamental processes taking place at the water
adsorbed surface may include: (1) adsorption ofSO2 on
the outer surface of water layer; (2) hydration ofSO2 to
formSO2·H2O; (3) di4usion ofSO2·H2Oinward; (4)
dis-solution ofCa(OH)2and C–S–H to formCa2+andOH−;
(5) di4usion of Ca2+ andOH− outward; (6) reaction of
OH− withSO
2· H2Oto formHSO−3 andSO2−3 ; (7)
reac-tion of Ca2+ withSO2−
3 to form calcium sul5te
precipi-tate. These processes have been considered to take place
in the absorption ofSO2into a lime slurry. In the present
case, the water is only several molecular layers thick, these processes may also take place, but the rates of the processes involving reactions (processes 2, 4, 6, and 7)
10 100 1 70 50 30 4 3 2 60oC 1000ppm 3000ppm 5000ppm k2 RH 10 100 1 70 50 30 4 3 2 70oC 1000ppm 3000ppm 5000ppm k2 RH 10 100 1 70 50 30 4 3 2 80 o C 1000ppm 3000ppm 5000ppm k2 RH (a) (b) (c)
Fig. 13. Relationship between k2 and relative humidity: (a) 60◦C,
(b) 70◦
Cand (c) 80◦
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 N-1=0.02 Calculated X Experimental X σ
Fig. 14. Comparison of the calculated and the experimental conversion values. 0 20 40 60 80 100 0 20 40 60 80 60oC 70oC 80oC H2 O Adsorbed, g/kg sorbent RH, %
Fig. 15. Water adsorption isotherms on Ca(OH)2=$y ash(70=30wt.
ratio, 38m2=g) sorbent.
may be greatly reduced, and the di4usion of species (pro-cesses 3 and 5) through such thin layer should be very fast.
Since the overall reaction rate of the sorbent decreases
with conversion and is slightly a4ected bySO2
concentra-tion in the early period of the reacconcentra-tion, the rate-controlling
step for the reaction may be that involving OH− or
Ca2+ ions, which are generated by the dissolution of the
sorbent surface where the surface is not covered by the product, i.e., process 6 or 7. Process 6 is the reaction of
OH− withSO
2· H2Oand process 7 is that of Ca2+ with
SO2−
3 .
The slight e4ect ofSO2 concentration may be due to
that theSO2 molecules adsorbed on the outer surface of
water layer is limited to a monolayer capacity, which
is insensitive to the SO2 concentration in the SO2
con-centration range used in this study like the behavior of the Langmuir adsorption isotherm (Smith, 1981). The
adsorbedSO2 molecule then react with water molecules
to form SO2 · H2O. If the reaction between OH− and
SO2 · H2O is the slowest step, the concentration of
SO2· H2O would be in equilibrium with the amount of
SO2 adsorbed, hence the rate would be a weak function
ofSO2concentration. If the reaction ofCa2+withSO2−3 is
the slowest step, the concentration ofSO2−3 would also be
in equilibrium with the amount ofSO2 adsorbed, and the
rate would also be insensitive to theSO2 concentration.
The rate of either process 6 or 7 will increase with increasing relative humidity because higher amount of adsorbed water can produce higher amounts of reactants to react.
The mild e4ect of temperature on the overall reaction rate may be due to that as the temperature increases, the chemical reaction rate constants increase, whereas the
amounts of water andSO2adsorbed decrease.
The value of k−1
2 or the maximum conversion
in-creases as the relative humidity inin-creases. This phe-nomenon may indicate that the thicker water layer provides a wider range for the product molecules to deposit. Thus at lower relative humidity the reaction product covers the sorbent surface more uniformly, and the overall reaction rate diminishes at a lower conver-sion; whereas at higher relative humidity the product builds up more clusterlike and covers less surface, and the reaction reaches a higher maximum conversion. Similar arguments have been set forth by Krammer, Brunner, Khinast, and Staudinger (1997) and Shih et al. (1999) to explain the e4ect of relative humidity on
the reaction rates of Ca(OH)2 with SO2 and with CO2,
respectively.
In the actual $ue gas, CO2; O2, and NOX are
present with SO2; CO2 has the ability to react with the
Ca(OH)2=$y ash sorbent, and O2 and NOX can oxidize
calcium sul5te to calcium sulfate. The e4ects of these
components on the reaction of theCa(OH)2=$yash
sor-bent with SO2 have been studied by the authors, and
the results will be reported in the future. The results of the present study can be extended to the more realistic case. Accurate modelling of the sulfation kinetics of
Ca(OH)2=$y ash sorbent is important to the design and
operation of the dry or semidry processes using this kind of sorbent.
4. Conclusion
The kinetics of the reaction of Ca(OH)2=$y ash (70=
studied at 60–80◦C by using a di4erential 5xed-bed
reactor.
The utilization of Ca content of Ca(OH)2=$y ash
sor-bent is incomplete. Both the initial sulfation rate and
the maximum conversion ofCa(OH)2=$yash sorbent are
higher than those of pureCa(OH)2. Relative humidity is
the most important factor a4ecting the reaction and the initial reaction rate and maximum conversion of sorbent increase with increasing relative humidity. Temperature
and SO2 concentration have slight e4ects on the initial
reaction rate and negligible e4ects on the maximum con-version.
The reaction kinetics ofCa(OH)2=$yash sorbent with
SO2can be well described by the surface coverage model
proposed by Shih et al. (1999) which assumes the sul-fation rate being controlled by chemical reaction on the sorbent grain surface and takes into account the surface coverage by product.
The results of this study are useful to the design and operation of the dry or semi-dry processes
us-ing Ca(OH)2=$y ash sorbent to remove SO2 from $ue
gas. Notation
f(y; T) function de5ned by Eq. (9)
g(T) function de5ned by Eq. (10)
kp constant de5ned by Eq. (3),m2=mol
ks initial reaction rate of solid,mol=(min m2)
k1 ksSg0M; min−1
k2 kp=Sg0M, dimensionless
M weight of solid per mole Ca, g=mol Ca
R gas constant, 8:314J=mol K
RH relative humidity, %
rs reaction rate of solid per unit initial surface
area,mol=(min m2)
Sg0 initial speci5c surface area of solid,m2=g
t time, min
T reaction temperature, K
X conversion, dimensionless
y SO2concentration, ppm
Greek letters
N−1 standard deviation, dimensionless
fraction of surface area which is not covered
by product, dimensionless Acknowledgements
This research was supported by the National Sci-ence Council of Republic of China under grant NSC 87-2211-E002-004.
References
Davini, P. (1995). Investigation of $ue gas desulfurization by $y ash and calcium hydroxide mixtures. Resources Conservation and Recycling, 15, 193–202.
Evans, U. R. (1981). An introduction to metallic corrosion (3rd ed.). London: Edward Arnold.
Garea, A., Fernandez, J., Viguri, J. R., Ortiz, M. I., Renedo, M. J., & Irabien, A. (1997a). Fly ash=calcium hydroxide mixtures for
SO2removal: Structural properties and maximum yield. Chemical
Engineering Journal, 66, 171–179.
Garea, A., Viguri, J. R., & Irabien, A. (1997b). Kinetics of the $ue gas desulfurization at low temperature: Fly ash=calcium (3=1) sorbent behavior. Chemical Engineering Science, 52, 715–732.
Ho, C. S. (1987). Reaction of Ca(OH)2 and 8y ash=Ca(OH)2
slurry with SO2. MS Engineering thesis, Department of Chemical
Engineering, National Taiwan University, Taipei, Taiwan, ROC. Ho, C. S., & Shih, S. M. (1992). Ca(OH)2=$y ash sorbents for
SO2removal. Industrial Engineering and Chemical Research, 31,
1130–1135.
Ho, C. S., & Shih, S. M. (1993a). Characteristic and SO2 capture
capacities of sorbents prepared from products of spray-drying $ue gas desulfurization. Canadian Journal of Chemical Engineering, 71, 934–939.
Ho, C. S., & Shih, S. M. (1993b). Factors in$uencing the reaction ofCa(OH)2 withSO2. Journal of Chinese Institute of Chemical
Engineering, 24, 187.
Ho, C. S., Shih, S. M., & Lee, C. D. (1996). In$uence ofCO2 and
O2 on the reaction of Ca(OH)2 under Spraying–Drying $ue gas
desulfurization conditions. Industrial Engineering and Chemical Research, 35(11), 3915–3919.
Irabien, A., Cortabitarte, F., & Ortiz, M. I. (1992). Kinetics of $ue gas desulfurization at low temperature: Nonideal surface adsorption model. Chemical Engineering Science, 47, 1533–1543.
Ishizuka, T., Tsuchiai, H., Murayama, T., Tanaka, T., & Hattori, H. (2000). Preparation of active absorbent for dry-type $ue gas desulfurization from calcium oxide, coal $y ash, and gypsum. Industrial Engineering and Chemical Research, 39, 1390–1396.
IUPAC, (1985). Reporting physisorption data for gas=solid systems with special reference to the determination of surface area and porosity. Pure and Applied Chemistry, 57, 603–619.
Jozewicz, W., Chang, J. C. S., Sedman, C. B., & Brna, T. (1988a). Silica-enhanced sorbents for dry injection removal of SO2 from
$ue gas. Journal of Air Pollution Control Association, 38, 1027–1034.
Jozewicz, W., Jorgensen, C., Chang, J. C. S., Sedman, C. B., & Brna, T. (1988b). Development and pilot plant evaluation of silica-enhanced lime sorbents for dry $ue gas desulfurization. Journal of Air Pollution Control Association, 38, 796–805. Jozewicz, W., & Rochelle, G. T. (1986). Fly ash recycle in dry
scrubbing. Environmental Progress, 5, 219–224.
Krammer, G., Brunner, Ch., Khinast, J., & Staudinger, G. (1997). Reaction of Ca(OH)2 with SO2 at low temperature. Industrial
Engineering and Chemical Research, 36, 1410–1418.
Lee, D. C., & Georgakis, C. (1981). A single, particle-size model for sulfur retention in $uidized bed coal combustors. A.I.Ch.E. Journal, 27, 472–481.
Liu, C. F. (1999). Kinetics of the reaction of Ca(OH)2=silica fume
and Ca(OH)2=8y ash sorbents withSO2. MS. Engineering thesis,
Department of Chemical Engineering, National Taiwan University, Taipei, Taiwan, ROC.
Martinez, J. C., Izquierdo, J. F., Cunill, F., Tejero, J., & Querol, J. (1991). Reactivation of $y ash and Ca(OH)2 mixtures for
SO2 removal of $ue gas. Industrial Engineering and Chemical
Sanders, J. R., Keener, T. C., & Wang, J. (1995). Heated $y ash=hydrated lime slurries for SO2 removal in spray dryer
absorbers. Industrial Engineering and Chemical Research, 34, 302–307.
Scully, J. C. (1975). The fundamentals of corrosion (2nd ed.). Oxford: Pergamon Press.
Shih, S. M., Ho, C. S., Song, Y. S., & Lin, J. P. (1999). Kinetics of the reaction ofCa(OH)2 withCO2 at low temperature. Industrial
Engineering and Chemical Research, 38(4), 1317–1322.
Smith, J. M. (1981). Chemical engineering kinetics (3rd ed.). New York: McGraw-Hill.
Taylor, H. F. W. (1964). The chemistry of cement. London: Academic Press.
