500 M. S. Hatamipour et al.

Leila Kavoshi Research Article

Mohammad S. Hatamipour
Amir Rahimi Kinetic Modeling of Reactive Absorption
University of Isfahan, Chemical of Carbon Dioxide in a Spray Dryer
Engineering Department,
Isfahan, Iran.
A semi-empirical kinetic correlation was obtained through a shrinking core mod-
el assumption for reactive absorption of CO2 with NaOH solution by applying
response surface method analogy in a laboratory-scale spray-dryer absorber. The
effect of approach temperature, absorbent concentration, nozzle diameter, and
L/G ratio on the kinetic coefficient was studied and the optimum operating
conditions to reach the maximum absorption were determined.

Keywords: Carbon dioxide, Kinetic equation, Response surface methodology, Shrinking core
model, Spray-dryer absorber
Received: June 01, 2012; revised: December 05, 2012; accepted: December 07, 2012
DOI: 10.1002/ceat.201200304

1 Introduction Up to now, its performance has been improved noticeably and

high removal efficiency rates are reported by this company [2].
Carbon dioxide is the most abundant greenhouse gas in the SDAs are being used in flue gas desulfurization (FGD) process-
atmosphere and plays an important role in plant and animal es and are installed as complementary facility in coal-burning
growth processes, such as photosynthesis and respiration, but power plants and for incinerators in many countries in Europe,
during the last 150 years the amount of CO2 released in the USA, and Asia for treatment of acidic flue gases. The structure
atmosphere has been significantly growing. The development of an SDA is based on a simple concept: hot, untreated flue/
of new industries and the irregular use of fossil fuels have kiln gas is fed into a drying chamber where a fine alkaline solu-
increased the amount of CO2 emission in the atmosphere. Cat- tion is sprayed on. The acidic components in the flue gas are
astrophic crises in many countries are the results of huge emis- absorbed by the alkaline droplets and converted into neutral-
sions of greenhouse gases, especially of CO2. Controlling the ized materials through chemical reactions. Due to the high
emission of these greenhouse gases should be the most impor- temperature, water is evaporated simultaneously. Experimental
tant environmental challenge at global level especially for and theoretical studies on the desulfurization process using
China, USA, and European Union, the main producers of SDAs are reported by Hill and Zank [3] and Scala et al. [4].
CO2. In recent years, CO2 production in Iran has been rising Experimental CO2 removal by a spray dryer was described
to a point that in 2009 Iran was declared the 7th in annual by Chen et al. [5] and Kavoshi et al. [6] but the lack of detailed
CO2 production [1]. Considering the importance of the mat- studies in this field promoted the idea of conducting this
ter, special effort is required in CO2 emission control. Different research. In our previous work, the reactive absorption of CO2
processes including reactive absorption, adsorption, mem- was studied experimentally and the most effective parameters
brane process, and cryogenic separation could be adapted for were determined through an experimental design procedure
post-combustion CO2 removal. New technologies, including [6]. In a recent work by Völkl et al. [7], the physical as well as
photocatalytic degradation, biological processes, and chemical the chemical solubility of CO2 in several ionic liquids were
synthesis are also under development. Selecting a CO2 removal predicted with quantum chemical a priori methods. In their
process greatly depends on the type of the fossil fuel burning work, by combining reaction and phase equilibrium, the CO2
unit from which CO2 is released. Among different processes, loading of the reactive solvents were predicted and the results
reactive absorption has undergone extensive studies and is an validated with literature data. An approach was developed for
applicable and economically feasible method for CO2 capturing. a quantitative a priori description of both physical and chemi-
The spray-dryer absorber (SDA) as a new facility for cal solvation of CO2.
removal of acidic gases was patented by GEA Niro Co. in 1980. Here, through conceptual combination of a theoretical
kinetic model accompanied with a set of experimental data, a
– semi-empirical kinetic model is obtained for reactive absorp-
Correspondence: Dr. M. S. Hatamipour ([email protected]), Uni-
tion of CO2 in a spray-dryer setup and the optimum operating
versity of Isfahan, Chemical Engineering Department, Hezar jarib st., conditions for maximum absorption efficiency are deter-
Isfahan 81746-73441, Iran. mined.

www.cet-journal.com © 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim Chem. Eng. Technol. 2013, 36, No. 3, 500–506
 Spray-dryer absorber 501

2 Experimental signed and installed at the exit line before the measuring
point.
2.1 Experimental Setup Each experiment was carried out within about 30 min. In
[6], the liquid/gas ratio (L/G) was fixed at 0.01 mL L–1 in order
The experiments were carried out in a laboratory-scale spray to achieve complete drying, while in this study different L/G
dryer (Armfield SD-05, LAB-PLANT UK Ltd.). The schematic ratios were examined in order to study the effect of this ratio
diagram of the process as reported in our previous work [6] is on the CO2 removal efficiency. However, some of these values
illustrated in Fig. 1. The equipment consisted of a main glass led to intensive weeping and accumulation of injected solution
chamber with 215 mm inner diameter and 500 mm height, in the collecting vessel located at the bottom of the spray
equipped with a fan and a feed pump. Feed and hot gas chamber.
streams flowed cocurrently through the drying chamber.
Clean air and pure CO2 were mixed in a static mixer and
used as the inlet gas. This mixture was heated through an elec- 2.2 Reaction Analysis by the Shrinking Core Model
trical heater before entering the drying chamber. A valve was
installed between static mixer and main heater for sampling When NaOH solution is sprayed into the hot gas stream, it
and measuring the inlet concentration of CO2. The absorbent breaks down to very fine liquid droplets and the reaction takes
was NaOH (Merck, Germany) purchased from a chemical place between CO2 and NaOH. The water content of droplets
trading company in Iran. For preparation of the absorbent so- is evaporated from the droplets surface simultaneously. The
lution, solid NaOH was weighed and dissolved in distilled overall reaction is as follows:
water. After the temperature in the spray dryer reached its pre- CO2 (g) + H2O + 2NaOH → Na2CO3 + 2H2O (1)
defined value and the inlet concentration of CO2 became
stable, the feed pump was turned on and the prepared absor- As mentioned before, due to simultaneous reaction and
bent was pumped and sprayed by a pressure-type nozzle. A evaporation processes, a solid crust is formed rapidly around
CO2 analyzer (Guardian Plus- D600, USA) measured the inlet each droplet. For the remainder of the drop’s residence time,
and outlet CO2 concentration in the gas stream. Since the CO2 the CO2 must diffuse through the solid layer to reach the un-
analyzer was sensitive to high temperatures and high velocity reacted NaOH. For this fluid-fluid reaction, since evaporation
of the gas stream, a cooling and expansion chamber was de- coincides with chemical reaction, the overall reaction could be

Figure 1. Schematic diagram of

the experimental spray-dryer
system.

Chem. Eng. Technol. 2013, 36, No. 3, 500–506 © 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.cet-journal.com
502 M. S. Hatamipour et al.

simulated like a fluid-solid reaction. Experimental observa- and mass transfer, well mixing of two existing phases, and
tions indicate that the shrinking core model (SCM), first de- using mean values for variable parameters, n_ A may be consid-
veloped by Yagi and Kunii [8], could be adopted for describing ered as a constant value. Thus, by substitution of the unreacted
the behavior of this reaction (see Fig. 2). core radius and KCB with the following assumed expressions,
Eq. (5) can be rewritten as Eq. (6):

R
rˆ
2

KCB ˆ K ′ ˆ cte

CAb
n_ A ˆ (6)
4=3pR3 4=3pR3 4=3pR3
‡ ‡
4pR2 kC 4pRDe pR2 K ′

Eq. (6) indicates that n_ A could be presented as Eq. (7):

Figure 2. Shrinking core model (SCM).

n_ A ˆ Kapp CAb (7)

For this special situation, the instantaneous reaction rate of where

gas A (CO2) with liquid B (NaOH solution) occurs during 1
three steps. The overall rate of reaction is obtained as follows. Kapp ˆ (8)
4=3pR3 4=3pR3 4=3pR3
CO2 transfer rate from bulk of the gas to droplet surface: ‡ ‡
4pR2 kC 4pRDe pR2 K ′
4pR2
n_ A ˆ k C CAR † (2)
4=3pR3 C Ab It is not possible to measure directly the existing parameters
(R, kc, De, K′) in Eq. (8). Therefore, measurable parameters
where R is the droplet radius, kC is the gas-phase mass transfer that have a great effect on R, kc, De, and K′ were distinguished
coefficient, CAb is the bulk concentration of CO2 in the gas experimentally and a correlation that relates Kapp to these
phase, and CAR is the concentration of CO2 on the droplet sur- parameters is obtained. Here, R depends on the nozzle diame-
face. ter, K′ on the NaOH concentration and reaction temperature,
CO2 diffusion through the solid layer to reach the unreacted and kc mostly on the temperature and L/G ratio.
core: Selected parameters were inlet gas temperature, T, NaOH
mass concentration, CB, nozzle diameter, D0, and L/G ratio. In
4pDe CAR CAr †
n_ A ˆ   (3) order to obtain a correlation that relates the nozzle diameter,
1 1
4=3pR3 NaOH saturation concentration, temperature, and L/G ratio to
r R
Kapp through experimental results, response surface methodol-
where De is the diffusion coefficient of CO2 through the solid ogy (RSM) was applied. Since the approach temperature
layer, CAr is the CO2 concentration on the unreacted core, and parameter which is the difference between the exit temperature
r is the radius of the unreacted core. and the adiabatic saturation temperature of the inlet gas,
Reaction of CO2 with NaOH: DTapp, is more illustrative than the inlet temperature, it is used
instead of the inlet temperature in the correlation.
4pr 2 Tab. 1 summarizes the selected parameters which were
n_ A ˆ KCB CAr (4)
4=3pR3 applied as operating conditions. Three levels were selected for
each parameter. Experiments were conducted in the absence of
where K is the rate constant for reaction (1) and CB is the
other acidic gases, in spite of the fact that their presence may
absorbent concentration.
have a noticeable effect on CO2 removal efficiency [9]. In all
Considering a quasi-steady-state condition, the right-hand
experiments, the CO2 volume concentration was 15 %.
sides of Eqs. (2)–(4) are equal. By eliminating the interfacial
concentrations CAR and CAr, the following equation is ob-
tained: Table 1. Factors and their levels in experimental design.
CAb
n_ A ˆ   (5) Factor Level 1 Level 2 Level 3
1 1
4=3pR3 T [°C] 100 150 200
4=3pR3 r R 4=3pR3
‡ ‡ –3 –3
4pR2 kC 4pDe 4pr 2 KCB CB ×10 [mol m ] 1.32 2.77 3.96
D0 [mm] 0.5 0.7 1
Generally, n_ A is not constant, but in this case, due to the low L/G [mL L ] –1
0.01 0.03 0.05
residence time of liquid droplets, high interfacial area for heat

www.cet-journal.com © 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim Chem. Eng. Technol. 2013, 36, No. 3, 500–506
 Spray-dryer absorber 503

2.3 Proposing a Correlation for Kapp by RSM Table 2. Designed experiments and calculated kinetic coefficients.

RSM was applied for experimental design. Through this Run T CB ×10–3 D0 L/G DTapp g Kapp
[°C] [mol m–3] [mm] [mL L–1] [°C] [%]
method, the correlation between a response variable and a
set of quantitative experimental factors or variables could 1 100 1.32 0.5 0.01 42 52.7 0.661342
be obtained. Application of RSM has gained prominence
2 200 1.32 0.5 0.01 95 38.7 0.354254
in process design and optimization tasks due to its simplic-
ity and possibility of correlating the model parameters by 3 100 3.96 0.5 0.01 38 38.6 0.410671
means of the measured responses [10, 11]. Response sur- 4 200 3.96 0.5 0.01 104 27.4 0.209158
face design is selected because it provides a reasonable dis-
tribution of data points throughout the region of interest 5 100 1.32 1 0.01 38 40.7 0.442315
and the sequence of the most important factors. Response 6 200 1.32 1 0.01 108 26.0 0.200458
surface designs do not require a large number of runs and
too many levels of the independent variables [10]. 7 100 3.96 1 0.01 40 25.3 0.224678
The response can be related to the selected factors by lin- 8 200 3.96 1 0.01 107 14.0 0.048502
ear or quadratic models. A quadratic model, including the
9 100 1.32 0.5 0.05 15 59.3 0.724483
linear model, is given as:
10 200 1.32 0.5 0.05 83 46.0 0.448285
X
k X
k k X
X
Z ˆ b0 ‡ b j xj ‡ b jj xj ‡
2
i<j
b ij xi xj ‡ e (9) 11 100 3.96 0.5 0.05 32 47.3 0.537926
jˆ1 jˆ1 jˆ1
12 200 3.96 0.5 0.05 88 32.0 0.236719

where Z is the response, xi and xj are the design variables, k 13 100 1.32 1 0.05 29 47.3 0.529719
is the number of independent variables (factors), b0 is the 14 200 1.32 1 0.05 87 34.7 0.273658
constant coefficient, bj, bjj, and bij are the coefficients of
linear, quadratic, and interaction terms, respectively, and e 15 100 3.96 1 0.05 22 32.7 0.274315
is the error value. The first sigma term is related to the 16 200 3.96 1 0.05 93 20.7 0.095791
main effects of variables, the second to the curvature of the
17 100 2.77 0.7 0.03 25 59.3 0.750745
response surface, and the third to the interaction between
variables. 18 200 2.77 0.7 0.03 93 46.0 0.4715
According to the above description, Kapp is correlated as
19 150 1.32 0.7 0.03 59 58.0 0.709476
a combination of the main factors and their squares and
interactions: 20 150 3.96 0.7 0.03 69 47.3 0.528971

L 21 150 2.77 0.5 0.03 71 72.0 1.058954

Kapp ˆ a0 ‡ a1 DTapp ‡ a2 CB ‡ a3 D0 ‡ a4 ‡ a5 DTapp 2
G 22 150 2.77 1 0.03 61 61.3 0.785205
 2
L
‡ a6 CB 2 ‡ a7 D0 2 ‡ a8 ‡ a9 DTapp CB ‡ a10 DTapp D0 23 150 2.77 0.7 0.01 89 59.3 0.803381
G
L L L 24 150 2.77 0.7 0.05 49 66.0 0.864922
‡ a11 DTapp ‡ a12 CB D0 ‡ a13 CB ‡ a14 D0 (10)
G G G 25 150 2.77 0.7 0.03 59 66.0 0.886885

Here, for determination of the levels of factors that lead required in many design procedures, and an average response
to maximum absorption, the desirability function technique for each trial is reported.
(also called utility transfer function) is applied. In this tech-
nique, a target value and allowable minimum and maximum
values are selected for optimization. In the case of maximiza-
3 Results and Discussion
tion, the responses that are smaller than the allowable mini-
For a quadratic model for the kinetic coefficient (Kapp), coeffi-
mum values of desirability, d, are zero. The closer the response
cients of Eq. (10) were estimated through nonlinear regression.
to the target, the closer the desirability is to 1 [10].
Tab. 3 illustrates the results of the regression.
Tab. 2 lists the conditions of experiments designed by RSM
In Eq. (10), the terms with P-values less than 0.05 are signifi-
(central composite design) and the calculated Kapp for each of
cant. Values of less than 0.01 indicate that the term exerts a
them. The removal efficiency for each run is also reported in
highly significant influence on all dependent variables, whereas
this table. The removal efficiency is calculated as follows:
values greater than 0.1000 denote no significant terms. After
Qin Cin Qout Cout regression it was concluded from P-values that some terms
gˆ × 100 (11)
Qin Cin should be omitted from the equation, and finally regression re-
sults to the following equation:
The software package Minitab® Release 15.1.0.0. (Minitab
Kapp ˆ 0:098 ‡ 0:0253DTapp ‡ 0:232 CB × 10 3 †
Inc.) was used for statistical modelin and analyses. The experi- (12)
ments were repeated at least twice in randomized order as is 0:288D0 0:000229DTapp 2 0:0541 CB × 10 3 †2

Chem. Eng. Technol. 2013, 36, No. 3, 500–506 © 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.cet-journal.com
504 M. S. Hatamipour et al.

Table 3. Estimated regression coefficients. approach temperature beyond the maximum. The approach
temperature increase, i.e., the increase of the difference be-
Term Coefficient Standard error P-Value tween exit gas temperature and adiabatic saturation tempera-
coefficient
ture of the inlet gas, corresponds to the inlet gas temperature
Constant 0.098 0.2432 0.69 increase. This result is similar to that obtained in our previous
study [6], i.e., by increasing the inlet temperature the removal
DTapp 0.0253 0.006061 0.000
efficiency decreases. The reason can be explained as follows:
–3
CB ×10 0.232 0.1018 0.031 although the increasing temperature enhances the reaction
D0 –0.288 0.1435 0.050
rate, reaction (1) is a neutralization reaction which proceeds in
the presence of water. When water is rapidly evaporated, the
2
DTapp –0.000229 0.00004663 0.000 reaction does not proceed to a considerable extent. Also, at
(CB ×10 ) –3 2
–0.0541 0.01861 0.008 high temperatures a solid crust is formed faster around the
droplet, CO2 in the gas stream has to pass through the crust to
2 2
R = 69.3 %; Radj = 63.1 % reach the alkaline solution, and the solid resistance decreases
the reaction rate. Hence, high temperatures have a negative
where the variables are DTapp (°C), CB (mol m–3), and D0 impact on the reaction rate.
(mm). The plot of Kapp versus approach temperature at different
It should be emphasized that Eq. (12) is valid only in the absorbent concentrations together with results of some addi-
mentioned ranges of the involved factors for the specified set- tional tests which were carried out to check the accuracy of
up. In other ranges and conditions, different model equations Eq. (12) are presented in Fig. 5. Experimental data agree well
should be obtained for the system response. The model accu- with the correlation.
racy was determined by residual plots. A normality assump-
tion checking may take place by constructing a normal prob-
ability plot of the residual. If the trend of the residual plot is 3.2 Effect of Absorbent Concentration
approximately a straight line, then the normality assumption
is satisfied [10]. Fig. 3 displays the residual plot for the present For absorbent concentration it can be concluded that medium
study. From kinetic characteristics of this reaction, the effect of values are a little better, similar to the results obtained in our
different parameters could be evaluated. previous work [6]. Fig. 6 presents the plot of Kapp versus absor-
bent concentration at different approach temperatures. Re-
garding Eq. (5), three resistances, i.e., external mass transfer
3.1 Effect of Approach Temperature resistance, diffusion resistance through the intermediate solid
layer, and the reaction resistance affect the overall reaction rate.
For the approach temperature, there is a positive coefficient As a matter of fact, increasing the absorbent concentration has
for the main effect and a negative coefficient for the square two opposite simultaneous effects: (i) it increases the density
term. A surface plot of correlated Kapp versus approach tem- and viscosity of the reaction medium which leads to a lower
perature and NaOH mass concentration is presented in Fig. 4. diffusion rate of CO2 in the liquid phase; (ii) it increases the
This surface indicates a maximum in an approach temperature reaction rate due to providing a higher concentration for the
range of about 55–65 °C (corresponding to an inlet tempera- solid reactant. The first effect increases the diffusion resistance
ture of about 150 °C) and a decreasing trend with increasing through the solid layer and the second one reduces the reac-
tion resistance simultaneously. The increasing-
Normal Probability Plot decreasing trend in Fig. 6 may be attributed to this
reason. In addition, the results of some additional
99
tests which were done to check the accuracy of
Eq. (12) are illustrated in Fig. 6. An agreement
95
exists between experimental data and correlation.
90

80
70
3.3 Effect of Nozzle Diameter
Percent

60
50
40 The nozzle diameter can affect the droplet diame-
30 ters and exhibits a negative coefficient in the pro-
20 posed correlation. A nozzle with smaller diameter
10 leads to higher efficiency. It should be noted that
5 the reaction of CO2 and NaOH occurs rapidly
through a film of liquid, not in the bulk of liquid.
1 For this kind of reactions, the surface area is an
-0.2 -0.1 0.0 0.1 0.2 important factor affecting the rate-controlling step.
Residual
Smaller droplet diameters increase the surface area
Figure 3. Normal probability plot. and consequently the reaction rate.

www.cet-journal.com © 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim Chem. Eng. Technol. 2013, 36, No. 3, 500–506
 Spray-dryer absorber 505

creasing the gas film resistance at higher

Hold Values temperatures. In Fig. 7, the surface plot of
D 0.75 Kapp versus nozzle diameter and inlet tem-
L/G 0.03 perature is presented.
1.0

3.4 Effect of L/G Ratio

0.5
Kapp
In Eq. (12), the terms containing the L/G
0.0 ratio have been omitted from the correla-
120
tion because of high P-values. This may be
90 related to the limitations of the setup. In
-0.5
60 delta Tapp all experiments, the gas flow rate was set at
0 30 its maximum possible value and by in-
2
4 creasing the liquid flow rate the desired
CB*1000 6
values of L/G were obtained. In some
Figure 4. Surface plot of Kapp versus absorbent concentration and approach tempera- experiments, inadequate atomization of
ture. liquid and formation of a liquid film on
the dryer wall was observed. Therefore, the
For reactions occurring in the film, the phase distribution accurate determination of the L/G effect on experimental
coefficient H (Henry’s constant) can indicate whether the gas- results was not clearly possible. In spite of this, when the sur-
phase resistance is likely to be important or not. For slightly face plot of Kapp versus L/G ratio and inlet temperature is con-
soluble gases, the coefficient H is high and the liquid film resis- sidered (Fig. 8), it can be stated that higher L/G ratios are
tance controls the rate equation, while for highly soluble gases slightly better than lower ones.
like CO2 this coefficient is almost small and the gas-phase re- Applying the desirability function technique and consider-
sistance will be dominant [12]. CO2 is known as a highly solu- ing all factors, the optimum operating conditions for maxi-
ble gas (H = 6300 Pa m3mol–1 at 60 °C for CO2 in water [12]) mum absorption are as follows: inlet gas temperature 142 °C,
and, therefore, the gas film resistance controls the reaction. absorbent concentration 2.13 mol L–1, nozzle diameter 0.5 mm,
The negative effect of higher temperature is also due to in- L/G ratio 0.05 ml L–1 with desirability of 0.97. The predicted

0.9
0.8
0.7
0.6
0.5 corrola
on (at NaOH 2 M)
Kapp
0.4 experiment (at NaOH 2M)
0.3 corrora
on (at NaOH 3 M)
0.2
experimental (at NaOH 3 M)
0.1
0 Figure 5. Effect of approach
0 20 40 60 80 100 120 temperature on Kapp at differ-
ent absorbent concentrations
Approach temperature (DNozzle = 0.7, L/G = 0.03).

0.7
0.6

0.5 corrola
on (at approach

temp.=85 C)
0.4
Kapp experiment (at
0.3 approachtemp.=85 C)

0.2 corrola
on (at approach

temp.=100 C)
0.1
experiment (at approach
Figure 6. Effect of absorbent
0 temp.=100 C)
concentration on Kapp at differ-
1.3 1.8 2.3 2.8 3.3 3.8
ent approach temperatures
absorbent molar Conc. (DNozzle = 0.7, L/G = 0.03).

Chem. Eng. Technol. 2013, 36, No. 3, 500–506 © 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.cet-journal.com
506 M. S. Hatamipour et al.

Hold Values
Symbols used
NaOH 2.64
L/G 0.03 CAb [mol m–3] CO2 concentration in
the bulk of the gas
1.2 CAR [mol m–3] CO2 concentration on
the droplet surface
0.9 CAr [mol m–3] CO2 concentration on
Kapp the unreacted core
0.6 CB [mol m–3] absorbent concentra-
200
tion
0.3
Cin/out [mol m–3] CO2 inlet/outlet
150
T concentration
0.3 D0 [mm] nozzle diameter
0.6 100
0.9
1.2 De [m2s–1] diffusion coefficient in
D0
the solid crust
Figure 7. Surface plot of Kapp versus nozzle diameter and inlet temperature. K [m4mol–1s–1] reaction coefficient
kC [m s–1] mass transfer coefficient
Hold Values in the gas phase
NaOH 2.64 n_ A [mol m–3S–1] absorbed CO2
D 0.75 Qin/out [m3s–1] inlet/outlet gas flow rate
R [m] droplet radius
r [m] unreacted core radius
0.8

Kapp 0.6 References

0.4
0.06 [1] CO2 Emissions from Fuel Combustion Highlights,
0.2 0.04 2011 edition, International Energy Agency, Paris
0.02 L/G 2011.
100 0.00
150
[2] Spray Drying Absorption: The Easy Way to Clean
T 200 the Flue Gas from Waste Incinerators, GEA Niro,
Soeborg 28, 2012.
Figure 8. Surface plot of Kapp versus L/G ratio and inlet temperature. [3] F. F. Hill, J. Zank, Chem. Eng. Process. 2000, 39 (1),
45–52. DOI: 10.1016/S0255-2701(99)00077-X
conversion and kinetic coefficient under optimum conditions [4] F. Scala, M. D’Ascenzo, A. Lancia, Sep. Purif. Technol. 2004,
are g = 73.3 %, Kapp = 0.981 and the experimental values of 34 (1–3), 143–153. DOI: 10.1016/S1383-5866(03)00188-6
these parameters are g = 75.1 %, Kapp = 0.954, thus indicating a [5] J. Chen, G. Fang, J. Tang, L. Liu, Chemosphere 2005, 59 (1),
satisfying agreement. This also confirms the results found in 99–105. DOI: 10.1016/j.chemosphere.2004.09.076
our previous study, in which by means of the Taguchi method [6] L. Kavoshi, M. S. Hatamipour, A. Rahimi, Int. J. Chem. React.
the approximate optimum conditions to reach the maximum Eng. 2011, 9, A60. DOI: 10.1515/1542-6580.2395
removal efficiency were: inlet temperature 150 °C, absorbent [7] J. Völkl, K. Müller, L. Mokrushina, W. Arlt, Chem. Eng.
concentration 2.77 mol L–1, D0 = 0.5 mm. Technol. 2012, 35 (3), 579–583. DOI: 10.1002/ceat.201100319
[8] S. Yagi, D. Kunii, in Proc. of the 5th Int. Symp. on Combustion,
Reinhold, New York 1955.
4 Conclusions [9] E. B. Rinker, S. S. Ashour, O. C. Sandall, Ind. Eng. Chem. Res.
2000, 39 (11), 4346–4356. DOI: 10.1021/ie990850r
In order to find a correlation for kinetics of reactive absorption [10] R. H. Myers, D. C. Montgomery, C. M. Anderson-Cook,
of CO2 with NaOH, the shrinking core model was applied and Response Surface Methodology: Process and Product Optimiza-
the parameters that could be changed were selected. Using tion Using Designed Experiments, 3rd ed., John Wiley & Sons,
results of experiments designed by RSM, a semi-empirical cor- New York 2009.
relation was obtained. The effect of different operating param- [11] D. C. Montgomery, Design and Analysis of Experiments, 5th
eters on the absorption rate was studied and the optimum ed., John Wiley & Sons, New York 2001.
operating conditions to reach the maximum absorption were [12] O. Levenspiel, Chemical Reaction Engineering, 3rd ed., John
determined. Wiley & Sons, New York 1999.

The authors have declared no conflict of interest.

www.cet-journal.com © 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim Chem. Eng. Technol. 2013, 36, No. 3, 500–506