core shell sorbent.pdf

A novel method to make regenerable core-shell calcium-based sorbents
Abstract
F.J. Liu a, *, K.S. Chou b , Y.K. Huang b
a Department of Chemical Engineering, National United University, 1 Lien Da, Kung-Ching Li, Miao-Li 36003, Taiwan, ROC
b Department of Chemical Engineering, National Tsing Hua University, Hsinchu 30013, Taiwan, ROC
Received 16 August 2004; revised 24 May 2005; accepted 30 May 2005
Available online 19 September 2005
A sorbent having a calcium oxide core and a clay shell was prepared and shown to be capable of reusable applications in absorption and
desorption processes for carbon dioxide. The novelty of this sorbent is that only calcium carbonate and clay are used for its preparation with
water as a binder. A two-step granulation procedure is used to get the core and then another step to coat the shell layer with the clay powder. A
repeated wet-and-dry procedure probably makes the core porous yet strong enough to serve as a sorbent. The pellet is then calcined at
1200 8C for 2 h to reach its final structure. The core-shell pellets have an overall diameter of 4.4 mm with average shell thickness of 0.45 mm,
crush load of 35 N and attrition index of 0.035 wt%/h. These results indicate that the pellets will probably be capable of withstanding the
stress in future applications. Carbon dioxide absorption at or below 300 8C showed a maximum weight gain of 38% for our pellets. Finally,
desorption in nitrogen at 800 8C can restore the pellet to its original state and hence it is ready for re-use as a sorbent.
q 2005 Elsevier Ltd. All rights reserved.
Keywords: Core-shell; Calcium carbonate; Sorbent
1. Introduction
There are many occasions, such as in the case of coalfired
power plants utilizing gas turbines, where both
particulates and gaseous pollutants need to be removed
from the hot gas. One proposal for doing this is to use a
moving granular bed filter (Henriquez and Macias-Machin,
1997; Hsiau et al., 1999). The success of this process
depends on the quality of the granular sorbent. Some of the
desired characteristics of such a sorbent include at least:
sufficient crushing strength to maintain integrity and
resistance to attrition during operation; capable of operating
at high temperatures; capable of reacting with H2S or other
corrosive gases and also regenerable after use.
When the purpose is limited to the removal of particles,
i.e. acting only as a filter, various materials can be used to
make the granules. However, if gaseous pollutants must also
be removed, the materials to choose from are greatly
limited. For H2S removal, calcined limestone is usually used
(van der Ham et al., 1996; Akiti et al., 2002a, b; Hasler et al.,
* Corresponding author. Fax: C886 37 332397.
E-mail address: liu@nuu.edu.tw (F.J. Liu).
0301-4797/$ - see front matter q 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.jenvman.2005.05.013
Journal of Environmental Management 79 (2006) 51–56
www.elsevier.com/locate/jenvman
2003). Here, CaO is first converted to CaS during
absorption. Then, for its regeneration, calcium sulfide can
be oxidized to CaSO4 and then be reduced with CO or H2
back to CaO for re-use. Wheelock and colleagues (Akiti
et al., 2002a, b; Hasler et al., 2003) have studied the various
compositions to prepare this sorbent with the desired
characteristics. They propose a core-shell structured pellet
for this purpose. They tried to use either limestone or plaster
of Paris to make the core. In the first case calcium aluminate
cement was used as a binder to hold the limestone particle
together. As for the shell, limestone particles mixed with
either alumina or calcium aluminate cement were used. The
basic idea was that the shell should be porous but strong
enough to offer some protection for the core material.
Integrated coal gasification combined cycle (IGCC)
power generation systems hold great promise for producing
electric power efficiently and economically. Limestone is
extensively used in the in situ removal of acid gas species
such as SO2, CO2 from pulverized and fluidized-bed coal
combustors and H2S from advanced combined-cycle
gasification systems. Usually, the SO 2 removal process
takes place at higher temperatures and is more dangerous.
Hence, in the preliminary experimental stage, only the CO 2
removal process was investigated. In this article, an
alternative method is proposed for making the core-shell

52
pellets using only limestone and clay particles. The results
show that it possesses all the required characteristics to
serve as a sorbent in moving bed operations.
2. Experimental
2.1. Materials and methods
Calcium carbonate (densityZ2.7 g/cm 3 , purityZ
99.3 wt%, B201, Zu-hsin, Taiwan) and clay (densityZ
2.7 g/cm 3 , hydrated aluminum silicate, 37.8 wt% Al2O3,
62.1 wt% SiO 2, No. 26, Tao-yi, Taiwan) were the only two
materials used in this work. The purity of the calcium
carbonate was 99.3%, and it had a BET surface area of
5.44 m 2 /g, or an equivalent particle size of around 0.408 mm
(for spherical particles, dZ6/(rs), d is the equivalent
particle size, ris the density, and s is the BET surface
area). The clay on the other hand had a BET surface area of
20.74 m 2 /g, or an equivalent size of 0.115 mm. The method
used for estimating the equivalent diameter of the small
particles is subject to two major errors. One error is due to
the assumption that the particles were spherical. The second
error arises from using the BET surface area, which
measures both the external particle surface area and the
surface area of micropores to calculate the equivalent
diameter. Only the external surface area should be used for
calculating the equivalent particle diameter.
To obtain core-shell structured pellets, Akiti et al. (Akiti
et al., 2002a, b; Hasler et al., 2003) tried to use either plaster
of Paris or limestone with calcium aluminate cement as a
binding agent. Yet, as found in this study, calcium carbonate
core pellets can be made by simply using water as a binder.
Pelletization was carried out in a planetary ball mill
(centrifugal ball mill pulverisette 6, Fritsch, Germany),
which had been modified by replacing the grinding bowl
with a 1000 ml polypropylene beaker without the cover. The
beaker is mounted on a horizontal rotating disc with the
longitudinal axis of the beaker at right angles to the disc and
off center. Consequently, when the disc is rotated, the
beaker operates in a planetary motion. Forty grams of
CaCO3 powder was placed in a beaker on a planetary ball
mill with a rotating speed of 150 rpm, 5 ml de-ionized water
was sprayed intermittently into the beaker to ball up the
calcium carbonate particles. At the same time, a glass rod
was used to break up any large granules that might be
formed due to the non-uniform distribution of water. After
about 10 min, a sieve was used to collect pellets within the
size range of 1.41–2.36 mm.
These pellets were then put back into the beaker to which
another 20 g of CaCO3 powder had already been added, and
the above procedure was continued to finally get pellets of
two size ranges: 4.0–4.76 mm (labeled as CS2-L) and 2.36–
4.0 mm (as CS2-S). The pellets were then sprayed with
small amounts of water (merely making them moist) and
then dried in an oven at 50 8C for 1 h. This wetting
F.J. Liu et al. / Journal of Environmental Management 79 (2006) 51–56
and drying procedure was repeated three times to make the
pellets strong enough for further processing steps. During
this procedure, the green body of the core pellets retained its
integrity and appeared more dense having sufficient
strength. After obtaining the CaCO3 core pellets, they
were then placed into clay powder (5 g) to form shells using
the same procedure. These core-shell pellets were then
heated at a rate of 10 8C/min to 1200 8C and held there for
2 h to obtain our sorbent having a porous CaO core and clay
shell.
2.2. Properties and tests
The sorbents were examined for their mechanical
strength against compression and attrition. The compressive
strength of the pellets was determined by measuring the
force required to fracture a single pellet between two plates
of a universal testing instrument (Shimadzu AGS-2000G)
with the upper plate lowered at a rate of 2 mm/min. About
15 pellets from each sample were tested. For comparison,
the strength of some CaO cores (i.e. without the clay shell)
of several different sizes was also measured. Several
attrition test methods are available to measure the attrition
tendency of fluidized solids or moving solids in a drum. In
this work, the Peter-Spencer method was adopted (Deng and
Lin, 1997). The testing equipment consisted of a variable
speed drive motor and a stainless steel tube (315 mm long,
10 mm ID) mounted on the drive shaft at a right angle to the
shaft and at a point which was 80 mm from one end of the
tube. The ends of the tube were closed so that as the tube
revolved the pellets being tested would flow back and forth
from one end of the tube to the other end. The tube was
rotated about an axis normal to the length with a controlled
speed ranging from 20 to 200 rpm. Briefly, about 5 g of
pellets were placed in a stainless steel tube and the tube was
rotated at 60 rpm for 24 h before measuring the weight
changes caused by attrition. At the end of the attrition test
the material had to be screened (Tyler standard screen mesh
#10) to separate the fines, which were generated from the
remaining pellets. The weight loss percentage rate (wt%/h)
calculated by the following equation was used as attrition
index (Deng and Lin, 1997):
Attrition index Z ðinitial weight
Kremaining weightÞ=initial weight=time
!100%:
To gain some knowledge about the porosity and pore size
distribution of these sorbents, a mercury porosimeter (PMI
60 KA2) was used.
The absorption characteristics of a sorbent pellet was
determined by employing a thermogravimetric analyzer
(TGA, SSC 5000, Seiko, Japan) to determine the weight
gain as a function of time at temperatures ranging from 50 to
300 8C under flowing gas of 95 vol% N2 and 5 vol% CO2.

However, before each absorption run, the pellet was first
heated under flowing nitrogen to 800 8C to remove any CO2
or moisture previously picked up by the pellet from the
atmosphere. It was then cooled down to the desired
temperature and held for 5 min before switching to the
mixture gas of N 2 and CO 2 for absorption experiments. In
order to demonstrate the ability for reuse of these pellets, we
would heat the pellet up to 800 8C again in nitrogen after
absorption to remove absorbed CO2 and then cool it down to
300 8C for a repeated absorption experiment. It was repeated
three times.
3. Results and discussion
3.1. Basic properties
Shown in Fig. 1 are the SEM pictures of the original
CaCO 3 powder before and after having been calcinated
at 1200 8C for 2 h. Quite different morphology was
observed between them. Next, in Fig. 2 a photograph is
shown both of the pellet made in this work and a cut-off
view of the core-shell structure. The thickness of the
shell was about 0.45 mm. Due to different shrinkage of
the core and shell materials, there was also some space
between the core and shell after calcination, which is
F.J. Liu et al. / Journal of Environmental Management 79 (2006) 51–56 53
Fig. 1. SEM pictures of (a) the original CaCO 3 powder; and (b) after 1200 8C calcinations.
very beneficial to the absorption since there is always a
volume increase accompanied with either CO2 or H2S
absorption. Simple calculation indicates that when CaO
changes into CaCO3, the volume increases by about
117%. The pellet would crumble if there were no room
for such an expansion.
Next shown in Fig. 3 are the compressive strengths (in
terms of crush load) of several different pellets made in this
work. On the left, for comparison purposes, are the crushing
loads of the CaO cores of three different sizes, ranging from
1.41 to 4.0 mm. On the right side are the corresponding
values of some CaO core-clay shell pellets. The pellets used
here were all tested after calcination at 1200 8C for 2 hours.
The increase in strength due to the clay shell is very
obvious. Also noticeable is that a large pellet often
possesses slightly higher strength than a small one. In this
figure, CS1 refers to the pellets made from one pelletizing
step, while CS2 refers to those made from two pelletizing
steps. The overall diameters of CS1 and CS2-L are 4.4 mm
and CS2-S is 3.2 mm. The average shell thickness for these
pellets is about 0.45 mm. The strengths of various pellets are
not very much different. However, the CS1 pellets cracked
after the CO2 adsorption experiments. It seems that the CaO
core in the CS2 pellets may contain more porosity due to the
two steps of the pelletizing procedure. Porosity inside the
CaO core as well as the space between the core and shell are
Fig. 2. Photographs of (a) pellet (after calcined at 1200 8C for 2 h, CS2-L) made in this work; (b) cut-off view to exhibit core-shell structure.

54
Crush Load (N)
45
40
35
30
25
20
15
10
5
two key factors for the success of our pellets as regenerable
sorbents.
Another mechanical property of concern is the pellet’s
ability to withstand attrition, if it is to be used in a moving
bed operation. Exhibited in Table 1 are the attrition indexes
of our pellets along with corresponding data of some
commercial sorbents from the literature (Deng and Lin,
1997) for comparison. The performance of our pellets is
reasonable, indicating again the proper function of a porous
clay shell.
Next, displayed in Fig. 4, are the pore size distributions
of the calcined core-shell pellet measured from the mercury
porosimeter. The curve exhibits two uniform peaks, located
at 0.009 and 0.011 mm, representing the pore size
distributions of core and shell, respectively. The pore size
distributions from pure core and pure shell (obtained from a
broken pellet) coincide with the data obtained here. Gaseous
transport through these pores would probably be by
Table 1
Attrition index compared with commercial sorbents from literature (Deng
and Lin, 1997)
Sample Dia (mm) Attrition
index
(wt%/h)
Pure CaCO 3
Core-shell
1.41-1.68 1.68-2.36 2.36-4 CS1 CS2-L CS2-S
Fig. 3. The average crush load and its standard deviation of different pellets:
on the left are CaO cores of three sizes (mm); and on the right are CaO
core-clay shell pellets.
Attrition
weight loss
(wt%)
Note
AL-S-2 1.5–2.0 0.033 0.785 From
literature
(Deng and
Lin, 1997)
AL-LD-390
(Alcoa)
4–4.6 0.177 4.25
Wessalith
DAY
(Degussa
AG)
3.5–3.7 0.073 1.75
CS2-L 4–4.76 0.035 0.844 This work
F.J. Liu et al. / Journal of Environmental Management 79 (2006) 51–56
dV/dlog(Diameter)
600
500
400
300
200
100
0
core-shell pellet after 1200˚C and 2 hr calcination
0.001 0.01 0.1 1 10 100
Knudsen diffusion and thus be at orders of magnitude
higher than any diffusion through solid products (Hsia et al.,
1993). Detailed analysis of the kinetics of this absorption
reaction is now in progress and will be reported later.
3.2. Sorbent characteristics
Pore Diameter, µm
Fig. 4. The pore size distributions of a core-shell pellet after 1200 8C for 2 h
calcination.
Carbon dioxide and TGA were utilized to test the
absorption capacity and regenerative ability of these coreshell
sorbents. The absorption results at various temperatures
are shown in Fig. 5. Two things can be noticed here. First,
regardless of the operating temperature, the weight gains
reached a plateau value of about 38%. It suggests an
equilibrium result for a complete conversion of CaO into
CaCO3. The theoretical amount of increase in the weight of
CaO when converted to CaCO3 is 78.6%. Yet, after deducing
the weight of clay that is inert to CO2, the rough estimate
indicated that the weight gain should be around 46%. The
small difference may be attributed to over-estimating the size
Weight Gain(%)
40
30
20
10
300˚ C
250˚ C
200˚ C
150˚ C
100˚ C
50˚ C
0
0 20 40 60 80 100 120 140
Time(min)
Fig. 5. Effect of temperature on absorption characteristics of the core-shell
pellets (CS2-L). Experiments were conducted in nitrogen with 5% CO2.

Weight Gain (%)
40
30
20
10
0
0 50 100 150 200 250 300 350
Time(min)
Fig. 6. Results from three cycles of absorption (300 8C) and desorption
(800 8C) experiments (CS2-L).
of the cores. Nevertheless, based on a 38% measured weight
gain for CO2 absorption, a weight gain about 13.8% can be
expected if the adsorbate is H 2S, which is about the same as the
result reported by Akiti et al. (Akiti et al., 2002a, b). The
increase in the rate of absorption with temperature indicated
by Fig. 5 is about what could be anticipated for a mass transfer
limited process.
It would be very economical if the sorbent could be
regenerated for multiple uses. To demonstrate this capability
of the sorbent, the results from three cycles of
absorption–desorption runs are now shown in Fig. 6. Here,
parts of the original curves (the leveling part) were removed
for easy presentation of the similarity between these three
runs. In all cases, the CaO was first completely converted to
Intensity
F.J. Liu et al. / Journal of Environmental Management 79 (2006) 51–56 55
CaCO 3 (012)
CaO(111)
CaCO 3 (104)
CaO(200)
2θ
CaCO3 and then back to CaO after the desorption in
nitrogen at 800 o C. At 5% CO 2, the CaCO 3 started
decomposing, roughly around 635 8C. After these experiments,
the pellet was still in good shape without any visible
cracks in its shell. In short, this preliminary test showed that
the calcium-based core-shell pellet can be a good candidate
as a regenerable sorbent for high temperature applications.
Finally, in Fig. 7, the XRD patterns of the CaO core before
and after absorption are shown. It clearly suggests complete
conversion to calcium carbonate after absorption with no
trace of CaO peaks.
4. Conclusions
CaCO3 (110) CaCO3 (202)
CaCO3 (018)
CaCO3 (113)
CaCO3 (116)
A core-shell sorbent consisting of a pure CaO core with a
pure clay shell was shown to be capable of repeated
absorption and desorption of CO 2 gas. The novelty of this
work is that the core-shell structure could be made
successfully using only pure calcium carbonate and clay
powder with water as a binder. By a two-step granulation
procedure, sufficient porosity was apparently introduced
into the core making it capable of withstanding the volume
expansion during absorption. Another important characteristic
of this pellet is that there exists some space between the
core and the shell, which also contributes to its resistance to
volume expansion. Absorption of carbon dioxide probably
reaches an equilibrium value below the temperature of
300 8C converting the CaO core completely into CaCO 3.
And finally, desorption at 800 8C can also convert the
CaCO3 back to regenerate CaO for repeated use.
Core before reaction with carbon dioxide
CaO(220)
CaO(311) CaO(222) CaO(400) CaO(331)
Core after reaction with carbon dioxide
10 20 30 40 50 60 70 80 90
Fig. 7. XRD patterns of the core, before and after the CO2 absorption.

56
References
Akiti Jr., T.T., Constant, K.P., Doraiswamy, L.K., Wheelock, T.D., 2002a.
An improved core-in-shell sorbent for desulfurizing hot coal gas. Adv.
Environ. Res. 6, 419–428.
Akiti Jr.., T.T., Constant, K.P., Doraiswamy, L.K., Wheelock, T.D., 2002b.
A regenerable calcium-based core-in-shell sorbent for desulfurizing hot
coal gas. Ind. Eng. Chem. Res. 41, 587–597.
Deng, S.G., Lin, Y.S., 1997. Granulation of sol-gel-derived nanostructured
alumina. AIChE J. 43, 505–514.
Hasler, D.J., Doraiswamy, L.K., Wheelock, T.D., 2003. A plausible model
for the sulfidation of calcium-based core-in-shell sorbent. Ind. Eng.
Chem. Res. 42, 2644–2653.
F.J. Liu et al. / Journal of Environmental Management 79 (2006) 51–56
Henriquez, V., Macias-Machin, A., 1997. Hot gas filtration using a
moving bed heat exchanger-filter (MHEF). Chem. Eng. Process. 36,
353–361.
Hsia, C., St Pierre, G.R., Raghunathan, K., Fan, L.S., 1993. Diffusion
through CaSO 4 formed during the reaction of CaO with SO 2 and O 2.
AIChE J. 39, 698–700.
Hsiau, S.S., Smid, J., Wang, C.Y., Kuo, J.T., Chou, C.S., 1999. Velocity
profiles of granules in moving bed filters. Chem. Eng. Sci. 54, 293–
301.
van der Ham, A.G.J., Heesink, A.B.M., Prins, W., van Swaaij, W.P.M.,
1996. Proposal for the regenerative high-temperature process for coal
gas clean-up with calcined limestone. Ind. Eng. Chem. Res. 35, 1487–
1495.