Journal of the Ceramic Society of Japan 119 [6] 460-463 2011 Paper

Fabrication of 12CaO·7Al 2 O 3 powders with high specific

surface area by sol–gel and ball-milling method
Kotaro OZAWA, Naonori SAKAMOTO,³ Naoki WAKIYA and Hisao SUZUKI*
Department of Materials Science and Chemical Engineering, Shizuoka University, 3–5–1 Johoku, Naka-ku, Hamamatsu 432–8561
*Graduate School of Science and Technology, Shizuoka University, 3–5–1 Johoku, Naka-ku, Hamamatsu 432–8561

12CaO·7Al2O3 (called as C12A7) is composed of [Ca24Al28O64]4+ cages in which oxide anions are entrapped. Recent studies have
revealed that the outstanding characteristics appear by exchanging the free oxide anions in the cages by various other anion
species. In the present report, in order to enhance the oxidative catalytic activity of C12A7 by the encaged oxygen radicals, we
synthesized C12A7 fine powders with high-specific surface area by sol­gel method and subsequent ball-milling. Aluminum sec-
butoxide, calcium nitrate tetrahydrate, absolute ethanol, 1 M hydrochloric acid, and ethyl 3-oxobutanate were used as raw
materials. The obtained precursor solution was dried at 100°C for 24 h, and the product obtained was annealed at 900°C in O2
atmosphere. Planetary ball milling was subsequently applied to the C12A7 powders. The XRD measurements revealed that the
synthesized sample was a single phase C12A7. The BET specific surface area of the ground samples was 48.9 m2/g (24 h ground)
and 17.3 m2/g (36 h ground) for those prepared by sol­gel method and by solid state reaction, respectively. The possibility was
thus demonstrated on the synthesis of C12A7 fine powders with higher specific surface area by sol­gel method and subsequent
planetary ball-milling.
©2011 The Ceramic Society of Japan. All rights reserved.

Key-words : C12A7, Catalytic ceramic, Sol–gel method, High oxidative power, Fine powder, High specific surface area, Planetary
ball-milling

[Received February 14, 2011; Accepted April 4, 2011]

the solid state reaction rate, the grains also grow and the grain
1. Introduction size increases. Thus it can be said that the solid state reaction
12CaO·7Al2O3 (called as C12A7) has attracted attention as method is not appropriate for obtaining powders with high
functional ceramics. The crystal structure of C12A7 is cubic with specific surface area. One of the potential methods for fabricating
lattice constant 1.199 nm. The unit cell of the crystal is composed powders with high specific surface area is the solution process,
of the “cages”, [Ca24Al28O64]4+, and the “oxide anions clathrated such as a sol­gel method, which does not require high annealing
in the cages”.1)­4) Though the crystal is a stable ceramic well temperatures. The sol­gel method can achieve homogeneous
known as a constituent of aluminum cements, the recent studies compositions at low temperatures, and can control the morphol-
have revealed that the outstanding characteristics appear by ogy of the products (fine particle, thin firm, etc.).12),13)
exchanging the oxide anions in the cages (called as free oxide In the present paper, in order to enhance an ability of C12A7 as
ions) by various other anion species.5)­8) an oxidation catalyst by caused by the encaged oxygen radicals,
When oxygen radicals, O¹, encaged in the C12A7 cage, we synthesized C12A7 fine powder with high-specific surface
C12A7 shows a high oxidative power. Dong et al. reported that a area by sol­gel method. The powder showed much higher
direct oxidation of benzene using C12A7­O¹ yields phenol.9) specific surface areas than the products by solid state reaction by
Wang et al. also reported a production of hydrogen by catalytic subsequent ball-milling.
steam reforming of bio-oil using C12A7­O¹.10) Yang et al.
reported partial oxidation of methane using C12A7­O¹.11) Even 2. Experimental
though C12A7 has a high potential for oxidation catalysis, much Aluminum sec-butoxide [Al(OBus)3, Tokyo Chemical Industry
work has been done on the fabrication route for C12A7 as Co., Ltd., Tokyo, Japan] and calcium nitrate tetrahydrate
catalysts, i.e. nano/microparticles with high specific surface area. [Ca(NO3)2·4H2O, 99.9%, Wako Pure Chem. Ind. Ltd., Tokyo,
Most of publications regarding C12A7 oxidizing catalysts have Japan] were used as starting reagents for C12A7 precursor
been based on bulk or powders basically fabricated by solid state solutions. Ethyl 3-oxobutanate (CH3COCH2COOC2H5, 98.0%,
reaction (and subsequent ball-milling). The solid state reaction is Wako Pure Chem. Ind. Ltd., Tokyo, Japan) was used as a
a conventional and simple method but needs high reaction chelating agent.14) Absolute ethanol was used as a solvent. 1 M
temperatures above 1200°C and long reaction time.9)­11) Such hydrochloric acid (HCl, Kanto Chem. Ind. Ltd., Tokyo, Japan)
high temperatures are undesirable from the viewpoint of energy was used a deflocculant.
consumption, and also possibly cause a crucial problem: The aluminum precursor solution was prepared by the
drastically reducing specific surface area of the C12A7 catalysts. following procedure. Aluminum sec-butoxide and ethyl 3-
When the powders are heated at high temperatures for increasing oxobutanate were mixed, and the mixture obtained was stirred
for 30 min at 50°C (the solution became uniform and clear). The
³
Corresponding author: N. Sakamoto; E-mail: tnsakam@ipc. solution thus obtained was added in absolute ethanol. Distilled
shizuoka.ac.jp water and 1 M hydrochloric acid were added, and the resulting

460 ©2011 The Ceramic Society of Japan

Journal of the Ceramic Society of Japan 119 [6] 460-463 2011 JCS-Japan

solution was stirred for another 2 h at 60°C. The calcium

precursor solution was prepared by the following procedure.
Calcium nitrate tetrahydrate was dehydrated at 150°C for 1 h.
The products were subsequently dissolved in absolute ethanol,
and the mixture was stirred for 1 h. The aluminum and calcium
precursor solutions thus prepared were mixed to obtain the
C12A7 precursor solution. The obtained precursor solution was
dried at 100°C for 24 h and annealed at 700­900°C in O2
atmosphere.
C12A7 powders were also prepared by solid state reaction for
comparison, where the mixture of CaCO3 and Al2O3 powders
with a molar ratio of 12:7 was annealed at 1300°C for 6 h in air.
Planetary ball milling (puluerisette-7, FRITSCH GmbH,
Germany) was subsequently applied to the C12A7 powders.
The milling was performed with using zirconia balls of 0.5 mm
in diameter in absolute ethanol. In order to estimate the optimum
milling time duration, the milling time duration was varied from
12 to 48 h with a fixed grinding rate of 150 rpm. Fig. 2. X-ray diffraction patterns of the powders synthesized by the
sol­gel method and annealed at (a) 900, (b) 800 and (c) 700°C, and
A Fourier transform infrared spectrometer (FTIR8500,
(d) those synthesized by the solid state reaction at 1300°C.
SHIMADZU Ltd., Kyoto, Japan) was used to investigate the
chemical bonds formed in the dried gel. The assignment of
absorption bands was based on Ref. 14. Figure 2 show the XRD patterns of the powder samples
The thermal behavior and the weight change during annealing synthesized by the sol­gel method at various annealing temper-
were investigated on the dried gel by thermogravimetry and atures and by the solid state reaction. The samples synthesized by
differential thermal analysis (TG­DTA) (Model Thermo plus either method [Figs. 2(a) and 2(d)] were single phase C12A7.
TG8120, Rigaku Ltd., Tokyo, Japan). The crystal structure of the The powder sample annealed at 900°C was single phase C12A7,
obtained powders was identified by X-ray diffraction (XRD) whereas the those at 700 and 800°C contained CaO and
(Model D8 Advance, Brucker AXS GmbH, Germany). The amorphous phases [Figs. 2(b) and 2(c)]. These results indicate
microstructure of the obtained powders was investigated by that the crystallization temperature of C12A7 is below 800°C. In
scanning (Model JSM-5510SV, JEOL Ltd., Tokyo, Japan) and most cases, the solid state reaction requires high temperatures
transmission electron microscopy (TEM) (Model JEM-2100F, above 1200°C for obtaining single phase C12A7. The present
JEOL Ltd., Tokyo, Japan). The particle size distribution was results indicate that the sol­gel method drastically reduces the
measured by dynamic light scattering (Model Photal SELS-800Y, temperature for synthesizing C12A7 powders. Based on these
Otsuka Electronics Co., Ltd., Osaka, Japan) using absolute results, the optimum annealing temperature was determined to be
ethanol as solvent. The specific surface area of the powders was 900°C for the sol­gel method. In the following sections, the
measured by BET analysis using nitrogen gas adsorption.15) annealing temperature was fixed at 900 and 1200°C for the sol­
gel method and the solid state reaction, respectively.
3. Result and discussion Figures 3(a) and 3(b) show the XRD patterns of the samples
Figure 1 shows the FTIR spectra of the dried gel. Character- prepared by the sol­gel method and the solid state reaction,
istic absorption bands were detected at 1525 and 1180 cm¹1, respectively, measured after the ball milling. It was confirmed
which are assigned to the C­C vibration of a six membered ring that both samples obtained by the sol­gel method and the solid
of the complex and to the ­C­C­O stretching vibration, respec- state reaction held the C12A7 crystal structure even after the ball
tively. This obviously shows that the aluminum sec-butoxide was milling, although amorphous phases seem to be slightly increas-
chelated by ethyl 3-oxobutanate, by which the dried gel became ed in amount. The amorphous phases might be decreased by
stable compared to that without using chelating agent. milling under milder conditions at lower rotation speeds, with
lighter weight balls and so on.
Figure 4 shows the BET specific surface area of the samples
obtained by the sol­gel method and solid state reaction at various
milling time duration. The specific surface area increased with
increasing milling time for both samples. The maximum specific
surface area was 48.9 m2/g (sol­gel derived sample with 24 h
ground) and 17.3 m2/g (sample by solid state reaction method
with 36 h ground). It is worth noting that the specific surface
area stayed at the maximum values even when further grinding
was applied. This result indicates that the highest specific surface
area of the sol­gel derived C12A7 powders is not achievable
by the solid state reaction. The highest specific surface area,
48.9 m2/g, is comparable to those reported on other oxidation
catalysts (about 30­130 m2/g).16)­19) In the following descrip-
tion, “ground samples” indicate the samples ground under the
optimum conditions, i.e. ground for 24 and 36 h for the samples
prepared by the sol­gel method and by the solid state reaction,
Fig. 1. FTIR spectra of the dried gel. respectively.

461
JCS-Japan Ozawa et al.: Fabrication of 12CaO·7Al2O3 powders with high specific surface area by sol–gel and ball-milling method

Fig. 3. XRD patterns of the C12A7 powders before and after the ball
milling obtained by (a) the sol­gel method and (b) the solid state reaction
(Milling time: 24 h for the sol­gel method sample, 36 h for the solid state
reaction sample). No obvious peak appeared after the ball milling for both
of the samples.

Fig. 5. Particle size distribution in the C12A7 powders (a): synthesized

by the sol­gel method and measured before the ball-milling (24 h),
(b): synthesized by the sol­gel method and measured after the ball-
milling, (c): synthesized by the solid state reaction and measured before
Fig. 4. Specific surface area of the samples with various milling times. the ball-milling, (d): synthesized by solid state reaction and measured
after the ball-milling (36 h).

Figure 5 shows the particle size distribution in the powder particle size for both samples. It is worth noting that the powders
samples. The particle size of the C12A7 powders synthesized by synthesized by the sol­gel method had the small average particle
the sol­gel method was distributed in 150­1200 nm (the average size and the relatively narrow particle size distribution compared
particle size: 357.5 nm) before the ball-milling [Fig. 5(a)], to those prepared by the solid state reaction. These results imply
whereas that after ball-milling [Fig. 5(b)] became narrower in that the sol­gel method is suitable for synthesizing C12A7
distribution in 100­450 nm (the average particle size: 193.1 nm). powders with small particle size.
On the other hand, the C12A7 powders synthesized by the solid In order to investigate why the different grinding behavior
state reaction had two average particle sizes, 279 and 2180 nm, appeared in the two samples, the microscopic and crystallo-
before ball-milling [Fig. 5(c)], which were reduced to 193 and graphic analyses were applied. Figure 6 shows the TEM images
879.2 nm, respectively, by the ball milling [Fig. 5(d)]. It is of the powder samples. The C12A7 powders synthesized by the
obviously shown that the ball-milling reduces the average sol­gel method consisted of aggregates of tiny primary particles

462
Journal of the Ceramic Society of Japan 119 [6] 460-463 2011 JCS-Japan

4. Summary
Single phase C12A7 powders were successfully synthesized
by the sol­gel method through annealing at 900°C. The primary
particle size (29.3 nm estimated by Rietveld analysis) of the
C12A7 powders obtained by the sol­gel method was several
hundred nanometers and they were composed of agglomerates of
several tens nanometers. The agglomerate was deagglomerated
by subsequent ball-milling and the highest specific surface area
of 48.9 m2/g (24 h ground) was achieved. On the other hand, the
highest specific surface area of the C12A7 powders synthesized
by the conventional solid state reaction was 17.3 m2/g. The TEM
and XRD analyses implied that the difference in the grinding
effect between these two samples lies in the different primary
particle sizes; the deagglomeration for tiny primary particles (the
sol­gel derived sample) and the crushing for large primary
particles (the sample prepared by the solid state reaction). It was
indicated that the sol­gel method is a desirable way to synthesize
C12A7 powders with high specific surface area.
Fig. 6. TEM images of the C12A7 powders (a) synthesized by the sol­
gel method and measured before the ball-milling, (b) synthesized by the References
sol­gel method and measured after the ball-milling, (c) synthesized by 1) B. Hallstedt, J. Am. Ceram. Soc., 73, 15­23 (1990).
the solid state reaction and measured before the ball-milling, and (d) 2) H. Hosono and Y. Abe, J. Non-Cryst. Solids, 95–96, 725­732
synthesized by the solid state reaction and measured after the ball-milling. (1987).
3) W. R. Nurse, O. H. J. Welch and J. A. Majumdar, Trans. Br.
Ceram. Soc., 64, 409­418 (1965).
Table 1. Average crystallite size calculated by Rietveld analysis using 4) B. H. Bartl and T. Scheller, Neues Jahrb. Mineral. Monatsh.,
TOPAS software 35, 547­552 (1970).
Crystallite diameter (nm)
5) K. Hayashi, M. Hirano, S. Matsuishi and H. Hosono, J. Am.
Method Chem. Soc., 124, 738­739 (2002).
Before ballmiling After ballmiling 6) K. Hayashi, S. Matsuishi, T. Kamiya, M. Hirano and H.
Sol­Gel method 29.3 26.0 Hosono, Nature, 419, 462­465 (2002).
Solid state reaction 275.5 136.6 7) S. Matsuishi, Y. Toda, M. Miyakawa, K. Hayashi, T. Kamiya,
M. Hirano, I. Tanaka and H. Hosono, Science, 301, 626­629
(2003).
8) J. Jeevaratnam, P. F. Glasser and D. S. L. Glasser, J. Am.
of several tens nm before the ball-milling [Fig. 6(a)]. The ball-
Ceram. Soc., 47, 105­106 (1964).
milling reduced the size of the aggregates drastically with the 9) T. Dong, J. Li, L. Huang, F. Wang, J. Tu, Y. Torimoto, M.
crystallite size maintained although the primary particles still Sadakata and Q. Li, Chem. Commun. (Camb.), 21, 2724­2726
formed small aggregates of 100 to 500 nm in size as shown in the (2005).
Fig. 6(b). On the other hand, the C12A7 powders synthesized by 10) Z. Wang, Y. Pan, T. Dong, X. Zhu, T. Kan, L. Yuan, Y.
the solid state reaction consisted of large grains with several Torimoto, M. Sadakata and Q. Li, Appl. Catal., 320, 24­34
micrometers before the ball-milling [Fig. 6(c)] whereas the (2007).
grains were cracked and the size decreased by the ball-milling 11) S. Yang, N. J. Kondo, K. Hayashi, M. Hirano, K. Domen and
[Fig. 6(d)]. These behaviors were confirmed by the XRD and H. Hosono, Appl. Catal., A, 277, 239­246 (2004).
subsequent Rietveld analysis as well (Table 1).20) The average 12) H. Miyazaki, Y. Hoshi, Y. Tanimura, T. Kunieda, N. Sakamoto
and H. Suzuki, J. Soc. Powder Technol. Jpn., 46, 86­89
crystallite size of the sol­gel derived sample before ball-milling
(2009).
was 29.3 nm, which was much smaller than that by the solid state 13) H. Suzuki, T. Mori, T. Fujinami, T. Ota, M. Fuji and M.
reaction, 275.5 nm. The grinding effect on the crystallite size was Takahashi, Ceram. Int., 30, 1811­1813 (2004).
almost negligible in the sol­gel derived sample, whereas the 14) S. Kurajica, G. Mali, T. Gazivoda, J. Sipusic and V. Mandic,
crystallite size of the sample by the solid state reaction was J. Sol-Gel Sci. Technol., 50, 58­68 (2009).
decreased to almost 50% by the grinding. The main reason for 15) C. Papelis, W. Um, E. C. Russell and B. J. Chapman, Colloids
this different grinding effect seems to lie in the primary particle Surf., A, 215, 221­239 (2003).
size. The agglomerates composed of tiny primary particles seen 16) M. Haneda, N. Aoki, M. Sasaki, H. Hamada and M. Ozawa,
in the sol­gel derived sample can be deagglomerated by the Appl. Catal., A, 394, 239­244 (2011).
grinding, whereas the large grains seen in the solid state reaction 17) B. Li, S. Kado, Y. Mukainakano, M. Nurunnabi, T. Miyao, S.
Naito, K. Kunimori and K. Tomishige, Appl. Catal., A, 304,
sample were ground rather than deagglomerated. It is easily
62­71 (2006).
understood that the agglomerates composed of tiny particles are 18) H. M. Youn, G. J. Seo, S. Park, C. J. Jung, R. D. Park and K. I.
readily deagglomerated into small particles, which results in the Song, Int. J. Hydrogen Energy, 33, 7457­7463 (2008).
high specific surface area. This result definitely indicates that the 19) S. Qiujie, L. Chengwei and C. Weiqing, J. Rare Earths, 27,
sol­gel method has a quite high advantage for obtaining the high 948­954 (2009).
specific surface area C12A7 powders. 20) A. A. Coelho, J. Appl. Crystallogr., 33, 899­908 (2000).

463