Synthesis of thermally stable χ-alumina by thermal decomposition of aluminum isopropoxide in toluene

Article abstract of DOI:10.1111/j.1551-2916.2004.01543.x

Thermal decomposition of aluminum isopropoxide in toluene at 315°C resulted in χ-alumina that had high thermal stability, whereas the reaction at lower temperatures resulted in formation of an amorphous product. The χ-alumina thus obtained directly transf

Full text of DOI:10.1111/j.1551-2916.2004.01543.x

J. Am. Ceram. Soc., 87 [8] 1543–1549 (2004)
journal
Synthesis of Thermally Stable ␹-Alumina by Thermal Decomposition of
Aluminum Isopropoxide in Toluene
Okorn Mekasuwandumrong,†,§ Hiroshi Kominami,*,†,¶ Piyasan Praserthdam,‡ and
Masashi Inoue*,†
Department of Energy and Hydrocarbon Chemistry, Graduate School of Engineering, Kyoto University,
Kyoto 606-8501, Japan
Department of Chemical Engineering, Faculty of Engineering, Chulalongkorn University, Bangkok 10330, Thailand
Thermal decomposition of aluminum isopropoxide in toluene
at 315°C resulted in ␹-alumina that had high thermal stability,
whereas the reaction at lower temperatures resulted in forma-
tion of an amorphous product. The ␹-alumina thus obtained
directly transformed to ␣-alumina at ϳ1150°C, bypassing the
other transition alumina phases, whereas the amorphous prod-
uct transformed to ␥-alumina and then to ␪-alumina before
final transformation to ␣-alumina. When the ␹-alumina, sol-
vothermally synthesized at 315°C, was recovered by the re-
moval of the solvent at the reaction temperature, thermal
stability of the product was improved further. This procedure
is convenient because it avoids bothersome work-up processes
that yield large-surface-area and large-pore-volume alumina.
the nucleation of ␣-alumina occurs along the {220} crystallo-
graphic plane of ␥-alumina thin films synthesized by radio-
frequency (rf) reactive-sputtering deposition. Johnston et al.18
have reported that ␥-alumina prepared by laser ablation of an
aluminum target in an oxygen atmosphere directly transforms to
␣-alumina at 1200°C, and they have suggested that this result is
due to the lack of water in the as-synthesized powders. Shek et
al.19 have reported that ␥-alumina prepared by oxidation of pure
aluminum metal transforms directly to ␣-alumina at 1370 K. They
have attributed this result to facilitation of nucleation of ␣-alumina
by the strain relaxation of the transition alumina lattice. Simpson
et al.20 have reported that ␥-alumina prepared by electron-beam
evaporation of alumina onto a sapphire substrate transformed to
␣-alumina without formation of the other intermediate phases.
They have attributed this result to the epitaxial growth of
␣-alumina on the sapphire substrate. Ogihara et al.21 have found
that monodispersed, spherical alumina (amorphous) prepared by
the controlled hydrolysis of aluminum alkoxide crystallized to
␥-alumina at 1000°C, which converted to ␣-alumina at 1150°C
without formation of intermediate phases. It also has been reported
that ␥-alumina forms by thermal decomposition of aluminum
sulfate transformed to ␣-alumina directly.22–24
I. Introduction
ANY industrial solid catalysts are made up of active centers
M
anchored on supports that have high porosity, large surface
area, and good mechanical strength as well as sufficient thermal
stability.1 Transition aluminas are most widely used in industry as
support materials, because a variety of aluminas that have these
requisites are commercially available.2
␹-alumina is a modification of transition alumina3–8 and is
characterized by the appearance of a diffraction peak at d ϭ 0.212
nm that cannot be explained by the spinel structure proposed for
other transition aluminas, such as ␥- and ␩-alumina.3,8,9 ␹-alumina
is the dehydrated phase of gibbsite (Al(OH)3). Dollimoer et al.10
have found that ␹-alumina is formed by thermal decomposition of
aluminum oxalate. Because it has been generally believed that
␹-alumina cannot be prepared by routes other than dehydration of
gibbsite, it has been concluded that aluminum oxalate has a crystal
structure similar to that of gibbsite.
Boehmite is one of the polymorphs of AlO(OH). So-called
pseudoboehmite is microcrystalline boehmite that has extra water
content because of surface hydroxyl groups3–5 and is widely used
as a precursor of alumina for catalysis uses. Dehydration of
pseudoboehmite yields ␥-alumina, which transforms to ␦-, then ␪-,
and finally ␣-alumina with increasing calcination temperatures.3–6
These phase transformations are accompanied by a decrease in
surface area and variation of surface properties. The phase trans-
formation of alumina is the cause for the loss of surface area.3 For
catalysts used for high-temperature reactions, such as catalytic
combustion, damage caused by sintering and phase transformation
of alumina must be avoided.1,2 Therefore, much attention has been
given to improve the thermal stability of alumina supports.
In the literature, direct transformation of low-temperature tran-
sition alumina (i.e., ␥-, ␩-, and ␹-alumina) to ␣-alumina without
formation of high-temperature transition alumina (i.e., ␦-, ␪-, and
␬-alumina) has been reported. Chou and Nieh17 have reported that
In a previous paper from our group,11 the thermal decomposi-
tion of aluminum isopropoxide (AIP) in toluene was examined,
and ␹-alumina was formed by this reaction, which directly
transformed to ␣-alumina at 1150°C. In this article, we reveal the
detailed results for ␹-alumina formation and ␣-alumina trans-
formation. Effect of the product recovery procedure also was
examined, and pore structures of the thus-obtained products are
reported.
II. Experimental Procedures
J. E. Blendell—contributing editor
(1) Sample Preparation
AIP (8 or 25 g) (Nacali Tesque, Kyoto, Japan) was dissolved in
60 mL of toluene (Wako, Osaka, Japan) in a test tube, which
served as an autoclave liner, and then the test tube was placed into
a 200 mL autoclave. An additional 40 mL of toluene was placed in
the gap between the autoclave wall and the test tube. The autoclave
was completely purged with nitrogen, heated to a desired temper-
ature at a rate of 2.5°C/min, and maintained at that temperature for
2 h, unless otherwise specified. The reaction temperature was
measured using a thermocouple held in the autoclave wall. The
Manuscript No. 10196. Received May 27, 2003; approved April 13, 2004.
Support for O.M. provided by Thailand Research Fund (TRF) and Association of
International Education in Japan (AIEJ).
*Member, American Ceramic Society.
†Kyoto University.
‡Chulalongkorn University.
§Visiting student from Chulalongkorn University.
¶Present address: Department of Applied Chemistry, Faculty of Science and
Engineering, Kinki University, Osaka 577-8502, Japan.
1543

1544
Journal of the American Ceramic Society—Mekasuwandumrong et al.
Vol. 87, No. 8
Fig. 1. Processing steps for the preparation of ␹-alumina.
Fig. 2. XRD patterns of the products obtained by thermal decomposition
of AIP in toluene under various conditions. Al(25 g)315C2h is the product
obtained by charging a large quantity (25 g) of aluminum isopropoxide.
actual temperature in the reaction vessel was ϳ15°C lower than
the measured temperature. In some experiments, the reaction
mixture was stirred during the reaction. As schematically illus-
trated in Fig. 1, the reaction product was recovered by two
methods. In the first method, after the autoclave was cooled to
room temperature, the resulting product was repeatedly washed
with acetone by vigorous mixing and centrifuging, and then it was
dried in air. In the second method, after the reaction, the autoclave
valve was slightly opened and the fluid phase was evaporated at
the reaction temperature.12,13
The products recovered by the first method are designated as
Al***C*h, where the first three asterisks represent the reaction
temperature and the fourth asterisk represents the holding time at
the reaction temperature. The product recovered by the second
method is specified by adding “P” to the designation. For the
products obtained by reaction with stirring, “S” also is added.
Therefore, SPAl315C2h means the product obtained by the reac-
tion with stirring at 315°C for 2 h followed by the removal of the
solvent at the reaction temperature.
although weak peaks due to pseudoboehmite were observed. On
the other hand, the XRD pattern of Al315C2h clearly showed a
peak at 2␪ ϭ 42.5°. This peak corresponded to the prohibited
321 diffraction of the spinel structure and clearly indicated the
formation of ␹-alumina.3,8,9 An amorphous product was ob-
tained when the reaction was quenched just after the autoclave
temperature reached 315°C (Al315C0h). This result suggested
that crystallization of ␹-alumina took place through a solid-state
transformation mechanism from the amorphous product depos-
ited from the solution.
To examine the stoichiometry of the reaction, the recovered
solvent was analyzed using gas chromatography. Only toluene was
detected (no 2-propanol was found), which indicated that the
organic moiety of AIP decomposed to propene. Therefore, the
overall reaction can be written as follows:
toluene
A part of the product was calcined in a box furnace by heating
to a desired temperature at a rate of 10°C/min and holding at that
temperature for 30 min.
Al͑OCH͑CH3͒ ͒
¡ 0.5Al2O3 ϩ 3C3H6 ϩ 1.5H2O
2
3
A small amount of water (0.7 mL) intentionally added to the gap
between the autoclave wall and the test tube14 did not affect the
formation of ␹-alumina, which indicated that the water formed by
the reaction did not alter the reaction but presumably facilitated the
crystallization of the product. However, addition of a large excess
of water resulted in formation of pseudoboehmite.
Physical properties of the as-synthesized products are summa-
rized in Table I. Amorphous products (Al300C2h and Al315C0h)
had higher surface areas and lower bulk densities, and they exhibited
larger weight losses when compared with the ␹-alumina products.
(2) Characterization
Powder X-ray diffractometry (XRD; Model XD-D1, Shimadzu,
Kyoto, Japan) measurements were made using CuK␣ radiation and
a carbon monochromator. Crystallite size was calculated from the
half-height width of a peak at 67.5° 2␪ using the Scherrer equation.
The composition of the medium after the reaction was analyzed
using gas chromatography (Model GC 9A, Shimadzu, Kyoto,
Japan) with a FID detector. Specific surface area of the product
was calculated using the BET single-point method based on
nitrogen uptake measured at 77 K. Nitrogen adsorption isotherms
were determined using a gas-sorption system (Model Quanta-
chrome Autosorb-1, Yuasa-Ionics, Tokyo, Japan). Pore-size dis-
tribution was calculated using the BJH method with a desorption
branch of the isotherm. Thermogravimetric/differential thermal
analyses (TGA/DTA) were performed (Model TG-50, Shimadzu,
Kyoto, Japan) at a heating rate of 10°C/min in a 40 mL/min flow
of dry air.
Table I. Physical Properties of Products Obtained by the
Thermal Decomposition of AIP in Toluene
BET
surface area Crystallite Bulk density Ignition loss
Product
(m2/g)
size (nm)
(g/cm3)
(%)
Al300C2h
292
291
201
211
213
244
0.34
0.35
0.6
0.64
0.67
0.6
28
36
12
18
11
16
Al315C0h
Al315C2h
10.3
10.7
10.8
9.9
III. Results and Discussion
Al325C2h
(1) Effect of Reaction Temperature and Holding Time
Al315C4h
The XRD patterns of products obtained by the thermal decom-
position of AIP in toluene under various conditions are shown
in Fig. 2. The Al300C2h product was essentially amorphous,
Al(25 g)315C2h†
†Product obtained by the reaction of 25 g of aluminum isopropoxide in toluene at
315°C for 2 h.

August 2004
Synthesis of Thermally Stable ␹-Alumina
1545
Fig. 3. TG and DTA data of the products: (A) Al315C4h; (B) Al315C h; (C) Al(25 g)315C2h; (D) Al325C2h; (E) Al315C0h; and (F) Al300C2h.
Surface areas and crystallite sizes of the latter products fell within a
relatively narrow range irrespective of the reaction conditions.
If the primary particle of ␹-alumina is a sphere that has a
diameter identical to the crystallite size (10 nm) and if the true
density of ␹-alumina is 3.0, then the specific surface area of the
sphere is calculated to be 200 m2/g. This value is in good
agreement with the observed surface areas of the products, which
indicates that each primary particle (crystallite) of the product
exposes its entire surface to the adsorbate molecules (nitrogen).
Results of thermal analyses of the products are shown in Fig. 3.
Two weight-loss processes were detected, both of which were
accompanied by endothermic peaks in DTA. The first weight-loss
was observed at Ͻ160°C and was due to the desorption of
physisorbed water. The second weight loss occurred at 300°–
600°C and was attributed to the dehydration of surface hydroxyl
groups, because the IR spectra of the products (data not shown)
showed no indication of the presence of organic moieties. We
assumed that the whole surface (200 m2/g) of the product was
covered with surface hydroxyl groups, each of which occupied
7.8 ϫ 10Ϫ2 nm2 (calculated from the cell dimension of gibbsite).
Therefore, the weight loss caused by the dehydration of the surface
hydroxyl groups that yielded surface anion vacancies and surface
oxide ions was calculated to be 3.83%. The observed weight loss
at 300°–600°C was 4.2%–5.6% and exceeded the calculated value.
Therefore, the products seemed to contain structural hydroxyl
groups or strongly bound chemisorbed water.
The XRD patterns of Al300C2h calcined at various tempera-
tures are shown in Fig. 4. The peak at 42.5° 2␪ was not observed
during the thermal transformation of the amorphous products,
which indicated that crystallization in the inert organic solvent was
essential for the formation of ␹-alumina.
Figure 5 shows the XRD patterns of Al315C4h calcined at
various temperatures. Direct phase transformation of ␹-alumina to
␣-alumina started at ϳ1100°C and essentially completed at
1150°C. A small amount of ␪-alumina was detected for the sample
calcined at 1150°C, and this phase seemed to be formed from
contaminated pseudoboehmite or amorphous alumina. Al315C2h
exhibited phase transformation behavior similar to that of
Al315C4h.
Fig. 4. XRD patterns of the product obtained by the thermal decompo-
sition of AIP in toluene at 300°C for 2 h (Al300C2h) and the samples
obtained by calcination thereof at various temperatures.
Fig. 5. XRD patterns of the product obtained by the thermal decompo-
sition of AIP in toluene at 315°C for 4 h (Al315C4h) and the samples
obtained by calcination thereof at various temperatures.

1546
Journal of the American Ceramic Society—Mekasuwandumrong et al.
Vol. 87, No. 8
Table II. BET Surface Areas of the Samples Obtained by
The BET surface areas of the calcined products are summarized
in Table II. Although the BET surface areas of the amorphous
products (Al300C2h and Al315C0h) were higher than those of the
␹-alumina products, they rapidly decreased, and ϳ50 m2/g of the
surface area remained when the products were calcined at 1100°C.
On the other hand, Al325C2h exhibited higher thermal stability
and maintained a BET surface area of 85 m2/g after calcination at
the same temperature.
It is generally believed that the ␣ transformation occurs by the
nucleation and growth mechanism.25–27 Although phase transfor-
mation of an oriented ␥-alumina film has been reported to proceed
by a topotactic process,17 this result is considered to be a rather
special case,9,28 because topotactic transformation of transition
aluminas in powder or pellet form never has been reported.
Because formation of ␣-alumina from transition alumina requires
rearrangement of a cubic oxygen-sublattice to a hexagonal
oxygen-sublattice, a large activation energy is required. Nucleation
of ␣-alumina is the rate-determining step, and explosive crystal
growth follows.25–27 Because of the rapid crystal growth that
accompanies ␣-alumina transformation, the surface area of the
product calcined at 1150°C abruptly decreases, as shown in Table
II.
Calcination of the Products at Various Temperatures
Surface area (m2/g)
Product
Phase
As-synthesized 1000°C 1100°C 1150°C
Al300C2h Amorphous
Al315C0h Amorphous
Al315C2h ␹-alumina
Al325C2h ␹-alumina
Al315C4h ␹-alumina
292
291
201
211
213
117
107
137
105
100
47
55
85
75
80
25
10
11
43
15
In general, ␹-alumina obtained by the dehydration of gibbsite
(Ͻ200 nm) transformed to ␣-alumina through ␬-alumina:3–8
gibbsite ¡ ␹-alumina ¡ ␬-alumina ¡ ␣-alumina
However, ␹-alumina obtained by the thermal decomposition of AIP in
toluene transformed to stable ␣-alumina directly, bypassing the
formation of the high-temperature transition alumina, i.e., ␬-alumina.
Although we cannot draw a common feature for the aluminas
that directly transform to ␣-alumina, the following two points are
addressed.
(1) Except for those formed from aluminum sulfate, the
transition aluminas that exhibited direct transformation to
␣-alumina were prepared from aluminum metal or alkoxides, and,
therefore, they were free of foreign ions.
Imamura et al.29 have compared the catalytic activities of
␹-alumina prepared by the present method and ␥-alumina prepared
by the hydrolysis of AIP, and they have found that these two
catalysts have essentially identical activities for the isomerization
of 1-butene, although the latter material has a higher surface area
(261 m2/g). Moreover, they have reported that calcinations at
1000°C of the ␹-alumina maintain 56% of the surface area of the
as-synthesized product, while that of the ␥-alumina decreases to
37%.29 This result is reasonable, because the former compound
does not suffer from the phase transformation, whereas the latter
transforms to high-temperature transition aluminas.
(2) ␥-alumina crystallized at high temperatures had a ten-
dency to transform to ␣-alumina without formation of high-
temperature transition aluminas.
The first point can be explained using either a thermodynamics or
a kinetics reason. When foreign ions thermodynamically destabilize
the low-temperature transition alumina structure, transformation to
high-temperature transition aluminas does not occur at lower temper-
atures. Foreign ions may facilitate the nucleation of high-temperature
transition aluminas, and this is kinetic enhancement of the formation
of high-temperature transition aluminas. On the other hand, absence
of foreign ions can disturb the crystallization of high-temperature
transition alumina, which thereby facilitates direct crystallization of
␣-alumina. For the second point, ␥-alumina that crystallizes at higher
temperatures contains a lesser number of crystal defects than that
crystallized at lower temperatures, because sufficient thermal energy
is supplied at the crystallization stage. This also lowers the energy
level of the transition alumina and prohibits the formation of high-
temperature transition aluminas.
(2) Effect of Product Recovery and Stirring
The XRD patterns of the products obtained from the reaction of
AIP in toluene at 315°C with stirring (SAl315C2h) and/or with the
removal of the solvent at the reaction temperature (SPAl315C2h
and SAl315C2h) showed that all the products were ␹-alumina. The
physical properties of the products are summarized in Table III.
When the solvent was removed at the reaction temperature, the
BET surface area slightly increased, and the total pore volume
significantly decreased and bulk density of the product decreased.
These results were attributed to the fact that this procedure avoided
coagulation of the particles caused by surface tension of the liquid
remaining among the particles,30 which took place during the
drying stage of the ordinary procedure.
Although the reaction temperature (315°C) was slightly lower
than the critical point of toluene (318°C), the fluid near the critical
point was similar to that of the supercritical fluid, and, therefore,
this product recovery procedure was considered as a supercritical
drying method.30,31 Using this procedure, we conducted the
solvothermal reaction and supercritical drying in one pot, which
avoided bothersome work-up processes and coagulation of the
products.12,13
Gibbsite is usually contaminated with a small amount of Naϩ
ions,3,4 and thermal dehydration of gibbsite that yields ␹-alumina
cannot eliminate the Naϩ ions from the matrix. Therefore, the
␹-alumina exhibits very broad peaks in the XRD pattern because of
the low crystallinity and many defects in the structure. These defects
in the ␹-alumina structure may promote the formation of ␬-alumina.
On the other hand, ␹-alumina obtained by the thermal decomposition
of AIP in toluene is not contaminated with Naϩ ions. Moreover, the
presence of a small amount of water in the solvothermal reaction
possibly increases the crystallinity of the product. These situations
stabilize the ␹-alumina structure and disturb the formation of
␬-alumina, which facilitates direct transformation to ␣-alumina.
The XRD patterns of SAl315C2h calcined at various tempera-
tures are shown in Fig. 6. The phase transformation started at
Table III. Effect of Stirring and Product Recovery Procedure on the Physical Properties of the As-Synthesized Product
Obtained by the Thermal Decomposition of AIP in Toluene
BET
surface area
Crystallite
size (nm)
Bulk density
(g/cm3)
Total pore
Average pore
Mode pore
Product
(m2/g)
St (m2/g)†
volume (cm3/g)‡
diameter (nm)§
diameter (nm)¶
Al315C2h
201
285
255
281
10.3
12.4
8.3
203
283
243
281
0.6
0.709
0.626
1.899
2.321
14.1
8.8
23.5
33.1
13.2
7.4
SAl315C2h
PAl315C2h
SPA1315C2h
0.57
0.1
9.3
0.12
†Calculated from the initial slope of the t plot. ‡Total nitrogen uptake at relative pressure of 0.98. §Calculated from total pore volume and BET surface area assuming that the
pore is tubular. ¶Calculated from the desorption branch of the isotherm using the BJH method.

August 2004
Synthesis of Thermally Stable ␹-Alumina
1547
Fig. 6. XRD patterns of the product obtained by the reaction of AIP in
toluene with stirring (SAl315C2h) and the samples obtained by calcination
thereof at various temperatures.
Fig. 7. XRD patterns of the product obtained by the reaction of AIP in
toluene with stirring and with the product recovery by the supercritical
drying method (SPAl315C2h) and the samples obtained by calcination
thereof at various temperatures.
ϳ1100°C and was completed at 1150°C. ␪-alumina was not
detected during the phase transformation. This result suggested
that SAl315C2h was more homogeneous than Al315C2h. Because
the water content in the gas phase was much higher than that in the
reaction medium, particles formed near the surface of the medium
were affected by the water in the gas phase. Stirring during the
reaction avoided this effect, and, therefore, a more homogeneous
product was obtained.
BET surface areas of the calcined products are summarized in
Table IV. Stirring during the reaction had a minor effect on the
physical properties of the calcined products, although SAl315C2h
had a larger surface area than Al315C2h. As discussed above, the
reacation proceeded by initial formation of an amorphous phase
through the deposition from a homogeneous solution followed by
transformation to ␹-alumina with the aid of adsorbed water. The
surface area of the ␹-alumina products seemed to be determined by
the particle size of the amorphous phase, which was, in turn,
determined by the degree of supersaturation of the system.
Therefore, stirring little affected the surface areas of the calcined
products.
This result was attributed to the fact that these products trans-
formed to ␣-alumina at higher temperatures.
(3) Morphology of the Products
Scanning electron microscopy results of SPAl315C2h are
shown in Fig. 8 and are representative of all the products. The
product particles are irregularly shaped and composed of fine
primary particles. The other products have similar morphologies.
(4) Pore Structure of the Products
The nitrogen adsorption isotherms (data not shown) of the
products recovered by the ordinary method (Al315C2h and
SAl315C2h) exhibited a hysteresis-loop characteristic of type
E, as classified by de Boer,32 which was explained by the
presence of tabular or ink-bottle pores of varying radius.32
Because this type of hysteresis loop is commonly observed for
spheroidal cavities or voids between close-packed spherical
particles, the observed hysteresis loop seemed to be due to
Figure
7 shows the phase transformation behavior of
SPAl315C2h. The phase transformation started at ϳ1150°C and
was almost completed at ϳ1200°C. PAl315C2h showed essen-
tially the same phase transformation behavior as SPA315C2 h. The
transformation temperatures of the products obtained by supercriti-
cal drying were Ͼ50°C higher than those recorded using the
ordinary method. This point is discussed later.
Table IV shows that SPAl315C2h and PA315C2h maintained
high surface areas of ϳ75 m2/g, even after calcinations at 1150°C.
Table IV. Effects of Stirring and Product Recovery
Procedure on the BET Surface Areas of Samples Calcined at
Various Temperatures
Surface area (m2/g)
Product
As-synthesized
1000°C
1100°C
1150°C
1200°C
Al315C2h
201
285
255
281
137
132
142
148
85
85
108
115
11
13
73
77
7
8
10
14
SAl315C2h
PAl315C2h
SPAl315C2h
Fig. 8. Scanning electron micrograph of SPAl315C2h.

1548
Journal of the American Ceramic Society—Mekasuwandumrong et al.
Vol. 87, No. 8
Al315C2h and SAl315C2h had mode pore diameters in the
mesopore range.
When supercritical drying was performed for product recovery,
the type-E hysteresis loop disappeared (data not shown), which
was ascribed to enlargement of pore sizes. The mode pore
diameters of PAl315C2h and SPAl315C2h (Fig. 10) were much
larger than those observed for Al315C2h or SAl315C2h, which
indicated that the primary particles were packed quite loosely. The
supercritical drying method prevented the coagulation of the
primary particles, and, therefore, PAl315C2h and SPAl315C2h
had a macropore system that formed between the loosely packed
primary particles.
The XRD and BET data showed that the products prepared by
one-pot synthesis (PAl315C2h and SPAl315C2h) had higher
thermal stabilities. This was explained by the increased pore size
and the total pore volume (Table III). Transformation from
transition alumina to ␣-alumina took place through a nucleation
and growth mechanism,25–27 and it was generally believed that
nucleation occurred at the contact point between the particles.19
Because the product obtained using supercritical drying had a
lesser number of contact points, the transformation to ␣-alumina
required higher temperature. Moreover, mass transfer for a long
distance was required for the sintering of these products, which
also contributed to their thermal stabilities.35
Although sintering of the particles inevitably occurs, thermal
treatment of the present products causes a relatively small decrease
in the surface area, because the particles do not suffer from the
phase transformation to high-temperature transition aluminas.
Therefore, the present product has a potential use as a catalyst
material for use at relatively high temperatures.
Fig. 9. t plots of (a) Al315C2h, (b) SAl315C2h, (c) PAl315C2h, and (d)
SPAl315C2h.
IV. Conclusions
capillary condensation into the voids between the primary
particles of ␹-alumina.
Thermal decomposition of AIP in toluene at 315°C gave
␹-alumina, which directly transformed to ␣-alumina. This trans-
formation behavior of nanocrystalline ␹-alumina was attributed to
the absence of Naϩ ions and the less defected structure. At lower
reaction temperatures, an amorphous product was obtained, from
which ␥-alumina was formed at lower temperature. The transfor-
mation temperature of the ␹-alumina products was recovered by
the supercritical drying method, which was Ͼ50°C higher than
those recovered by washing and drying in air. The former products
maintained a surface area Ͼ70 m2/g, even after calcination at
1150°C. The increased pore volume and pore size, brought about
by the prevention of the coagulation of the alumina particles,
contributed to the higher thermal stability of these ␹-alumina
products.
The t plots33,34 derived from the isotherm are presented in Fig.
9. In all the cases, the first part of the plot was a straight line going
through the origin, and, from the slope of the line, total surface
area, St, could be calculated. When a sample had a micropore
system, an abrupt decrease in the slope was detected.33,34 The
present products showed no sign of a decrease in the slope in the
t plots, which indicated the absence of a micropore system. Good
agreement between SBET and St (Table III) also supported this
argument. At higher relative pressures (higher t values), deviation
from the straight line occurred, because capillary condensation in
the mesopore system formed between the primary particles of
␹-alumina. The pore-size distributions are shown in Fig. 10.
References
1C. N. Satterfield, Heterogeneous Catalysis in Practice. McGraw-Hill, New York,
1980.
2R. K. Oberlander, “Aluminas for Catalysts—Their Preparation and Properties”; in
Applied Industrial Catalysis. Edited by B. E. Learch. Academic Press, London, U.K.,
1984.
3W. H. Gitzen, Alumina as a Ceramics Material; pp. 14–54. American Ceramic
Society, Columbus, OH, 1970.
4K. Wefers and G. M. Bell, “Oxides and Hydroxides of Alumina,” Tech. Paper No.
19, Alcoa, Bauxite, AR, 1972.
5K. Wefers, “Nomenclature, Preparation, and Properties of Aluminum Oxides,
Oxide Hydroxides, and Trihydroxides”; in Alumina Chemicals: Science and Tech-
nology Handbook. Edited by L. D. Hart. American Ceramic Society, Westerville, OH,
1990.
6H. C. Stumpf, A. S. Russell, J. W. Newsome, and C. M. Tucker, “Thermal
Transformations of Aluminas and Alumina Hydrates,” Ind. Eng. Chem., 42 [7]
1398–403 (1950).
7H. Saalfeld, “Structure Phases of Dehydrated Gibbsite”; pp. 310–15 in Reactivity
of Solids. Edited by J. H. de Boer. Elsevier, Amsterdam, The Netherlands, 1961.
8G. W. Brindley and J. O. Choe, “The Reaction Series, Gibbsite to chi-Alumina to
kappa-Alumina to Corundum,” Am. Mineral., 46 [7–8] 771–85 (1961).
9I. Levin and D. Brandon, “Metastable Alumina Polymorphs: Crystal Structures
and Transition Sequences,” J. Am. Ceram. Soc., 81 [8] 1995–2012 (1998).
10D. Dollimore, J. Dollimore, and P. D. Perry, “The Thermal Decomposition of
Oxalates. Part VIII. Thermalgravimetric and X-ray Analysis Study of the Decompo-
sition of Alumina Oxalate,” J. Chem. Soc. A, [3] 448–50 (1967).
Fig. 10. Pore-size distributions of as-synthesized products: (A)
Al315C2h; (B) SAl315C2h; (C) PAl315C2h; and (D) SPAl315C2h.

August 2004
Synthesis of Thermally Stable ␹-Alumina
1549
11M. Inoue, H. Kominami, and T. Inui, “Thermal Transformation of ␹-Alumina
Formed by Thermal Decomposition of Aluminum Alkoxide in Organic Media,”
J. Am. Ceram. Soc., 75 [9] 2597–98 (1992).
21T. Ogihara, H. Nakagawa, T. Yanagawa, N. Ogata, and K. Yoshida, “Preparation
of Monodisperse, Spherical Alumina Powders from Alkoxides,” J. Am. Ceram. Soc.,
74 [9] 2263–69 (1991).
22D. W. Johnson and F. J. Schnettler, “Characterization of Freeze-Dried Al2O3 and
Fe2O3,” J. Am. Ceram. Soc., 53 [8] 440–44 (1970).
12P. Pratherthdam, M. Inoue, O. Medkasuvandumrong, W. Thanakulrangsan, and
S. Phatanasri, “Effect of Organic Solvents on the Thermal Stability of Porous
Silica-Modified Alumina Powders Prepared via One Pot Solvothermal Synthesis,”
Inorg. Chem. Commun., 3 [11] 671–76 (2000).
23E. Kato, K. Daimon, and M. Nanbu, “Decomposition of Two Aluminum Sulfates
and Characterization of the Resultant Aluminas,” J. Am. Ceram. Soc., 64 [8] 436–45
(1981).
13S. Iwamoto, K. Saito, M. Inoue, and K. Kagawa, “Preparation of the Xerogels of
Nanocrystalline Titanias by the Removal of the Glycol at the Reaction Temperature
after the Glycothermal Method and Their Enhanced Photocatalytic Activities,” Nano
Lett., 1 [8] 417–21 (2001).
24M. D. Sacks, T.-Y. Tseng, and S. Y. Lee, “Thermal Decomposition of Spherical
Hydrated Basic Aluminum Sulfate,” Am. Ceram. Soc. Bull., 63 [2] 301–10 (1984).
25F. W. Dynys and J. W. Halloran, “␣-Alumina Formation in Alum-Derived
␥-Alumina,” J. Am. Ceram. Soc., 65 [9] 442–48 (1982).
14M. Inoue, H. Kominami, and T. Inui, “Novel Synthesis Method for Thermally
Stable Monoclinic Zirconia. Hydrolysis of Zirconium Alkoxides at High Tempera-
tures with a Limited Amount of Water Dissolved in Inert Organic Solvent from the
Gas Phase,” Appl. Catal. A General, 121 [1] L1–L5 (1995).
26K. J. Morrissey, K. K. Czanderna, C. B. Carter, and R. P. Merrill, “Growth of
␣-Alumina within a Transition Alumina Matrix,” J. Am. Ceram. Soc., 67 [5]
C-88–C-90 (1984).
27D. S. Tucker, “Gamma-to-Alpha Transformation in Spherical Aluminum Oxide
Powders,” J. Am. Ceram. Soc., 68 [7] C-163–C-164 (1985).
15J. J. Fripiat, H. Bosmans, and P. G. Rouxhet, “Proton Mobility in Solids. I.
Hydrogenic Vibration Modes and Proton Delocalization in Boehmite,” J. Phys.
Chem., 71 [4] 1097–11 (1967).
28R. B. Bagwell, G. L. Messing, and P. R. Howell, “The Formation of ␣-Alumina
from ␪-Alumina: The Relevance of a ‘Critical Size’ and Diffusional Nucleation or
‘Synchro-Shear’?,” J. Mater. Sci., 36 [7] 1833–41 (2001).
16A. B. Kiss, G. Keresztury, and L. Farkas, “Raman and IR Spectra and Structure
17
2h
of Boehmite (␥-AlOOH), Evidence for the Recently Discarded D Space Group,”
29S. Imamura, T. Kitao, H. Sasaki, and K. Utani, “Properties of ␹-Alumina as
Related to Catalysis,” React. Kinet. Catal. Lett., 55 [1] 19–24 (1995).
30J. N. Armor and E. J. Carlson, “Variables in the Synthesis of Unusually High
Pore Volume Aluminas,” J. Mater. Sci., 22 [7] 2549–56 (1987).
31G. M. Pajonk, “Aerogel Catalysts,” Appl. Catal., 72 [2] 217–66 (1991).
32J. H. de Boer, The Structure and Properties of Porous Materials. Butterworth,
London, U.K., 1958.
Spectrochim. Acta, 36A, 653–58 (1980).
17T. C. Chou and T. G. Nieh, “Nucleation and Concurrent Anomalous Grain
Growth of ␣-Al2O3 during ␥–␣ Phase Transformation,” J. Am. Ceram. Soc., 74 [9]
2270–79 (1991).
18G. P. Johnston, R. Muenchausen, D. M. Smith, W. Fahrenholtz, and S. Foltyn,
“Reactive Laser Ablation Synthesis of Nanosize Alumina Powder,” J. Am. Ceram.
Soc., 75 [12] 3293–98 (1992).
33B. C. Lippens and J. H. de Boer, “Studies on Pore Systems in Catalysts. I. The
Adsorption of Nitrogen, Apparatus, and Calculation,” J. Catal., 3 [1] 32–37 (1964).
34B. C. Lippens and J. H. de Boer, “Studies on Pore Systems in Catalysts. V. The
t Method,” J. Catal., 4 [3] 319–23 (1965).
19C. H. Shek, J. K. L. Lai, T. S. Gu, and G. M. Lin, “Transformation Evolution and
Infrared Absorption Spectra of Amorphous and Crystalline Nano-Al2O3 Powders,”
Nanostruct. Mater., 8 [5] 605–10 (1997).
20T. W. Simpson, Q. Wen, N. Yu, and D. R. Clarke, “Kinetics of the Amor-
phous3␥3␣ Transformation in Aluminum Oxide: Effect of Crystallographic Ori-
entation,” J. Am. Ceram. Soc., 81 [1] 61–66 (1998).
35T. Ishikawa, R. Ohashi, H. Nakabayashi, U. Kakuta, A. Ueno, and A. Furuta,
“Thermally Stabilized Transitional Alumina Prepared by Flame Pyrolysis of Boeh-
mite Sols,” J. Catal., 134 [1] 87–97 (1992).
Ⅺ