Synthesis of Short Fibrous Boehmite Suitable for Thermally Stabilized Transition Aluminas Formation

Article abstract of DOI:10.1006/jcat.1998.2144

Transmission electron microscopy showed that xerogel of long and short fibrous boehmite samples, AlO(OH), were respectively produced via hydrolysis of aluminium alkoxide at slow and high stirring rates. Thermogravimetry, differential scanning calorimetry,

Full text of DOI:10.1006/jcat.1998.2144

JOURNAL OF CATALYSIS 178, 198–206 (1998)
ARTICLE NO. CA982144
Synthesis of Short Fibrous Boehmite Suitable for Thermally
Stabilized Transition Aluminas Formation
Kamal M. S. Khalil
Chemistry Department, Faculty of Science, South Valley University, Sohag 82524, Egypt
Received December 29, 1997; revised April 20, 1998; accepted April 28, 1998
phous nature of the parent alumina. Murrell and Tauster
Transmission electron microscopy showed that xerogel of long (18) concluded that an alumina sol of a critical size range
and short fibrous boehmite samples, AlO(OH), were respectively (not less than ca 20 nm) can be transformed into the fi-
produced via hydrolysis of aluminium alkoxide at slow and high
stirring rates. Thermogravimetry, differential scanning calorime-
nalalumina structure without significant aggregation, which
otherwise (for smaller particles) produce a dense alumina
try, and X-ray diffractometry, showed the two xerogels to produce
structure of lower surface area. Moreover, the phase tran-
�
�
similar calcination products, at 400 C and up to 1000 C, for the
chemical and phase composition of the material bulk. In contrast,
nitrogen adsorption measurement revealed different surface textu-
ral consequences of the boehmite fiber length. It has been concluded
that the short fiber boehmite produces on calcination aluminas that
are more resistant to particle sintering (thermally stabilized) than
�
sition of � -alumina into �-alumina (at ca 1000 C), which
leads to a drastic decrease in specific surface area, may be
slowed down by adding small amounts of a different oxide
(
7, 8). However, catalysis on these modified aluminas might
be different from that on pure aluminas. Thus, thermal sta-
bilization of � -alumina is welcomed if achieved using pure
boehmite (7).
the long fiber boehmite.
�c 1998 Academic Press
Key Words: alkoxide derived boehmite; stabilized alumina; high
surface area alumina.
Different methods have been reported for the prepara-
tion of thermally stable ultra pure alumina derived from
aluminum tri-isopropoxide. These include, fume pyrolysis
of fibrous boehmite sol (7), where the achieved stability was
ascribed to suppression of the rate of phase transformation
1
. INTRODUCTION
Surface texture and thermal stability largely determine to �-alumina caused by the fact that alumina particles were
the performance of catalytic aluminas (1). In fact, these pro- of crude structures assembled from fibrils. Another attempt
perties are influenced strongly by the preparation method has been done by heat treatment of hydrolyzed Aluminum
applied and depend on the major phase present (1–5). iso-propoxide sol of fibrous particles mixed with carbon
Alkoxide derived fibrous boehmite is an important precur- matrix (saccharose) (8). Further improvement of the ther-
sor for preparation of high surface area single (6–10) and mal, texture, and structural properties of alumina may be
mixed aluminum oxide (11–13). Alkoxide-derived boeh- achieved via a super critical drying method (19–21). De-
mite is normally produced by the Yolads process (14, 15). creasing the zeta-potential of the boehmite particles during
The process involves hydrolysis of aluminum iso-propoxide aging gives rise to better structural stability, too (22). It has
(
or sec-butoxide) in excess of water, followed by peptiza- been reported (23) that high surface area alumina may be
�
tion via acid addition. Aging over a period of � 24 h at 80 C synthesized by applying an inert solvent, polar super critical
yields the production of boehmite, whereas aging at room fluids, and large dilution of aluminum alkoxide.
temperature results in bayerite, Al(OH)3. When porous
The above survey highlights that short fibrous boehmite
texture and high surface areas are preferred, it is necessary of a certain range of size is the base point for stabilized
to maintain the textural properties of the wet-hydrolysed texture formation of alumina. Our previous investigations
gel by adopting the appropriate drying method (16).
(24–25) have indicated that fibrous boehmite can be
The structure of the hydrolyzed gel largely controls the produced through a single-stage preparation method. The
thermal properties of the produced alumina. An alkoxide- method utilizes a 1 : 1 water/alkoxy ratio in a nonpolar
�
derived alumina calcined at 500 C (and still amorphous) ex- solvent. Thus it is different from the Yoldas method
2
hibits a surface area as 600 m /g (17). However, this alumina which utilizes high water/alkoxy ratio in polar solvents.
calcination product completely transfers to �-alumina of The present work describes the controlling of the size
�
very low surface area (17) at a temperature as low as 800 C. of the produced fibrous boehmite. Examining structural,
This poor structural stability was attributed to the amor- textural, and thermal stability of the produced xerogel in
198
0021-9517/98 $25.00
Copyright �c 1998 by Academic Press
All rights of reproduction in any form reserved.

SHORT FIBROUS BOEHMITE
199
relation to the preparation conditions. These properties tification purposes diffraction patterns thus obtained were
were determined by X-ray powder diffractometry (XRD), matched with JCPDS standards (26).
transmission electron microscopy (TEM), thermogravime-
try (TG), Differential scanning calorimetery (DSC), and 2.4. Transmission Electron Microscopy (TEM)
N2 adsorption volumetery.
TEM micrographs were obtained, using a model JEM-
1
010 Jeol microscope (Japan). Test samples were prepared
2
. EXPERIMENTAL METHODS
by ultrasonic dispersion of the solid in isopropanol. A drop
of the resulting suspension was doped onto a carbon-coated
grid and allowed to dry at 60 C. A number of grids were
2
.1. Boehmite Preparation
�
Two boehmite xerogel samples were prepared using the prepared for each sample by this method and investigated
�
following procedure. Aluminum tri-isopropoxide (99.99% , by TEM at 100 kV. The 400 C calcination products were
Aldrich) 0.05 molar solution in n-heptane was hydrolyzed selected as representative for the morphology of the pre-
by a dropwise addition of a calculated amount of water to cursor materials. This was because low calcination temper-
furnish the 1 : 1 water/alkoxy ratio. The solution was stirred ature samples are less stable under the electron beam of the
magneticallyat 100rpm (slow-stirred sample), or at 800rpm microscope and take advantage of the topotactic nature of
(
fast-stirred sample). Stirring was allowed to continue for the boehmite fiber.
3
0 min in each case; then the obtained gel was allowed to
2
.5. Nitrogen Gas Adsorption
stand for 3 days without any further stirring. Subsequently,
the gel was filtered off from its solution and allowed to dry
at room temperature for an overnight period and then at
N2 gas adsorption/desorption isotherms were measured
�
at � 196 C using a model ASAP 2010 (Micromertics Instru-
�
6
0 C for 24 h. To examine the surface and the texture for
ment Corporation, USA). Test samples were thoroughly
the two xerogels, portions of the xerogels were calcined in a
Muffelfurnace for 3h at 400, 800, and 1000 C. An additional
calcination for 10h at 1000 C for the two xerogelswasmade.
Moreover, a group of materials (for the two xerogels) were
calcined at 400–1000 C for 30 h and examined by XRD (see
XRD results below).
�
outgassed for 2 h at 250 C. The specific surface area, SBET,
�
was calculated applying the BET equation (27). Porosity
distribution was generated by DFT plus V1.00 (2010) soft-
ware of the instrument indicated implementing the classical
Kelvin equation, slit like pore shape, and Halsey thickness
curve (28).
�
�
2
.2. Thermal Analysis
3
. RESULTS
A thermal analyst 2000 TA instrument controlling a 2050
thermogravimetric analyzer and a 2010 differential scan- 3.1. Characteristics of the Dry Hydrolysis Products
ning calorimeter was used. For thermogravimetry (TG) a
Figure 1 shows the XRD patterns of the slow and
platinum sample boat was used with the sample size being
fast stirred dry hydrolysis products (xerogels). Clearly the
two patterns are largely similar, in indexing aluminum
�
1
0.0 � 0.1 mg. The heating rate was 5 C/min in flow of nitro-
gen gas at the flowrate of 30ml/min. Transition temperature
analysis was carried out for each region in each TG curve
using the TGA 5.1 software of the instrument indicated.
For differential scanning calorimetry (DSC) measure-
ment, aluminum sample pans were used with the sample
size 3.0 � 0.1 mg. Each sample pan was covered with a lid
to improve sample-to-pan contact, but without crimping of
the pan to allow volatile materials to skip out. Measurement
was carried out against an empty covered and crimped alu-
minum pan as a reference, following cell constant determi-
nation and temperature calibration. The heating rate was
�
5
C/min. Transition temperature analysis was carried out
for each region in each DSC curve using the DSC 4.0 soft-
ware of the instrument indicated.
2
.3. X-Ray Powder Diffractometry (XRD)
XRD powder diffractograms were recorded for the test
samples by means of a model JSX-60 PA Jeol diffractome-
FIG. 1. XRD patterns obtained for the slow- and the fast-stirred
ter, equipped with a Ni-filtered CuK� radiation. For iden- xerogels.

2
00
KAMAL M. S. KHALIL
FIG. 3. DSC curves recorded for the slow- and the fast-stirred xero-
gels.
FIG. 2. TG and DTG curvesrecorded for the slow-and the fast-stirred
xerogels.
TEM micrographs, shown in Figs. 4a and 4b, indicate the
textural differences between the two xerogels. It is obvious
that the slow stirring resulted in long (ca 200 nm) fibrous
boehmite of crumpled sheet-like morphology (a), whereas
the fast stirring resulted in short (ca 50 nm)—relatively—
straight ones (b).
�
oxyhydroxide, boehmite. TG curvesrecorded at 5 C/min for
the two xerogels are shown in Fig. 2, along with their DTG
curves. The figure shows that dehydration occurs gradually,
with no sign of weight stable steps. However, four weight
loss regions may be characterized for each sample. These
weight loss regions are indicated by “start” and “stop” tem-
3
.2. Characteristics of the Xerogels Calcination Products
�
XRD patternswere recorded for the 400, 800, and 1000 C
peratures in Table 1. Results of the transition temperature (3 h) calcination products of the two xerogels. The patterns
analysis for each start–stop temperature region as “onset, reveal that the calcination products were amorphous. Nev-
�
maximum, and end” of transition step are cited in Table 1, ertheless, the patterns recorded for the 800 and 1000 C cal-
along respective chemical formula.
cination products (of the two xerogels) show very small
Figure 3 shows the DSC curves recorded at 30–500 C, peaks which could be related to the main d-spacing of
and 5 C/min for the two xerogels. Three endothermic peaks � - and �-alumina, respectively. To explore the ultimate
�
�
are observed for each sample. The onset and maximum tem- change that would occur upon prolonged calcination (i.e.,
peratures for each start–stop region indicated are cited in structural stability) of the above materials. Portions of
�
Table 2, along with the area of each peak measured in J/g. the two xerogels were calcined for 30 h at 400–1000 C.
TABLE 1
TG Results for the Two Xerogels as the Onset, Maximum, End, Loss Percentage for Each Start–Stop Range Indicated
for the Different Dehydration Steps
�
�
�
�
�
Formulaa
Precursor dehydration step
Start ( C)
Onset ( C)
Max ( C)
End ( C)
Stop ( C)
W (% )
Slow
Fast
I
II
III
IV
44
181
277
496
62
235
343
541
65
248
388
560
77
257
434
660
181
277
496
890
88.47
83.74
72.65
70.58
AlO(OH) � 0.78H2O
AlO(OH) � 0.58H2O
1/2Al2O3 � 0.26H2O
1/2Al2O3 � 0.09H2O
AlO(OH) � 0.75H2O
AlO(OH) � 0.37H2O
1/2Al2O3 � 0.23H2O
1/2Al2O3 � 0.02H2O
I
II
III
IV
31
192
280
480
32
231
335
534
56
244
377
553
77
252
423
651
191
280
480
893
87.97
82.61
72.30
69.71
a
�
H2O values were calculated based on considering that the weight recorded (69.46% ) for the fast-stirred xerogel at 1000 C as Al2O3 �0.00H2O.

SHORT FIBROUS BOEHMITE
201
Figure 5 shows XRD patterns for the 30-h calcination prod-
ucts at 600, 800, and 1000 C for the slow stirred xerogel.
TABLE 3
�
Surface Area, cBET, and Volume in Pores of the Materials Calcined
at Different Temperatures
The patterns are rather similar to those obtained for the
long fibers. Indexing of the patterns (26) reveals that the
�
3
calcination products at 400 C, of the two xerogels, were
Volume in pores (cm /g)
Surface area
completelyamorphous(not shown in Fig. 5). Just slight crys- Calcination
2
(
m /g)
cBET
<1.80 nm
Fast
>200 nm
tallinity into gamma alumina commenced to appear after
temp.
( C)
�
�
�
Slow Fast Slow Fast Slow
Slow Fast
calcination at 600 C and developed considerably at 800 C.
�
The 1000 C calcination resulted in a phase transition into
400/3 h
389
240
182
139
376
257
227
171
218
197
94
311 0.107 0.111 0.343 0.31
121 0.073 0.072 0.186 0.21
141 0.052 0.068 0.145 0.17
113 0.047 0.057 0.088 0.11
weakly crystalline �-alumina.
800/3 h
1000/3 h
1000/10 h
Nitrogen gas adsorption was measured at liquid nitro-
�
�
88
gen temperature on the 400, 800, 1000 C (3 h), and 1000 C
(
10 h) calcination products of the two xerogels.
The N2 adsorption/desorption isotherms for the 3-h cal-
cination products for the slow- and fast-stirred xerogels
Figures 9 and 10 show the pore volume distribution
PVD) curves for the different calcination products of the
are shown in Figs. 6 and 7, respectively. Isotherms of the
(
�
4
00 C/3 h calcination (curves indicated “a” on the corres-
slow and the fast stirred xerogels, respectively. PVD for the
�
ponding figures) are of type IV of isotherm indicating meso-
porous surfaces (29). They exhibit type H3 hysteresis loops,
which are normally observed for aggregates of plate-like
particles giving rise to slit-shaped pores (29). Isotherms
4
00 C/3 h calcinations of the slow and the fast-stirred ma-
terials are shifted toward the micropore limit and smoothly
distributed throughout the mesopore range (2–50 nm).
Maxima appear near 2, 5, and 15 nm for the slow, and near
�
for the 800 and 1000 C/3 h calcination products assume
2
, 7, and 30 nm for the fast stirred material. PVD for the
�
similar type and hysteresis loops. The 1000 C/10 h calci-
higher calcinations are smoother for the slow-stirred mate-
rial. Similar PVD was obtained for the fast-stirred material;
nevertheless, a new maximum emerged around 7 nm for the
nation product of the slow stirred xerogel (isotherm “d”
on Fig. 6) indicate a change in the isotherm type, namely
from type IV to type I, and the complete disappearance
of the hysteresis loop. A slight change in the isotherm
type (into type I) is shown for the fast-stirred materials
� �
1
000 C/3 h and 1000 C/10 h calcination products. Generally
speaking, porosity of any slow-stirred materials is less than
the corresponding fast-stirred one, as the area below the
distribution curves indicated.
Finally, N2 volume (ccm/g) in pores<1.8nm (micropores)
and in pores >200 nm for the different calcinations were
calculated (see Table 3) and represented in Fig. 11. The
figure indicates that reasonably high porosity (either <1.8
or >200 nm) was preserved at high calcinations, especially
for the group prepared from the fast-stirred material.
(
isotherm “d” on Fig. 7), and the hysteresis loop still some
what recognizable. These results may imply that the slow-
�
stirred material is more sintered at 1000 C/10 h than the
fast one.
The specific Surface area, SBET, and cBET values are
shown in Table 3. Figure 8 shows the change of SBET
as a function of calcination. The results indicate that the
�
4
00 C/3 h calcination products of the slow stirred material
has got higher surface area than the corresponding calcina-
tion products of the fast-stirred material. However, the lat-
ter maintains higher surface area upon calcination at higher
temperatures.
4
. DISCUSSION
The results presented above may be outlined as follows:
. Boehmite xerogels were produced at room tempera-
1
ture, as evidenced by XRD results. Slow-stirring rate re-
sulted in the formation of long fibrous (ca 200 nm long)
boehmite, whereas fast-stirring rate resulted in the forma-
tion of short fibrous (ca 50 nm long) boehmite.
TABLE 2
DSC Results for the Two Xerogels as the Onset, Maximum, and
Peak Area (1H ) for Each Start–Stop Range Indicated for the Dif-
ferent Dehydration Steps
2
. Calcination products of the long and the short fibrous
boehmite exhibited good structural and textural stability.
Precursor step Start ( C) Onset ( C) Max ( C) Stop ( C) 1H (J/g) However, the short fibrous boehmite phase stability leads
�
�
�
�
to better textural stability upon calcination at higher tem-
Slow
I
II
III
50
194
279
51
234
325
85
254
402
160
272
472
101.0
13.5
170.1
peratures. Weakly crystalline � -alumina commenced to ap-
�
�
pear at 600 C and developed at 800 C; phase transition into
�
weakly crystalline �-alumina occurs at 1000 C.
Fast
I
II
III
52
194
295
53
238
319
81
256
394
160
283
467
67.5
48.8
143.7
Thus, the following discussion will compare the present
results in relations to those of other published work.

2
02
KAMAL M. S. KHALIL
FIG. 4. TEM micrographs for the slow (top) and the fast (bottom) stirred xerogels, indicating long (ca 200 nm) and short (ca 50 nm) fibrous
morphology.
�
4
.1. The Formation of Fibrous Boehmite
at Room Temperature
duced by aging at 80 C. But, due to the high water/alkoxy
ratio employed, there appears to be little evidence that
under this aggressive hydrolysis conditions the oligomeric
According to Yoldas process (14, 15), which utilizes high structure of Al(OR)3 precursor is maintained. Thus it is
water/alkoxy ratio, boehmite, AlO(OH), can only be pro- not surprising that many similarities do exist between this

SHORT FIBROUS BOEHMITE
203
FIG. 5. XRD patterns of the short fibrous boehmite calcined for 30 h
at the indicated temperatures.
route and the aqueous one (from inorganic precursors)
(
30).
Assih et al. (31) postulated that a boehmite-like struc-
FIG. 7. N2 adsorption/desorption isotherms of alumina obtained by
the fast stirred xerogel calcination at the 400 C/3 h (a), 800 C/3 h (b),
� �
000 C/3 h (c), and 1000 C/10 h (d).
ture forms during Yoldas sol concentration. Furthermore,
they suggested that under these conditions hydrolysis may
occur without destroying the original oligomeric struc-
ture. Hydrolysis at the interface between a limited amount
of water/nonaqueous aluminum sol media (the present
conditions) led to partial hydrolysis of the aluminum tri-
isopropxide. The polar nature of the partially hydrolyzed
species will force them to accumulate and condense at the
�
�
1
interface. It isimportant that under these conditionshydrol-
ysis may occur without destroying the original oligomeric
structure due to the concentration of the hydrolyzed species
at or in the small droplets of water slowly stirred in the
nonpolar media. Thus, condensation between partially hy-
drolyzed species retaining their polymeric structure should
lead to fibrils structure. Increasing the stirring rate would
lead to further dispersion of water droplets, resulting in
smaller ones containing, of course, less content of the pre-
cursor material. Thus, finer particles will form upon con-
densation. Thus despite the oligomeric structure which
could be maintained (i.e., boehmite is produced), there
FIG. 6. N2 adsorption/desorption isotherms of alumina obtained by
�
�
the slow stirred xerogel calcination at the 400 C/3 h (a), 800 C/3 h (b),
FIG. 8. Specific surface area of the calcination products of the slow-
and the fast-stirred xerogels as a function of calcination.
�
�
1
000 C/3 h (c), and 1000 C/10 h (d).

2
04
KAMAL M. S. KHALIL
�
�
�
FIG. 9. Pore size distribution of alumina obtained with the slow-stirred xerogels calcination at the 400 C/3 h (a), 800 C/3 h (b), 1000 C/3 h (c), and
000 C/10 h (d).
�
1
�
�
�
FIG. 10. Pore size distribution of alumina obtained with the fast-stirred xerogels calcination at the 400 C/3 h (a), 800 C/3 h (b), 1000 C/3 h (c),
and 1000 C/10 h (d).
�

SHORT FIBROUS BOEHMITE
205
to remove half the oxygen from the layers, cause the col-
lapse and rearrangement ofoxygen into cubicclose-packing
(
33).
In lights of the XRD results obtained for both xerogels
and their calcination products for a period as long as 30 h,
the following structural transformation route may be assign
to occur as a function of temperature:
�
4
00 C/30 h
Boehmite ����→ amorphous � -alumina
�
8
00 C/30 h
�
���→ weakly crystalline � -alumina
�
000 C/30 h
1
�����→ � -alumina.
In terms of the surface area results the prepared xerogels
offer the formation of high surface area transitional alu-
minas upon calcination. However, the transformation of
boehmite to � -alumina and the subsequent transformation
�
to �-alumina at 1000 C are accompanied by coarsening of
the micro structure. This is evidenced by the decreasing val-
ues of SBET and increasing values of the average pore size
(
34). It has been reported that the � → � → �-Al2O3 trans-
formation sequence does not involve loss of water (32). But,
the present results reported some weight loss upon heating
up to 900 C, however small (see TG, Fig. 1). Moreover, TG
FIG. 11. Volume in pores <1.8 nm and >200 nm for the calcined prod-
ucts of the slow- and the fast-stirred xerogels as a function of calcination.
�
resultsshowed that the longfiberspreserve more water than
the short fibers at the end of each dehydration stage, prob-
ably due to the little folding morphology of the former (see
TEM micrographs). Pierre and Uhlmann (35) have shown
that large folding of a boehmite layers may result in the for-
mation of superamorphous gel (in presence of high amount
of water). Folding occurs due to stacking faults and leads to
intercalation of water, which causes swelling. However, in
the present work where a limited amount of water only was
available at the stage of boehmite formation. The observed
little folding may lead to a kind of local amorphous surface
formation within the folded fibers.
Thus, the present results agree well with the model pro-
posed bySoled (36), who proposed a dynamicchange model
to describe coalescence of� -alumina particles. Thisresulted
in a surface area decrease and coarsening of the micro struc-
ture (36). The dynamic model implies that during coalesce
via –OH groups, a terminaloxide ion isformed and becomes
is a slight chance for the formation of long fibrous struc-
tures.
Furthermore, a slow stirring rate slows down the partial
hydrolysis process, allowing for the condensation reaction
to take place. In contrast, fast stirring may result in fast ini-
tial hydrolysis, giving rise to more nuclei and consequently
smaller particles than the previous case.
Evidence for the above mechanism, which relates the
morphology difference to the initial droplet size, can be
drawn from the fact that the hydrolyzed gel was allowed
to age for three days before drying. Even so, two distinctly
different textures were obtained. Thus, the texture is deter-
mined essentially by the droplet size (which is related to the
stirring rate), and long aging could not cause the conversion
of the small particles to larger fibers.
4
.2. Thermal Stability of the Short Fibrous Boehmite
incorporated as the bulk anion bridging the two particles.
The question now is why the amount of such a dynamic
change is so small for the short fibrous particles, i.e. in-
creasing their thermal stability.
To answer the above question, let us first describe the de-
hydration process in terms of the phase formed, according
to Soled (36), formally and stoichiometrically as:
Results showed that the fibrous size of both xerogels ex-
ceeds the critical limit suggested by Murrell and Tauster
18), which promises the formation of stabilized alumina.
(
The dried short and long fibrous boehmite gels exhibit
weight loss over the range of the dehydration stages I,
�
II, and III (room, � 400 C) due to loss of physically ad-
sorbed, structural, and intercalated water (32). Thus they
are gradually transformed into � -alumina, in a topotactic
transformation that is accomplished by internal conden-
sation of protons and hydroxyls between boehmite layers
Al2O3 � H2O (boehmite) → Al2O3 � 1/5H2O (� -alumina)
→ �-Al2O3

2
06
KAMAL M. S. KHALIL
and
11. Machida, M., Eguchi, K., and Arai, H., Bull. Chem. Soc. Jpn. 61, 3659
1989).
2. Mizushima, Y., and Hori, M., J. Mater. Res. 9, 2272 (1994).
3. Boccaccini, A. R., Trusty, P. A., and Teller, R., Mater. Lett. 29, 171
(1996).
(
1
1
AlO(OH) (boehmite) → Al2.5ᮀ0.5O3.5(OH)0.5 (� -alumina)
→
�-Al2O3.
1
1
1
1
4. Yoldas, B. E., Amer. Ceram. Soc. Bull. 54, 286 (1975).
5. Yoldas, B. E., J. Mat. Sci. 10, 1856 (1975).
6. Pajonk, G. M., Rev. Chem. Appl. 24, 13 (C4) (1989).
7. Murrell, L. L., Dispenziere, N. C., Jr., and Kim, K. S., ACS Symp. Ser.
If the hydroxide ions populate only the surface and the
oxide ions comprise the bulk of � -alumina, the fraction of
anions on the surface is 1/8 and the rest is in the bulk (36).
Thus, one can speculate (for the range of size studied) that
anions exist in a smaller concentration on the surface of
the short fibers than that on the longer ones, since the sur-
face/mass ratio is larger for the former. Thus, if we consider
that surface OH anions are the driving force for the model
described. It follows that a weaker driving force for coales-
cence is present in the case of the short particles than for the
long ones, and consequently, slower sintering occurs upon
calcinations.
437, 97 (1990).
18. Murrell, L. L., and Tauster, S. J., in “Proc. 2nd Int. Congress on Cata-
lysis and Automotive Pollution Control, CAPoC, Brussels, Belgium,
1990,” p. 547.
1
2
2
2
9. Pajonk, G. M., Catal. Today 35, 319 (1997).
0. Mizushima, Y., Hori, M., and Saaki, M., J. Mater. Res. 8, 2109 (1993).
1. Mizushima, Y., and Hori, M., J. Non-Cryst. Solids 167, 1 (1994).
2. Horiuchi, T., Osaki, T., Sugiyama, R., Masuda, H., Horio, M., Suzuki,
K., Mori, T., and Sago, T., J. Chem. Soc. Farad. Trans. 90, 2573 (1994).
3. Ponthieu, E., Payen, E., and Grimblot, J., in “Sol-Gel Processing and
Applications” (Y. A. Attia, Ed.), p. 221. Plenum, New York, 1994.
4. Khalil, K. M. S., Baird, T., Zaki, M. I., and El-Samahy, A. A., “Electron
Microscopy,” Vol. 2, p. 421. Eurem 92, Granada, Spain, 1992.
2
2
REFERENCES
25. Khalil, K. M. S., Ph.D. thesis, Assiut Univ., Sohag, Egypt, 1994.
6. JCPDS, International Center for Diffraction Data, CD, 1996.
2
1
2
. Trimm, D. L., and Stanisluas, A., Appl. Catal. 21, 215 (1986).
. Nortier, P., Fourre, P., Mohammed Saad, A. B., Saur, O., and Lavalley,
J. C., Appl. Catal. 61, 141 (1990).
27. Brunauer, B., Emmett, P. H., and Teller, E., J. Am. Chem. Soc. 60, 309
(1938).
28. Refer to “Micromertics DFT Plus, Operator’s Manual.” V1.00, 1996.
29. International Union of Pure and Applied Chemistry, IUPAC, Pure &
Appl. Chem. 57, 603 (1985).
30. Brinker, C. J., and Scherer, G. W., in “Sol-Gel Science, The Physics and
Chemistry of Sol-Gel Processing,” p. 78. Academic Press, New York/
London, 1989.
3. Fukui, T., and Hori, M., J. Mater. Sci. 31, 3245 (1996).
4. Kawamura, K., and Endo, H., J. Ceram. Soc. Jpn. 104, 734 (1996).
5. Murrell, L. L., Catal. Today 35, 225 (1997).
6
. Utiyama, M., Hattori, H., and Tanabe, K., Bull. Chem. Soc. Jpn. 54,
2
521 (1981).
7
8
9
. Ishikawa, T., Ohashi, R., Nakabayashi, H., Kakuta, N., Ueno, A., and
Furuta, A., J. Catal. 134, 87 (1992).
. Jaworska-Galas, Z., Mista, W., Wrzyszcz, J., Zawadzki, M., and
Machnikowski, J., J. Catal. 145, 450 (1994).
. Segal, D., “Chemical Synthesis of Advanced Ceramic Materials.”
Cambridge Univ. Press, Cambridge, 1989.
31. Assih, T., Ayral, A., Abenoza, M., and Phalippou, J., J. Mat. Sci. 23,
3326 (1988).
32. Dwivedi, R. K., and Gowda, G., J. Mat. Sci. Lett. 4, 331 (1985).
33. Wilson, S. J., and Stacey, M. H., J. Coll. Int. Sci. 82, 507 (1981).
34. Yang, X., Pierre, A. C., and Uhlmann, D. R., J. Non-Cryst. Solids 100,
371 (1988).
1
0. Janosovite, U., Ziegler, G., Scharf, U., and Wokaun, A., J. Non-Cryst.
Solids 210, 1 (1997).
35. Pierre, A. C., and Uhlmann, D. R., J. Non-Cryst. Solids 82, 271 (1986).
36. Soled, S., J. Catal. 81, 252 (1983).