Effect of Boehmite crystallite size and steaming on alumina properties

Article abstract of DOI:10.1021/jp001024v

Boehmite and pseudoboehmite were studied in order to find the effect of their individual properties upon the properties of transitional alumina and ?±-Al₂O₃ produced thereof; the influence of steaming on these properties was also considered. Samples were characterized by X-ray powder diffraction, crystalline structure refinement, scanning electron microscopy, thermoanalysis, and nitrogen adsorption. Boehmite and pseudoboehmite had the same crystalline structure but different crystallite dimensions, Differences between lattice parameters (only 0.0052 nm along the b axis) are so small that the assumption of many authors that pseudoboehmite has intercalated water in its crystalline structure cannot be supported. Small crystallites with larger active areas, increased by crystal faces perpendicular to (020) planes, explain the larger amounts of desorbed water that pseudoboehmite exhibits compared to boehmite while drying. The hydrogen-bond length in boehmite depends on the crystallite size and determines the temperature where transformation into transitional alumina takes place. The boehmite crystallite size also defined pore size and distribution in transitional aluminas as well as their grain and crystallite size. Since the transformation of the transitional alumina into ?±-Al₂O₃ depends on hydroxyls diffusion, the corresponding temperature is higher for the transitional alumina with a large grain size. Steaming promoted crystal growth, because water molecules favored migration of aluminum atoms. Nitrogen adsorption-desorption isotherms of transitional aluminas derived from pseudoboehmite exhibited hysteresis loops corresponding to less ordered structures associated with swelling of platelets during desorption of nitrogen molecules. Consequences on the properties of transitional aluminas used as cracking catalysts were inferred. As texture and acidity evolve in opposite directions, there is a need for controlling them independently.

Full text of DOI:10.1021/jp001024v

J. Phys. Chem. B 2001, 105, 2099-2106
2099
Effect of Boehmite Crystallite Size and Steaming on Alumina Properties
M. L. Guzma´n-Castillo,†,* X. Bokhimi,†,‡ A. Toledo-Antonio,† J. Salmones-Bla´squez,† and
F. Herna´ndez-Beltra´n†
Instituto Mexicano del Petro´leo, Eje Central L. Ca´rdenas 152, C.P. 07730 Me´xico D.F., Me´xico, and Institute
of Physics, The National UniVersity of Mexico (UNAM), A. P. 20-364, C.P. 01000 Me´xico D.F., Me´xico
ReceiVed: March 17, 2000; In Final Form: September 16, 2000
Boehmite and pseudoboehmite were studied in order to find the effect of their individual properties upon the
properties of transitional alumina and R-Al2O3 produced thereof; the influence of steaming on these properties
was also considered. Samples were characterized by X-ray powder diffraction, crystalline structure refinement,
scanning electron microscopy, thermoanalysis, and nitrogen adsorption. Boehmite and pseudoboehmite had
the same crystalline structure but different crystallite dimensions. Differences between lattice parameters
(only 0.0052 nm along the b axis) are so small that the assumption of many authors that pseudoboehmite has
intercalated water in its crystalline structure cannot be supported. Small crystallites with larger active areas,
increased by crystal faces perpendicular to (020) planes, explain the larger amounts of desorbed water that
pseudoboehmite exhibits compared to boehmite while drying. The hydrogen-bond length in boehmite depends
on the crystallite size and determines the temperature where transformation into transitional alumina takes
place. The boehmite crystallite size also defined pore size and distribution in transitional aluminas as well as
their grain and crystallite size. Since the transformation of the transitional alumina into R-Al2O3 depends on
hydroxyls diffusion, the corresponding temperature is higher for the transitional alumina with a larger grain
size. Steaming promoted crystal growth, because water molecules favored migration of aluminum atoms.
Nitrogen adsorption-desorption isotherms of transitional aluminas derived from pseudoboehmite exhibited
hysteresis loops corresponding to less ordered structures associated with swelling of platelets during desorption
of nitrogen molecules. Consequences on the properties of transitional aluminas used as cracking catalysts
were inferred. As texture and acidity evolve in opposite directions, there is a need for controlling them
independently.
Introduction
yields that constrains unit operation. Therefore, besides bottoms
upgrading and relative low coke and gas production capabilities,
Transitional aluminas are widely used as catalysts and catalyst
supports in several processes related to petroleum refining and
the petrochemical and chemical industries.1-4 The extended
usage of transitional aluminas in catalysis is explained by their
particular properties that combine high specific areas, high
thermal stability, and mechanical strength and by the fact that
their acid-base properties and porosity may be varied by a
relatively large extent.
Application of transitional aluminas is of special interest in
fluid catalytic cracking (FCC),5-13 which is one of the most
important oil refining processes worldwide. The active com-
ponent of a fluid cracking catalyst is a faujasite “Y” zeolite
supported on a matrix that can have different compositions. The
fluid cracking catalysts which contain transitional aluminas
exhibit higher matrix activity compared to catalysts containing
untreated clay or large amounts of silica.
One of the current approaches in FCC heavy and reside feeds
processing is to design catalysts with moderate zeolite contents
and a significant nonzeolitic large pore activity.6 Both, meso-
and macropores coupled to a relatively high acidity are necessary
for efficient cracking of heavy feeds.11 Those types of feeds,
however, are highly contaminated with metals (nickel and
vanadium are the most abundant), which disperse on the matrix
and act as dehydrogenating centers, increasing gas and coke
alumina-based matrixes should also feature high metals tolerance
in order to minimize the deleterious contaminant selectivity.
EXAFS and fixed fluid bed testing studies carried out on nickel-
contaminated FCC catalysts suggest that matrices promoting
formation of nickel aluminate exhibit superior metals tolerance.12
Since high temperatures and steam atmospheres are prevailing
conditions in the operating cycles in FCC units, hydrothermal
stability is also needed for a robust alumina matrix, specially
in reside cracking operations.13
The properties that govern the performance of aluminas as
cracking catalyst are interdependent and evolve with opposite
trends. It is well-known that the higher the specific surface, the
higher the acidity, the lesser the stability, and the lesser the
pore size. Therefore it becomes of paramount importance to
have a better understanding of the origin of changes in properties
of alumina induced by thermal and hydrothermal conditions and
how the properties of precursors are concerned. Although a great
number of studies have been published through the years on
these subjects, a full picture of the atomic structure of boehmites
is still missing. A better comprehension of the boehmite structure
and the mechanism that controls its transformation through
transitional alumina will help to control catalytic cracking
properties.
This work dealt with the study of structural aspects of
aluminas and their precursors. Fine details of the structure of
boehmite and pseudoboehmite were studied with the aim of
† Instituto Mexicano del Petro´leo.
‡ The National University of Mexico.
10.1021/jp001024v CCC: $20.00 © 2001 American Chemical Society
Published on Web 02/22/2001

2100 J. Phys. Chem. B, Vol. 105, No. 11, 2001
Guzma´n-Castillo et al.
understanding differences between the transitional aluminas
produced thereof. The interdependency among porosity, crystal
size, and structure at the atomic level was addressed. The effects
of steaming upon these properties were also studied.
Experimental Section
Sample Preparation. Powdered samples of aluminum oxy-
hydroxide, namely pseudoboehmite (PSB) and boehmite (BOE),
were studied. BOE was kindly provided by Grace-Davison,
whereas PSB was obtained from Condea (Catapal). Samples
were annealed at 788 °C during 4 h in static air (PSB-C, and
BOE-C) or steamed (PSB-S, and BOE-S). Steaming took place
with 100% steam using 20 g of powdered material in a vertical
quartz tubular reactor where distilled liquid water was vaporized
and entered from the bottom at 0.2 mL min-1
.
Characterization Techniques. X-ray diffraction measure-
ments were performed at room temperature with Cu KR
radiation in a Bruker Advance D-8 diffractometer having θ-θ
configuration and a graphite secondary-beam monochromator.
Sample powders were packed in a glass holder. Diffraction
intensity was measured by step scanning in the 2θ range between
10 and 110°, with a 2θ step of 0.02° for 8 s per point. Crystalline
structures were refined with the Rietveld technique by using
DBWS-941114 and FULLPROF-V3.5d15 programs; peak profiles
were modeled with a pseudo-Voigt function16 that depends on
average crystallite size.17 Standard deviations, which show the
last figure variations of a number, are given in parentheses.
When these deviations correspond to Rietveld refined param-
eters, their values are not estimates of the probable error in the
analysis as a whole but only of minimum possible probable
errors based on their normal distribution.18
Figure 1. X-ray diffraction patterns of the samples derived from the
boehmite with a crystallite thickness of 26.1(5) nm. The tick marks
under the pattern of BOE-C and BOE correspond to θ-alumina and
boehmite, respectively.
Nitrogen adsorption isotherms were obtained at -196 °C with
a Micromeritics ASAP-2100 apparatus. The specific area of
solids was calculated by the multipoint BET method. The pore
size distribution and shape are closely related to the hysteresis
loop classified by IUAPAC,19 BET,20 or BDDT.21 The analysis
of the adsorption-desorption isotherms was complemented
following criteria on the inspection of the hysteresis loop shape
based on the so-called descending boundary curve (DBC).22
Weight loss and temperatures associated with phase trans-
formation were determined by thermogravimetry (TG) and
differential thermal analysis (DTA) using a Perkin-Elmer TG-7
and a Perkin-Elmer 1700 apparatus, respectively. The sample
(10 mg) was heated in flowing dried air (20 mL min-1) from
room temperature up to 1000 °C at 10 °C min-1; DTA was
carried out using platinum crucibles and R-alumina as reference.
Micrographs of fresh, calcined, and steamed samples were
obtained using a JEOL 35 CF scanning electron microscope
(SEM).
Figure 2. X-ray diffraction patterns of the samples derived from the
boehmite with a crystallite of 3.42(4) nm. The tick marks under the
pattern of PSB-C and PSB correspond to θ-alumina and boehmite,
respectively.
difference between boehmite and pseudoboehmite cannot be
produced by molecules of water intercalated in (020) planes,
as has been assumed in the literature.23,24
Results and Discussion
Crystallography and Crystallite Morphology. X-ray dif-
fraction patterns of fresh, calcined, and steamed samples are
presented in Figures 1 and 2. Boehmite and pseudoboehmite
samples contained only one phase (Figures 3 and 4), which was
refined with an orthorhombic unit cell having the symmetry
described by the space group Cmcm and atoms in the positions
given in Tables 1 and 2. Atom positions for BOE and PSB were
very similar, differing only in the third decimal digit. The lattice
parameters in both samples were also similar (Table 3). Only a
difference of 0.0052 nm was found along the b axis. The most
open planes of the structure, which correspond to the (020)
reflection (Figure 1), had a difference of only 0.0026 nm, in
which no atom can be intercalated. This means that the
Crystallite morphology in both boehmite samples was aniso-
tropic. Crystallites were thin platelets with their shortest
dimension along the b axis. This anisotropy produced diffraction
peak widths that depended on Miller indices. Since the thinnest
part of the crystallite was along the b axis, the broadest
diffraction peaks were those produced by the planes perpen-
dicular to this axis, i.e., those corresponding to (0k0) reflections.
FULLPROF code15 can model this effect, allowing us to refine
correctly the crystalline structure. From refinements a crystallite
thickness of 26.1(5) nm for BOE and 3.32(4) nm for PSB (Table
3) was obtained.
Figures 3 and 4 show their corresponding Rietveld refinement
plots. Although experimental data fit was good, we tried to

Boehmite Crystallite Size and Steaming
J. Phys. Chem. B, Vol. 105, No. 11, 2001 2101
TABLE 3: Lattice Parameters and Average Crystallite
Thickness
sample
a (nm)
b (nm)
c (nm)
d(020) (nm)
BOE
PSB
0.28683(1)
0.28679(5)
1.2221(6)
1.2273(5)
0.36958(1)
0.37021(7)
26.1(5)
3.32(4)
TABLE 4: Lattice Parameters and Average Crystallite Size
of the Transitional Aluminas produced from boehmite
sample
a (nm)
b (nm)
c (nm)
â (deg)
d (nm)
BOE-C 1.215(2)
0.2799(4) 0.5576(8) 102.9(1)
6.6(2)
BOE-S 1.2030(8) 0.2817(2) 0.5574(4) 103.21(6) 7.0(2)
PSB-C 1.218(1) 0.2789(3) 0.5547(5) 102.99(7) 3.52(6)
PSB-S 1.1979(6) 0.2852(1) 0.5575(3) 103.14(6) 6.03(8)
Therefore, microstrain effects were excluded from the refine-
ment of boehmite’s crystalline structure.
The above results suggest that, from the crystallographic point
of view, the only notable difference between boehmite and
pseudoboehmite is the crystallite size. For that reason, instead
of referring to boehmite and pseudoboehmite as two different
phases, from here on we will refer to them as boehmite, but
with different crystallite size which reflects the real properties
of this system (sample designations were kept for clearness).
This conclusion will certainly conflict with the accepted and
traditional way of distinguishing these materials; however, the
evidence obtained in the present work leads us to conclude that
they correspond to the same crystalline phase.
Figure 3. Rietveld refinement plot of BOE (Rwp ) 0.14). Experi-
mental data are indicated with cross marks, and the calculated one and
the difference between them, with continuous lines. Tick marks
correspond to boehmite (RBragg ) 0.045).
As will be evident, the transitional alumina properties depend
on the crystallite size of the precursor to a great extent.
Therefore, for practical meaning, it becomes of great interest
to characterize it in a relatively simple way. Boehmite crystallites
are thin plates whose smallest dimension is along the b axis.
By consequence, the (020) X-ray reflectionsthe one at the
lowest 2θ angle (Figure 1) that is strong and not superimposed
with other reflections (Figures 3 and 4)scan be used to
characterize it. The full width at half-maximum (fwhm) of this
peak is related to the average crystallite thickness and thus might
be used to characterize this dimension. This will only require
the measurement of the diffraction pattern in the 2θ region
between 9 and 19°, which might be easily implemented from a
practical standpoint.
Figure 4. Rietveld refinement plot of PSB (Rwp ) 0.13). Experimental
data are indicated with cross marks, and the calculated one and the
difference between them, with continuous lines. Tick marks correspond
to boehmite (RBragg ) 0.031).
The crystalline structure of the transitional aluminas in
samples BOE-C, BOE-S, PSB-C, and PSB-S was modeled with
the monoclinic unit cell reported for θ-alumina,26 which is
derived from the spinel structure. The atom positions in the
monoclinic unit cell have the symmetry described by the space
group C2/m. The residue obtained after the refinement with the
above unit cell was 10% smaller than with the nondeformed
cubic unit cell having a spinel structure. Table 4 gives the lattice
cell parameters and the average crystallite sizes obtained after
refinement.
TABLE 1: Boehmite, Space Group Cmcm: Atom Fractional
Coordinatesa
atom
site
x
y
z
Al
O
OH
4c
4c
4c
0.0
0.0
0.0
yAl
yO
yOH
0.25
0.25
0.25
a Since hydrogen X-ray diffraction is negligible, it was not consid-
ered. OH is the atom position of the oxygen atoms associated with
hydroxyls.
Aluminas derived from BOE had larger crystallites than those
derived from PSB (Table 4). Differences were more pronounced
when alumina was prepared in air (Table 4). For example, the
alumina derived from the boehmite (BOE) with the thick
crystallite (Figure 1), 26.1(5) nm, had an average crystallite size
of 6.6(2) nm, while the one derived from the boehmite (PSB)
with thin crystallites (Figure 2), 3.32(4) nm, had an average
crystallite size of 3.52(6) nm.
Boehmite’s crystallite dimensions also had an effect upon
the transformation rate of the corresponding transitional alumina
into the nonhydroxylated R-alumina (Figure 5). For instance,
when transitional aluminas were annealed for 20 h at 1100 °C,
samples derived from PSB were completely transformed into
R-alumina, whereas those derived from BOE transformed only
TABLE 2: Atom Fractional Coordinates y for the Different
Atoms
sample
yAl
yO
yOH
BOE
PSB
-0.3180(2)
-0.3199(3)
0.2907(2)
0.2979(4)
0.0784(2)
0.0779(3)
improve it by modeling microstrains. FULLPROF code15 has
several models for this microstrain, some of them related to
crystalline unit cells with orthorhombic symmetry,25 which
corresponds to the symmetry actually observed in our samples.
None of these models, however, gave rise to a better fit of the
experimental data; on the contrary, when they were applied,
the temperature factors were negative, which is meaningless.

2102 J. Phys. Chem. B, Vol. 105, No. 11, 2001
Guzma´n-Castillo et al.
Figure 6. TG curves of BOE and PSB.
TABLE 5: Weight Loss Stages (weight %) by TGA
sample
stage 1
stage 2
stage 3
BOE
PSB
1.281
10.63
14.44
14.76
1.42
1.45
Figure 5. X-ray diffraction patterns of BOE and PSB after calcining
at 1100 °C: (A) PSB calcined for 20 h, (B) BOE calcined for 20 h,
(C) BOE calcined for 69 h. The tick marks under the boehmite pattern
correspond to the monoclinic phase; those under the pattern of the
pseudoboehmite correspond to R-alumina.
ature up to 90 °C for BOE and up to 120 °C for PSB (Figure
6 and Table 5). This stage corresponded to the desorption of
water molecules adsorbed on crystallites surface. The weight
loss in PSB was comparatively higher due to its smaller
crystallite size and thus higher surface areas available for water
adsorption.
SCHEME 1: Pseudomorphic Evolution of Boehmite
Crystallites into Those of Transitional Alumina and
These into r-Al2O3
The above argument seems obvious but can be detailed further
from structural considerations. Boehmite crystal structure has
double layerssparallel to the (020) planessof octahedra sharing
edges along the a axis and vertexes along the c axis, with an
aluminum atom near their center and two hydroxyls and four
oxygen atoms in their vertexes.29 The interaction within a double
layer is strong, but between double layers it is weak (via
hydrogen bonds). Therefore, a boehmite crystallite ends in the
interface between two double layers, thus producing surfaces
that are full of hydroxyls. The crystallite surfaces perpendicular
to double layers have oxygen atoms in low coordination. These
oxygen atoms can easily react with hydroxyls and hydrogen
ions from their environment, thus allowing these surfaces to
get fully covered with hydroxyls. For small crystallite size, the
number of these low coordination oxygen sites increases, thus
leading more water molecules to be adsorbed compared to larger
crystallite sizes.
partially into R-alumina. Transformation was completed only
after 69 h at this temperature.
The observed crystallite size dependence of transitional
aluminas transformation (into R-alumina) can be explained in
A second weight loss stage showed by TG analysis was found
between 90 and 550 °C for BOE and between 120 and 460 °C
for PSB. This stage corresponded to the transition, via an
endothermic reaction (Figure 7), into the corresponding tran-
sitional alumina. The third weight loss stage, which was not so
clearly defined as the two precedent, corresponded with a
continuous and gradual loss of residual hydroxyls, which is
characteristic of transitional aluminas.23,30
The temperatures associated with transformation into tran-
sitional alumina depended on boehmite’s crystallite dimensions
(Tables 3 and 6). According to DTA, this temperature was 448
°C for boehmite (PSB) with small crystallites and 508 °C for
boehmite (BOE) with large crystallites. This difference is related
to the hydrogen-bond strength between double layers (Table
6).31 Although, for the present study the difference between
hydrogen-bond lengths was only 0.0012 nm, a systematic study29
showed that atom bonds in boehmite depend on the crystallite
size and that the transition into γ-alumina is related to the
hydrogen bond.31
terms of the pseudomorphic character of the process.27,28
A
boehmite crystallite is transformed into several alumina crys-
tallites, forming a single grain with the dimensions of the
original boehmite crystallite (Scheme 1). Therefore, BOE-C
exhibited a larger crystal grain compared to PSB-C. Transfor-
mation of transitional alumina into R-alumina is determined by
the dehydroxylation process where hydroxyl diffusion is a
controlling step below 1100 °C. Since the path hydroxyls should
go through is shorter for small grains (Scheme 1), the complete
transformation of these grains into R-alumina occurs faster than
for large grains, where hydroxyls takes longer to leave (Figure
5C). This explains the fact that DTA or any other heating
treatment at fixed heating rate or fixed temperature show that
transitional aluminas with large crystallites require either high
temperatures or long times before a complete transformation
into R-Al2O3 (Figure 5B,C) takes place.23
Water Desorption and Phase Transformation. Boehmite
samples exhibited three weight loss stages as a function of
temperature. The first stage was observed from room temper-
Crystal Aggregation. When transitional aluminas were
subjected to steaming (samples BOE-S and PSB-S), they

Boehmite Crystallite Size and Steaming
J. Phys. Chem. B, Vol. 105, No. 11, 2001 2103
Figure 7. DTA curves of BOE and PSB.
TABLE 6: Transition Temperature of Boehmite into
γ-alumina and Some Characteristic Atom Bond Lengthsa
dOH···O
(nm)
dAl-OH
(nm)
dAl-O
(nm)a
dAl-O
(nm)c
sample T (°C)
BOE
PSB
448
508
0.2678(1) 0.1907(1) 0.1952(1) 0.1878(1)
0.2800(1) 0.1905(1) 0.1990(1) 0.1879(1)
a Hydrogen-bond length ) dO-H···O. Bond lengths of the interaction
of aluminum with oxygen atoms (which determine unit cell dimension
along lattices parameters) ) dAl-O along the a axis, dAl-OH along the
b axis, and dAl-O along the c axisa.
c
produced samples with larger crystallites (Table 4). It seems
that water molecules favored aluminum transport for crystal
growth, which explains the absence of small crystal aggregates
in the steamed samples (Figures 8A and 9A). BOE samples
were homogeneously formed by spherical crystal aggregates
with diameters between 5 and 50 µm (Figure 10A) and
composed by oriented platelet-like crystals. In contrast, PSB
samples did not show a specific grain form (Figure 10B). The
grain morphology in BOE did not change upon transformation
into transitional alumina (Figure 8), which agrees with the
pseudomorphic character of this process, as has been already
discussed. A similar but more pronounced effect was observed
for PSB (Figure 9). Under steaming conditions, the initial
irregular grain aggregates were transformed into nonspherical
grains with similar dimensions. Thus the effect of annealing in
air was only to disaggregate the initial boehmite grains (Figures
8B and 9B).
Pore Size Distribution. Analysis of the hysteresis loop shape
by using the descending boundary curve (DBC)20-22 technique
provides information about pore blocking and pore structure
rigidity in alumina samples. The hysteresis loop shape of
aluminas derived from BOE (Figure 11) corresponds to struc-
tures of medium strength without pore blockage.22 Therefore,
steaming had only a weak effect on pore size distribution (Figure
12). Aluminas from BOE had mesopores with diameters
between 10 and 100 nm, though diameters between 50 and 100
nm dominated (Figure 12).
The initial slopping character on the descending boundary
curves observed for PSB-C and PSB-S (Figure 13) reflects the
porous structure shrinkage caused by stress change of the
condensate during depressurization. This is due to swelling of
parallel grain plates by adsorbed nitrogen,22 because the
structures in these samples were less rigid than those found for
aluminas derived from BOE. The plateau of the descending
boundary curve of PSB-C and PSB-S indicates an appreciable
pore blockage. Moreover, the hysteresis loop type would suggest
that a mixture between ink-bottle and interstitial pores between
aggregates according to De Boer and Lippens32 classification
Figure 8. Scanning electron micrographs of annealed boehmite: (A)
BOE-S and (B) BOE-C.
is present. The small pore size in these samples (Figure 14)
gave rise to high specific areas and pore volumes (Table 7);
the centroid of their distributions moved from 5 nm for PSB-C
to around 11 nm when PSB was annealed under steaming
conditions (Figure 14).
Although BOE-C and PSB-S had similar crystallite sizes, their
adsorption loop hysteresis were very different (Figures 11 and
12). This shows that porosity is not determined by the alumina
crystallite size but by its grain size. According to Scheme 1,
alumina grains would be composed by a number of crystallites;
this number decreases as the precursor crystallite size dimin-
ishes. Therefore, the grain size of transitional alumina derived
from BOE would be larger than the one derived from PSB. A
wider crystallite size distribution of PSB compared to BOE
would create higher potential for crystal growth,30 thus produc-
ing poorly ordered and less rigid structures possessing less
strength.
Aluminas derived from BOE had smaller BET surface areas
than those derived from PSB. This would be explained by the
fact that crystallites in BOE were larger compared to crystallites
in PSB. In both systems, however, steaming led to a decrease
in surface area that is accompained by crystal growth (Table
7). The surface area was comparatively less affected in BOE-
S, thus indicating a more stable phase that is related to a larger
grain size.
Implications on the Properties of Transitional Aluminas
as Potential Active Matrixes for Cracking Catalysts. Upon

2104 J. Phys. Chem. B, Vol. 105, No. 11, 2001
Guzma´n-Castillo et al.
Figure 9. Scanning electron micrographs of annealed pseudoboeh-
mite: (A) PSB-S and (B) PSB-C.
Figure 10. Precursors scanning electron micrographs: (A) BOE and
(B) PSB.
annealing, small crystallite boehmite produced transitional
alumina possessing a less ordered structure and larger surface
areas with smaller and irregular pores. Less hydrothermal
stability, in terms of surface area loss and crystal growth, was
also found. While this trend confirms what is already known,
fundamental aspects based on structural considerations that
explain such properties have been found in this work. The
structural analysis of boehmite suggests that the surface of these
solids will present different type of sites and that their relative
population will depend on the crystallite size. From the bonds
between aluminum and oxygen atoms (Table 6), it is observed
that the bond lengths determining unit cell dimensions are
different (Table 3), thus giving rise to different sites on the
crystal faces perpendicular to the [100], [010], and [001]
directions. Faces perpendicular to the [010] direction will
contribute the most to surface properties in large crystallites,
while faces perpendicular to [100] and [001] directions will have
a considerable contribution in small crystallites. Transformation
to transitional alumina is a pseudomorphic process, and
therefore, these relative contributions will prevail in the annealed
samples. On the other hand, as texture will depend on the
crystallite ordering, less ordered surfaces will be obtained with
materials possessing small crystallites. Implications from the
catalytic standpoint concern more heterogeneous pore sizes and
geometries found in aluminas derived from PSB.
Figure 11. Nitrogen adsorption-desorption isotherms of BOE-C and
BOE-S.
molecules, but pore blockage would also produce less coke
tolerance. In contrast, BOE aluminas have fewer acid sites which
is a required property for cracking activity. It follows that a
control of porosity is probably needed in the case of PSB derived
aluminas, while increasing the number of active sites will
probably improve the cracking activity of BOE aluminas.
Not only is the smaller pore size found in PSB derived
aluminas less desirable for the cracking of large hydrocarbon

Boehmite Crystallite Size and Steaming
J. Phys. Chem. B, Vol. 105, No. 11, 2001 2105
eters and atom positions in the unit cell were similar. However,
differences in crystallite size were determined, with pseudo-
boehmite having shown smaller crystallite size. The largest
structural difference was observed for the b axis magnitude.
The calculated value of 0.0052 nm is not enough to assume
that pseudoboehmite would have intercalated water in its
crystalline structure. The higher water content always observed
for pseudoboehmite compared to boehmite during drying should
be ascribed to crystallites with larger active areas and associated
with water molecules adsorbed in faces perpendicular to (020)
planes. The transition temperature of boehmite into transitional
alumina depended on the hydrogen bond between boehmite
double-octahedra layers. Boehmite crystallite dimensions also
defined pore size and distribution of the transitional alumina as
well as its crystallite size. Since the transformation of transitional
alumina into R-Al2O3 depends on the diffusion of hydroxyls,
the temperature associated with structural transformation is
higher for transitional alumina with a larger grain size. Anneal-
ing under steaming conditions hindered formation of small
crystallite agglomerates and gave rise to transitional aluminas
with larger crystallites, because it favored transporting of
aluminum atoms; this effect was larger for pseudoboehmite.
Nitrogen adsorption-desorption isotherms showed that in the
transitional alumina derived from pseudoboehmite, platelet-like
grains were less ordered with water intercalated between the
platelets.
Figure 12. Pore size distribution of BOE-C and BOE-S.
Consequences on the properties of transitional aluminas used
as cracking catalysts were inferred. As texture and acidity evolve
in opposite directions, there is a need for controlling them
independently. A deeper comprehension on the origin of these
properties will certainly help to reach this goal.
Acknowledgment. We would like to thank Mr. A. Morales
and Mr. M. Aguilar for technical assistance; we also tank Mr.
B. Handy for his help in reviewing the manuscript. This work
was financially supported by IMP projects D.01024 and D.01234
Figure 13. Nitrogen adsorption-desorption isotherms of PSB-C and
PSB-S.
References and Notes
(1) Ismagilov, Z. R.; Shkrabina, R. A.; Koryabkina, N. A. Catal. Today
1999, 47, 51.
(2) Hietala, J.; Root, A. R.; Knuutila, P. J. Catal. 1994, 150, 46.
(3) Barath, F.; Turki, M.; Keller V.; Maire, G. J. Catal. 1999 185, 1.
(4) Grange, P. Catal. ReV.- Sci.- Eng. 1980, 21, 135.
(5) Lerner, B. A. US Patent 5,559,067, 1996, assignee Engelhard Corp.
(6) O’Connor, P.; Verlaan, J. P. J.; Yanik, S. J. Catal. Today 1998,
43, 305.
(7) Cheng, W.-Ch.; Rajagopalan, K. US Patent 5,547564, 1996,
assignee W. R. Grace & Co.-Conn.
(8) Lussier, R. J. US Patent 4,894,143, 1990, assignee W. R. Grace &
Co.-Conn.
(9) Upson, L.; Van de Gender, P. J.; Van Dijk, W. US Patent 5,173,-
174, 1992, assignee UOP
(10) Otterstedt, J. E.; Gevert, S. B.; Jaras, S. G.; Menon, P. G. Appl.
Catal. 1988, 38, 143.
(11) O’Connor, P.; Gevers, A. W.; Humpries, A.; Gerrisen, L. A.; Desai,
P. H. In Fluid Catalytic Cracking II: Concepts in Catalyst Design; Occelli,
M., Ed.; ACS Symposium Ser. 452; American Chemical Society: Wash-
ington, DC, 1991; p 318.
(12) Woolery, G. L.; Farnos, M. D.; Quinones, A. R.; Chin, A. In Fluid
Cracking Catalysts; Occelli, M. L., O’Connor, P., Eds.; Marcel Dekker:
New York, 1998; Vol. 74, p 48.
Figure 14. Pore size distribution of PSB-C and PSB-S.
TABLE 7: Textural Properties of the Transitional Aluminas
specific area
(m2/g)
pore volume
(cc/g)
average pore
size (nm)
sample
(13) Alearasool, S.; Doolin, P. K.; Hoffman, J. F. In Fluid Cracking
Catalysts; Occelli, M. L., O’Connor, P., Eds.; Marcel Dekker: New York,
1998; Vol. 74, p 96.
(14) Young, R. A.; Sakthivel, A.; Moss, T. S.; Paiva-Santos, C. O. J.
Appl. Crystallogr. 1995, 28, 366.
BOE-C
BOE-S
PSB-C
PSB-S
69.2
48.7
218.4
124.0
0.32
0.25
0.52
0.51
40.9
30.4
6.6
12.1
(15) Rodr´ıguez-Carbajal, J. Laboratoire Leon Brillouin (CEA-CNRS),
France. Tel: (33) 1 6908 3343. Fax: (33) 1 6908 8261. E-mail:
juan@llb.saclay.cea.fr.
(16) Thompson, P.; Cox, D. E.; Hasting, J. B. J. Appl. Crystallogr. 1987,
20, 79.
Conclusions
Using X-ray analysis and refinement techniques of the
crystalline structures of boehmite and pseudoboehmite samples,
the same crystalline structure was observed. The lattice param-
(17) Young R. A.; Desai, P. Arch. Nauki Mater. 1989, 10, 71.

2106 J. Phys. Chem. B, Vol. 105, No. 11, 2001
Guzma´n-Castillo et al.
(18) Prince, E. J. Appl. Crystallogr. 1981, 14, 157
(19) IUPAC, Manual of Symbols and Terminology for Physicochemical
Quantities and Units; Appendix II, Definitions, terminology and symbols
in colloid and surface chemistry; Pure Appl. Chem. 1972, 31, 578.
(20) Brunauer, S.; Emmett, P. H.; Teller, E. J. Am. Chem. Soc. 1938,
60, 309.
(21) Brunauer, S.; Demming, L. S.; Demming, W. E.; Teller, E. J. Am.
Chem. Soc. 1940, 62, 1723.
(22) Salmones, J.; Garciafigueroa E.; Mayagoitia, V.; Rojas, F.; Korn-
hauser I. Adsorption Science and Technology; Multi-Sc Pub. Co. LTD.:
UK, 1997; 15 N 9, p 661.
(25) Rodr´ıguez-Carvajal, J.; Ferna´ndez, M. T.; Mart´ınez, J. L. J. Phys:
Condens. Matter. 1991, 3, 3215.
(26) Husson, E.; Repelin, Y. Eur. J. Solid State Inor. Chem. 1996, 33,
1223.
(27) Lippens, B. C.; de Boer, J. H. Acta Crystallogr. 1964, 17,
1312.
(28) Tikhov, S. F.; Fenelonov, V. B.; Zaikovskii, V. I.; Potapova, Yu.
V.; Sadykov, V. A. Microporous Mesoporous Mater. 1999, 33, 137.
(29) Bokhimi, X.; Toledo-Antonio, J. A.; Guzma´n-Castillo, M. L.;
Herna´ndez-Beltra´n, F. J. Solid State Chem., accepted.
(30) Mamchick, A. I.; Kalinin, S. V.; Vertegel, A. A. Chem. Mater.
1998, 10, 3548.
(31) Toledo-Antonio, J. A.; Bokhimi, X.; Guzma´n-Castillo, M. L.; Mar-
Mar, B.; Herna´ndez-Beltra´n, F. Chem. Mater. Submitted.
(32) Lippens, B. C.; Linsen, B. G.; De Boer, J. H. J. Catal. 1964, 3, 32.
(23) Tsukada, T.; Segawa, H.; Yasumori, A.; Okada, K. J. Mater. Chem.
1999, 9, 549.
(24) Ismagilov, Z. R.; Shkrabina, R. A.; Koryabkina, N. A. Catal. Today
1999, 47, 51.