Influence of synthesis methods on tungsten dispersion, structural deformation, and surface acidity in binary WO3-ZrO2 system

Article abstract of DOI:10.1021/jp053255d

WO₃-ZrO₂ catalysts were synthesized by precipitating the aqueous solutions of zirconium oxynitrate and ammonium metatungstate with ammonium hydroxide. The white slurry precipitate was treated under three different conditions. In the as-made materials, the amorphous phase was formed in the aged and refluxed samples, while well-crystallized tetragonal and monoclinic phases were obtained in the hydrothermally treated sample. The real amount of tungsten loaded in the samples was similar for the three samples, independently of the treatments; however, the tungsten surface atomic density in the annealed WO₃-ZrO₂ samples varied between 6 and 9 W atoms/nm². Two different contrast types of aggregates were determined by scanning electron microscopy, the white particles which are rich in W, and the gray ones which are rich in zirconium; both of them were formed in the calcined solids prepared under aging or reflux condition. A very high dispersion of tungsten species on the zirconia surface was achieved in the hydrothermally treated sample. The degree of the interaction between WO_x and ZrO₂ surface strongly modified the Zr-O bond lengths and bond angles in the structure of tetragonal zirconia as proved by X-ray diffraction analysis and the Rietveld refinement. The catalyst obtained under hydrothermal condition exhibited the highest dispersion of tungsten species in the zirconia, which in turn causes strong structural deformation of the tetragonal ZrO₂ phase responsible of the strongest surface acidity and, consequently, the optimum catalytic activity for n-hexane isomerization.

Full text of DOI:10.1021/jp053255d

22730
J. Phys. Chem. B 2005, 109, 22730-22739
Influence of Synthesis Methods on Tungsten Dispersion, Structural Deformation, and
Surface Acidity in Binary WO3-ZrO2 System
M. A. Corte´s-Ja´come,* J. A. Toledo, and C. Angeles-Chavez
Instituto Mexicano del Petro´leo. Prog. Ingenier´ıa Molecular, Eje Central L. Ca´rdenas # 152,
07730 Me´xico D. F., Me´xico
M. Aguilar
UniVersidad Nacional Auto´noma de Me´xico. Instituto de F´ısica, AP 20-364, 01000 Mexico D. F., Mexico
J. A. Wang
Laboratorio de Cata´lisis y Materiales, SEPI-ESIQIE, Instituto Polite´cnico Nacional, Col. Zacatenco,
07738 Mexico D. F., Mexico
ReceiVed: June 16, 2005; In Final Form: September 2, 2005
WO3-ZrO2 catalysts were synthesized by precipitating the aqueous solutions of zirconium oxynitrate and
ammonium metatungstate with ammonium hydroxide. The white slurry precipitate was treated under three
different conditions. In the as-made materials, the amorphous phase was formed in the aged and refluxed
samples, while well-crystallized tetragonal and monoclinic phases were obtained in the hydrothermally treated
sample. The real amount of tungsten loaded in the samples was similar for the three samples, independently
of the treatments; however, the tungsten surface atomic density in the annealed WO3-ZrO2 samples varied
between 6 and 9 W atoms/nm2. Two different contrast types of aggregates were determined by scanning
electron microscopy, the white particles which are rich in W, and the gray ones which are rich in zirconium;
both of them were formed in the calcined solids prepared under aging or reflux condition. A very high dispersion
of tungsten species on the zirconia surface was achieved in the hydrothermally treated sample. The degree of
the interaction between WOx and ZrO2 surface strongly modified the Zr-O bond lengths and bond angles in
the structure of tetragonal zirconia as proved by X-ray diffraction analysis and the Rietveld refinement. The
catalyst obtained under hydrothermal condition exhibited the highest dispersion of tungsten species in the
zirconia, which in turn causes strong structural deformation of the tetragonal ZrO2 phase responsible of the
strongest surface acidity and, consequently, the optimum catalytic activity for n-hexane isomerization.
1. Introduction
has been extensively studied, less attention has been paid on
the defective characteristics and distorted crystalline structure
of the WOx species doped zirconia nanophases, which are
usually acted as catalytic active sites involving in many catalysis
reactions.15,16
Solid acid catalysts play an important role in hydrocarbon
conversion reactions in the chemical and petroleum industries.1,2
In such cases, zirconia is frequently used as catalyst or support.
It is well known that zirconia has three stable crystalline
phases: tetragonal, monoclinic, and cubic zirconia, and their
concentration and transformation between each other depend
on zirconia doping and/or thermal treatment. The zirconia doped
by a variety of compounds such as sulfate,3 phosphate, and
heteropolyacids4 creates additional electron-deficient regions that
may generate new acid sites and increase the strength of
Bro¨nsted acidity responsible for an enhanced catalytic activity
in alkanes isomerization reactions.1-6 In addition, modifying
by these anions mentioned above, tetragonal zirconia phase can
be stabilized at high temperature,6-12 which is reported to be
favorable in catalytic reactions such as alkanes isomerization.
The acidic nature and crystalline structure of zirconia can be
also modified by other oxides such as tungsten oxide.9-13 Hino
and Arata reported that WO3-ZrO2 mixing oxide is an
alternative material in the acid-catalyzing reactions requiring
strong acidity.5,14 Although the crystalline structure of zirconia
The crystalline structures, acidity, and catalytic properties of
zirconia-tungsten oxides are greatly dependent on synthesis
method. Many synthetic techniques, such as conventional
precipitation with different impregnation techniques (evapora-
tion, incipient wetness), microemulsion, and sol-gel, have been
employed for the preparation of WOx-ZrO2 oxides.10,12,17 The
thermal treatments suggested are reflux,12 aging,17 and hydro-
thermal treatment under different temperatures.9 When a
precipitation or sol-gel method is used for activation of the
solids, a higher annealing temperature than that used for
impregnated sample is usually required.10,18 This is because sol-
gel solids usually consist of homogeneous phase composition;
high temperature favors the formation of active WO3 phases
expelled from the bulk of ZrO2 on the zirconia surface to form
active species responsible for the catalytic activity. It is reported
that the WOx domain depends only on the WOx surface density
and not on the WO3 concentration in the material.1,19 The high
activity appears to require the presence of WOx domains on
ZrO2 surfaces in the binary WOx-ZrO2 oxides, however, is
* To whom correspondence may be addressed. E-mail: macortes@imp.mx.
Phone: 52 55 91 75 84 33. Fax: + 52 55 91 75 62 39.
10.1021/jp053255d CCC: $30.25 © 2005 American Chemical Society
Published on Web 11/04/2005

Influence of Synthesis Methods
J. Phys. Chem. B, Vol. 109, No. 48, 2005 22731
TABLE 1: Atomic Fractional Coordinates of Tetragonal
Zirconia (P42/ nmc)
TABLE 2: Limits of the Atomic Fractional Coordinates of
Monoclinic Zirconia (P21/c)a
atom
site
x
y
z
atom site
x
y
z
Zr
O
2a
4d
0.75
0.25
0.25
0.25
0.75
ua
Zr 4e 0.259(1)-0.269(2) 0.031(1)-0.038(1) 0.208(1)-0.213(1)
O(1) 4e 0.065(7)-0.134(7) 0.326(6)-0.367(4) 0.345(5)-0.365(3)
O(2) 4e 0.468(7)-0.492(4) 0.748(2)-0.774(4) 0.437(5)-0.464(6)
a u varied between 0.441(1) and 0.452(2).
a The first value is the lower limit, and the last value is the upper
limit.
usually affected by thermal treatment conditions. Comparative
studies of the influences of the synthetic conditions on the active
structure, surface acidity and catalytic property of the binary
WOx-ZrO2 oxides are quite interesting.
Table 2 and have been reported elsewhere.24 The WO3 structure
was modeled with an orthorhombic unit cell that had the
symmetry of Pmnb space group.25 Standard deviations, which
show the variation of the last digit, are given in parentheses
and have been reported elsewhere.26
BET surface areas, pore volumes, and pore size distribution
were obtained by nitrogen physisorption in an ASAP-2000
analyzer from Micromeritics.
Samples were characterized by FT-IR spectroscopy using a
170 SX Nicolet spectrometer. Self-supported wafers (∼14 mg)
were mounted in a glass cell equipped with KBr windows, which
allowed following spectrum variations with thermal treatments.
The sample was outgassed at 273 K under 60 mL/min dried
airflow for 1 h before the IR spectrum measurement. For the
acidity measurements, the sample was treated in an IR cell under
vacuum conditions (10-4 Torr) at 673 K for 2 h and then cooled
to room temperature; the IR spectra of the calcined solid were
recorded too. Pyridine was adsorbed on the wafer at room
temperature, evacuated and heated at various temperatures.
In the present work, syntheses of a series of WOx-ZrO2
prepared by use of the precipitation technique are reported. The
resultant materials were treated through three different routes:
aging, refluxing, and hydrothermal conditions at a desired
temperature. The crystalline structure of the materials were
characterized with the X-ray diffraction (XRD) technique and
carefully refined by the Rietveld method; the surface nature and
morphological features were studied by using back-scattered
scanning electron microscopy (SEM) and Fourier transform
infrared (FT-IR) spectroscopic methods, and the textural proper-
ties, including surface areas and pore size distribution, were
determined by the Brunauer-Emmett-Teller (BET) method.
Finally, the catalytic activity and selectivity of the WOx-ZrO2
catalysts for n-hexane isomerization reaction are reported.
2. Experimental
2.1. Synthesis of Materials. A mixed zirconium tungsten
hydroxide was prepared from aqueous solution of zirconium
oxynitrate, ZrO(NO3)2‚6H2O (99%) 0.2 M, and ammonium
metatungstate hydrate, [(NH4)6W12O39.xH2O, 99% + 80 mesh]
0.007 M. The precipitation was carried out at a pH between
9.5 and 10 with an aqueous solution of ammonium hydroxide
14 vol % of NH4OH (28 vol %) as a precipitating agent. The
ammonium metatungstate solution was placed into a 4-L
container, and the other two solutions were added slowly to
the former one. A constant and vigorous agitation was main-
tained during the precipitation procedure. The obtained white
slurry was split into three parts; one of them was aged for 24
h at room temperature (hereinafter referred to as WZA), another
part was refluxed at 373 K for 24 h (hereinafter referred to as
WZR), and the last part was hydrothermally treated in a stainless
steel autoclave at 468 K, with autogenous pressure (hereinafter
referred to as WZH). All samples were dried at 383 K for 18 h;
before characterization, dried powders were annealed at 1073
K for 4 h in anhydrous air flow.
2.2. Characterization. The atomic tungsten concentration in
the annealed materials was determined by atomic absorption
spectroscopy with a Perkin-Elmer 5000 spectrophotometer.
XRD patterns of the samples packed in a glass holder were
recorded at room temperature with Cu KR radiation in a Bruker
Advance D-8 diffractometer having θ-θ configuration and a
graphite secondary-beam monochromator. Diffraction in-
tensity was measured in the 2θ range between 15 and 110° with
a 2θ step of 0.02° for 8 s per point. Crystalline structures were
refined with the Rietveld technique using DBWS-941120 and
FULLPROF-V3.5d codes;21 peak profiles modeled with a
pseudo-Voigt function contained average crystallite size as one
of its characteristic parameters.22,23 To refine the crystalline
structures of the solids by the Rietveld method, the tetragonal
unit cell symmetry was described with the P42/nmc space group,
with atomic positions given in Table 1. Monoclinic unit cell
symmetry was modeled with a unit cell that had symmetry of
P21/c space group and the positions of the atoms are given in
The samples were analyzed in an environmental scanning
electron microscope XL30, which was attached to an X-ray
energy dispersive spectroscope (EDAX). The SEM micrographs
were taken with backscattering electrons in order to obtain
compositional images.
2.3. Catalytic Evaluations. The catalytic activity was
measured in n-hexane isomerization reaction, and it was carried
out at 2 kg/cm2 pressure and 533 K in a continuous flow fixed-
bed microrreactor. Powdered solids (0.1 g) were loaded into
the reactor, and 0.2 mL/min of n-hexane and 3.12 mL/min of
hydrogen were fed. n-hexane molar flow was 3.19 × 10-5 mol/
h, and the H2/n-hexane ratio was 8.19. Thus, the weight hourly
space velocity (WHSV) was 1.92 h-1. The product gas mixture
was analyzed online with a Varian 3400 gas chromatograph
equipped with a flame ionization detector (FID) and a 50-m
methyl silicon gum capillary column. The conversion was kept
below 10 mol % at 5 min of time on stream. Before each
reaction test, the catalysts were activated in situ at 623 K in 80
mL/min dry N2 flow for 2 h, cooled at 533 K under flowing
N2, then n-hexane reactant was admitted to start the reaction.
3. Results
3.1. XRD Analysis. The XRD patterns of as-synthesized
WZA and WZH samples, obtained under different conditions,
are shown in Figure 1. Only an amorphous phase corresponding
to the hydroxylated zirconia was observed in the sample
obtained by aging precipitation at room temperature (WZA) and
refluxed at 373 K (WZR) (not shown). Well-crystallized
tetragonal and monoclinic zirconia phases were obtained when
the precipitate was treated under hydrothermal conditions
(WZH), Figure 1. When these as-made materials were calcined
at 1073 K in an air flow, both tetragonal and monoclinic zirconia
phases were formed (Figure 2). The presence of WO3 ortho-
rhombic phase was evident only in WZR and WZA samples,
whereas the material prepared by hydrothermal treatment, WZH,

22732 J. Phys. Chem. B, Vol. 109, No. 48, 2005
Corte´s-Ja´come et al.
Figure 3. Rietveld refinement plot of the sample WZR annealed at
1073 K (Rwp ) 0.147). Crosses correspond to the experimental data
and the continuous line to the calculated ones. The difference between
experimental and calculated data is also shown with a continuous line.
Upper tick marks correspond to tetragonal zirconia phase (RB ) 0.017),
middle tick marks to monoclinic zirconia phase (RB ) 0.039), and the
lower ones to the orthorhombic WO3 phase (RB ) 0.066).
Figure 1. XRD patterns of WZC and WZH as-synthesized samples
dried at 373 K. Upper tick marks correspond to the tetragonal zirconia
phase and the lower ones to the monoclinic zirconia phase. Miller
indices are indicated with “t” for tetragonal and with “m” for monoclinic
phases.
TABLE 3: Crystallite Size and Phases Compositions
Determined by Rietveld Method
crystallite size
(nm)
phase composition
(wt. %)
annealing
temperature
(K)
sample
t-ZrO2 m-ZrO2 WOx t-ZrO2 m-ZrO2 WOx
ZrO2
WZA
WZR
WZH
973
1073
1073
1073
12.0(3) 18.3(4)
23.2(4) 18(1)
15.7(2) 16.1(5) 48 (6) 59(2)
16.6(2) 19(1) 74.28
30.46
69.54
5.3(4) 63.4(4) 25(2) 11.3(5)
33.65
25.72
7.16
TABLE 4: Lattice Parameters of Tetragonal Zirconia
sample
T (K)
a )b (nm)
c (nm)
ZrO2
973
1073
1073
1073
0.35979(6)
0.35953(3)
0.3594(2)
0.35928(3)
0.5177(1)
0.51885(5)
0.51823(5)
0.52013(6)
WZA
WZR
WZH
WZA, 15.7 nm in WZR, and 16.6 nm in WZH solid. However,
the crystallite size of monoclinic zirconia was approximately
18 nm, independent of the aging treatment.
The lattice parameters of tetragonal zirconia are reported in
Table 4. It was noted that the structure of tetragonal zirconia
was distorted in comparison with that in the pure zirconia solid;
it is expanded in the c axis but contracted in a direction, when
tungsten was incorporated into the zirconia structure.
Figure 2. XRD patterns of WO3-ZrO2 samples annealed at 1073 K.
Upper tick marks correspond to the tetragonal zirconia phase, their
corresponding Miller indices are marked with a “t”, and the middle
tick marks correspond to the monoclinic zirconia phase, their corre-
sponding Miller indices are marked with “m”. The lower ones
correspond to orthorhombic WO3 phase, their Miller indices are
indicated with “w” in the amplified inset from 2θ ) 22-26.
3.2. SEM Images and Energy-Dispersive Spectrometry
(EDS) Analysis. To observe the surface morphologies and
tungsten distributed on the surface of zirconia solids, the three
samples, annealed at 1073 K, were analyzed by SEM using
backscattering electron images. Figure 4 shows compositional
images and chemical analysis of white and gray contrasts
presented in each sample. Small white particles on the surface
of the big gray particles were usually observed in WZA and
WZR samples (see parts a and b of Figure 4). However, small
gray particles with a similar size as the white particles are also
presented in the surface of the big particles. The sizes of the
gray particles in WZA are bigger than those in WZR sample;
however, the white particles in the WZA sample are smaller
than those presented in the WZR sample. In the case of the
WZH sample (see Figure 4c), it is constituted mainly by big
and small particles with only one kind of contrast. Similarly,
the smaller particles are also deposited on the surface of the
bigger particles in this solid.
did not show any characteristic XRD peaks corresponding to
orthorhombic WO3 phase.
The crystalline phases in the calcined materials were refined
by using the Rietveld method. Figure 3 shows a typical plot of
the Rietveld refinement corresponding to WZR sample. The
results of the Rietveld refinements (see Table 3) confirmed that
the concentration of tetragonal zirconia phase in the tungsten
containing samples is higher than 60 wt. %. Whereas, pure ZrO2
solid, annealed at 973 K, consisted of only 30 wt. % of
tetragonal phase. This result suggests that the presence of
tungsten stabilizes the metastable tetragonal phase at 1073 K
and inhibits the phase transformation to monoclinic crystals.6-12
On the other hand, formation of the tetragonal phase also
depends on the aging treatments of the precipitate: higher
concentration of this phase can be retained when the precipitate
was crystallized under hydrothermal condition (74 wt %).
The SEM images suggest that the gray particles are mainly
constituted by zirconia, while the white particles are rich in W.
This is because the image formation process is based on the
The crystallite size of the tetragonal zirconia in the three
samples varied with preparation conditions, it was 23 nm in

Influence of Synthesis Methods
J. Phys. Chem. B, Vol. 109, No. 48, 2005 22733
Figure 4. Backscattering SEM micrographs of the samples (a) WZA, (b) WZR, and (c) WZH annealed at 1073 K and EDS spectra corresponding
to (d) white particles, (e) big gray particles, and (f) small gray particles.
backscatter electrons of the atoms present in the sample: W
atoms scatter more electrons than Zr atoms, due to W greater
atomic number compared to that of Zr. This condition must
produce white contrast images where high W concentration is
present.
contrast agglomerates from Figure 4b for WZR sample, sug-
gesting a more homogeneous dispersion of WOx on the zirconia
surface.
To obtain stronger evidence about WOx dispersion on the
surface of zirconia, chemical mapping of the compositional
images was carried out at higher magnification. Figure 5 shows
the chemical distribution of O, W, and Zr atoms on the surface
of particles in WZR sample. Oxygen atoms are distributed
uniformly on the surface of the particles. The W atoms showed
high concentration in the areas where white particles are
deposited, whereas mapping of Zr atoms were deficient in those
regions; this can be appreciated as holes in the corresponding
chemical Zr map. From this analysis, it can be concluded that
the small gray particles have the same chemical concentration
as the big ones on which they are deposited, because it did not
show an appreciable variation on the chemical distribution on
the surface analyzed. Similar results were obtained in the WZA
sample (not shown). The chemical elemental maps of WZH
sample are shown in Figure 6. Although this sample showed
Typical EDS spectra, recorded from the white particles in
the sample WZA (see Figure 4d), showed a higher W peak
intensity than those observed for O and Zr atoms. The chemical
composition determined by this analysis showed that the white
particles are basically composed of 61.05 wt. % W atoms, 22.27
wt. % Zr atoms, and 16.68 wt. % O atoms. The EDS obtained
from gray agglomerates (see Figure 4e), showed that the largest
peak corresponds to Zr atoms, which means that these gray
particles are composed mainly of Zr atoms (14.48 wt. % of W
atoms, 62.36 wt. % of Zr atoms, and 23.16 wt. % of O atoms).
Chemical composition of the more homogeneous gray ag-
glomerates in WZH sample (see Figure 4f), showed a similar
concentration of O and Zr atoms, but the W atoms concentration
was slightly higher, 19.00 wt. %, than that observed in the gray

22734 J. Phys. Chem. B, Vol. 109, No. 48, 2005
Corte´s-Ja´come et al.
Figure 5. Chemical mapping of oxygen, zirconium, and tungsten atoms in WZR sample annealed at 1073 K.
Figure 6. Chemical mapping of oxygen, zirconium, and tungsten atoms in WZH sample annealed at 1073 K.
particles with only gray contrast type in backscatter image, it
exhibited inhomogeneous distribution of tungsten and oxygen
atoms on the surface of the agglomerates. The concentration of
W and O atoms slightly increased in some areas, suggesting
that there are some regions on the particles having high
concentration of WOx species, even though they remain in low
crystallite size, that is, less than 3 nm.
The chemical composition between the particles of different
contrasts observed in compositional images was investigated
by line scan images. Their composition profiles are shown in
Figure 7. The W composition profile of WZA sample (see
Figure 7a) shows a low W concentration in the big gray particle
surface but a high W concentration in the white particles. In
contrast, the Zr composition profile showed a decrease in the

Influence of Synthesis Methods
J. Phys. Chem. B, Vol. 109, No. 48, 2005 22735
Figure 7. Chemical composition profile of oxygen, zirconium, and tungsten atoms along the line drawn in the corresponding micrographs of the
samples (a) WZA, (b) WZR, and (c) WZH annealed at 1073 K.
TABLE 5: Textural properties of the samples calcined at
corresponding regions where the particles are rich in W atoms.
The oxygen composition profile exhibited a slight increase in
1073 K.
surface area
porous volume
mean porous size
(Å)
the regions where the white particles are deposited. The WZR
sample presented a low O atom concentration profile in
comparison with that of Zr and W atoms (Figure 7b), and it
did not vary along the white and gray particles. Variations of
the profiles of W and Zr atoms in WZH are shown in Figure
7c, Zr atoms concentration is rich in the gray regions, while W
atoms is slightly rich in the small gray particles. These results
confirm that the gray particles are related to ZrO2 phase and
that the white particles are related to WOx species. For the
hydrothermally treated sample, i.e., WZH, the O atom concen-
tration profile showed a similar behavior as that in Figure 7a;
its concentration increased in the regions where W atoms
concentration increased. It is interesting to note that this behavior
in the O atoms concentration was not observed in WZR sample,
where the WO3 crystallite size was very large. Therefore, it
appears that the increase in oxygen concentrations in the regions
enriched with W atoms is dependent on the crystallite dimen-
sions of WOx species.
sample
(m2 g-1
)
(cm3 g-1
)
WZA
WZR
WZH
48
66
54
0.073
0.694
0.164
61
226
77
3.3. Textural Properties. The interaction between WOx
species and ZrO2 in the solids prepared with different aging
conditions not only stabilizes the tetragonal zirconia phase and
produces a significantly deformed structure but also affects the
specific surface area and pore size distribution. The specific
surface areas were not quite different between the samples,
which varied from 48 to 66 m2/g for ZWA and ZWR samples,
respectively, as shown in Table 5. When examining the
hysteresis loop shape of the nitrogen adsorption-desorption
isotherms in Figure 8, information about pore blocking and pore
structure can be obtained. The isotherms presented by the three
materials correspond to type IV in the Brunauer, Deming,
Deming, and Teller (BDDT) classification.27 The initial sloping

22736 J. Phys. Chem. B, Vol. 109, No. 48, 2005
Corte´s-Ja´come et al.
TABLE 6: Tungsten Concentration and Density on the
Zirconia Surface of the Samples Calcined at 1073 K
W
(wt. %)
W
W
sample
(atoms/nm2)a
(atoms/nm2)b
WZA
WZR
WZH
11.6
15
14.5
7.92
6.55
8.79
2.32
4.45
8.79
a Determined taking into account the total amount of W atoms loaded
determined by AAS. b Determined by subtracting the amount of W
atoms segregated as WO3 (determined by Rietveld analysis) to the total
amount of W atoms determined by AAS.
Figure 8. The hysteresis loops of the nitrogen adsorption-desorption
isotherms of the different samples calcined at 1073 K.
Figure 10. FTIR spectra of the different samples calcined at 1073 K.
Figure 9. Pore diameter distributions of the different samples calcined
at 1073 K.
samples and that, at pH ) 10.0, tungsten is present mainly as
WO4) species in solution,13 one can conclude that when aging
the precipitate at room temperature only 64% of the total WO4
oxoanions interact with the zirconium hydroxide precipitate;
however, when the samples were thermally refluxed or crystal-
lized under hydrothermal conditions, almost all WO4 oxoanions
were retained by the zirconium hydroxide or by the crystallized
zirconium oxide. Then, thermal treatment aging at 373 K is
required to obtain a solid in which tungsten ions completely
incorporate into the crystallites of zirconium precipitate.
The density of surface atomic tungsten was determined by
taking into account the specific surface area (see Table 5), and
the amount of tungsten loaded was determined by atomic
absorption (see Table 6, column 3), assuming that all tungsten
atoms loaded on zirconia are homogeneously distributed.
However, this assumption is not valid for WZR and WZA,
where the WO3 phase was segregated forming aggregates on
the surface of zirconia. Then, to estimate this value, the amount
of W atoms that in fact plays the role of stabilizing the zirconia
phases was estimated by the difference between the total
amounts of W atoms (determined by AAS) minus the amount
of W atoms segregated as WO3 (determined by Rietveld
analysis), this value is reported in the last column of the Table
6.
characteristic of the descending boundary curves reflects porous
structure shrinkage caused by the stress change of the condensate
during depressurization, and the sharp decline in the desorption
curve at relative pressure of 0.4, 0.6, and 0.7 for the WZA,
WZR, and WZH, respectively, indicates the mesoporous
characteristics of the material, with different pore diameters (see
Table 5 and Figure 9). This type of hysteresis loop has been
attributed to either ordered or disordered packed tubes in M41S
mesostructures or to wormholelike morphologies shown by very
small crystallite size material.28-30 In the WZR sample, an
additional hysteresis loop can be observed at between 0.9 and
1.0 relative pressure, with relatively high adsorbed volume,
which is presumably attributed to the surface of large crystallites
of WO3 phase deposited on the zirconia phases, in agreement
with our SEM and XRD results.
The pore volume and pore diameter of the WZA solid is lower
than those for WZR and WZH samples. The samples aged at
high temperature, that is, WZR and WZH, developed very
narrow and homogeneous pore size distributions (see Figure
9), while WZA aged at room temperature showed a broad pore
size distribution with relatively low pore volume. Therefore,
aging the precipitate by reflux or crystallizing under hydro-
thermal condition is necessary to develop high porosity in the
resulting materials.
3.5. FTIR Analysis. Figure 10 shows the different FTIR
spectra of the samples calcined at 673 K in the range of 1500-
400 cm-1. Regardless of the preparation methods, all the
samples showed a band at 970 cm-1, which was assigned to
polytungstate structure with tungsten in octahedral coordina-
tion.31 Therefore, symmetrical WdO stretching vibration was
assigned to octahedrally coordinated WOx species on the surface
of tetragonal ZrO2. This unique vibration mode was observed
on the WZH sample, because the WOx species were more
3.4. Atomic Absorption Spectroscopy (AAS) Analysis.
Chemical composition of the samples determined by AAS
showed that the lowest tungsten content (11.6 wt %) was
obtained in the room temperature aged sample, WZA, as shown
in Table 6; and the high W content (about 15 wt %) appeared
in the refluxed and hydrothermal aging samples. By consider-
ation that the calculated W amount was 15% loaded in all the

Influence of Synthesis Methods
J. Phys. Chem. B, Vol. 109, No. 48, 2005 22737
TABLE 7: Activity and Selectivity Results of the Samples
Calcined at 1073 K
selectivity (mol %)
conversion
sample
(mol %)
cracking
2,3DMB
2MP
3MP
ZrO2
0
0.0
0.0
0.0
0
0
0
WZA
WZR
WZH
4.23
6.98
8.16
15.3
15.4
14.66
52.62
52.85
51.3
32.08
31.75
32.09
1.94
3.7. Catalytic Properties. Catalytic activity and selectivity
of n-hexane isomerization over the three catalysts are reported
in Table 7. The highest conversion was observed in the
hydrothermally treated WZH sample, which corresponded to
the sample with the highest tungsten dispersion or the highest
surface atomic tungsten density, consequently the highest density
of Bro¨nsted and Lewis acid sites. Small amount of cracking
products were identified when WZH was used as catalyst, which
could be associated to the strongest surface acidity. The
selectivity to the isomers on the other two samples reached
100% without any cracking products. The main products in the
n-hexane isomerization were 2,3-dimethyl butane (2,3DMB),
2-methyl pentane (2MP), and 3-methyl pentane (3MP) for all
the samples. The selectivity to these products was quite similar
in all the samples.35,36
Figure 11. In situ FTIR spectra of pyridine adsorption on the samples
calcined at 1073 K. The samples were thermally treated in the IR cell
at different temperatures: (a) ambient temperature and (b) 573 K.
homogeneously distributed on the surface of zirconia. This band
appeared at lower wavenumbers than those reported in other
works, which are between 1000 and 1020 cm-1 10,32,33
The shift
.
of the band position toward lower wavenumbers (970 cm-1) in
our samples could be attributed to a strong tungsten-zirconia
interaction,34 which could be in turn explained by the neutraliza-
tion of the basic sites in ZrO2 by tungstate anions; this aspect
will be later discussed in this paper.
4. Discussion
4.1. Effect of Different Treatments on the Phase Forma-
tion and Crystallite Size. XRD analysis shows that the
materials obtained under different treatment conditions had quite
different phase compositions. For instance, in the as-synthesized
WZH solid, 74 wt. % of the tetragonal zirconia was already
formed, then WO4) species interacted with the surface of well-
crystallized zirconia phases during the hydrothermal treatment.
However, in samples prepared by aging the slurry at room
temperature and refluxing at 373 K, the resultant solids are in
Two additional FTIR bands at 508 and 743 cm-1 were
observed in the WZA sample, in which the WO3 phase was
segregated with a 5.23(4) nm crystallite size. These two
additional peaks were also observed in the WZR solid, and they
could be assigned to the W-O vibration of tungsten species
bonded on the surface of ZrO2, as observed on the samples of
10 wt. % W/ZrO2 prepared by precipitation or sol-gel, where
they were assigned to a W-O vibration of crystalline and
noncrystalline tungsten species.10 However, from this study it
can be established that the FTIR bands at 743 and 508 cm-1
appear only when the WO3 phase presents some crystalline
order with crystallite size larger than 5 nm. FTIR analysis of
the WZR presented three additional bands at 1400, 1260, and
1100 cm-1, which are assigned to the well crystallized WO3
phase.9
)
amorphous state, then WO4 species must interact with the
amorphous hydroxylated zirconia phase, where tungstate species
stabilize the material in 100 wt % tetragonal phase even after
annealing at 973 K, as was reported elsewhere.16 When the
samples were annealed at 1073 K, a certain amount of WOx
crystallites, which are then partially transformed into the
orthombic phase, segregates from tetragonal zirconia. These
results can be explained by assuming that, during precipitation
and aging treatment (but not under hydrothermal crystallization
)
3.6. Surface Acidity. The FTIR spectra of adsorbed pyridine
evacuated at 298 K (see Figure 11a) showed Lewis and Bro¨nsted
acid sites in all the samples and the intensities of the corre-
sponding vibration bands on the WZH sample were higher than
those on the WZR and WZA solids. Hence, the sample prepared
by hydrothermal treatment showed high acid density, which is
related to its high density of surface atomic tungsten.
When the samples were heated at 573 K under vacuum
conditions, the vibration bands of the adsorbed pyridine on
Lewis and Bro¨nsted acid sites remained with considerable
intensity only in the WZH sample (see Figure 11b). In the
refluxed WZR sample, some broad peaks, which could be
confused with the background, suggest that a little amount of
Bro¨nsted and Lewis acid sites were still retained in this sample,
whereas in the WZA sample, practically all the pyridine was
desorbed at 573 K. From these results, one can conclude that
the acid strength on WOx-ZrO2 catalysts obviously depends on
the tungsten dispersion and, more precisely, on the surface
atomic tungsten density, as indicated in Table 6. In other words,
the higher atomic tungsten density on the surface, the higher
the Bro¨nsted and Lewis acidity.
condition), WO4 species are trapped inside the zirconia
hydroxide polymeric species [Zr8(OH)20(H2O)24]12+, followed
by transformation of the octamer to larger aggregates, forming
bidimensional sheets or three-dimensional agglomerates with
rapid nucleation.37 Finally, a solid WxZr1-xO2 solution was
formed by isomorphic cationic substitution of zirconium atoms
by tungsten atoms in the lattice of the tetragonal zirconia, when
annealing at 973 K.16 After annealing at 1073 K, the tetragonal
structure partially collapses segregating WO3 crystallites.
The tungsten incorporation also affects the crystallite size of
the tetragonal zirconia phase (see Table 3). In pure ZrO2 sample,
low concentration of tetragonal zirconia was present; however,
its crystallite size was (12 nm) smaller than that of the tungsten
containing samples (16-23 nm). The largest crystallite size of
tetragonal zirconia was formed in the ZWA sample annealed
at 1073 K. In all cases, with and without tungsten, the tetragonal
zirconia phase has been stabilized at crystallite sizes between
12 and 23 nm, which is lower than that obtained under
nonhydrostatic stress as reported by Garvie.38
The crystallite size of the orthorhombic phase of WO3 loaded
on the different solids was quite different, strongly depending

22738 J. Phys. Chem. B, Vol. 109, No. 48, 2005
Corte´s-Ja´come et al.
Figure 12. Representative polyhedron of the tetragonal crystalline
structure of zirconia.
TABLE 8: Atomic Bond Lengths and Bond Angles of
Tetragonal Zirconia
T (°C) T (K)
ZrO2 973 0.21242(6) 0.23162(6) 115.74(2) 113.38(2)
WZA 1073 0.20806(6) 0.23715(6) 119.53(2) 115.18(2)
WZR 1073 0.20799(6) 0.23691(6) 119.53(2) 115.13(2)
d1 (nm)
d2 (nm)
φ
δ
Figure 13. Lattice of the tetragonal zirconia phase projected perpen-
dicular to the c axis.
WZH 1073 0.20536(6) 0.24094(6) 122.04(2) 116.37(2)
segregated from tetragonal zirconia, decreasing the interaction
between tetragonal ZrO2 and WO3 phase, thus, as a result, their
d1 bond lengths remained between the values of pure tetragonal
ZrO2 and WZH solids. The Zr-O bond lengths, d2, increased
from 0.23162(6) nm for pure tetragonal ZrO2 to 0.24094(6) nm
for WZH. On the other hand, WZR and WZA samples showed
a Zr-O bond length (d2) quite similar, that is, 0.23715(6) and
0.23691(6) nm, respectively. All these data clearly show that
the crystalline structure of tetragonal zirconia was strongly
deformed after tungsten incorporation, depending on the degree
of the interaction between tetragonal ZrO2 and WO3 phases.
The structural deformation of the tetragonal ZrO2 structure
was also reflected in the variations O-Zr-O angles. Both, φ
and δ angles increased as the interaction degree between WOx
species and tetragonal ZrO2 increased in WZH sample, remain-
ing constant for WZA and WZR samples, where WO3 phase
was segregated.
All the above results point out structural parameters of the
tetragonal zirconia are very sensitive to the interaction degree
between WOx species and the surface of tetragonal ZrO2 phase.
The expansion of the tetragonal structure only along the c
direction and the increase in the Zr-O (d2) bond lengths
suggests that WOx species are bonded to the oxygen atoms
exposed on the surface perpendicular to the c axis (see Figure
13), because all the oxygen atoms exposed at the top layer of
the (001) plane are bonded to the zirconium atoms with a d2
Zr-O bond length. Hence, the increase in the d2 dimension,
Zr-O bond length, as the interaction degree between WOx
species and the surface of tetragonal ZrO2 increases could be
related to the electron effect of the WOx species (attractor or
acceptor). The strong interaction between WOx species and the
surface of tetragonal ZrO2 leads to electronic deficiency around
the Zr4+ nuclei, thus, the d1 bond length must decrease in order
to compensate such an electronic deficiency around the Zr4+
nuclei caused by the elongation of the d2 bond.
on the aging treatment. Large crystallite size (48 nm) was
obtained in the WZR sample. However, in the WZA sample its
crystallite size was 5.3 nm. In the WZH sample, there were not
any WO3 crystals detected by XRD analysis, indicating that WOx
species are highly dispersed on the surface of zirconia with a
crystallite size less than 3 nm. It is very likely that in this sample
WOx species are strongly bounded to the surface of the zirconia.
Therefore, the various aging treatments on the zirconia precipi-
tate result in different degrees of interaction between WOx
species and zirconia surface, leading to WO3 crystallites with
different dimensions.
4.2. Effect of Different Treatments on the Crystalline
Structural Deformation. Noteworthy, the unit cell of tetragonal
zirconia in the tungsten-containing solids, relative to the pure
zirconia structure, was expanded only along the c direction,
while along the other directions, a and b, it remained almost
constant. This fact must be indicative of strong variations in
the tetragonal local order structure. In pure ZrO2, the tetragonal
lattice parameter c was 0.5171(1) nm. The refluxed WZR
sample, where larger WO3 crystallite size was formed, suggests
a weak interaction between tetragonal ZrO2 and WOx species,
showing a lattice c parameter of 0.51823(5) nm, only slightly
larger than that in pure tetragonal ZrO2. However, the WZH
sample, where the WOx species were well dispersed on the
tetragonal ZrO2 phase with crystallite size less than 3 nm,
showed a lattice c parameter of 0.52013 (6) nm, an increment
of 3% was obtained. In this sample, strong interaction between
WOx and tetragonal ZrO2 must exist. Hence, the changes in
lattice c parameters in the structure of tetragonal zirconia may
be indicative of the interaction degree between WOx and
tetragonal ZrO2.
In the unit cell of tetragonal zirconia, there are two different
Zr-O bonds with different length, referred to as d1 and d2 (see
Figure 12), with two different O-Zr-O bond angles, φ, the
angle formed between two d1 bond lengths, and δ, the angle
formed between two d2 bond lengths.16 All the data related to
Zr-O bond lengths and bond angles are reported in Table 8. It
is interesting to examine how Zr-O bond lengths and bond
angles change as a function of the interaction degree between
tetragonal ZrO2 and WOx species. Rietveld refinements showed
that the Zr-O (d1) bond length was 0.21242(6) nm for pure
tetragonal ZrO2, and it decreased to 0.20536(6) nm in the WZH
sample, calcined at 1073 K, where WOx species were well
dispersed. In the aged and refluxed samples (WZR and WZA),
the d1 bond lengths remain quite similar (0.20806(6) and
0.20799(6) nm, respectively). In these two samples, WO3
4.3. Effect of Different Treatments on the Surface Acidity
and Catalytic Activity. It has been established that segregation
of large crystallites of WO3 on the zirconia surface improves
the acidity and catalytic properties of the WOx-ZrO2 cata-
lysts;12,18 however, large amount of the tungsten species are not
exposed to the surface in well-crystallized WO3, therefore, a
question can be formulated: are well-crystallized WO3 phase
or dispersed WOx species that stabilize the tetragonal ZrO2 phase
responsible for the high acidity? This question can be answered
by analyzing the acidity and catalytic property of the three
samples treated in a variety of conditions. Figure 11 shows that
the WZH solid had not only the largest number of both Bro¨nsted

Influence of Synthesis Methods
J. Phys. Chem. B, Vol. 109, No. 48, 2005 22739
and Lewis acid sites but also the strongest acidity. Because the
dispersion of WOx species in the WZH solid was very high
and uniform, it is reasonable to conclude that highly dispersed
WOx species are mainly responsible for the generation of acidity.
As discussed above, the crystalline structure of the zirconia
phase in WZH sample shows the strongest deformation that
resulted from the strongest interaction between the WOx species
and zirconia, due to the highest dispersion of WOx species,
therefore, the deformation of the crystalline structure is also
correlated well with acidity of the solid.
The catalytic activity showed the same trend as that observed
in the population of Bro¨nsted and Lewis acid sites. The relatively
highest conversion of n-hexane was achieved on the sample
with the highest surface tungsten density. WOx species homo-
geneously distributed on the surface of the tetragonal ZrO2 are
the active species that participate in the reaction, and that well
crystallized WO3 phase with large particles size are less active
for the n-hexane isomerization reaction.
Acknowledgment. The financial support from the IMP-
D00237 Project is greatly appreciated. We thank Mrs. Rosa Isela
Conde Velazco and Mr. Gabriel Pineda Vela´zquez for their
technical assistance.
References and Notes
(1) Barton, D. G.; Shtein, M.; Wilson, R. D.; Soled, S. L.; Iglesia, E.
J. Phys. Chem. B 1999, 103, 630.
(2) Iglesia, E.; Soled, S. L.; Kramer, J. Catal. 1993, 144, 238.
(3) Pe´rez, M.; Armendariz, H.; Toledo, J. A.; Va´zquez, A.; Navarrete,
J.; Montoya, A.; Ga´rcia, A. J. Mol Catal., A 1999, 149, 69.
(4) Lo´pez-Salinas, E.; Herna´ndez-Corte´z, J. G.; Corte´s, M. A.;
Navarrete, J.; Llanos, M. E.; Va´zquez, A.; Armenda´riz, H.; Lo´pez, T. Appl.
Catal. A 1998, 175, 43.
(5) Hino, M.; Arata, K.; J. Chem. Soc., Chem. Commun. 1987, 1259.
(6) Iglesia, E.; Barton, D. G.; Soled, S. L.; Miseo, S.; Baumgartner, J.
E.; Gates, W. E.; Fuentes, G. A.; Meitzner, G. D. Surf. Sci. Catal. 1996,
101, 533.
(7) Barton, D. G.; Soled, S. L.; Meitzner, G. D.; Fuentes, G. A.; Iglesia,
E. J. Catal. 1999, 181, 57.
(8) Scheithauer, M.; Grasselli, R.; Kno¨zinger, K. Langmuir 1998, 14,
3019.
(9) Armenda´riz, H.; Corte´s-Ja´come, M. A.; Herna´ndez, I.; Navarrete,
J.; Va´zquez, A. J. Mater. Chem. 2003, 13, 143.
5. Conclusions
Binary WO3-ZrO2 solids with different phase compositions
and structural distortion were synthesized by use of the
coprecipitation method under a variety of aging conditions. The
solids aged at room temperature or refluxed at 373 K for 24 h
showed an amorphous phase, which may segregate nanocrystals
of WO3 with a size larger than 5 nm from the phase of tetragonal
zirconia after 1073 K of calcination. However, when the sample
was hydrothermally treated in an autoclave, well-crystallized
phases of tetragonal and monoclinic zirconia were formed in
the dried material, and even when it was calcined at 1073 K,
not any WO3 nanocrystals larger than 3 nm were segregated,
indicating a very high dispersion of tungsten species. All the
materials showed mesoporous textures with surface area between
48 and 66 m2/g and average pore diameter between 5 and 10
nm.
The crystalline structure of tetragonal zirconia in the solid
obtained in hydrothermal conditions was strongly deformed, as
indicated by clear expansion of the lattice c parameter and Zr-O
bond lengths (d2) and variations in bond angles (φ and δ) that
resulted from the strong interaction between WOx and ZrO2.
However, the structure of zirconia in the materials obtained by
aging the precipitate at room temperature and refluxing at 373
K showed only slight distortion, which resulted from weak
interaction between WOx and ZrO2 because of segregation of
large WO3 crystals.
(10) Boyse, R. A.; Ko, E. I. J. Catal. 1997, 171, 191.
(11) Gazzoli, D.; Valigi, M.; Dragone, R.; Marucci, A.; Mattei, G. J.
Phys. Chem. B 1997, 101, 431.
(12) Santiesteban, J. G.; Vartuli, J. C.; Han, S.; Bastian, R. D.; Chang,
C. D. J. Catal. 1997, 168, 431.
(13) Loridant, S.; Feche, C.; Essayem, N.; Figueras, F. J. Phys. Chem.
B 2005, 109, 5631.
(14) Hino, M.; Arata, K. Proc. 9th Int. Congr. Catal. 1988, 1727.
(15) Bokhimi, X.; Morales, A.; Garc´ıa, A.; Xiao, T. D.; Chen, H.; Strutt,
P. R. J. Solid State Chem. 1999, 142, 409.
(16) Corte´s-Ja´come, M. A.; Toledo, J. A.; Armenda´riz, H.; Hernandez,
I.; Bokhimi, X. J. Solid State Chem. 2002, 164, 339.
(17) Chuah, G. H. Catal. Today 1999, 49, 131.
(18) Vaudagna, S. R.; Canavese, S. A.; Comelli, R. A.; F´ıgoli, N. S.
Appl. Catal. A 1998, 168, 93.
(19) Iglesia, E.; Barton, D. G.; Biscardi, J. A.; Guines, M. J. L.; Soled,
S. L. Catal. Today 1997, 38, 339.
(20) Young, R. A.; Sakthivel, A.; Moss, T. S.; Paiva-Santos, C. O. J.
Appl. Crystallogr. 1995, 28, 366.
(21) Rodr´ıguez-Carbajal, J. Laboratoire Leon Brillouin, France. Phone:
(33) 1 6908 3343. Fax: (33) 1 6908 8261. E-mail: juan@llb.saclay.cea.fr.
(22) Thompson, P.; Cox, D. E.; Hasting, J. B. J. Appl. Crystallogr. 1987,
20, 79.
(23) Young, R. A.; Desai, P. Arch. Nauki Mat. 1989, 10, 71.
(24) Bokhimi, X.; Morales, A.; Novaro, O.; Portilla, M.; Lo´pez, T.;
Tzompantzi, F.; Go´mez, R. J. Solid State Chem. 1998, 135, 28.
(25) Salje, E. Acta Crystallogr. B 1982, 24, 1968.
(26) Prince, E. J. Appl. Crystallogr. 1981, 14, 157.
(27) Brunauer, S.; Deming, L. S.; Deming, W. E.; Teller, E. J. Am. Chem.
Soc. 1940, 38, 1723.
The sample prepared under hydrothermal treatment conditions
exhibited the highest density of surface acid sites, which was
responsible for the best catalytic activity in n-hexane isomer-
ization; however, due to its stronger acidity, 1.9% of cracking
products were produced during the reaction, this differs from
the other two catalysts on which selectivity to 100% isohexanes
was achieved but with slightly low conversion.
This work provides new experimental information on how
to obtain WOx-ZrO2 binary oxides with different WOx disper-
sions, surface acidity, and catalytic behaviors by treatment the
precipitating materials under controllable conditions. It also
illustrates how computer simulation (the Rietveld refinement)
combined with experimental techniques contributes to interpret
the variations of crystalline structure and catalytic properties
of the solids.
(28) Cabrera, S.; Haskouri, J. E.; Alamo, J.; Beltran, A.; Beltran, D.;
Mendioroz, S.; Marcos, M. D.; Amoros, P. AdV. Mater. 1999, 11, 379.
(29) Zhang Z.; Pinnavaia T. J. J. Am. Chem. Soc. 2002, 124, 12294.
(30) Gutie´rrez, R.; Toledo, J. A.; Cortes-Jacome, M. A.; Va´zquez, A.;
Sebastia´n, P. J. Langmuir 2004, 20, 4265.
(31) Baertsch, C. D.; Soled, S. L.; Iglesia, E.; J. Phys. Chem. B 2001,
105, 1320.
(32) Yamagushi, T.; Tanabe, K. Mater. Chem. Phys. 1986, 16, 67.
(33) Shon, J. R.; Park, M. Y. Langmuir 1998, 14, 6140.
(34) Dongare, M. K.; Ramaswamy, V.; Gopinath, C. S.; Ramaswamy,
A. V.; Scheurell, S., Brueckner, M.; Kemnitz, E. J. Catal. 2001, 199, 209.
(35) Vu, T. N.; van Gestel, J.; Wilson, J. P.; Mollet, C.; Dath, J. P.;
Duchet, J. C. J. Catal. 2005, 231, 468.
(36) Duchet, J. C.; Guillaume, D.; Monnier, A.; Dujardin, C.; Wilson,
J. P.; van Gestel, J.; Szabo, G.; Nascimento, P. J. Catal. 2001, 198, 328.
(37) Singhal, A.; Toth, L. M.; Lin, J. S.; Affholter, K. A. J. Am. Chem.
Soc. 1996, 118, 11529.
(38) Garvie, R. C. J. Phys. Chem. 1965, 69, 1238.