Development of acidic sites in WOx/ZrO2

Article abstract of DOI:10.1039/b102867g

Evidence is shown of the appearance of Bronsted acidic sites in WO_x/ZrO₂ at a given W loading and of a direct correlation between the abundance of strong Bronsted acidic sites in these catalysts and propan-2-ol decomposition activity.

Full text of DOI:10.1039/b102867g

Development of acidic sites in WO /ZrO
x
2
Thomas Onfroy, Guillaume Clet and Marwan Houalla*
Laboratoire de Catalyse et Spectrochimie (UMR CNRS 6506), ISMRA-Université de Caen, 6 Bd. du
Maréchal Juin, 14050 Caen (cedex), France. E-mail: marwan.houalla@ismra.fr
Received (in Cambridge, UK) 28th March 2001, Accepted 20th June 2001
First published as an Advance Article on the web 6th July 2001
Evidence is shown of the appearance of Brønsted acidic sites
in WO /ZrO at a given W loading and of a direct correlation
between the abundance of ‘strong’ Brønsted acidic sites in
these catalysts and propan-2-ol decomposition activity.
equilibrium pressure of 133 Pa. 2,6-dimethylpyridine (lutidine)
was introduced at room temp. (Pequilibrium = 133 Pa) followed
by thermodesorption from 373 to 573 K.
The catalytic conversion of propan-2-ol was measured in a
fixed bed flow reactor. A mass of 100 mg of sample was pre-
x
2
It is well established that addition of W onto zirconia induces
the formation of acid sites. With suitable preparation methods
solid acid catalysts can be obtained which can isomerise n-
hexane. However, the genesis of acidity and the actual nature
of the acid sites have not been fully investigated. Previous work
treated at 723 K in N
atmospheric pressure with 120 ml min
2
for 2 h. The reaction was carried out at
2
1
2
N as carrier gas
(Ppropan-2-ol = 1.23 kPa) at 413 K. Reactants and products were
analysed with an on line Gas Chromatograph (HP 5890 Serie II)
equipped with a capillary column (CP WAX 52 CB) and an FID
detector. Catalytic study was conducted in non diffusional
conditions, with conversions typically below 10%. Neither an
induction period nor deactivation were observed during the
analysis time (2 h). The rate of propene formation was
calculated from eqn. (1), assuming a first order reaction for
propan-2-ol:
1
x 2
with WO /ZrO obtained by the incipient wetness impregnation
method, indicated that a threshold of W loading was required for
2
any activity for propan-2-ol decomposition to develop. Similar
behaviour was already reported for solids obtained by impreg-
nation of zirconium hydroxide, for different reactions such as o-
3
4,5
2
xylene and n-pentane isomerisation. It was proposed, but
not verified, that the origin of this behaviour is due to the
formation of different types of acid sites at a given W loading.
Furthermore, the simultaneous presence of polymeric surface
F0 Ê100 - Cˆ
r = - ln
(1)
W
Ë
100
¯
where r is the rate of propene formation (mol h g21), F
propan-2-ol molar flux (mol h ) and C is the % conversion to
propene.
Raman spectra of WO
support showed intense peaks between 100 and 700 cm
characteristic of the monoclinic phase of zirconia. No Raman
2
1
is the
WO
x
and bulk WO
3
in active catalysts complicated any attempt
0
2
1
to elucidate the nature of the active sites.
In the present work, a series of WO
x
/ZrO
2
catalysts was
6
prepared by the equilibrium adsorption method. This proce-
dure was shown to prevent WO formation. A direct correlation
x 2
/ZrO catalysts and the zirconia
2
1
3
8
between the surface structure and the catalytic activity for
propan-2-ol decomposition is reported.
Tungsten surface species were monitored by Raman spec-
troscopy. The number, the strength and the type of acidity were
characterised by IR spectroscopy. Propan-2-ol decomposition
was used as a probe reaction to assay the acidity of the solids.
9
3
peaks which can be ascribed to bulk WO were detected. A
2
1
band at 935 cm attributed to the W surface phase appears on
W deposition and increases in intensity with W loading. A shift
in the position of this band to higher wavenumbers is observed
1
0
with W content in accordance with previous results. This is
indicative of monomeric to polymeric transformation of W
species.
Catalysts were prepared by adsorption of W from aqueous
6
solutions of ammonium metatungstate on zirconia (ZrO
2
). The
support was prepared by hydrolysis of zirconium n-propoxide
Zr(OPr) , 70% in n-propanol; Aldrich]. The precipitate thus
The IR spectra of tungstated zirconias show significant
dehydroxylation which increases with W loading. This in-
dicates that W deposition occurs via replacement of the
hydroxyl groups.
The presence of Lewis acid sites was monitored by FTIR and
CO adsorption. The amount of CO required to saturate the sites
decreased on W addition but it remained the same for all the
tungstated solids.
[
4
obtained was washed free of propanol, dried and calcined at 823
7
K in air for 24 h. The original zirconia had a BET surface area
2
21
of 43 m g . A series of WO
suspending a known amount of ZrO
4 6 2 12 40
aqueous ammonium metatungstate solution [(NH ) H W O ,
x
/ZrO
2
catalysts was prepared by
2
in a large volume of
Aldrich] at a given pH. The samples were filtered off, dried and
calcined at 773 K in air for 24 h. Solids with various W loadings
ranging from 1.1 to 4.6 wt% W were obtained by modifying
either the pH of the adsorption solution (adsorption time = 96
h) or the adsorption time (at pH = 12). No significant surface
area loss was observed on W deposition.
The adsorption of lutidine at room temp. followed by
evacuation at 423 K evidenced the presence of Brønsted acidic
sites on the surface of WO
Fig. 1(a) by the appearance of a doublet at 1645 and 1628 cm
x 2
/ZrO catalysts. This is illustrated in
2
1
attributed to lutidinium species and the subsequent increase of
its intensity with increasing W loading. The peaks at 1610 and
Raman spectra were recorded with a Nicolet FT-Raman
spectrometer attachment for a Nicolet Nexus FTIR spectrome-
2
1
1580 cm are attributed to Lewis acid sites. After desorption of
2
1
21
ter between 100 and 1100 cm (resolution: 8 cm ). The
samples were analysed in powder form and under ambient
conditions without any pre-treatment. The technique was used
to monitor the structure of the support (monoclinic; tetragonal),
lutidine at 423 K, Brønsted acid sites were still observed for a W
2
2
loading of 1 atom nm . However, following desorption at 523
K only the Brønsted acid sites of solids containing more than 1.4
2
2
atom W nm were sufficiently strong to retain the probe
molecule on the surface [Fig. 1(b)]. These results are compara-
ble to those reported by Wachs11 concerning the appearance of
Brønsted acidity on molybdenum, vanadium and niobium oxide
supported on alumina.
WO
IR spectra were recorded with a Nicolet 710 FT-IR
3
formation and the evolution of surface W species.
2
1
spectrometer (resolution: 4 cm ). Samples were pressed into
2
discs (ca. 20 mg; 2 cm ) and activated alternatively in vacuum
and O
2
at 723 K. Acidic properties (type and abundance) of the
The catalysts were tested for the reaction of propan-2-ol
decomposition. Only two products, propene and diisopropyl
ether were observed. For all catalysts, the selectivity to propene
was high (4 85%) and increased with temperature as expected
samples were characterised by two basic probe molecules (CO
and 2,6-dimethylpyridine). CO was introduced at room tem-
perature by doses (from 2.6 to 664 mmol g21) up to an
1378
Chem. Commun., 2001, 1378–1379
This journal is © The Royal Society of Chemistry 2001
DOI: 10.1039/b102867g

Fig. 2 Correlation between propene formation rate (:) and abundance of
strong” Brönsted sites as evidenced by lutidinium peaks area after
“
desorption at 523 K (.).
x
a minimum degree of polymeric WO units may be necessary
for the formation of acidic sites. A similar behaviour was
observed for WO /Al catalysts and was attributed to an
enhanced stabilisation of the proton by delocalisation among
x
2 3
O
adjacent W atoms.1
3,14
This interpretation is consistent with the
observed lack of activity for low loading catalysts where W is
reportedly present in a monomeric form. These results highlight
the importance of extended W coverage for the development of
Brønsted acid sites on supported catalysts.
Notes and references
Fig. 1 Evolution of the lutidinium band. (a) IR spectra of lutidine evacuated
at 423 K. (b) Evolution of the area of the band (1620–1660) after desorption
of lutidine at 423 K (-) and 523 K (.). [Numbers in (a) refer to the W
loading in W atom nm2 ].
1 M. Hino and K. Arata, J. Chem. Soc., Chem. Commun., 1987, 1259.
2 N. Vaidyanathan, Ph. D. Thesis, University of Pittsburgh, 1998.
3 D. G. Barton, S. L. Soled, G. D. Meitzner, G. A. Fuentes and E. Iglesia,
J. Catal., 1999, 181, 57.
2
from thermodynamics. These results are consistent with acid
catalysis. In Fig. 2 the rate of propene formation was plotted vs.
4 J. G. Santiesteban, J. C. Vartuli, S. Han, R. D. Bastian and S. D. Chang,
J. Catal., 1997, 168, 431.
7
5
6
7
M. Scheithauer, T.-K. Cheung, R. E. Jentoft, R. K. Grasselli, B. C. Gates
and H. Knözinger, J. Catal., 1998, 180, 1.
D. Gazzoli, M. Valigi, R. Dragone, A. Marucci and G. Mattei, J. Phys.
Chem. B, 1997, 101, 11129.
C. Lahousse, F. Maugé, J. Bachelier and J. C. Lavalley, J. Chem. Soc.,
Faraday Trans., 1995, 91, 2907.
W content. The results show a very low rate up to 1.4 atom W
2
2
nm , followed by a steep increase for higher loadings. Similar
behaviour was observed for measurements performed with
different flows and temperatures. Fig. 2 compares the evolution
of catalytic activity with that of the intensity of lutidinium band.
Note that the onset of catalytic activity coincides with the
appearance of ‘strong’ Brønsted sites. In addition, the evolution
of the rate of decomposition parallels that of the abundance of
Brønsted acid sites. The decomposition of propan-2-ol appears,
thus, directly related to the presence and abundance of ‘strong’
Brønsted acid sites. Note that in the present study Brønsted acid
sites are formed prior to any reductive treatment. This is in
variance with previous studies12 where strong Brønsted acid
8 B. K. Kim, J. W. Hahn and K. R Han, J. Mater. Sci. Lett., 1997, 16,
669.
9 B. Zhao, X. Xu, J. Gao, Q. Fu and Y. Tang, J. Raman Spectrosc., 1996,
2
7, 549.
0 M. A. Vuurman, I. E. Wachs and A. M. Hirt, J. Phys. Chem., 1991, 95,
928.
1
9
1
1
1 I. E. Wachs, Catal. Today, 1996, 27, 437.
2 S. Kuba, P. Concepción Heydorn, R. K. Grasselli, B. C. Gates, M. Che
and H. Knözinger, Phys. Chem. Chem. Phys., 2001, 3, 146.
3 S. L. Soled, G. B. McVicker, L. L. Murrell, L. G. Sherman, N. C.
Dispenziere, S. L. Hsu and D. Waldman, J. Catal., 1988, 111, 286.
x 2
sites were reportedly formed on reduction of WO clusters in H
or by the reactant mixture. The appearance and the evolution of
the abundance of the Brønsted sites on our catalysts suggest that
1
14 D. G. Barton, S. L. Soled and E. Iglesia, Top. Catal., 1998, 6, 87.
Chem. Commun., 2001, 1378–1379
1379