Aldol condensation reactions of acetone and formaldehyde over vanadium phosphate catalysts: Comments on the acid-base properties

Article abstract of DOI:10.1039/b108794k

The self-aldol condensation of acetone and the condensation of acetone with formaldehyde over vanadium phosphate catalysts was studied. Catalysts containing only acid sites, i.e., the V⁴⁺ phases (VO)₂P₂O₇ and VOHPO₄·0.5H₂O were selective for the formation of isophorone from acetone alone and methyl vinyl ketone from the reaction of acetone and formaldehyde. Catalysts containing V⁵⁺ phases showed basic and acidic sites and were not selective to these products and only formed hydrocarbons. However, for the Aldol condensation of acetone and the decomposition of 2-methyl-3-butynol (MBOH), the catalyst activity was short lived and deactivation was observed due to coke formation on the catalyst surface. For the reaction of MBOH, the data gave a linear relationship for a Cremer-Constable plot due to the high heat of adsorption of MBOH dominating the energies of the overall catalyzed reaction.

Full text of DOI:10.1039/b108794k

View Article Online / Journal Homepage / Table of Contents for this issue
Aldol condensation reactions of acetone and formaldehyde over
vanadium phosphate catalysts: Comments on the
acid–base properties
a
Richard Tanner,a Philip Gill,a Richard Wells,a Jillian E.Bailie, Gordon Kelly,b
b
S.David Jackson and Graham J.Hutchings *a
a
Department of Chemistry, Cardiff University, P.O. Box 912, Cardiff, UK CF10 3TB
E-mail: hutch@cardiff.ac.uk
Synetix, P.O. Box 1, Billingham, Cleveland, UK TS23 1LB
b
Received 27th September 2001, Accepted 3rd December 2001
First published as an Advance Article on the web 10th January 2002
The acid–base properties of vanadium phosphate catalysts are investigated using the aldol condensation of
acetone and the reactions of 2-methylbut-3-yn-2-ol (MBOH). Three well characterised samples of
VOHPO4ꢀ0.5H2O were prepared using the reaction of V2O5 and H3PO4 with aqueous hydrochloric acid or
isobutanol as reducing agents, or from the reaction of VOPO4 .2H2O with isobutanol. (VO)2P2O7 , prepared by
heating VOHPO4ꢀ0.5H2O in He (8 h, 750 ꢁC), before and following partial oxidation in air or butane=air, and
aI-VOPO4 were also investigated. The reaction of MBOHwas used to probe the nature of the acid–base
properties of the vanadium phosphates. The V4þ phases (VOHPO4ꢀ0.5H2O and (VO)2P2O7) exhibited only
acidic active sites, whereas the V5þ phases (aI-VOPO4 and oxidised (VO)2P2O7) exhibited some basic sites in
addition to the acid sites. For the aldol condensation reactions of acetone, the V4þ phases were found to be
selective for the formation of isophorone from acetone alone and methyl vinyl ketone from the reaction of
acetone and formaldehyde. In contrast, vanadium phosphate catalysts containing V5þ phases are not selective
to these products and only form hydrocarbons (typically isobutane and isobutene). For all these reactions, the
catalyst activity is short lived and the deactivation that is observed is due to the surface becoming fouled by the
adsorption of products of polymerisation of the reaction products. However, the catalyst reactivity can be
restored by a simple oxidation treatment. The nature of active sites in n-butane oxidation to maleic anhydride is
also discussed and it is concluded that basic sites are required in addition to acidic surface sites for the selective
formation of maleic anhydride. For the reaction of MBOH, the data are found to give a linear relationship for a
Cremer–Constable plot and this is discussed in terms of the enthalpy of adsorption of MBOH.
observation that well defined single phase vanadium phos-
Introduction
phates can readily be prepared, and most interest has focused on
VOHPO4ꢀ0.5H2O, (VO)2P2O7 and aI-VOPO4 . Vanadium
phosphate catalysts have also been shown by Ai35 to be effective
catalysts for the aldol condensation of formaldehyde with acetic
acid to form methyl vinyl ketone and the aldol condensation of
acetone and formaldehyde to form methyl vinyl ketone.36 Ai
considers this to be due to the combination of acid and base
properties of vanadium phosphates.37 However, the vanadium
phosphate prepared by Ai comprises a complex mixture of
vanadium phosphate phases, and it is possible that the catalytic
activity observed in these studies was a composite of the activ-
ities of the component phases. In view of this, we have studied
the self-aldol condensation of acetone and the condensation
of acetone with formaldehyde over well defined vanadium
phosphate catalysts. In addition, we have characterised the
properties of vanadium phosphates using the reaction of
2-methylbut-3-yn-2-ol38–40 as a probe for acid–base sites.
Heterogeneous acid–base catalysts are extensively used in
industrial processes,1–6 including alkylation, isomerisation,
amination, cracking, aldol condensations and ether formation.
A recent review by Tanabe and Holderich7 has classified these
industrial processes and it is noted that acid catalysed pro-
cesses are predominant at the present time; since of 127
industrial processes using acid–base heterogeneous catalysts,
103 used acid catalysts. Recently, there has been considerable
interest in extending the use of heterogeneous catalysis in the
fine chemicals industry, and a particular reaction of interest
has been aldol condensations, since these can be catalysed
effectively by both acids and bases. In this respect, aldol con-
densation reactions using acetone as a substrate, are viewed as
model reactions for the study of acid and base catalysts. These
reactions have been extensively studied using base cata-
lysts, such as MgO and alkali doped MgO,8–13 alkali doped
silica,14–16 acid catalysts, such as zeolites and related micro-
porous materials17–21 and layered double hydroxides,12,22–29
since these have acid–base properties.
Experimental
Vanadium phosphate catalysts represent a class of materials
that are amongst the most well studied and characterised het-
erogeneous catalysts. They have been extensively used for the
oxidation of alkanes,30–34 in particular n-butane oxidation to
maleic anhydride. Part of the research interest is due to the
Preparation of vanadium phosphates
aI-VOPO4 was prepared according to a literature method41
and the structure was confirmed using powder X-ray diffrac-
688
Phys. Chem. Chem. Phys., 2002, 4, 688–695
This journal is # The Owner Societies 2002
DOI: 10.1039/b108794k

View Article Online
tion and laser Raman spectroscopy. (VO)2P2O7 was prepared
by heating VOHPO4ꢀ0.5H2O in dry He at 750 ꢁC for 8 h,40
again the structure was confirmed using powder X-ray dif-
fraction and laser Raman spectroscopy. A sample of a catalyst,
previously prepared and described by Ai35–37 was prepared
with a V : P atomic ratio of 1 : 1.06 according to the following
procedure. V2O5 (15.45 g, Strem) was refluxed in benzyl
alcohol (60 ml, Aldrich) and isobutanol (40 ml, Aldrich) for 2
h. H3PO4 (20.76 g, 85%, Aldrich) was added and the mixture
was refluxed for 2 h. The solid was recovered by vacuum fil-
tration and washed with acetone (100 ml) and water (100 ml)
and dried in air (2 h, 110 ꢁC). The solid was calcined in air (6 h,
300 ꢁC; 6 h, 450 ꢁC) prior to use as a catalyst.
standard laboratory microreactor. The combined acetone :
formaldehyde feed rate was maintained at 0.2 g gÀ1 (catalyst)
h
À1. The products were analysed using on-line gas chromato-
graphy and the carbon mass balance was typically 97 Æ 5%
depending on the level of coke deposition.
Reaction of 2-methylbut-3-yn-2-ol (MBOH). MBOH(Aldrich)
was fed using a calibrated syringe pump at a rate of 2.5 g hÀ1 to a
vaporiser and mixed with helium (15 ml minÀ1). The reactants
were fed via heated lines to a laboratory microreactor containing
heated catalyst (1.0 g) diluted with glass beads (Sigma, 710–1180
mm). The products were collected following cooling of the exit
gases and analysed using gas chromatography and the carbon
mass balance was typically 97 Æ 5% depending on the level of
coke deposition.
VOHPO4ꢀ0.5H2O was prepared according to the following
three methods.
VPA method. V2O5 (7.26 g, Aldrich) was dissolved in aqu-
eous HCl (80 ml, 37%, Aldrich), and refluxed for 3 h. H3PO4
(10.47 g, 85%, Aldrich) was added and the mixture refluxed for
3 h. The solvent was removed by distillation and the eventual
slurry dried in air (120 ꢁC, 16 h) to give a blue-green solid. The
solid was refluxed with water (10 ml gÀ1) for 3 h, filtered hot
and the blue-green solid was dried in air (120 ꢁC, 16 h).
Results and discussion
Aldol condensations of acetone over vanadium phosphate
catalysts
VOHPO4ꢀ0.5H2O was prepared according to the VPO method.
The material was then heated in helium at 750 ꢁC for 8 h to
form (VO)2P2O7 , which was confirmed by X-ray powder dif-
fraction, laser Raman spectroscopy and electron microscopy
and the material gave identical characterisation data to that we
have published previously for this preparation method.34
Acetone conversion was then investigated over the sample of
(VO)2P2O7 and the results are shown in Fig. 1. Below 300 ꢁC,
there was very little conversion of acetone. However, above
300 ꢁC acetone was converted primarily to isophorone and
selectivities >90% were observed, but at 300 ꢁC only iso-
phorone was observed. Traces of mesityl oxide and light
hydrocarbons, e.g. isobutane and isobutene, together with
carbon monoxide, were also observed.
Slow deactivation of the catalyst was observed at 400 ꢁC but,
during this deactivation, the selectivity did change markedly,
indicating the deactivation is due primarily to the loss of active
surface=active site. Characterisation of the catalysts after
reaction using powder X-ray diffraction and laser Raman
spectroscopy confirmed that the material was unchanged fol-
lowing the reaction and was still well crystalline (VO)2P2O7 .
However, analysis revealed that 2–3 wt.% carbon was present
on the used catalyst. This carbonaceous material was probably
the result of further reactions of acetone with the reaction
products on the catalyst surface. Hence, the deactivation is
considered to be the result of the surface being deactivated by
reaction products. To examine this, the catalyst that had been
reacted with acetone at 400 ꢁC was treated in flowing air (30 ml
VPO method. V2O5 (11.8 g, Aldrich) and H3PO4 (16.49 g,
85%, Aldrich) were refluxed in isobutanol (250 ml) for 16 h.
The solid was recovered by filtration, washed with isobutanol
(200 ml) and ethanol (250 ml), refluxed with water (10 ml gÀ1
)
for 3 h, filtered hot and the blue solid was dried in air (120 ꢁC,
16 h).
VPD method. V2O5 (11.8 g, Aldrich) was refluxed for 8 h
with H3PO4 (115 g, 85%, Aldrich) and water (300 ml). The
yellow VOPO4 .2H2O was recovered by hot filtration, washed
with water and acetone and dried in air (110 ꢁC, 16 h).
VOPO4 .2H2O (8 g) was refluxed in isobutanol (160 ml) for 21
h. The blue solid was recovered by hot filtration, washed with
isobutanol (200 ml) and ethanol (250 ml), dried in air (110 ꢁC,
2 h) and refluxed in water (10 ml gÀ1) for 3 h, filtered hot,
washed with water and dried in air (120 ꢁC, 16 h).
Characterisation methods
Powder X-ray diffraction was performed using an ENRAF
Nonius FR590 X-ray generator with a Cu-Ka source fitted
with an Inel CPS 120 hemispherical detector. BET surface area
measurements using nitrogen adsorption were carried out
using a Micromeritics ASAP 2000 instrument. Laser Raman
spectra were obtained using a Renishaw Ramascope Specto-
graph fitted with a green Ar+ laser (l ¼ 514.532 nm) typically
operated at 20 mW. Scanning electron microscopy (SEM) was
performed using a Hitachi S-2460N SEM operating at an
accelerating voltage of 25 kV.
Catalyst testing
Acetone reactions. Acetone (Fischer, 99%) was fed using a
calibrated syringe pump to a vaporiser. The acetone vapour
was mixed with argon as a carrier gas (30 ml minÀ1) and fed to
a laboratory microreactor and reacted with the heated catalyst
(0.5 g). The acetone flow rate was 0.2 g (acetone) gÀ1 (catalyst)
h
À1. The products were analysed using on-line gas chromato-
graphy. When acetone was reacted with formaldehyde, triox-
ane (Aldrich, 99%) was used as a source of formaldehyde. In
this case, a mixture of acetone and trioxane was fed via a
calibrated syringe pump, together with argon (30 ml minÀ1), to
a cracker at 380 ꢁC where trioxane decomposed to for-
maldehyde. The acetone : formaldehyde molar ratio was 5 : 1.
The acetone : formaldehyde vapour was cooled to the desired
reaction temperature and reacted with a catalyst (0.5 g) in a
&
Fig.1 Aldol condensation of acetone over (VO)2P2O7. Key:
400 ꢁC; m 300 ꢁC; * 400 ꢁC reactivated; closed symbols ¼ acetone
conversion; open symbols ¼ isophorone selectivity.
Phys. Chem. Chem. Phys., 2002, 4, 688–695
689

View Article Online
minÀ1) for 16 h at 500 ꢁC, conditions which are known to
remove the carbonaceous material. Analysis confirmed that
the coke had been removed but laser Raman spectroscopy
indicated that a small amount of oxidation had occurred and
some VOPO4 phases were also formed. The material was
subsequently heated in helium at 750 ꢁC for 8 h to reform with
crystalline (VO)2P2O7 . When the reactivated (VO)2P2O7 was
used as a catalyst for acetone conversion (Fig. 1), a lower
conversion was observed compared to fresh (VO)2P2O7 ,
however, the selectivity was very similar to that of the first
material and isophorone was observed as the only major
product.
aI-VOPO4 was prepared and used as a catalyst for acetone
conversion (Fig. 2). aI-VOPO4 was found to be significantly
more active for acetone conversion when compared with
(VO)2P2O7 , and conversions were observed initially at 150 ꢁC.
However, the catalyst was not selective for isophorone for-
mation and only traces (<5%) of mesityl oxide were observed.
The major products were hydrocarbons and carbon oxides. aI-
VOPO4 deactivated rapidly under these reaction conditions
and the used catalysts were found to contain typically 55
wt.% carbon. X-ray powder diffraction and laser Raman
spectroscopy confirmed that the material remained aI-VOPO4
following reaction and, consequently, the rapid deactivation is
again considered to be due to polymerisation of acetone with
reaction products on the catalyst surface. The catalyst samples
were reactivated by heating in air (30 ml minÀ1, 16 h, 500 ꢁC)
and were then re-used as catalysts for acetone conversion (Fig.
2). In this case, the catalyst activity was almost restored to the
original levels, and the reaction products were also identical to
those of the fresh catalysts.
A sample of a vanadium phosphate catalyst (V : P ¼ 1 : 1.06)
was prepared according to the method of Ai35–37 and calcined
in air (450 ꢁC, 1.5 ml minÀ1, 6 h) as described by Ai36,37 in the
previous studies in which vanadium phosphorus catalysts have
been investigated as catalysts for the aldol condensation of
acetone. Powder X-ray diffraction of this material showed that
the material comprised a mixture of phases, mainly (VO)2P2O7
and d-VOPO4 , together with amorphous material. Acetone
conversion was investigated over this material (Fig. 3) and, at
400 ꢁC, acetone was converted mainly to hydrocarbons (iso-
butane and isobutene), and some carbon oxide. In addition, ca.
20% mesityl oxide was formed initially. The catalyst rapidly
deactivated and was found to contain ca. 5% carbon following
reaction.
Fig.3 Aldol condensation of acetone over vanadium phosphate (V :
P ¼ 1 : 1.06) at 400 ꢁC. Key:
&
acetone conversion; m mesityl oxide
selectivity.
this oxidation treatment, the catalyst was investigated for
acetone conversion at 400 ꢁC and the results are given in Fig. 4.
It is apparent that even a short oxidation treatment (0.5 h)
of the vanadium phosphate significantly increases the cata-
lyst activity for this reaction, when compared with the
untreated (VO)2P2O7 (Fig. 1). Longer oxidation treatments of
(VO)2P2O7 give more active catalysts. However, all the
oxidised catalysts show a marked deactivation due to deposi-
tion of carbonaceous materials. All oxidised catalysts are
initially selective only for the formation of hydrocarbons and
carbon monoxide, as noted for aI-VOPO4 (Fig. 2). However,
as the catalysts deactivate, mesityl oxide and, subsequently,
In
(VO)2P2O7 , prepared using the VPO method as described
a subsequent set of experiments, well crystalline
earlier, was oxidised in flowing air in the reactor (30 ml minÀ1
)
at 400 ꢁC for a range of times (0.5, 1, 2 and 24 h). Following
Fig.4 Aldol condensation of acetone over (VO)2P2O7 oxidised in air
at 400 ꢁC for different times, (a) acetone conversion; (b) selectivity.
&
Fig.2 Aldol condensation of acetone over aI-VOPO4. Key:
400 ꢁC; m 300 ꢁC; * 150 ꢁC; closed symbols ¼ fresh catalyst; open
Key:
0.5 h; m 1.0 h; * 2.0 h; ^ 24 h. For selectivity, open sym-
&
symbols ¼ reactivated catalyst.
bols ¼ isophorone selectivity; closed symbols ¼ mesityl oxide.
690
Phys. Chem. Chem. Phys., 2002, 4, 688–695

View Article Online
absence of acetone resulted in extremely rapid deactivation
and no volatile reaction products were observed. The acetone–
formaldehyde reaction was also investigated over aI-VOPO4
and the results at 400 ꢁC are also shown in Fig. 5. aI-VOPO4
was much less active for this reaction, when compared with
(VO)2P2O7 , but methyl vinyl ketone was observed initially as
the major product.
Catalyst characterisation
It is apparent that the oxidation of (VO)2P2O7 , derived from
the VPO method, at 400 ꢁC significantly affected the catalyst
activity and selectivity for the reaction of acetone to iso-
phorone. The VOHPO4ꢀ0.5H2O, (VO)2P2O7 and (VO)2P2O7
treated with air were characterised using scanning electron
microscopy, powder X-ray diffraction and laser Raman
spectroscopy and the results are shown in Fig. 6–8. It is
apparent that the morphology of the VOHPO4ꢀ0.5H2O pre-
cursor is retained in the (VO)2P2O7 prepared by heating the
precursor in helium at 750 ꢁC. Oxidation of the sample also
does not change the morphology of the material. In addition,
oxidation of the VPO-derived (VO)2P2O7 does not sig-
nificantly affect the bulk structure of the material as there are
no significant changes in the powder X-ray diffraction pattern
or the laser Raman spectrum. In a previous study, we have
shown42 that (VO)2P2O7 derived from the VPO method is
significantly more resistant towards oxidation when com-
pared with VPA- and VPD-derived (VO)2P2O7 . It is, there-
fore, apparent, that the oxidation treatment must just affect
the surface of the catalyst. Indeed, Volta and co-workers43,44
have completed a detailed study of the effects of oxidation of
(VO)2P2O7 derived from the VPO preparation method. Using
a range of techniques, principally UV–vis, Raman and XPS
spectroscopy, they showed that oxidation of the surface
occurred and aI-VOPO4 was formed. Previous studies45,46
also confirmed that the surface of vanadium phosphate
crystallites can be readily oxidised by short exposure to
oxygen at elevated temperatures.
Fig.5 Aldol condensation of acetone and formaldehyde over vana-
ꢁ
&
dium phosphates at 400 C. Key:
(VO)2P2O7 ; * aI-VOPO4 ; closed
symbols ¼ methyl vinyl ketone selectivity.
isophorone are observed. At long reaction times, the iso-
phorone selectivity returns to a level similar to that observed
for the untreated (VO)2P2O7 for the samples oxidised for a
short time (42 h). This is not observed with the catalyst
subjected to the 24 h oxidation pre-treatment.
In a subsequent experiment, a sample of well crystalline
(VO)2P2O7 derived from the VPO method, as described earlier,
was used as a catalyst for the reaction of acetone with for-
maldehyde (acetone : formaldehyde ¼ 5 : 1 mol ratio) and the
results are shown in Fig. 5 for the reaction at 400 ꢁC and, at
this temperature, the formaldehyde conversion was 100%.
Similar trends in conversion and selectivity are observed at
150 ꢁC and 300 ꢁC, although the initial acetone conversions are
lower, 11% and 9% respectively. For all reaction con-
ditions, methyl vinyl ketone was formed, almost exclusively.
Interestingly, no isophorone or mesityl oxide were observed
when formaldehyde was co-fed with acetone. As observed with
the reaction of acetone alone over (VO)2P2O7 , the cata-
lyst activity decreased with time-on-stream. However, the
deactivation rate was more marked when formaldehyde was
present. Analysis of the catalyst following the reaction con-
firmed that carbonaceous material was formed (ca. 5 wt.%).
Since the formaldehyde conversion remained at 100%
throughout the timescale of the experiment, whereas the
acetone conversion and yield of methyl vinyl ketone
decreased, it is considered that the enhanced deposition of
carbonaceous material results mainly from the self con-
densation of formaldehyde on the catalyst surface. Indeed,
experiments in which formaldehyde was reacted in the
The effect of surface oxidation is further confirmed by the
changes in selectivity observed for the oxidised samples for
the reaction of acetone (Fig. 4). Initially, only hydrocarbons
(i.e. isobutane and isobutene) were observed and these
products are typical of aI-VOPO4 , a V5þ vanadium phos-
phate (Fig. 2). However, as the catalyst deactivates, the
selectivity gradually becomes that expected for untreated
(VO)2P2O7 and isophorone is observed as the major pro-
duct. This indicates that the effect of the oxidation treat-
ments for times 42 h are short lived for the acetone
conversion reaction.
Fig.6 Scanning electron micrographs of vanadium phosphates prepared using VPO method, (a) VOHPO4ꢀ0.5H2O; (b) (VO)2P2O7 ; (c)
(VO)2P2O7 oxidised in air for 2 h at 400 ꢁC.
Phys. Chem. Chem. Phys., 2002, 4, 688–695
691

View Article Online
Fig.7 Powder X-ray diffraction patterns of vanadium phosphates
prepared using the VPO method, (a) VOHPO4ꢀ0.5H2O; (b)
(VO)2P2O7 ; (c) (VO)2P2O7 oxidised in air for 2 h at 400 ꢁC.
Scheme 1
this indicates that water formed during the dehydration of
MBOHon the acidic sites has no effect on the catalyst struc-
ture during the short timescale of the experiments. It is known
that, during the selective oxidation of n-butane to maleic
anhydride, water is also formed and the catalyst structure is
relatively resistant to this product.30–33 However, the catalysts
all contained carbon (ca. 1–2 wt.%). Since the catalyst selec-
tivity did not change during the deactivation, it is concluded
that the deactivation is due to the deposition of carbonaceous
material on the surface of the vanadium phosphate. This is
probably due to acid catalysed polymerisation of the MBOH.
It was subsequently shown that the carbon could be readily
removed by an oxidation treatment. A sample of (VO)2P2O7
prepared using the VPO method, and deactivated by reaction
with MBOHat 180 ꢁC, was oxidised in flowing air at 500 ꢁC
for 2 h. Following the oxidation treatment, the material was
heated in helium at 750 ꢁC for 8 h to reform (VO)2P2O7 , which
was confirmed by powder X-ray diffraction and laser Raman
spectroscopy. When this reactivated (VO)2P2O7 was reacted
with MBOH, similar results to the fresh catalyst were
obtained.
Fig.8 Laser Raman spectra of vanadium phosphates prepared using
the VPO method, (a) VOHPO4ꢀ0.5H2O; (b) (VO)2P2O7 ; (c) (VO)2P2O7
oxidised in air for 2 h at 400 ꢁC.
In a subsequent experiment, (VO)2P2O7 derived from the
VPO method was oxidised in air (400 ꢁC, 30 ml minÀ1, 2 h),
and the oxidised sample was reacted with MBOHat 180 ꢁC,
and the results are shown in Fig. 10. Again, a rapid deacti-
vation is observed but, in this case, a low selectivity is observed
for base-catalysed products (acetone, acetylene) together with
Characterisation of the acid–base nature of vanadium phosphate
catalysts using the reaction of 2-methyl-but-3-yn-2-ol (MBOH)
Lauron-Pernot et al.39 have shown that the reactions with
MBOHcan be used to characterise the acid–base sites on
catalyst surfaces (Scheme 1). Interestingly, although a number
of acidic and basic catalysts have been characterised using this
method, vanadium phosphates have not been investigated to
date. Consequently, in this study, we present the data for
VPA-, VPO- and VPD-derived vanadium phosphates. In
addition, the reaction profiles for the reaction of MBOHwith
time-on-stream has not been discussed in previous papers and
this aspect is addressed in this study.
MBOHwas reacted at 180 ꢁC over the three VOH-
PO4ꢀ0.5H2O samples prepared using the VPA, VPO and VPD
methods and the results are shown in Fig. 9. In addition, the
results for the reaction of MBOHat 180 ꢁC (VO)P2O7 derived
from VOHPO4ꢀ0.5H2O on heating in helium at 750 ꢁC for 8 h,
are also shown in Fig. 9. Only the products characteristic of
the acid catalysed decomposition of MBOHare observed. For
all samples, a rapid deactivation in the MBOHconversion was
observed, but this did not affect the product selectivity. Ana-
lysis of the samples following reaction confirmed that the
catalyst structure, composition and morphology were
unchanged for the VOHPO4ꢀ0.5H2O and (VO)2P2O7 materials,
Fig.9 Conversion of MBOHover vanadium phosphates at 180 ꢁC.
&
&
Key: VOHPO4ꢀ0.5H2O:
VPA, *VPO, m VPD; (VO)2P2O7:
VPA, ꢂ VPO, 4 VPD.
692
Phys. Chem. Chem. Phys., 2002, 4, 688–695

View Article Online
Fig.10 Conversion of MBOHover vanadium phosphate catalysts at
Fig.11 Conversion of MBOHover vanadium phosphate catalysts at
180 ꢁC pre-activated by heating in 1.5% butane=air, 400 ꢁC, 72 h. Key:
180 ꢁC, (a) conversion, (b) selectivity. Key:
(VO)2P2O7 oxidised in
&
air at 400 ꢁC for 2 h;* aI-VOPO4 ; m VOPO4ꢀ0.5H2O; open sym-
bols ¼ acid catalysed products, closed symbols ¼ base catalysed pro-
ducts.
&
VPA; * VPO; m VPD; open symbols ¼ acid catalysed products;
closed symbols ¼ base catalysed products.
comparative purposes, the vanadium phosphate catalyst pre-
pared by the method of Ai,35–37 as described earlier, was also
reacted with MBOHand the results are also given in Fig. 10.
Again, the products of basic, as well as acid, decomposition of
MBOHare observed together with rapid deactivation.
the products for the acid catalysed decomposition of MBOH.
This indicates that the oxidised catalyst surface43,44 now con-
tains a low concentration of basic sites, together with a higher
concentration of acidic sites. These basic sites are probably
associated with V5þ–V4þ sites that are formed on oxidation,
and are known to be present in the surface of oxidised vana-
dium phosphate catalysts.43,44,47,48
Comments on reactivity of acid–base sites on vanadium
phosphates
aI-VOPO4 and VOPO4ꢀ2H2O were reacted with MBOHat
180 ꢁC and the results are also given in Fig. 10. Again, a rapid
deactivation is observed with these V5þ vanadium phosphates.
However, it is apparent that aI-VOPO4 gives a significant level
of product of base catalysed decomposition of MBOH, in
addition to the products of the acid catalysed reaction.
From these studies, it is clear that oxidised vanadium phos-
phates have surfaces that comprise both acidic and basic sites;
whereas V4þ vanadium phosphates (i.e. well crystalline VOH-
PO4ꢀ0.5H2O, (VO)2P2O7) exhibit only acidic sites. Vanadium
phosphate catalysts are well known for the oxidation of n-
butane to maleic anhydride30–34 and, in most studies, the cat-
alysts are stabilised and reacted in 1.5% n-butane in air.33
Under these oxidising conditions, it is known that the surface
becomes partially oxidised30,43,44,47 and, hence, we have studied
the reaction of MBOHover vanadium phosphate catalysts
activated in this way. Three samples of VOHPO4ꢀ0.5H2O pre-
pared using the VPA, VPO and VPD methods, were pre-treated
in 1.5% n-butane in air at 400 ꢁC for 72 h. Under these con-
ditions, it is known34,49 that VOHPO4ꢀ0.5H2O transforms to
(VO)2P2O7 and VOPO4 phases, although the intrinsic activities
The reaction of MBOHover the vanadium phosphates reveals
very interesting differences for catalysts that are selective for
the oxidation of n-butane or aldol condensation reactions of
acetone. For the aldol condensation of acetone to form iso-
phorone or the reaction of acetone with formaldehyde to form
methyl vinyl ketone, (VO)2P2O7 is the most selective catalyst.
Indeed, oxidation of (VO)2P2O7 significantly decreases the
selectivity initially. (VO)2P2O7 , when characterised by the
reaction of MBOH, is found to comprise only acid sites. In
contrast, aI-VOPO4 or material prepared by the oxidation of
(VO)2P2O7 , also comprise base sites in combination with the
acid sites. It is clear that the basic sites are introduced on
oxidation of (VO)2P2O7 , but these sites are totally unselective
for the aldol condensation reactions. Hence, for aldol con-
densation reactions over vanadium phosphate catalysts, acid
sites are the preferred active site. This is in contrast to the
earlier studies of Ai36,37 in which it was concluded that a
combination of acid=base sites was important for these reac-
tions. It is interesting to comment further on the origin of the
hydrocarbon products obtained from aI-VOPO4 and oxidised
(VO)2P2O7 catalysts. Previously, we have shown51 that acetone
can be converted to isobutane and isobutene, together with the
formation of carbon oxides, over the acidic zeolite H-b at
400 ꢁC. In this case, we showed that the acetone was converted
(mol MA mÀ2 À1) are similar for the three catalysts.50 The
h
activated materials were then investigated for the reaction of
ꢁ
MBOHat 180 C and the results are given in Fig. 11. Again, the
catalysts all deactivate rapidly, but the products of both acid
catalysed and base catalysed reactions are observed. For
Phys. Chem. Chem. Phys., 2002, 4, 688–695
693

View Article Online
initially to phorone and isophorone, and this was the precursor
for the formation of butene and isobutene. It is possible that
formation of hydrocarbons from acetone over the vanadium
phosphate occurs via a similar reaction pathway.
experimental factors were determined for a range of oxides
(MgO and lanthanides) which are known to give only the
products associated with base-catalysed decomposition of
MBOH.54 These additional data were all found to agree with
the linear plot for the vanadium phosphates (Fig. 12). The
linear relationship between ln A and Eact is known as a Cre-
mer–Constable relation and is sometimes referred to as a
compensation effect. The Cremer–Constable relation has been
the subject of intense debate since the effect was first observed
in 192555 and the literature has recently been comprehensively
reviewed by Bond.56 The effect has been observed for the same
reaction (i.e. conversion of a single reactant to a specific pro-
duct) over a range of catalysts (e.g. the hydrogenolysis of
ethane over various silica-supported metals,56–59 or the
hydrogenation of CO2 over supported metal catalysts56). The
effect has also been observed for the groups of reactants over a
single catalyst, e.g. the hydrogenolysis of alkanes over
Pt=SiO2 .56,60,61 However, this is the first observation of the
relationship for the more complex catalysed processes of dif-
ferent reactions of a substrate over a broad range of oxide and
phosphate catalysts. Bond56 indicates that a large number of
explanations have been put forward to explain the origin of the
Cremer–Constable relation. In the present case, the common
step for the acid and base catalysed reactions is the adsorption
of the MBOHmolecule at the active sites. The adsorption of
the large MBOHmolecule probably dominates the energetics
of the overall catalysed process. This is consistent with the
observation of Bond, and indicates that care needs to be
exercised when using the reactions of complex probe molecules
to characterise heterogeneous catalyst. Indeed, the use of
smaller, less complex molecules may be preferred since, for
these molecules, the adsorption enthalpy may not dominate
the apparent activation energy of the reactions.
As stated earlier, the partial oxidation of n-butane to maleic
anhydride is known to require vanadium phosphate catalysts
that are partially oxidised.30-34 For example, it has been shown
that oxidation of (VO)2P2O7 with air at 400 ꢁC, in a similar
manner to that carried out in this study, significantly increases
the selectivity to maleic anhydride when these catalysts are
subsequently studied for n-butane oxidation.43 In addition, we
have shown that combinations of aI-VOPO4 and (VO)2P2O7
can give improved yields of maleic anhydride in this reaction.52
The in situ activation of VOHPO4ꢀ0.5H2O in n-butane=air
transforms the material to a mixture of V4þ and V5þ phases
and, during this process, the selectivity to maleic anhydride is
significantly increased.49 The reaction of MBOHover these
oxidised vanadium phosphates has shown that basic sites are
present, together with acidic sites, and it can be concluded that
the presence of these basic sites appears to be crucial with
respect to the formation of maleic anhydride on these surfaces.
This is probably related to the enhanced adsorption of n-
butane at the active site of the catalyst. On electronic grounds,
V5þ would be expected to be more acidic than V4þ since it has
a higher electron affinity and, consequently, a higher Lewis
acidity. The basic sites must arise from combination of V4þ
and V5þ and are probably associated with defect sites formed
on oxidation.53
Cremer–Constable relation for the reactions of MBOH
In the preceding studies of MBOHdecomposition on acidic
and basic materials,38–40 the activation energy for these reac-
tions has not been discussed. In the present study, the activa-
tion energies were determined from the initial reaction rates for
the reaction of MBOHover catalysts for a range of tempera-
tures (160–200 ꢁC). It is important that the initial rates are
used, since the catalysts rapidly deactivate during the reaction,
even at lower reaction temperatures. The activation energies
for the vanadium phosphates are found to be in the range 80–
200 kJ molÀ1. Interestingly, the plot of ln A versus the acti-
vation energy gives a linear plot (Fig. 12) for all the VOH-
PO4ꢀ0.5H2O, (VO)2P2O7 and the oxidised samples, even
though, in some cases, both acid- and base-catalysed reactions
are observed. In view of this, the activation energies and pre-
Conclusion
Vanadium phosphate oxide catalysts based on (VO)2P2O7 and
aI-VOPO4 have been investigated for aldol condensation of
acetone. In addition, the reaction of MBOHhas been used to
characterise the nature of the acid–base properties of the cat-
alysts. Catalysts containing only acid sites, i.e. the V4þ phases
(VO)2P2O7 and VOHPO4ꢀ0.5H2O, were found to be selective
for the formation of isophorone from acetone alone and
methyl vinyl ketone from the reaction of acetone and for-
maldehyde. Catalysts containing V5þ phases exhibit basic as
well as acidic sites and are not selective to these products and
only form hydrocarbons. However, for the aldol condensation
of acetone and the decomposition of MBOH, the catalyst
activity is short lived and deactivation is observed due to coke
formation on the catalyst surface. Catalyst reactivity can be
readily restored by an oxidation treatment. The results
observed have been used to consider the nature of the acid–
base sites in the selective conversion of n-butane to maleic
anhydride and, in this case, it is considered that a mixture of
basic and acidic sites is required to observe high selectivity.
Finally, for the reaction of MBOH, the data are found to give
a linear relationship for a Cremer–Constable plot. It is con-
cluded that this is caused by the high heat of adsorption of
MBOHdominating the energetics of the overall catalysed
reaction. This suggests that smaller, less complex, molecules
may be preferred for the characterisation of acid–base
properties.
Fig.12 Cremer–Constable plot for the reaction of MBOHover
oxide and phosphate catalysts. Key: 1. VOHPO4ꢀ0.5H2O VPO; 2.
(VO)2P2O7 , VPO; 3. VOHPO40.5H2O, VPA; 4. (VO)2P2O7 , VPA; 5.
VOHPO4.0.5H2O, VPD; 6. (VO)2P2O7 , VPD; 7. MgO ex carbonate; 8.
MgO ex hydroxide; 9. La2O3 ; 10. CeO2 ; 11. Nd2O3 ; 12. Gd2O3 ; 13.
Dy2O3 ; 14. Ho2O3 ; 15. Er2O3 .
Acknowledgements
We thank Synetix and the EPSRC for financial support. We
also thank Geoff Bond for valuable discussions concerning the
origin of the compensation effect.
694
Phys. Chem. Chem. Phys., 2002, 4, 688–695

View Article Online
32 G. Centi, F. Trifiro, G. Busca, J. Ebner and J. Gleaves, Faraday
Discuss., 1989, 87, 215.
References
33 G. J. Hutchings, Appl. Catal., 1992, 72, 1.
34 C. J. Kiely, A. Burrows, G. J. Hutchings, K. E. Bere, J. C. Volta,
A. Tuel and M. Abon, Faraday Discuss., 1996, 105, 103.
35 M. Ai, J. Catal., 1987, 107, 201.
36 M. Ai, J. Catal., 1987, 106, 273.
37 M. Ai, J. Catal., 1986, 100, 336.
38 C. Lahouse, J. Bachelier, J. C. Lavalley, H. Lauron-Pernot and
A.-M. Le Govic, J. Mol. Catal., 1994, 87, 329.
39 H. Lauron-Pernot, F. Luck and J. M. Popa, Appl. Catal., 1991, 78,
213.
1
2
3
4
M. Misono and N. Nojiri, Appl. Catal., 1990, 64, 1.
N. Nojiri and M. Misono, Appl. Catal. A, 1993, 93, 103.
J. N. Armor, Appl. Catal., 1991, 78, 141.
A. Chauvel, B. Delmon and W. F. Holderich, Appl. Catal. A,
1994, 115, 173.
W. F. H ldoerich, Stud. Surf. Sci. Catal., 1993, 75, 127.
J. A. Cusumano, CHEMTECH, 1992, 482.
K. Tanabe and W. F. Holderich, Appl. Catal. A, 1999, 181, 399.
S. Lippert, W. Baumann and K. Thomke, J. Mol. Catal., 1991, 69,
199.
5
6
7
8
40 N. E. Fouad, P. Thomasson and H. Knozinger, Appl. Catal. A,
2000, 194, 213.
9
J. I. DiCosimo, V. K. Diez and C. R. Apesteguia, Appl. Catal. A,
1996, 137, 149.
41 F. Ben Abdelouahab, R. Olier, N. Guillaume, F. Lefebvre and J.
C. Volta, J. Catal., 1994, 134, 151.
10 G. C. Podrebarac, F. T. T. Ng and G. L. Rempel, Chem. Eng. Sci.,
1997, 52, 2991.
42 F. J. Cabello Sanchez, R. P. K. Wells, C. Rhodes, C. J. Kiely and
G. J. Hutchings, submitted for publication.
11 J. I. DiCosimo and C. R. Apesteguia, J. Mol. Catal. A, 1998, 130,
177.
43 K. Aıt-Lachgar, M. Abon and J. C. Volta, J. Catal., 1997, 171,
383.
12 D. Tichit, M. N. Bennani, F. Figueras, R. Tessier and J. Ker-
vennal, Appl. Clay Sci., 1998, 13, 401.
44 K. Aıt-Lachgar, A. Tuel, M. Brun, J. M. Herrmann, J. M. Krafft,
J. R. Martin, J. C. Volta and M. Abon, J. Catal., 1998, 177, 224.
45 G. Koyano, F. Yamaguchi, T. Okuhara and M. Misono, Catal.
L ett., 1996, 41, 149.
46 Y. Schuurman, J. T. Gleaves, J. R. Ebner and M. J. Mummey, in
New Developments in Selective Oxidation II, ed. V. Cortes Cor-
´
13 S. C. Luo and J. L. Falconer, J. Catal., 1999, 185, 393.
14 W. J. Ji and Y. Chen, Gaodeng Xuexiao Huaxue Xuebao, 1997, 18,
277.
15 M. R. Gogate, J. J. Spivey and J. R. Zoeller, Catal. Today, 1997,
36, 243.
16 J. J. Spivey, M. R. Gogate, J. R. Zoeller and R. D. Bolberg, Ind.
Eng. Chem. Res., 1997, 36, 4600.
beran and S. Vic Bellon, Elsevier, Amsterdam, 1994, p. 203.
´
´
47 G. W. Coulston, S. R. Bare, H. Kung, K. Birkeland, G. K.
Methke, R. Harlow, N. Herron and P. L. Lee, Science, 1997, 275,
191.
48 J. T. Gleaves, J. R. Ebner and T. C. Knechler, Catal. Rev.Sci. Eng.,
1988, 30, 49.
17 C. O. Veloso, J. L. F. Monteiro and E. F. Souzaaguiar, Stud. Surf.
Sci. Catal., 1994, 84, 1913.
18 A. G. Panov and J. J. Fripiat, Catal. L ett., 1999, 57, 25.
19 Y. C. Chuang and A. N. Ko, Appl. Catal. A, 2000, 190, 149.
20 C. Flego and C. Perego, Appl. Catal. A, 2000, 192, 317.
21 A. Philippou and M. W. Anderson, J. Catal., 2000, 189, 395.
22 A. Guida, M. H. Lhouty, D. Tichit, F. Figueras and P. Geneste,
Appl. Catal. A, 1997, 164, 251.
23 K. K. Rao, M. Gravelle, J. S. Valente and F. Figueras, J. Catal.,
1998, 173, 115.
24 C. P. Keller and A. A. Schutz, Appl. Clay Sci., 1998, 13, 417.
25 M. N. Bennani, D. Tichit, F. Figueras and S. Abouarnadasse, J.
Chim. Phys. Phys-Chim. Biol., 1999, 96, 498.
26 D. Tichit and F. Fajula, Stud. Surf. Sci. Catal., 1999, 125, 329.
27 F. Prinetto, D. Tichit, R. Teissier and B. Coq, Catal. Today, 2000,
55, 103.
28 J. C. A. A. Roelofs, A. J. van Dillen and K. P. de Jong, Catal.
Today, 2000, 60, 297.
29 M. del Ario, S. Gutierrez, C. Martin and V. Rives, Phys. Chem.
Chem. Phys., 2001, 3, 119.
30 G. Centi, Forum on Vanadium Pyrophosphate Catalysis, Catal.
Today, 1994, 16.
31 E. Bordes, Catal. Today, 1987, 1, 499.
49 G. J. Hutchings, A. Desmartin-Chomel, R. Olier and J. C. Volta,
Nature, 1994, 368, 41.
50 C. J. Kiely, A. Burrows, S. Sajip, G. J. Hutchings, M. T. Sananes,
A. Tuel and J. C. Volta, J. Catal., 1996, 162, 31.
51 M. A. Aramendia, V. Borau, I. M. Garcia, C. Jimenez, A. Mar-
inas, J. M. Marinas, A. Porras and F. J. Urbano, Appl. Catal. A,
1999, 184, 115.
52 G. J. Hutchings and R. Higgins, J. Catal., 1996, 162, 153.
53 P. L. Gai and K. Kourtakis, Science, 1995, 267, 661.
54 G. J. Hutchings, P. Johnston, D. F. Lee, A. Warwick, C. D.
Williams and M. Wilkinson, J. Catal., 1994, 147, 177.
55 F. H. Constable, Proc. R.. Soc. L ondon, Ser. A, 1925, 108, 355.
56 G. C. Bond, Catal. Rev.Sci. Eng., 2000, 42, 323.
57 J. H. Sinfelt, Catal. Rev., 1969, 3, 175.
58 J. H. Sinfelt, Catal. Rev., 1974, 9, 147.
59 J. H. Sinfelt, Adv. Catal., 1973, 23, 91.
60 G. C. Bond, Appl. Catal. A, 2000, 191, 23.
61 S. D. Jackson, G. J. Kelly and G. Webb, J. Catal., 1998, 176, 225.
Phys. Chem. Chem. Phys., 2002, 4, 688–695
695