α-Pinene isomerisation over heteropoly acid catalysts in the gas-phase

Article abstract of DOI:10.1016/j.apcata.2010.10.018

The isomerisation of α-pinene was studied in the gas phase over solid heteropoly acid (HPA) catalysts in a fixed-bed continuous flow reactor at 200 °C and ambient pressure. The catalysts included bulk and SiO ₂-supported H₃PW₁₂O₄₀ and bulk Cs_2.5H_0.5PW₁₂O₄₀ possessing strong Br?nsted acid sites, as well as composites prepared by supporting 15 wt% H₃PW₁₂O₄₀ on Nb₂O₅, ZrO₂ and TiO₂ possessing both Br?nsted and Lewis acid sites of moderate strength. The reaction yielded camphene as the main product in a mixture with monoterpene by-products such as limonene, terpinolenes, terpinenes, β-pinene, p-cymene and others. The HPA catalysts with strong Br?nsted acid sites exhibited high initial activities, but suffered from catalyst deactivation, resulting in low camphene yields. Conversely, the HPA catalysts supported on Nb₂O₅, ZrO ₂ and TiO₂, although weaker acids, showed more stable performance in α-pinene isomerisation. The HPA catalyst supported on TiO₂ gave a camphene yield of 51% and a total camphene and limonene yield of 58%.

Full text of DOI:10.1016/j.apcata.2010.10.018

Applied Catalysis A: General 390 (2010) 219–224
Contents lists available at ScienceDirect
Applied Catalysis A: General
journal homepage: www.elsevier.com/locate/apcata
␣
-Pinene isomerisation over heteropoly acid catalysts in the gas-phase
1
∗
Ali Alsalme , Elena F. Kozhevnikova, Ivan V. Kozhevnikov
Department of Chemistry, University of Liverpool, Liverpool L69 7ZD, UK
a r t i c l e i n f o
a b s t r a c t
Article history:
The isomerisation of ␣-pinene was studied in the gas phase over solid heteropoly acid (HPA) catalysts
in a fixed-bed continuous flow reactor at 200 C and ambient pressure. The catalysts included bulk and
Received 13 September 2010
Received in revised form 14 October 2010
Accepted 15 October 2010
Available online 23 October 2010
◦
SiO2-supported H3PW12O40 and bulk Cs2.5H0.5PW12O40 possessing strong Brønsted acid sites, as well
as composites prepared by supporting 15 wt% H3PW12O40 on Nb2O5, ZrO2 and TiO2 possessing both
Brønsted and Lewis acid sites of moderate strength. The reaction yielded camphene as the main product
in a mixture with monoterpene by-products such as limonene, terpinolenes, terpinenes, ␤-pinene, p-
cymene and others. The HPA catalysts with strong Brønsted acid sites exhibited high initial activities,
but suffered from catalyst deactivation, resulting in low camphene yields. Conversely, the HPA catalysts
supported on Nb2O5, ZrO2 and TiO2, although weaker acids, showed more stable performance in ␣-pinene
isomerisation. The HPA catalyst supported on TiO2 gave a camphene yield of 51% and a total camphene
and limonene yield of 58%.
Keywords:
Heterogeneous acid catalysis
Heteropoly acid
␣
-Pinene isomerisation
©
2010 Elsevier B.V. All rights reserved.
1
. Introduction
green benefits [12–15]. For example, Showa Denko [14] and BP
16] have commercialised the synthesis of ethyl acetate from
[
Acid-catalysed isomerisation of ␣-pinene, a naturally occurring
ethene and acetic acid in the gas phase over silica-supported HPA.
The majority of catalytic applications use the most stable and easy
available Keggin HPAs comprising heteropoly anions of the for-
abundant monoterpene, is of industrial value to produce cam-
phene and limonene, important intermediates and ingrediants in
cosmetics, food and pharmaceutical industries [1]. Commercially,
n−
5+
4+
mula [XM12O40
]
, where X is the heteroatom (P , Si , etc.) and
6+
6+
␣
-pinene isomerisation is carried out in liquid phase over an acid-
M the addendum atom (Mo , W , etc.). HPAs possess stronger
acidity than conventional solid acid catalysts such as acidic oxides
and zeolites. The acid strength of Keggin HPAs decreases in the
treated TiO2 catalyst. The drawback to this process is its low rate
and poor selectivity, hence there is strong demand for new catalysts
exhibiting higher activity and selectivity and a continuous process
order:
H PW12O40 > H SiW12O40 > H PMo12O40 > H SiMo12O40
3 4 3 4
[
1].
[12,13]. Usually, H PW12O40 is the catalyst of choice because of
3
␣
-Pinene isomerisation has been extensively studied with a
its stronger acidity and higher thermal stability [12–15]. Bulk
HPAs, however, have very low surface areas. Therefore, supported
HPA catalysts, usually HPA on silica, are often used. Recently,
H PW12O40 supported on SiO , TiO2 and ZrO2 has been studied in
wide range of solid acid catalysts in liquid and gas phases [1].
Amongst the catalysts are zeolites [2–4], activated clays [5,6], sul-
fated zirconia [7], ion exchange resins [8] and heteropoly acids
3
2
(
HPAs) [9–11]. Typically, 30–50% yields of camphene together with
␣-pinene isomerisation in the liquid phase [9–11]. In solvent-free
◦
a complex mixture of other monoterpene by-products have been
obtained. The reaction is suggested to proceed through protona-
tion of the ␣-pinene double bond to form pinanyl cation, followed
by two possible transformations: a ring opening to monocyclic
products (limonene, terpinolene, terpinene, p-cymene, etc.) and a
rearrangement to bi- and tricyclic products (camphene, ␤-pinene,
tricyclene, etc.) (Scheme 1) [1].
␣-pinene isomerisation over 20% H PW12O40/SiO2 at 100 C, a 45%
3
camphene yield has been obtained [11].
In recent years, there has been considerable activity focussing on
acidic composites comprising tungsten HPAs and Nb(V), Zr(IV) or
Ti(IV) oxides as an oxide matrix [17–26]. These composites are usu-
ally prepared by wet chemical synthesis, followed by calcination at
◦
500–850 C. These materials have been found active in a range of
Heterogeneous acid catalysis by HPAs has attracted much
interest because of its potential of great economic rewards and
acid-catalysed reactions in liquid phase, often with good catalyst
recycling. Their catalytic activity is often lower than that of the
“standard” solid HPA catalysts, such as bulk and silica-supported
H PW12O40 and bulk Cs2.5H0.5PW12O40 [26]. This, however, can be
3
compensated for by a better performance stability and recyclabil-
∗ Corresponding author. Fax: +44 151 794 3588.
ity of these composites. In contrast to the parent acid, H PW12O40
,
3
E-mail address: I.V.Kozhevnikov@liverpool.ac.uk (I.V. Kozhevnikov).
1
possessing strong Brønsted acid sites, the catalysts supported on
Present address. King Saud University, PO Box 2455, Riyadh 11451, Saudi Arabia.
0
926-860X/$ – see front matter © 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.apcata.2010.10.018

2
20
A. Alsalme et al. / Applied Catalysis A: General 390 (2010) 219–224
Scheme 1. Acid-catalysed isomerisation of ␣-pinene.
TiO , ZrO and Nb O5 have both Brønsted and Lewis acid sites,
with the latter mainly originating from the oxide support [26].
The strength of acid sites in these catalysts is weaker than that in
lowed by drying in a rotary evaporator. Finally, the catalyst was
calcined in air at 300 or 500 C for 3 h. Hereafter, the calcination
2
2
2
◦
temperature is noted in the catalyst formulation as, for example,
15%H PW12O40/Nb O5-300 and 15%H PW12O40/Nb O5-500.
H PW12O40 and Cs2.5H0.5PW12O40, decreasing in the order of sup-
3
3
2
3
2
ports: SiO > TiO > Nb O5 > ZrO [26]. Interaction of H PW12O40
Nb O5 was prepared as described in the literature [29]. NbCl5
2
2
2
2
3
2
with support has been found to increase in the same order of sup-
ports, leading to partial HPA decomposition [26].
The aim of this work is to study the isomerisation of ␣-pinene
in the gas phase over solid HPA catalysts. The catalysts under
powder (5 g) was dissolved in 10 ml of ethanol and slowly added
to 200 ml of 0.3 M NH3 aqueous solution to afford a white precip-
itate of niobic acid Nb O5·nH O. The precipitate was washed with
2
2
distilled water until chloride free, as tested with AgNO , and dried
3
◦
study are both the “standard” HPA catalysts, such as H PW12O40
,
in an oven at 100 C for 3 h. ZrO2 was prepared by adding drop-
3
H PW12O40/SiO and Cs2.5H0.5PW12O40, and the acidic composites
wise aqueous ammonium hydroxide (25%) to an aqueous solution
3
2
prepared by supporting H PW12O40 (15 wt%) on Nb O5, ZrO2 and
of ZrOCl at room temperature with intense stirring until a pH of 10
3
2
2
TiO with sub-monolayer HPA coverage. Previously, these catalysts
have been characterised at a gas–solid interface regarding their acid
properties and chemical structure of HPA on the catalyst surface
was reached. The hydrogel formed was aged at room temperature
for 24 h with stirring and then filtered through a Buchner funnel.
2
The hydrous ZrO2 thus obtained was washed with distilled water
◦
[
26].
until chloride free (AgNO ) and dried in an oven at 100 C for 24 h.
3
2.2. Techniques
2
. Experimental
The BET surface area and porosity of catalysts were determined
from nitrogen physisorption measured on a Micromeritics ASAP
2
.1. Chemicals and catalysts
◦
2
000 instrument at −196 C. Before the measurement, the samples
The organic and inorganic chemicals were purchased from
◦
were evacuated at 250 C for 1–2 h. Powder X-ray diffraction (XRD)
spectra of catalysts were recorded on a PANalytical Xpert diffrac-
tometer with a monochromatic Cu K␣ radiation (ꢀ = 0.154 nm).
XRD patterns were attributed using the JCPDS database. ICP-AES
elemental analysis was carried out on a Spectro Ciros emission
spectrometer. The amount of carbon deposited on post-reactor cat-
alysts was measured by combustion elemental analysis.
Aldrich unless stated otherwise and used as supplied without fur-
ther purification. ␣-Pinene was 98% pure.
The acidic heteropoly salt Cs2.5H0.5PW12O40 was prepared
according to the literature procedure [27] by adding dropwise the
required amount of aqueous solution of caesium carbonate (0.47 M)
to aqueous solution of H PW12O40 (0.75 M) at room temperature
with stirring. The precipitate obtained was aged in aqueous mix-
ture for 48 h at room temperature and dried in a rotary evaporator
at 45 C/3 kPa and after that in an oven at 150 C/0.1 kPa for 1.5 h.
The silica-supported catalyst 15% H PW12O40/SiO , was prepared
by wet impregnation of Aerosil 300 silica (300 m /g surface area)
with an aqueous solution of H PW12O40 [28]. The mixture was
stirred overnight at room temperature, followed by drying in a
3
◦
◦
2.3. Catalyst testing
3
2
2
The isomerisation of ␣-pinene was carried out in the gas phase
under atmospheric pressure in a vertical Pyrex continuous down-
flow fixed-bed reactor (9 mm internal diameter) fitted with an
on-line gas chromatograph (Varian Star 3400 CX instrument with
a 30 m × 0.25 mm × 0.25 ␮m ZB-1701 capillary column and a flame
ionisation detector). The temperature in the reactor was controlled
by a Eurotherm controller using a thermocouple placed at the top of
the catalyst bed. The gas feed contained ␣-pinene (1.95 kPa partial
pressure) in nitrogen obtained by passing nitrogen flow controlled
by a Brooks mass flow controller through a saturator, which held
3
◦
rotary evaporator. Finally, the catalyst was dried at 150 C/0.1 kPa
for 1.5 h.
The catalysts comprising H PW12O40 (15 wt%) supported onto
3
Nb O5, ZrO and TiO2 were prepared by wet impregnation of the
oxides with H PW12O40 from aqueous solution. The TiO2 was a
Hombikat purchased from Fluka, which, from manufacturer’s spec-
ification, is a fine-crystalline anatase modification with a surface
2
2
3
2
◦
area ≥300 m /g and a weight loss on drying ≤10%. The Nb O5 and
liquid ␣-pinene at 50 C to maintain the chosen reactant pres-
2
ZrO were in-house made hydrous oxides (see below). An aqueous
sure. The gas feed entered the reactor at the top at a flow rate of
20 ml/min. All gas lines were made of stainless steel. The down-
2
solution of H PW12O40 (1.5 g) was mixed with 8.5 g oxide sup-
port to form hydrogel with a minimum amount of aqueous phase.
It was then aged for 24 h at room temperature with stirring, fol-
3
◦
stream lines and sampling valves were heated at 175 C to prevent
condensation of reactants and products. The reactor was packed

A. Alsalme et al. / Applied Catalysis A: General 390 (2010) 219–224
221
Table 1
a
Catalyst texture from nitrogen adsorption.
b
Pore diameterc
(A)
d
Catalyst
SBET
(
Pore volume
(cm /g)
2
˚
3
m /g)
SiO2 (Aerosil 300)
300e
287
162
a
Nb2O5-250
46
23
0.33
0.09
ZrO2-250a
TiO2
e
≥300
TiO2-250a
H3PW12O40-250
44
2.0
90
81
204
52
80
24
0.10
0.04
1.17
0.22
0.19
0.07
0.08
0.16
0.23
0.07
0.88
a
1
1
1
1
1
1
1
5%H3PW12O40/SiO2-250a
5%H3PW12O40/Nb2O5-300
5%H3PW12O40/Nb2O5-500
5%H3PW12O40/ZrO2-300
5%H3PW12O40/ZrO2-500
5%H3PW12O40/TiO2-300
229
166
96
121
54
41
40
111
185
56
154
230
24
5%H3PW12O40/TiO2-500
a
Cs2.5H0.5PW12O40-250
2
0%Cs2HPW12O40/SiO2-250a
191
Fig. 1. XRD patterns (Cu K␣ radiation) for: (1) ZrO2-500, (2) 15%H3PW12O40/ZrO2-
3
00, (3) 15%H3PW12O40/ZrO2-500, (4) TiO2-500, (5) 15%H3PW12O40/TiO2-300,
a
◦
Prior to N2 adsorption the catalysts were pre-treated at 250 C in vacuum for
(
6) 15%H3PW12O40/TiO2-500, (7) Nb2–O5-500, (8) 15%H3PW12O40/Nb2O5-300, (9)
2
h.
15%H3PW12O40/Nb2O5-500, (10) H3PW12O40.
b
BET surface area.
Average BET pore diameter.
Single point total pore volume.
The value of surface area given by manufacturer.
c
d
e
2
◦
4
4 m /g upon pre-treatment at 250 C in vacuum. Supporting HPA
◦
on TiO2 followed by calcination at 300 and 500 C had no effect on
the catalyst surface area (41 and 40 m /g, respectively). The texture
of our bulk H PW12O40 and Cs2.5H0.5PW12O40 was typical of these
materials as reported in the literature [27,30].
2
with 0.100 g catalyst powder sieved to 45–180 ␮m. Prior to reac-
tion, the catalysts were pre-treated in situ at 300 C in a nitrogen
3
◦
flow (20 ml/min) for 1 h. At regular time intervals, the downstream
gas flow was analysed by the on-line GC. The response factors were
assumed to be the same for all monoterpenes in the product mix-
ture. The products were also collected in an ice trap and analysed
by the off-line GC. The products were identified by GC and GC–MS
using authentic samples. Carbon balance was monitored by GC as
the percentage of the reacted carbon atoms found in products.
The surface area of our supported HPA catalysts ranged from 40
2
to 229 m /g. Assuming for the H PW12O40 molecule a cross section
3
2
of 144 A [12,13], the 15% HPA loading would correspond to 45 m /g
monolayer HPA coverage. Therefore, these catalysts could be ade-
quately viewed as such having sub-monolayer HPA coverage, with
some reservation in case of titania supported catalysts which had
˚
2
a surface area of 40–41 m /g. The XRD spectra (Fig. 1) confirm fine
dispersion of HPA on the catalyst surface.
3
. Results and discussion
Fig.
1 shows XRD patterns for the bulk and supported
H PW12O40 catalysts as well as for catalyst supports. The
3
3.1. Catalyst characterisation
15%H PW12O40/Nb O5 and 15%H PW12O40/ZrO2 catalysts, pre-
3 2 3
pared from the amorphous Nb O5 and ZrO hydrous oxides, after
2
2
◦
The texture of catalysts and supports studied is shown in Table 1.
calcination at 300 C were XRD amorphous. After calcination at
500 C, both the supports and the catalysts became crystalline.
◦
All the catalysts were mesoporous materials with pore diameters
of 24–230 A˚ . Amongst them, silica and titania based catalysts pos-
Nb O5-500 and 15%H PW12O40/Nb O5-500 exhibited the pat-
2
3
2
sessed larger pores and zirconia based ones smaller pores, with
tern of hexagonal niobia. ZrO -500 showed the pattern of the
2
niobia catalysts inbetween. For HPA catalysts supported on Nb O5
tetragonal phase, with a little of monoclinic ZrO2 also present.
15%H PW12O40/ZrO -500 consisted of a mixture of tetragonal and
2
and ZrO2 hydrous oxides, the surface area reduced with increasing
3
2
◦
the calcination temperature from 300 to 500 C, as expected. The
monoclinic modifications of ZrO . 15%H PW12O40/TiO , prepared
2 3 2
commercial TiO2 support decreased its surface area from ∼300 to
from commercial anatase TiO , also exhibited the pattern of anatase
2
Table 2
␣
-Pinene conversion and product selectivity.a
d
Catalyst
Conversion (%)
Selectivity (mol%)
Camphene
Cc
−ꢁH [26]
Limonene
Othersb
(wt%)
(kJ/mol)
H3PW12O40-300a
Cs2.5H0.5PW12O40-300
80
48
100
94
88
88
85
74
75
60
2
55
67
27
39
56
32
36
55
51
45
55
47
9
10
0
3
7
36
23
73
58
37
61
56
26
31
34
0
0.7
4.9
195
164
154
143
130
132
121
121
127
a
1
1
1
1
1
1
1
5%H3PW12O40/SiO2-300a
5%H3PW12O40/TiO2-300
5%H3PW12O40/TiO2-500
5%H3PW12O40/Nb2O5-300
5%H3PW12O40/Nb2O5-500
5%H3PW12O40/ZrO2-300
5%H3PW12O40/ZrO2-500
5.4
9.6
3.0
7
8
19
18
21
45
28
Nb2O5-500
ZrO2-500
TiO2-500
1.5
3.0
20
25
a
◦
Reaction conditions: 200 C, 0.1 g catalyst, 2.0 vol.% ␣-pinene in N2, 20 ml/min flow rate, 10 h time on stream. Prior to reaction the catalysts were pre-treated in situ at
◦
3
00 C for 1 h in N2 (20 ml/min).
b
A mixture of terpinolenes, terpinenes, p-cymene, ␤-pinene and unidentified monoterpene isomers.
Carbon deposition on catalysts after reaction from combustion elemental analysis.
Initial enthalpy of ammonia adsorption at 100 C.
c
d
◦

2
22
A. Alsalme et al. / Applied Catalysis A: General 390 (2010) 219–224
◦
after calcination at 300 and 500 C. All supported catalysts were
100
XRD amorphous with regard to H PW12O40, which indicates fine
3
dispersion of HPA on the catalyst surface. XRD pattern of WO3 was
not observed in the catalyst samples.
Table 2 shows the strength of acid sites of HPA catalysts in terms
of initial enthalpy of ammonia adsorption, ꢁH, as determined pre-
viously [26]. It ranges from 195 to 121 kJ/mol and decreases
8
6
0
0
Conversion
Camphene
Limonene
in the order of catalysts: H PW12O40 > Cs2.5H0.5PW12O40
>
3
40
1
5%H PW12O40/SiO > 15%H PW12O40/TiO > 15%H PW12O40/
3 2 3 2 3
Nb O5 > 15%H PW12O40/ZrO2.
2
3
20
3.2. Isomerisation of ˛-pinene
0
Previous studies of the HPA-catalysed isomerisation of ␣-pinene
9–11] have been focussed on the reaction in liquid phase in a
0
5
10
15
20
[
Time, h
batch reactor. In our work, ␣-pinene isomerisation was studied
at the gas–solid interface in continuous mode, aiming at the pro-
duction of camphene and limonene. For camphene production, an
1
00
Conversion
◦
optimum temperature of 200 C had been identified from prelimi-
Camphene
Limonene
◦
nary experiments in the temperature range of 140–240 C, and the
80
results presented here were obtained at that temperature. Decreas-
ing reaction temperature led to an increase in limonene selectivity
at the expense of camphene, which may be explained by lower
activation energy for the formation of limonene compared to the
formation of camphene (Scheme 1).
Scheme 1 is a rather simplified representation of the reaction in
question. In fact, GC analysis of the product mixture found about 60
monoterpene components, a third of them in very small amounts
below the threshold of quantitative determination. Neither light
hydrocarbons, nor monoterpene oligomers were observed amongst
the products. This indicates that the HPA catalysts studied did not
cause monoterpene cracking, whereas oligomerisation, if occurred,
contributed to coke deposited on the catalyst surface.
6
4
2
0
0
0
0
0
5
10
15
20
Time, h
Table 2 gives a summary of catalyst performance after 10 h on
stream, displaying ␣-pinene conversion and the selectivity to cam-
phene and limonene. Camphene was the main reaction product,
with limonene, terpinolenes, terpinenes, ␤-pinene, p-cymene as
well as smaller amounts of a large number of unidentified monoter-
pene by-products also formed. As in previous studies [1], coke
deposition on catalysts was observed (up to 9.6%, Table 2). Coking
has been suggested to cause catalyst deactivation ([1] and refer-
ences therein). By monitoring the carbon balance, it was found
that during the first hour of reaction, only 50–60% of the carbon
fed ended up in the products formed, with the rest being coke
deposited on the catalyst. After 3 h time on stream, this value
became 90–100%. This shows that coke deposited mainly during
the first 2–3 h of reaction, and after that little coking occurred.
The reaction time-course is shown in Figs. 2–6. Generally, the
conversion of ␣-pinene decreased with the time on stream due to
catalyst deactivation. Usually, the selectivities to camphene and
limonene were increasing with the time on stream, along with cata-
lyst deactivation. Since coke is likely to block the strongest acid sites
in the catalyst, this result indicates that the formation of camphene
is favoured on moderate rather than strong catalyst acid sites.
1
00
8
6
4
2
0
0
0
0
0
Conversion
Camphene
Limonene
0
5
10
15
20
Time, h
Fig. 2. ␣-Pinene conversion and product selectivity over bulk H3PW12O40-300
(top), bulk Cs2.5H0.5PW12O40-300 (middle) and 15%H PW12O40/SiO -300 (bottom)
3
2
◦
at 200 C as a function of time on stream.
Bulk HPA and Cs2.5H0.5PW12O40
, possessing very strong
Brønsted acid sites with ammonia adsorption enthalpies of −195
and −164 kJ/mol, respectively (Table 2), showed high initial cat-
alytic activity (Fig. 2). However, their activity was quickly declining
arise from the strong catalyst acidity (Table 2) as well as from
the large surface area and pore diameter of this catalyst (Table 1).
Nevertheless, it probably also suffered from deactivation, which
is indicated by gradual increase in camphene selectivity with the
time on stream. Eventually, this catalyst gave a moderate selectiv-
ity of 27% to camphene, still being the main reaction product. Only
traces of limonene were formed in this case. The poor selectivities
to camphene and limonene may be the result of the high ␣-pinene
conversion. Therefore, these selectivities might improve at smaller
contact times, i.e., at a higher flow rate or a smaller amount of
catalyst.
with the time on stream, especially in the case of Cs2.5H0.5PW12O40
,
due to catalyst coking (Table 2). At the same time, their selectivity
to camphene was increasing to achieve 57% with H PW12O40 and
3
7
0% with Cs2.5H0.5PW12O40 in 17 h on stream, which was the high-
est camphene selectivity observed in this work. The selectivity to
limonene was also increasing to reach 11–14%.
Silica-supported H PW12O40 exhibited 100% ␣-pinene conver-
3
sion within 20 h time on stream (Fig. 2). The high activity could

A. Alsalme et al. / Applied Catalysis A: General 390 (2010) 219–224
223
100
100
Conversion
Camphene
Conversion
Camphene
Limonene
8
6
4
2
0
0
0
0
0
80
Limonene
6
4
2
0
0
0
0
0
2
4
6
8
10
0
2
4
6
8
10
12
14
Time, h
Time, h
100
100
8
6
4
2
0
0
0
0
0
8
6
4
2
0
0
0
0
0
Conversion
Camphene
Limonene
Conversion
Camphene
Limonene
0
5
10
15
20
Time, h
0
5
10
15
Time, h
Fig. 3. ␣-Pinene conversion and product selectivity over Nb2O5-500 (top) and
5%H3PW12O40/Nb2O5-500 (bottom) at 200 C as a function of time on stream.
◦
1
Fig. 5. ␣-Pinene conversion and product selectivity over TiO2-500 (top) and
5%H3PW12O40/TiO2-500 (bottom) at 200 C as a function of time on stream.
◦
1
The HPA catalysts supported on Nb O5, TiO and ZrO2 calcined
2
2
◦
at 500 C were all crystalline materials with finely dispersed HPA
species on the catalyst surface (Fig. 1). Although weaker acids
nating pyridine, as found by IR spectroscopy of adsorbed pyridine
[26].
(
Table 2), these catalysts showed higher camphene selectivity and
The pure niobium oxide, Nb O -500, which possesses a rela-
2
5
better deactivation resistance in ␣-pinene isomerisation than the
stronger acidic catalysts discussed above. This backs the sugges-
tion that the formation of camphene is favoured on moderate acid
sites, whereas strong catalyst acid sites enhance further isomeri-
sation and catalyst coking, hence reducing camphene selectivity.
tively strong acidity [26], exhibited stable activity providing 45%
camphene and 21% limonene selectivity at 60% conversion within
10 h on stream (Fig. 3). Loading of HPA on Nb O5 increased the
2
catalyst activity, albeit at the expense of camphene and limonene
selectivity, providing 40% camphene and 10% limonene at 77% con-
version in 20 h on stream.
Figs. 4–6 display the performance of HPA supported on Nb O5, TiO2
2
and ZrO in comparison with that of the corresponding pure oxides.
The pure zirconia, ZrO -500, showed only very low activity in ␣-
2
2
The Nb O5, TiO and ZrO2 oxides all possess Lewis acid sites and
pinene isomerisation, <2% conversion (Table 2). In contrast, the HPA
2
2
Nb O5 and ZrO also possess Brønsted acid sites capable of proto-
catalyst 15%HPA/ZrO -500 performed quite well: 54% camphene
2
2
2
and 20% limonene selectivity at 72% conversion in 20 h on stream
(
39% camphene yield). The catalyst exhibited very good resistance
Conversion
Camphene
Limonene
1
00
to deactivation, with only a slight decrease in conversion from 81%
to 72% in 20 h (Fig. 4).
The pure titania, TiO2-500, possessing only Lewis acidity [26],
showed a 100% initial ␣-pinene conversion, which, however,
dropped down to 4% in 13 h on stream (Fig. 5). This can be explained
by blocking the Lewis acid sites by coke or adsorbed terpene
molecules. The catalyst 15%HPA/TiO -500 exhibited the best per-
formance amongst the catalysts studied, giving 59% camphene and
8
8
6
4
2
0
0
0
0
0
2
% limonene selectivity at 81% conversion in 15 h on stream (51%
camphene yield), with good resistance to deactivation (Fig. 5). It
took, however, about 10 h for the catalyst to reach this selectivity.
It is worth noting that acid treated TiO is used as the catalyst in the
2
commercial liquid-phase process of ␣-pinene isomerisation [1].
0
5
10
15
20
The catalysts comprising HPA supported on Nb O5, TiO and
2
2
Time, h
◦
ZrO2 calcined at 300 C showed ␣-pinene conversions similar
to that for the catalysts calcined at 500 C (Fig. 6). In case
of 15%HPA/ZrO2 catalyst, the selectivity to camphene did not
◦
Fig. 4. ␣-Pinene conversion and product selectivity over 15%H3PW12O40/ZrO2-500
◦
at 200 C as a function of time on stream.

2
24
A. Alsalme et al. / Applied Catalysis A: General 390 (2010) 219–224
1
00
and supported heteropoly acid catalysts based on H PW12O40 in
3
a fixed-bed flow reactor to yield camphene as the main product
8
6
4
2
0
0
0
0
0
in a mixture with other monoterpenes such as limonene, ter-
pinolenes, terpinenes, ␤-pinene, p-cymene, etc. The performance
of these catalysts, regarding the product selectivity and catalyst
resistance to deactivation, is profoundly dependent on the catalyst
acidity. Bulk H3PW12O40 and Cs2.5H0.5PW12O40 possessing very
strong Brønsted acid sites suffer from deactivation. Conversely,
the catalysts prepared by supporting H3PW12O40 on Nb2O5, ZrO2
Conversion
Camphene
Limonene
and TiO , although weaker acids, exhibit more stable as well as
2
more selective performance in this reaction. The catalyst compris-
ing H PW12O supported on TiO2 gives a camphene yield of 51%
0
5
10
15
3
40
◦
and a total camphene and limonene yield of 58% at 200 C after 15 h
on stream.
Time, h
1
00
Acknowledgements
8
6
4
2
0
0
0
0
0
Financial support from the EPSRC, UK (grant EP/F014686/1) and
the scholarship for A. Alsalme from King Saud University, Saudi
Arabia are gratefully acknowledged.
Conversion
Camphene
Limonene
References
[1] A. Corma, S. Iborra, A. Velty, Chem. Rev. 107 (2007) 2411–2502.
[2] A. Severino, A. Esculcas, J. Rocha, J. Vital, L.S. Lobo, Appl. Catal. A 142 (1996)
2
55–278.
3] G. Gunduz, R. Dimitrova, S. Yilmaz, L. Dimitrov, Appl. Catal. A 282 (2005) 61–65.
[4] R. Rachwalik, Z. Olejniczak, J. Jiao, J. Huang, M. Hunger, B. Sulikwoski, J. Catal.
52 (2007) 161–179.
5] M.Kr. Yadav, C.D. Chudasama, R.V. Jasra, J. Mol. Catal. A 216 (2004) 51–59.
[
2
0
2
4
6
8
10
[
Time, h
[6] C. Volzone, O. Masini, N.A. Comelli, L.M. Grzona, E.N. Ponzi, M.I. Ponzi, Mater.
Chem. Phys. 93 (2005) 296–300.
[
[
7] N.A. Comelli, E.N. Ponzi, M.I. Ponzi, Chem. Eng. J. 117 (2006) 93–99.
8] O. Chimal-Valencia, A. Robau-Sanchez, V. Collins-Martinez, A. Aguilar-
Elguezabal, Bioresour. Technol. 93 (2004) 119–123.
1
00
[
9] N.A. Comelli, L.M. Grzona, O. Masini, E.N. Ponzi, M.I. Ponzi, J. Chil. Chem. Soc.
8
6
4
2
0
0
0
0
0
Conversion
Camphene
Limonene
49 (2004) 245–250.
[10] A.D. Newman, A.F. Lee, K. Wilson, N.A. Young, Catal. Lett. 102 (2005) 45–50.
[11] K.A. da Silva Rocha, P.A. Robles-Dutenhefner, I.V. Kozhevnikov, E.V. Gusevskaya,
Appl. Catal. A 352 (2009) 188–192.
12] T. Okuhara, N. Mizuno, M. Misono, Adv. Catal. 41 (1996) 113–252.
13] I.V. Kozhevnikov, Catalysis by Polyoxometalates, Wiley & Sons, Chippenham,
[
[
2002.
[
[
[
14] M. Misono, Catal. Today 100 (2005) 95–100.
15] I.V. Kozhevnikov, J. Mol. Catal. A 305 (2009) 104–111.
16] I.D. Dobson, Green Chem. 5 (2003) G78–G81.
[17] K. Okumura, K. Yamashita, M. Hirano, M. Niwa, J. Catal. 234 (2005) 300–307.
18] K. Okumura, K. Yamashita, K. Yamada, M. Niwa, J. Catal. 245 (2007) 75–78.
19] B.M. Devassy, S.B. Halligudi, J. Catal. 236 (2005) 313–323.
20] B.M. Devassy, F. Lefebvre, S.B. Halligudi, J. Catal. 231 (2005) 1–10.
21] B.M. Devassy, S.B. Halligudi, J. Mol. Catal. A 253 (2006) 8–15.
[
[
[
[
[
0
5
10
15
Time, h
Fig. 6. ␣-Pinene conversion and product selectivity over 15%H3PW12O40/Nb2O5-
22] D.P. Sawant, A. Vinu, F. Lefebvre, S.B. Halligudi, J. Mol. Catal. A 262 (2007)
3
00 (top), 15%H3PW12O40/ZrO2-300 (middle) and 15%H3PW12O40/TiO2-300 (bot-
9
8–108.
[23] S.M. Kumbar, G.V. Shanbhag, F. Lefebvre, S.B. Halligudi, J. Mol. Catal. A 256
2006) 324–334.
◦
tom) at 200 C as a function of time on stream.
(
[
[
24] N. Bhatt, A. Patel, J. Mol. Catal. A 264 (2007) 214–219.
25] A. Alsalme, E.F. Kozhevnikova, I.V. Kozhevnikov, Appl. Catal. A 349 (2008)
170–176.
depend on the calcination temperature. For the niobia- and titania-
◦
supported HPA catalysts calcined at 300 C, the selectivity to
◦
[26] A. Alsalme, P.V. Wiper, Y.Z. Khimyak, E.F. Kozhevnikova, I.V. Kozhevnikov, J.
Catal. 276 (2010) 181–189.
[
camphene was lower than for the catalysts calcined at 500 C
(
Table 2).
27] Y. Izumi, M. Ono, M. Kitagawa, M. Yoshida, K. Urabe, Microporous Mater. 5
1995) 255–262.
(
[
[
28] E.F. Kozhevnikova, I.V. Kozhevnikov, J. Catal. 224 (2004) 164–169.
29] N. Uekawa, T. Kudo, F. Mori, Y.J. Wu, K. Kakegawa, J. Colloid Interface Sci. 264
4
. Conclusion
(
2003) 378–384.
[30] T. Okuhara, H. Watanabe, T. Nishimura, K. Inumaru, M. Misono, Chem. Mater.
2 (2000) 2230–2238.
The acid-catalysed isomerisation of ␣-pinene is of industrial
1
value to produce camphene and limonene, important interme-
diates for organic synthesis. Here this reaction has been studied
at the gas–solid interface over a range of pre-characterised bulk