Selective vapour-phase α-pinene isomerization to camphene over gold-on-alumina catalyst

Article abstract of DOI:10.1016/j.apcata.2010.07.002

The vapour-phase isomerization of α-pinene for the first time was studied over a supported Au catalyst. α-pinene was isomerized to camphene over the 2.2% Au/γ-Al₂O₃ catalyst at 463-483 K using a solution of the reagent in n-octane as the initial reaction mixture and H₂ or N₂ as a carrier gas. Under these conditions, the selectivity to camphene reaches 60-80% at 99.9% conversion of α-pinene. The reaction is found to be first-order with respect to α-pinene, the apparent activation energy being similar to that observed with the conventional TiO₂ catalyst. The prominent catalyst deactivation has been observed at increased α-pinene concentrations in the inlet reaction mixture (≥4 vol% in n-octane solution). According to HRTEM and TPO results, the deactivated catalyst contains the carbonaceous deposits that may block the catalyst surface. Almost complete regeneration was done in flowing O₂ at temperature up to 923 K required to totally eliminate the coke deposits.

Full text of DOI:10.1016/j.apcata.2010.07.002

Applied Catalysis A: General 385 (2010) 136–143
Contents lists available at ScienceDirect
Applied Catalysis A: General
journal homepage: www.elsevier.com/locate/apcata
Selective vapour-phase ␣-pinene isomerization to camphene over
gold-on-alumina catalyst
I.L. Simakovaa,∗, Yu.S. Solkinaa,b, B.L. Moroza,b, O.A. Simakovaa,c, S.I. Reshetnikova, I.P. Prosvirina,
V.I. Bukhtiyarova,b, V.N. Parmona,b, D.Yu. Murzinc,∗
a G.K. Boreskov Institute of Catalysis SB RAS, Novosibirsk, Russia
b Novosibirsk State University, Russia
c Åbo Akademi University, Process Chemistry Centre, Turku/Åbo, Finland
a r t i c l e i n f o
a b s t r a c t
Article history:
The vapour-phase isomerization of ␣-pinene for the first time was studied over a supported Au catalyst.
␣-pinene was isomerized to camphene over the 2.2% Au/␥-Al2O3 catalyst at 463–483 K using a solution
of the reagent in n-octane as the initial reaction mixture and H2 or N2 as a carrier gas. Under these
conditions, the selectivity to camphene reaches 60–80% at 99.9% conversion of ␣-pinene. The reaction
is found to be first-order with respect to ␣-pinene, the apparent activation energy being similar to that
observed with the conventional TiO2 catalyst. The prominent catalyst deactivation has been observed at
increased ␣-pinene concentrations in the inlet reaction mixture (≥4 vol% in n-octane solution). According
to HRTEM and TPO results, the deactivated catalyst contains the carbonaceous deposits that may block
the catalyst surface. Almost complete regeneration was done in flowing O2 at temperature up to 923 K
required to totally eliminate the coke deposits.
Received 26 April 2010
Received in revised form 22 June 2010
Accepted 3 July 2010
Available online 30 July 2010
Keywords:
Gold catalysts
Isomerization
␣-Pinene
Camphene
© 2010 Elsevier B.V. All rights reserved.
1. Introduction
catalyst activated in situ by H2SO4 with rather low reaction rate
obtaining yields from 35 to 50% [12,13]. Other catalysts mainly
As it is known starting from classical Haruta’s works [1], metal-
lic gold particles with a size of several nanometers may exhibit
high catalytic activity and/or unusual selectivity in numerous
redox transformations of organic compounds [2–5]. Meanwhile,
only a few examples of isomerization reactions catalyzed by sup-
ported gold including double bond migration in allylbenzene [6]
or linoleic acid [7] and skeletal rearrangements (ring opening of
propylene oxide to acetone [8], cycloisomerization of ␥-acetylinic
carboxylic acids [9] etc.) were communicated as yet. Herein, we
report that the nanodispersed Au/␥-Al2O3 catalyst prepared by
fairly simple procedure reveals good catalytic activity and selec-
tivity for vapour-phase isomerization of ␣-pinene to camphene at
463–483 K with the use of H2 or N2 (1 bar) as a carrier gas. This
reaction representing alkyl group migration from one carbon to a
neighboring carbon (Wagner–Meerwein rearrangement) has high
practical significance as a first step in the industrial synthesis of
camphor from ␣-pinene which is the major constituent of tur-
pentine [10,11]. In industry, the transformation of ␣-pinene into
camphene is performed in liquid phase at 423–443 K over TiO2
of acidic type such as natural [10,14–16] and synthetic zeolites
[17–19], activated carbon [20], clays [21,22], ion exchange resins
[11] and silica-supported rare earth oxides [23] were also exam-
ined in this reaction. To the best of our knowledge, supported gold
metal catalysts have never been tested for ␣-pinene isomeriza-
tion.
2. Experimental
2.1. Materials
Hydrogen tetrachloroaurate HAuCl4·aq (49.47 wt% Au) was pur-
chased from Aurat (Russia) and used without further purification.
␣-Pinene used for the catalytic tests was isolated from turpentine
oil by vacuum distillation and contained the following impurities:
tricyclene—0.19 wt%, camphene—1.61 wt%, ␤-pinene—3.54 wt%, 3-
carene—1.30 wt%. n-Octane which is used as
a solvent was
purchased from Reakhim (Russia) and purified by vacuum distil-
lation.
As a catalyst support, ␥-alumina supplied by Ryazan Oil Refinery
Company (Russia) was used in the form of 200–500 ␮m granules
with SBET of 268 m2 g−1 and total pore volume of 0.69 cm3 g−1. It
was dried at 383 K in an oven for 8 h immediately before the catalyst
preparation.
∗
Corresponding authors.
E-mail addresses: simakova@catalysis.ru (I.L. Simakova), dmurzin@abo.fi
(D.Yu. Murzin).
0926-860X/$ – see front matter © 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.apcata.2010.07.002

I.L. Simakova et al. / Applied Catalysis A: General 385 (2010) 136–143
137
2.2. Catalyst preparation
The method for preparing the Au/␥-Al2O3 catalyst was similar
to those previously described [24]. A weighed amount of ␥-Al2O3
was incipient wetness impregnated with an aqueous 0.06 M HAuCl4
solution, the volume of the impregnating solution being 10% higher
than the total pore volume of the support. The resulting paste was
mixed for 1 h, allowed to stay in air at room temperature overnight
and dried at 323 K for 8 h with the use of a vacuum system. Upon
treatment in flowing H2 at 673 K for 4 h, the sample was agitated
with 2 portions of an aqueous 1 M NaOH solution (100 mL per 1 g
sample) for removal of Cl− impurity and neutralization of acidic
sites on the support surface. After that, it was thoroughly washed
with warm distilled water in order to remove Na+ impurity and
excess hydroxide ions. Finally, the sample was repeatedly dried as
described above and calcined in air at 673 K for 4 h.
2.3. Characterization of catalyst
Fig. 1. Diffuse reflectance UV–vis spectra: (a) ␥-Al2O3 support lacking gold; (b)
␥-Al2O3 after impregnation with HAuCl4 solution and drying; (c) sample b after
reduction with H2 at 673 K; (d) sample c treated with 1 M NaOH and calcined in air
at 673 K.
The Au and Cl contents of the catalyst were measured using
X-ray fluorescence (XRF) technique on a VRA-30 instrument
equipped with a Cr-anode. The diffuse reflectance UV–vis spec-
tra were recorded on a Shimadzu UV-2501 PC spectrometer
equipped with an ISR-240 integrating sphere attachment for diffuse
reflectance measurements, using BaSO4 as a reference, in the range
of 200–1000 nm and presented in Kubelka–Munk function F(R)-
wavelength coordinates. Transmission electron microscopy (TEM)
studies were carried out on a JEOL JEM-2010 electron microscope
operated at 200 kV and giving an information limit of 0.14 nm.
The diameters of the Au crystallites were determined from TEM
images taken with a medium magnification, at least 300 crystals
being included in the particle size distribution for a sample. The
temperature-programmed oxidation (TPO) of the catalyst samples
was carried out on a NETZSCH STA 409PC instrument equipped
with a quadrupole mass-spectrometer for determination of the
composition of oxidation products. The sample placed on a ther-
mobalance was heated in flowing air (40 cm3 min−1) at 323 K for
1 h, and thereafter the temperature was raised to 923 K for 1 h
(ramping 10 K min−1).
Photoelectron spectra were recorded using SPECS spectrome-
ter with PHOIBOS-150 hemispherical energy analyzer and MgK␣
irradiation (hꢀ = 1253.6 eV, 100 W). Binding energy scale was pre-
liminarily calibrated by the position of the peaks of Au4f7/2 (84.0 eV)
and Cu2p3/2 (932.67 eV) core levels. For recording of spectra the
samples were supported to the conductive scotch tape. The bind-
ing energy of peaks was corrected to take into account the sample
charging by referencing to the Al2p (74.5 eV) (internal standard).
This reference was applied since the spectrometer was equipped
with turbomolecular pumps, and calibration on the basis of the
position of C1s core level, typically used in the case of vacuum
systems with diffusion pumps, was difficult. The ratio of surface
atomic concentrations of the elements was calculated from the
integral intensities of photoelectron peaks corrected by corre-
sponding atomic sensitivity factors (ASF). In addition to the survey
photoelectron spectra, more narrow spectral regions Al2p, Au4f,
Al2s, C1s, and O1s have been recorded. For the survey spectra the
pass energy of the analyzer was 50 eV, while for the narrow spectral
regions the pass energy was 10 eV.
by passing successively through the columns filled with a reduced
Ni-Cr catalyst and silica. A catalyst sample (200 mg) was placed in a
U-shaped glass plug-flow reactor of 4 mm i.d. equipped with a ther-
mocouple which was fixed in the middle of the catalyst bed outside
the reactor tube. The carrier gas was passed through the catalyst
bed at the desired space velocity (SV), and the temperature was
raised up to 463–483 K. Then ␣-pinene solution in n-octane was
supplied via a heated pre-chamber of U-reactor to be vapourized
and mixed with the carrier gas before passing to the catalyst bed.
The residence time (ꢁ) in the reactor was fixed by controlling SV of
the carrier gas within the range of 1600–14500 h−1. The samples
(100–200 ␮L) of the outlet gas stream were taken into a condenser
connected to the effluent tube of the reactor. The products were
identified by GC–MS (Agilent 6890, the quartz capillary column
50 m/0.2 mm/0.5 ␮m, SE-30). The contents of the inlet and outlet
reaction mixtures were analyzed by a gas chromatograph («Tzvet-
500») equipped with a flame-ionization detector and a capillary
Carbowax-20 M column (50 m/0.2 mm/0.5 ␮m). The column was
hold at 393 K, both injection inlet and detector temperatures were
473 K. The conversion of ␣-pinene, X (%), and the selectivity, S (%), to
the determined product were calculated, using the corresponding
equations below:
[␣-pinene]inlet − [␣-pinene]outlet
X(%) =
× 100
[␣-pinene]inlet
[product]outlet
[␣-pinene]inlet − [␣-pinene]outlet
S(%) =
× 100
During all tests, the carbon balance between the feed and outlet
stream was within 100 5%.
3. Results and discussion
3.1. Characterization of catalyst
2.4. Isomerization procedure
The impregnation of ␥-alumina with an aqueous HAuCl4 solu-
tion followed by drying produced yellow samples which turned to
dark violet upon calcination in a H2 flow at 673 K. The loading of
Au was 2.2 wt%, as determined by XRF.
Fig. 1 shows the diffuse reflectance UV–vis spectra of the Au/␥-
Al2O3 catalyst at different stages of its preparation. The spectrum
of ␥-Al2O3 after impregnation with the HAuCl4 solution exhibits
Vapour-phase isomerization of ␣-pinene was carried out under
continuous flow conditions at atmospheric pressure using a solu-
tion of ␣-pinene in n-octane of the determined concentration as
the initial reaction mixture and H2 or N2 as a carrier gas. Just before
mixing with ␣-pinene vapour, H2 or N2 were deoxidized and dried

138
I.L. Simakova et al. / Applied Catalysis A: General 385 (2010) 136–143
Fig. 2. TEM images of the 2.2% ␥-Al2O3 catalyst: (a) as-prepared catalyst; (b) the catalyst after testing for ␣-pinene isomerization at 473 K and [␣-pinene]inlet = 20 vol%; (c)
HRTEM image of Au particle located at the surface of the spent catalyst (the arrow indicates the layer of disordered structure covering the surface of Au particles); (d) sample
b after TPO in flowing air at 323 K for 1 h and thereafter during the temperature rise up to 923 K (heating rate 10 K min−1). The insets in (a), (b) and (d) show Au particle size
distributions in the samples.
strong, broadened absorption bands at 230 and 320 nm, belonging
to [AuCl4]− complexes. Subsequent treatment of the sample with
H2 results in the appearance of the strong band at 525 nm derived
from the plasmon absorption by Au0 nanoparticles [25]. Simulta-
neously, the absorption bands at 230 and 320 nm disappear from
the spectrum. Washing with an aqueous 1 M NaOH solution fol-
lowed by calcination in air at 673 K leads neither to gold leaching
from the reduced sample nor to significant changes in the UV–vis
spectrum, but allows the Cl content to be decreased by an order
of magnitude (from 5000 to 100–200 ppm), according to XRF data.
From the TEM data shown in Fig. 2a, the final Au/␥-Al2O3 catalyst
comprises both small (di < 5 nm) and relatively large (di = 5–8 nm)
Au0 particles, while the former predominate in the number and
weight fraction. Particle sizes of the sample are fitted well by the
lognormal distribution with the maximum at around 4 nm (mean
linear diameter (dl) is equal to 4.0 0.85 nm).
3.2. Catalytic performance
Table 1 presents the results obtained with the nanosized 2.2%
Au/␥-Al2O3 catalyst in vapour-phase isomerization of ␣-pinene at
473 K and low reagent concentration in the inlet n-octane solu-
tion ([␣-pinene]inlet = 0.4 vol%) under H2 atmosphere. Under these
conditions, the gold catalyst exhibits high activity for ␣-pinene
isomerization, the ␣-pinene conversion being not significantly
changed during the whole testing run (typically, for 7–8 h) (Entries
1 and 2). At the residence time of 0.33 s (SV = 2200 h−1) the ␣-pinene
conversion reached 99.9% with fairly good and stable selectiv-
ity to camphene (S = 65.5%), the rest of the products were mainly
The XPS results of Au4f spectra for the fresh catalyst, the crushed
fresh catalyst and the catalyst after reaction are shown in Fig. 3. For
the samples the binding energy of Au4f7/2 was 83.9 eV, which was
assigned to metallic gold. Au4f7/2 binding energy value is typical
for metallic gold particles supported on various oxide carriers [26].
The Au/Al atomic ratio measured for the fresh Au/␥-Al2O3 sample
(0.022) does not evidence any change after crushing (0.021) and
the reaction (0.022), which is an indication of gold homogeneity
along support particle radius at least above 1 ␮m, which is the size
of grinded catalysts. The fact that no peaks are observed in the Eb
region of 84.5–88 eV characteristic of “ionic” gold species can be
accounted for by the rapid reduction of Au(III) or Au(I) species, if
any are contained in the sample, under the action of photoelec-
tron beam during recording the XP spectra [24]. The results of XPS
indicated that the major fraction of the gold species on the catalyst
surface is metallic gold, with a minor fraction of gold species that
can be attributed to an oxidized state (Au+) which is coordinated
with OH-groups [24].
Fig. 3. XP spectra of Au4f region of a 2.2 wt% Au/Al2O3 catalyst: (a) before reaction,
(b) after reaction, and (c) crushed catalyst before reaction.

I.L. Simakova et al. / Applied Catalysis A: General 385 (2010) 136–143
139
Table 1
Results of vapour-phase ␣-pinene isomerization over the 2.2 wt% Au/␥-Al2O3 catalyst (in comparison with ␥-Al2O3 lacking gold).a.
Entry
Catalyst
Total conversion (%)
Selectivity (%)
Camphene
Limonene
p-Cymene
Tricyclene
Others
1
Au/␥-Al2O3
Au/␥-Al2O3
Au/␥-Al2O3
99.9
99.9
99.8
2.7
65.5
66.3
66.2
28.3
1.2
1.5
1.8
14.6
14.1
14.0
11.1
8.2
8.3
7.7
10.5
9.8
10.3
29.2
2b
3c
4
d
␥-Al2O3
14.6
16.8
a
Reaction conditions: [␣-pinene]inlet 0.4 vol%, solvent n-octane, catalyst 0.2 g, 473 K, H2 atmosphere, SV 2200 h−1, 1 h.
In 6 h after the beginning of the catalytic run.
Tested with the use of N2 as a carrier gas.
b
c
d
Impregnated with 0.01 M HCl solution, outgassed at 323 K, calcined in a H2 flow at 673 K for 4 h, then washed with 1 M NaOH and distilled water, outgassed at 323 K and,
finally, calcined in air at 673 K for 4 h.
p-cymene and tricyclene (S = 14.5 and 8.2%). Even at ꢁ of 0.05 s that
was the minimum ꢁ value possible during our testing procedure,
␣-pinene conversion remains too high in order to measure the reac-
tion rate under differential conditions. The replacement of H2 as a
carrier gas by N2 does not cause any considerable changes in the
␣-pinene conversion and product distribution (Entry 3). A blank
experiment (Entry 4) was done with ␥-Al2O3 lacking gold but sub-
jected to the same kind of chemical and thermal treatment that
was used at different stages of the gold catalyst preparation (except
that a solution of 0.01 M HCl was taken instead of HAuCl4 solution
for ␥-alumina impregnation). As a result, only very low conver-
sion (X = 2.7%) was observed with a wide product distribution in
which camphene and limonene were the main products (S = 28.3
and 14.6%). The data obtained suggest that, firstly, the vapour-phase
␣-pinene isomerization reaction proceeds on the Au/␥-Al2O3 cat-
alyst without chemical participation of H2 and, secondly, the gold
is necessary for the catalysis.
Fig. 4 shows the results of ␣-pinene isomerization over the Au/␥-
Al2O3 catalyst at various [␣-pinene]inlet values. It is clearly seen
that increasing ␣-pinene concentration in the inlet reaction mix-
ture by a factor of 10 induces the prominent deactivation of the
catalyst, the deactivation rate rising sharply with a further increase
in the reagent concentration (Fig. 4a). At [␣-pinene]inlet ≥ 4 vol%,
selectivity to camphene during the initial period of testing arises
to ca. 80% that apparently is related to a drop in ␣-pinene conver-
sion, but then decreases notably with time-on-stream along with
the catalyst activity (Fig. 4b). Initial period of ␣-pinene isomeriza-
tion does not represent a single value being dependent on pinene
concentration (ca. 1–5 min). During this time approximately the
same amount of ␣-pinene passed through the catalyst bed (e.g.
for higher pinene concentration the initial period was lower) in
order to assure that sampling times correspond to approximately
the same extent of deactivation.
The effects of residence time and temperature on ␣-pinene iso-
merization over Au catalyst were further studied at the highest
concentration of the reagent ([␣-pinene]inlet = 20 vol%), which was
used in this work. Fig. 5a displays ␣-pinene conversion measured
during the initial period of testing versus ꢁ for three different reac-
tion temperatures. The experimental data are well described by the
equation:
X = 1 − e−kꢁ
,
(1)
where X is conversion and k is the rate constant. This indicates the
reaction is first-order with respect to ␣-pinene. The k values for
different reaction temperatures determined by fitting the experi-
mental data to Eq. (1) are presented in Table 2. Plotting the graph
of ln(k) versus 1/T results in a straight line, as shown in Fig. 5b.
The apparent activation energy, Ea, for ␣-pinene isomerization over
the 2.2 wt% Au/␥-Al2O3 catalyst (within the temperature range
from 463 to 483 K) calculated from the slope of Arrhenius plot
is 51.0 kJ mol−1 (with 95% confidence limits). This value is close
Table 2
First-order reaction rate constants (with 95% confidence limits) for ␣-pinene iso-
merization over the 2.2 wt% Au/␥-Al2O3 catalyst at different temperature.a.
Temperature (K)
Rate constant, k (s−1), and deviation
463
473
483
3.60 0.10
4.75 0.10
6.24 0.12
Fig. 4. Time dependence of ␣-pinene conversion (a) and selectivity to camphene
and p-cymene (b) over the 2.2% Au/␥-Al2O3 catalyst at various concentrations of
␣-pinene in the inlet n-octane solution. Reaction conditions: temperature 473 K,
a
Reaction conditions: [␣-pinene]inlet 20 vol%, solvent n-octane, catalyst 0.2 g, H2
catalyst 200 mg, carrier gas H2, SV 2200 h−1
.
atmosphere, residence time 0.05–0.45 s.

140
I.L. Simakova et al. / Applied Catalysis A: General 385 (2010) 136–143
Fig. 5. (a) Conversion of ␣-pinene during the initial period of reaction over the 2.2% Au/␥-Al2O3 catalyst as a function of residence time for different temperatures. The symbols
represent the experimental data, while the lines represent the model fit to the experimental data. (b) Arrhenius plot of rate constant (k, s−1) of ␣-pinene isomerization. Reaction
conditions: [␣-pinene]inlet 20 vol%, solvent n-octane, catalyst 200 mg, carrier gas H2.
to the one obtained for ␣-pinene isomerization over TiO2 cata-
lyst (46.1 kJ mol−1) [27], which suggests that the rate-determining
steps are similar in both cases.
products of ␣-pinene isomerization, but it may result from sub-
sequent polymerization with the formation of high molecular
weight compounds, which are not detected by GC. Such preferential
oligomerization/polymerization of limonene could be explained
by the presence of the terminal double bond. It is also seen from
Fig. 6 that the selectivity to camphene, p-cymene, tricyclene and
limonene over Au catalyst does not depend on reaction temper-
ature, at least, in the range of 463–483 K. Thus, the apparent
activation energies for their formation are approximately the same
and, therefore, the difference in the rate constants for the forma-
tion of various products is probably due to the difference in the
activation entropies.
At present, the reason for high activity and selectivity of
the Au/␥-Al2O3 catalyst in ␣-pinene isomerization to camphene
cannot be elucidated mainly due to the lack of available infor-
mation on the nature of the active site. However, taking into
account that cationic gold(I) complexes have recently evolved as
Fig. 6 presents selectivity to the main products measured during
the initial period of ␣-pinene isomerization over the Au/␥-Al2O3
catalyst at high reagent concentration ([␣-pinene]inlet = 20 vol%)
as a function of the reagent conversion for different reaction
temperatures. Selectivity to camphene, p-cymene and tricyclene
varies little with conversion indicating that these products are
formed in parallel pathways. It is in a good agreement with the
conventional reaction mechanism suggested for ␣-pinene isomer-
ization over acid catalysts [15,21], which is shown in Scheme 1.
According to this, pinylcarbonium ion that is initially formed from
␣-pinene is the general precursor of all the reaction products.
It is further rearranged into bornylcarbonium and terpenilcarbo-
nium ions, the former giving camphene and tricyclene in parallel
steps, while the latter leading to p-cymene and limonene in parallel
steps too. In fact, selectivity to limonene changes with conver-
sion stronger (from 17.6 to 1.6%) than selectivity to other main
Fig. 6. Selectivity to significant products during the initial period of ␣-pinene
isomerization over the 2.2 wt% Au/␥-Al2O3 catalyst as a function of the reagent con-
version for different reaction temperatures. The reaction conditions: [␣-pinene]inlet
20 vol%, solvent n-octane, catalyst 200 mg, carrier gas H2, residence time 0.05–0.45 s.
Scheme 1. Mechanism of ␣-pinene isomerization over acid catalysts.

I.L. Simakova et al. / Applied Catalysis A: General 385 (2010) 136–143
141
excellent Lewis acid catalysts for various Wagner–Meerwein shifts
such as the ring expansion of cyclopropanols to the corresponding
cyclobutanones [28,29], participation of Au+ species in the catalytic
cycle of ␣-pinene isomerization over the Au/Al2O3 catalyst can
be hypothesized; even though they were not detected by XPS in
the present study. Application of XPS did not allow to detect Au
cations because Au particle sizes in the investigated catalysts are
about 3–5 nm (TEM), which is comparable with the depth of XPS
beam penetration. It means that Au4f7/2 intensity corresponds to
total concentration of metallic Au. Obviously it is difficult by XPS to
reveal relatively small amounts of Au cations in comparison with
much larger amounts of metallic Au even if Au cations are indeed
present in the sample.
Au+ species were presumed [24,30–32] to be present on the sur-
face together with the gold in the metallic state. In particular in
[24] it was demonstrated (including also catalysts used in the cur-
rent work), that regardless of the preparation techniques catalysts
contained metallic Au particles together with the surface ionic Au
oxide species. Other authors using FTIR spectroscopy of adsorbed
CO also showed that heating of reduced metallic gold in oxygen at
673 K leads to formation of a small fraction of Au+ sites isolated on
the support and oxidation of the gold atoms on the surface of the
gold particles to Au+ [33]. In the current work Au/␥-Al2O3 catalyst
was not studied by FTIR spectroscopy of adsorbed CO, however,
the catalyst was treated at conditions comparable with [33]: under
reducing atmosphere at the temperature of 673 K for 4 h followed
by oxidizing at the temperature of 673 K for 4 h.
The role of cationic and metallic Au was discussed in [34] where
a unique role of cationic interfacial gold species has been proposed.
The elegant experiments of Fu et al. [35] revealed the critical role of
cationic gold in the water–gas shift reaction. The authors deposited
gold onto a La-doped cerium oxide support by either deposition
precipitation or co-precipitation and then calcined the material in
air at 400 ◦C. Most of the gold was reduced to metallic particles
by the thermal treatment. A basic sodium cyanide solution was
then used to leach the metallic gold from the catalyst surface in
a process similar to gold extraction during mining. X-ray photo-
electron spectroscopy showed that in a co-precipitated sample, the
gold remaining on the support after leaching was exclusively ionic.
Although the leaching process removed 85% of the gold, the cat-
alytic activity per surface area was unaffected by the loss of metallic
particles. These results suggest that ionic gold strongly interacts
with the support and is the active site for the reaction.
3.3. Catalyst deactivation and regeneration
TEM study of the spent Au/␥-Al2O3 catalyst demonstrated
that the mean linear diameter of Au particles does not sig-
nificantly change during ␣-pinene isomerization at 473 K and
[␣-pinene]inlet = 20 vol% (cf. Fig. 2a and b). Generally, gold parti-
cles of 3–6 nm in diameter are observed both before and after
testing, even though the particle size distribution of the spent cat-
alyst is found to be somewhat broader due to appearance of few
larger Au particles with the sizes between 10 and 20 nm. Never-
theless, it seems that sintering of metal particles is not the main
reason for the deactivation of the catalyst. At the same time, in
the micrographs of the spent Au/␥-Al2O3 catalyst recorded under
the high-resolution (HR) conditions (Fig. 2c) the surface of the Au
crystallites was seen to be covered with a thin (subnanometer)
layer of disordered structure. The HRTEM images of these overlay-
ers exhibit very low contrast, indicating that relatively light atoms
(such as carbon) are exclusively present in the overlayer. A simi-
lar layer seemingly covers the Al2O3 surface as can be suggested
on the basis of low contrast and high fuzziness of the TEM picture
observed with the spent catalyst in comparison with the fresh one.
Fig. 7. DTG/DTA curves of (a) as-prepared and (b) spent Au/␥-Al2O3 catalyst during
TPO in flowing air at heating rate 10 K min−1. (c) The outlet product pressure profiles
during TPO of the spent Au/␥-Al2O3 catalyst.
Fig. 7 displays the results of a TPO study on the 2.2 wt%
Au/␥-Al2O3 catalyst before and after vapour-phase ␣-pinene
isomerization, which was carried out in the equipment for ther-
mogravimetric and differential thermal analysis (TGA and DTA,
respectively). The differential thermogravimetric (DTG) curve of
the fresh catalyst (Fig. 7a) exhibits broad peaks at 385 and >800 K.
The former one is related, most probably, to removal of adsorbed
water, while the latter can be assigned to elimination of OH groups
from the alumina surface (total loss of the sample weight during

142
I.L. Simakova et al. / Applied Catalysis A: General 385 (2010) 136–143
heating rate of 10 K min−1 and cooled down to 298 K in flowing O2.
Thereafter the catalyst was allowed to stay at room temperature for
12 h before measuring is catalytic activity. Fig. 8 demonstrates that
such kind of treatment restores ␣-pinene conversion over the gold
catalyst at 473 K, [␣-pinene]inlet = 4 vol% and ꢁ = 0.33 s up to 90% of
its initial value, while selectivity to camphene even increased from
ca. 70 to ca. 80%.
4. Conclusion
In summary, this work demonstrates for the first time that the
nanodispersed Au/␥-Al2O3 catalyst, prepared from HAuCl4 as a gold
precursor by impregnation procedure including treatment of the
reduced sample with NaOH for neutralization of acidic sites on the
support surface, shows good catalytic activity and selectivity for
vapour-phase isomerization of ␣-pinene tocampheneat 463–483 K
with the use of H2 or N2 (1 bar) as a carrier gas. Activity signifi-
cantly exceeded the catalytic performance of neat Al2O3 pretreated
with NaOH, which indicates that gold is necessary for this kind of
Wagner–Meerwein skeletal rearrangements. At low ␣-pinene con-
centration in the inlet reaction mixture ␣-pinene conversion over
Au/␥-Al2O3 catalyst is not significantly changed during the catalytic
run, at least, for 7–8 h. However, increasing ␣-pinene concentration
in the inlet reaction mixture induces the intense deactivation of the
catalyst. HRTEM and TPO results show that deposition of carbona-
ceous species, which may block the active sites of the Au/␥-Al2O3
catalyst, is the main reason for the catalyst deactivation. The calci-
nation of the deactivated catalyst in flowing O2 at temperatures up
to 923 K restores its activity and selectivity in ␣-pinene isomeriza-
tion to camphene almost completely. In an ongoing study, we are
optimizing the composition and conditions of preparation of gold
catalysts for this reaction and elucidating the reason for high activ-
ity and selectivity of supported gold in isomerization of ␣-pinene
to camphene.
Fig. 8. Time dependence of ␣-pinene conversion and selectivity to camphene over
the 2.2% Au/␥-Al2O3 catalyst taken as-prepared and after regeneration in flowing
O2 at 323–923 K (heating rate 10 K min−1). Reaction conditions: temperature 473 K,
[␣-pinene]inlet 4 vol%, solvent n-octane, catalyst 200 mg, carrier gas H2, residence
time 0.33 s.
TPO is 6.39%). The DTA curve of the fresh sample shows no signs
of any exothermal reaction. For the spent catalyst, the DTG curve
measured under TPO conditions (Fig. 7b) contains additionally the
peak at 685 K that corresponds to loss of 5.42% of the sample weight.
Simultaneously, the DTA curve of the spent catalyst shows the over-
lapping exhothermic peaks at 635, 685, 745 and 790 K, the most
distinct peak coinciding with the maximum weight loss observed
by means of TGA equipment. This may be attributed to combustion
of carbonaceous species deposited onto the catalyst surface. Indeed,
the formation of CO2 as the main oxidation product during TPO
of the spent Au/␥-Al2O3 catalyst was observed by means of mass
spectrometry (Fig. 7c) starting from 625 K up to about 775 K with a
maximum at 700 K. Along with CO2, the traces of H2O appear with
a maximum yield at 682 K, indicating that the exhothermic peaks
at 635 and 685 K are related to oxidation of hydrogen-enriched car-
bonaceous species deposited onto the catalyst surface, whereas the
exhothermic peaks at 745 and 790 K are caused by surface carbon
(coke) combustion. One can suggest that coke is formed on the
catalyst surface owing to adsorption of ␣-pinene and its isomers
followed by their oligomerization and secondary transformations
of the oligomers. Apparently, deposition of carbonaceous species
blocks the active sites of the Au/␥-Al2O3 catalyst being the main
reason for the catalyst deactivation. The rate of coke formation
should decrease as ␣-pinene concentration in the inlet reaction
mixture decreases. Probably, this is the reason the gold catalyst to
show stable activity at ␣-pinene concentration of 0.4 vol%, at least,
for 7–8 h.
As shown in Fig. 2d, TPO of the spent catalyst in flowing air
conducted up to 923 K restores high contrast and sharpness of TEM
images of the catalyst surface, but has no considerable effect on the
mean Au particle size and size distribution. This is in agreement
with the previous observations of very high resistance of alumina-
supported gold nanoparticles against sintering, probably, caused
by the epitaxial interaction between Au crystallites and alumina
surface with the formation of Au–O–Al bonds [32,36,37].
The high sintering stability of Au/␥-Al2O3 catalyst used in this
work provides the possibility of restoring its activity in ␣-pinene
isomerization by burning the carbonaceous species deposited onto
the catalyst surface at temperatures ≥800 K that are required for
coke burning, as follows from TPO experiments. In order to check
this assumption, the deactivated catalyst was calcined in a quartz
flow reactor with an O2 flow (40 cm3 min−1) initially at 323 K for
1 h, then during the temperature rise up to 923 K with a constant
Acknowledgements
The authors are grateful to Tatiana V. Larina, Eugene Yu. Gerasi-
mov, George A. Filonenko, Irina L. Kraevskaya and Pavel A. Pyrjaev
for their help in carrying out this work. This project is supported
by the Russian Foundation for Basic Research (grants nos. 08-03-
91758 and 09-03-12272). One of the authors (Yu.S.) acknowledges
the financial support from the Foundation for Assistance to Small
Innovative Enterprises.
References
[1] M. Haruta, Catal. Today 36 (1997) 153–166.
[2] G.C. Bond, C. Louis, D.T. Thompson, Catalysis by Gold, IC Press, London, 2006.
[3] P. Claus, Appl. Catal. A 291 (2005) 222–229.
[4] B. Grzybowska-Swierkosz, Catal. Today 112 (2006) 3–7.
[5] D.T. Thompson, Nanotoday 2 (2007) 40–43.
[6] V.V. Smirnov, S.A. Nikolaev, G.P. Murav’eva, L.A. Tyurina, A.Yu. Vasil’kov, Kinet.
Catal. 48 (2007) 265–270.
[7] O.A. Simakova, A.-R. Leino, B. Campo, P. Mäki-Arvela, K. Kordás, J.-P. Mikkola,
D.Yu. Murzin, Catal. Today 150 (2010) 32–36.
[8] A. Fasi, I. Palinco, K. Hernadi, I. Kirisci, Catal. Lett. 81 (2002) 237–240.
[9] F. Neatu, P.Y. Toulic, V. Michelet, V. Parvulescu, Pure Appl. Chem. 81 (2009)
2387–2396.
[10] F. Özkan, G. Gündüz, O. Akpolat, N. Bes¸ ün, D.Yu. Murzin, Chem. Eng. J. 91 (2003)
257–269.
[11] O. Chimal-Valencia, A. Robau-Sanchez, V. Collins-Martinez, A. Aguilar-
Elguezabal, Biores. Technol. 93 (2004) 119–123.
[12] K.A.D. Swift, Top. Catal. 27 (2004) 143–155.
[13] P. Maki-Arvela, B. Holmbom, T. Salmi, D.Yu. Murzin, Catal. Rev. 49 (2007)
197–340.
[14] A.I. Allahverdiev, S. Irandoust, D.Yu. Murzin, J. Catal. 185 (1999) 352–362.
[15] A.I. Allahverdiev, B. Andersson, D.Yu. Murzin, Appl. Catal. A: Gen. 198 (2000)
197–206.
[16] O. Akpolat, G. Gündüz, F. Özkan, N. Bes¸ ün, Appl. Catal. A: Gen. 265 (2004) 11–22.

I.L. Simakova et al. / Applied Catalysis A: General 385 (2010) 136–143
143
[17] A.D. Stefanis, G. Perez, A.A.G. Tomlinson, Appl. Catal. A: Gen. 132 (1995)
353–365.
[28] F. Kleinbeck, F.D. Toste, J. Am. Chem. Soc. 131 (2009) 9178–9179 (and references
therein).
[18] A. Severino, A. Esculcas, J. Rocha, J. Vital, L.S. Lobo, Appl. Catal. A: Gen. 142
(1996) 255–278.
[19] C.M. Lopez, F.J. Machado, K. Rodriguez, B. Mondez, M. Hasegawa, S. Pekerar,
Appl. Catal. A: Gen. 173 (1998) 75–85.
[29] S.G. Sethofer, S.T. Staben, O.Y. Hung, F.D. Toste, Org. Lett. 10 (2008) 4315–4318.
[30] G.C. Bond, D.T. Thompson, Gold Bull. 33 (2000) 41–50.
[31] H.-S. Oh, C.K. Costello, C. Cheung, H.H. Kung, M.C. Kung, Stud. Surf. Sci. Catal.
139 (2001) 375–381.
[20] C.B. Davis, J.J. McBride, US Patent 3,824,135 (1974).
[21] G.A. Rudakov, Chemistry and Technology of Camphor, Lesnaya promyshlen-
nost’, Moscow, 1976 (in Russian).
[22] S. Kullaj, Bull. Shkencave Nat. 43 (1989) 81–85.
[23] T. Yamamoto, T. Matsuyama, T. Tanaka, T. Funabiki, S. Yoshida, J. Mol. Catal. 155
(2000) 43–58.
[32] B.L. Moroz, D.A. Zyuzin, V.I. Zaikovskii, A.N. Shmakov, P.A. Pyrjaev, E.M. Moroz,
V.I. Bukhtiyarov, Extended abstracts of the 9th European Conference on Cataly-
sis “Catalysis for a Sustainable World” (EuropaCat IX), Salamanca, Spain, 2009,
pp. P10–108.
[33] Tz. Venkov, Hr. Klimev, M.A. Centeno, J.A. Odriozola, K. Hadjiivanov, Catal.
Commun. 7 (2006) 308–313.
[24] B.L. Moroz, P.A. Pyrjaev, V.I. Zaikovskii, V.I. Bukhtiyarov, Catal. Today 144 (2009)
292–305.
[25] A.B.P. Lever, Inorganic Electronic Spectroscopy, second ed., Elsevier, Amster-
dam, 1984.
[26] J.F. Moulder, W.F. Stickle, P.E. Sobol, K.D. Bomben, in: J. Chastain (Ed.), Handbook
of X-Ray Photoelectron Spectroscopy, Perkin-Elmer, Eden Prairie, Minnesota,
1992.
[34] R.J. Davis, Science 301 (2003) 926–927.
[35] Q. Fu, H. Saltsburg, M. Flytzani-Stephanopoulos, Science 301 (2003) 935.
[36] B.L. Moroz, P.A. Pyryaev, V.I. Zaikovskii, S.Ph. Ruzankin, T.V. Larina, V.F.
Anufrienko, V.I. Bukhtiyarov, in: V.N. Parmon, V.I. Bukhtiyarov, Z.R. Ismagilov
(Eds.), Proceeding from the III International Conference “Catalysis: Fundamen-
tals and Application”, Part I, Boreskov Institute of Catalysis, Novosibirsk, 2007,
p. 140.
[27] F. Ebmeyer, J. Mol. Struct. (Theochem.) 582 (2002) 251–255.
[37] S. Ivanova, C. Petit, V. Pitchon, Gold Bull. 39 (2006) 3–8.