Ni nanoparticles supported on microwave-synthesised hectorite for the hydrogenation of styrene oxide

Article abstract of DOI:10.1016/j.apcata.2011.09.002

Three hectorites were synthesised at different preparation conditions by aging with microwaves. One more hectorite was aged by conventional heating for comparison. Synthesized hectorites were used as supports of nickel nanoparticles for the catalytic hydrogenation of styrene oxide to obtain 2-phenylethanol. Ni/hectorite obtained by impregnation of a microwaved-synthesised hectorite with the highest nickel content (10 wt%) resulted in high active, high selective to 2-phenylethanol and high resistant to deactivation catalyst. Catalytic results were related to the different NiO-hectorite interactions observed by TPR together with the accessibility to the acid sites present in the supports. Both factors mainly depended on the hectorite purity and the use of microwaves for hectorite synthesis.

Full text of DOI:10.1016/j.apcata.2011.09.002

Applied Catalysis A: General 408 (2011) 31–37
Contents lists available at SciVerse ScienceDirect
Applied Catalysis A: General
journal homepage: www.elsevier.com/locate/apcata
Ni nanoparticles supported on microwave-synthesised hectorite for the
hydrogenation of styrene oxide
∗
Isabel Vicente, Pilar Salagre , Yolanda Cesteros
Departament de Química Física i Inorgànica, Universitat Rovira i Virgili, C/Marcel·lí Domingo s/n, 43007 Tarragona, Spain
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 29 June 2011
Received in revised form 31 August 2011
Accepted 2 September 2011
Available online 8 September 2011
Three hectorites were synthesised at different preparation conditions by aging with microwaves. One
more hectorite was aged by conventional heating for comparison. Synthesized hectorites were used
as supports of nickel nanoparticles for the catalytic hydrogenation of styrene oxide to obtain 2-
phenylethanol. Ni/hectorite obtained by impregnation of a microwaved-synthesised hectorite with the
highest nickel content (10 wt%) resulted in high active, high selective to 2-phenylethanol and high resis-
tant to deactivation catalyst. Catalytic results were related to the different NiO–hectorite interactions
observed by TPR together with the accessibility to the acid sites present in the supports. Both factors
mainly depended on the hectorite purity and the use of microwaves for hectorite synthesis.
Keywords:
Hectorite
Microwaves
Hydrothermal treatment
Nickel nanoparticles, Styrene oxide
©
2011 Elsevier B.V. All rights reserved.
2
-Phenylethanol
1
. Introduction
-Phenylethanol (2-PEA), the main component of rose oils, is
2-ethoxy-2-phenylethanol (2-EPE), catalysed by Lewis and
Brønsted acid sites with medium strength, and condensation
reactions, which are responsible for catalyst deactivation and are
catalysed by strong Brønsted acid sites in the presence of Lewis
acid sites [13–15].
2
widely used as component in chemical perfumes, and as additive
in foods, due to its pleasant smell [1]. Catalytic hydrogenation of
styrene oxide to obtain selectively 2-phenylethanol is a cleaner
alternative to the economical, environmental and purification
problems shown by classical industrial production methods, such
as the Friedel–Crafts alkylation of benzene with ethylene oxide
Hectorite is a trioctahedral 2:1 mineral clay with formula
n+
(Si8.0) [Mg
Lix] (OH·F) O20M
·mH O, where some substi-
6.0−x
4
x/n
2
tution of Mg(II) by Li(I) in the octahedral (Oh) sheet, causes the
negative charge of layers, which must be compensated by inter-
layer exchangeable cations. When these cations are exchanged by
[
2], reacting chlorobenzene with Grignard-type reactants followed
transition metal cations (Ni2+, Pd , Cu ), metallic nanoparticles
can be obtained in the interlayer space by reduction. Hectorites
can be used in nanocomposites preparation [16,17], as heteroge-
neous catalysts due to their acidity and thermal stability [18,19],
for ionic adsorption, ionic exchange processes and in nanoparti-
cles synthesis. Although clays can be found in nature, in order
to obtain a reproducible composition they must be synthesised.
Unfortunately, preparation of hectorites often involves long time
hydrothermal treatments [20].
2+
2+
by several reaction steps [3], or by ring opening the epoxide with
reduction agents such as hydrides or alkaline metals [4–6].
Catalytic hydrogenation of styrene oxide has been studied from
5
0s to solve the important problems mentioned above [7]. Bulk
nickel, palladium and platinum catalysts have shown to be good
catalysts for this reaction [8,9]. Their activity and selectivity were
considerably improved with the addition of basic solutions to the
reaction medium [9]. However, the basic medium could also favour
condensation reactions that may deactivate the catalyst [10]. In
previous studies, we reported the preparation of high selective and
high resistant to deactivation catalysts for this reaction using basic
solids as supports of nickel catalysts [11,12].
In a previous study, we synthesised hectorite clays using
microwaves, obtaining a reproducible material up to 24 times faster
than by conventional heating. However, residual amounts of amor-
phous silica were observed for all samples [21].
The aim of this study was to evaluate the catalytic behaviour
of several hectorite-supported nickel catalysts for the hydrogena-
tion of styrene oxide to 2-phenylethanol. The influence of using
microwave or conventional heating during hectorite aging, the
effect of residual silica in synthesised hectorites and the presence
of acid sites in competition with metal sites were also studied.
Other transformation reactions of styrene oxide can take place
in the presence of acid sites: isomerisation to phenylacetaldehyde
(
PA) catalysed by Brønsted acid sites; ring-opening reaction to
∗ Corresponding author. Tel.: +34 977559571; fax: +34 977559563.
E-mail address: pilar.salagre@urv.cat (P. Salagre).
0
926-860X/$ – see front matter © 2011 Elsevier B.V. All rights reserved.
doi:10.1016/j.apcata.2011.09.002

3
2
I. Vicente et al. / Applied Catalysis A: General 408 (2011) 31–37
Table 1
Preparation conditions of the catalysts.
Catalyst
wt% Ni
Nickel introduction by
Reduction parameters
Nis5-SiO2
Nis5-H1
Nis10-H1
Nis5-H1ac
Nis5-H1conv
Nis5-H2
5
5
10
5
5
5
2
2
2
2
Impregnation
Impregnation
Impregnation
Impregnation
Impregnation
Impregnation
Exchange
Exchange
Exchange
Exchange
Calcination at 623 K + reduction with H2 at 623 K
Calcination at 623 K + reduction with H2 at 623 K
Calcination at 623 K + reduction with H2 at 623 K
Calcination at 623 K + reduction with H2 at 623 K
Calcination at 623 K + reduction with H2 at 623 K
Calcination at 623 K + reduction with H2 at 623 K
Reduction with ethylene glycol; elimination at 453 K
Reduction with ethylene glycol; elimination at 453 K. Treatment with NaOH
Reduction with hydrazine; elimination at 473 K until pH 8
Reduction with hydrazine; elimination at 473 K until pH 7
Nie/1-H1
Nie/1NaOH-H1
Nie/2pH8-H1
Nie/2pH7-H1
Ni/hectorite catalysts were obtained by impregnation with differ-
ent nickel contents or by cation exchange.
(Nie/1NaOH-H1) and Nie/2pH8-H1 was heated under pure nitro-
gen at 473 K for 1 more h until the pH of released vapours was 7
(Nie/2pH7-H1).
2
. Experimental
2.3. Characterisation of samples
2.1. Synthesis of hectorites
2.3.1. X-ray diffraction (XRD)
XRD measurements were made using a Siemens D5000 diffrac-
tometer (Bragg–Brentano parafocusing geometry and vertical
Hectorites were prepared following the method reported by
Vicente et al. [21,22] in which brucite sheets were proposed to
act as crystallisation nuclei of hectorite. Preparation was carried
ꢀ
–ꢀ goniometer) fitted with a curved graphite diffracted-beam
◦
monochromator and diffracted-beam Soller slits, a 0.06 receiving
out as follows: 50 ml of a slurry composed by SiO , freshly pre-
2
slit, and scintillation counter as a detector. The angular 2ꢀ diffrac-
pared brucite Mg(OH)2 and LiF, with a SiO :Mg(OH) :LiF molar
2
2
◦
◦
tion range was between 2 and 70 . Sample was dusted on to a low
ratio = 7.5:3:1, was vigorously stirred for 1 h. Fresh brucite was pre-
viously synthesised by precipitating an aqueous solution of MgCl2
with ammonia. The slurry, which contained a 3 wt% of solids, was
aged by autoclaving with microwaves (Milestone Ethos Touch Con-
trol laboratory microwave equipment) at 393 K for 8 h (H1) and by
autoclaving with conventional heating at 393 K for 8 days (H1conv).
background Si(5 1 0) sample holder. The data were collected with
◦
an angular step of 0.05 at 3 s per step and sample rotation. Cu
k␣
radiation was obtained from a copper X-ray tube operated at 40 kV
and 30 mA.
The X-ray pattern was analysed using the Fundamental Param-
eters Approach convolution algorithm [25] as implemented in the
program TOPAS 3.0 [26]. This approach calculates the contribu-
tion to the reflection width produced by a specific instrument
configuration. The crystallite size was calculated from the net inte-
H1 was exchanged with NH NO3 and calcined at 623 K for 6 h to
4
obtain an acid hectorite (H1ac). Lastly, in order to synthesise a purer
hectorite, the SiO2 content was decreased in the starting slurry
using a SiO :Mg(OH) :LiF molar ratio = 4:3:1 [22]. This suspension
2
2
gral breadth of the reflections, ˇ [27], according to the following
was aged by autoclaving with microwaves at 393 K for 8 h (H2). All
samples were washed by dialysis and dried overnight in an oven at
53 K.
i
formula that comes from the Scherrer expression: ˇ = ꢁ/εcos ꢀ
i
where ꢁ is the X-ray wavelength, ε is the crystallite size and
is the Bragg angle. 060 reflection was used to calculate the
3
ꢀ
crystallite size of hectorites. Nickel crystallite size was calculated
from 111 reflection of metallic nickel phase for supported nickel
catalysts.
2
.2. Preparation of hectorite-supported nickel catalysts
Table
1 shows the preparation conditions of hectorite-
supported nickel catalysts. Nickel was introduced by two
procedures: impregnation and cation exchange. All hectorites were
impregnated with nickel nitrate in the appropriate amounts to
have 5 wt% of Ni in the final catalysts. Samples were calcined in air
at 623 K (NiOs5-H1, NiOs5-H1ac, NiOs5-H1conv, NiOs5-H2), and
reduced under pure hydrogen at 623 K for 6 h (Nis5-H1, Nis5-H1ac,
2
.3.2. Infrared spectroscopy (FTIR)
Infrared spectra were recorded on a Bruker-Equinox-55 FTIR
spectrometer. The spectra were acquired by accumulating 32
−
1
−1
scans at 4 cm resolution in the range of 400–4000 cm . Sam-
ples were prepared by mixing the powdered solids with pressed
KBr disks in a mass ratio of 1:250, and dried in an oven before
measurements.
Nis5-H1conv, Nis5-H2). The SiO used as reagent for the synthesis
2
of hectorites was also impregnated, calcined and reduced at the
same conditions to provide a benchmark for comparison (Nis5-
2
.3.3. X-ray fluorescence (XRF)
Elemental analyses of the samples were obtained with a Philips
SiO ). One more catalyst was obtained by impregnating H1 with
2
nickel nitrate to have 10 wt% of Ni in the final catalyst, which was
obtained after calcination–reduction at the same conditions than
the above samples (Nis10-H1).
PW-2400 sequential XRF analyser with Phillips Super Q software.
Analyses were made by triplicate for each sample.
Ni2+-exchanged hectorites were prepared by stirring 35 ml of
Ni(NO ) ·6H O 0.25 M with 1 g of H1 at 338 K for 2 h (NiOe-H1).
2.3.4. N2 adsorption
3
2
2
N -adsorption–desorption isotherms were recorded at 77 K
2
This sample was reduced by two procedures: (a) reduction with
ethyleneglycol and later elimination of reducing agent excess at
using a Micromeritics ASAP 2000 surface analyser. Prior to anal-
ysis samples were outgassed at 353 K. Specific surface areas were
calculated from the BET method.
4
53 K for 5 h under vacuum (Nie/1-H1) following the method
reported by De Lima et al. [23], and (b) reduction with hydrazine
24] and later elimination of the hydrazine excess under pure
nitrogen at 473 K for 1 h until the pH of released vapours was
(Nie/2pH8-H1). In order to improve the activity of the cata-
lysts, Nie1-H1 was treated with a 0.01 M NaOH solution for 10 min
[
2.3.5. Temperature-programmed reduction
TPR experiments were carried out using a TPD/R/O 1100
(Thermo Finnigan) equipped with a programmable temperature
8

I. Vicente et al. / Applied Catalysis A: General 408 (2011) 31–37
33
Fig. 1. XRD patterns oh samples H1, H1conv, H1ac and H2.
furnace and a calibrated TCD detector. Experiments were per-
were smaller. In the case of H1ac, this may be due to some
layer disaggregation occurred during the exchange step. The lower
crystallite size of H1conv could be associated with its aging by
conventional heating, which seems to be less effective than with
formed with 5% H /Ar flowing through the sample, that was heated
2
from 303 K to 1173 K at 10 K/min.
microwaves.
2.3.6. Transmission electronic microscopy (TEM)
The basal spacing was higher for sample H2 (14.2 A˚ ) than for the
TEM images were collected using a JEOL 1011 Transmission
rest of samples (12.6–13.3 A˚ ) (Table 2). This may be related to the
Electron Microscope operating at 80 kV and magnification values
of 200–800k. Samples were dispersed in ethanol, and a drop of
resultant suspensions were poured on carbon coated-copper grids.
presence of higher amounts of interlamellar cations, in agreement
with its higher CEC value (Table 2).
Fig. 2 shows the FTIR spectra taken for samples H1, H1conv
and H1ac. IR bands were assigned taking into account the char-
acterisation data reported for clay samples [20,29,30]. The band
2
.3.7. Cation exchange capacity (CEC)
Cation exchange capacity (CEC) of hectorites was determined by
the method reported by Bergaya and Vayer [28]. The same method
was followed for silica (CEC < 5 meq/100 g sample).
−
1
around 1027 cm , which corresponds to the Si–O–Si ordered in
the clay mineral layer, was observed for all samples, with higher
intensity for H2 in agreement with XRD results. The appearance
−
1
of one intense band around 1105 cm together with a shoulder
2
.4. Catalytic activity determination
−1
at 1200 cm was related to the presence of the amorphous silica
phase, being more intense for H1, according again to XRD results.
This also agrees with the higher CEC value observed for the sample
with less amorphous silica content (H2, Table 2) since CEC of silica
Catalytic hydrogenation of styrene oxide was performed in the
liquid phase, using, for all tests, 1 g of catalyst, 150 mL of absolute
ethanol (Panreac, 96%) and 2 mmol of styrene oxide (Aldrich, 97%)
with a hydrogen flow of 2 mL/s and mechanic stirring of 500 rpm.
The reaction was performed at room temperature. Sample was
taken at 5 min, 1, 2, and 6 h of reaction. Impregnated catalysts were
reused at least for 3 times. The reaction products were analysed
by gas chromatography, using a chromatograph model Shimadzu
GC-2010 equipped with a 30 m capillary column DB-1 coated with
phenylmethylsilicon and a FID detector. They were quantified by
adding an internal standard and using calibration lines.
−1
was ∼0. The other IR band observed at 666 cm for all samples was
associated to vibrations related to Mg–OH librations.
From XRF results, Si/Mg and Mg/Li ratios were determined
(Table 2). The Si/Mg ratios of samples H1 and H1conv were higher
than that expected for a theoretical pure hectorite (Si/Mg = 1.7,
when x = 1.2) due to the presence of the amorphous silica. H1 had
much higher Mg/Li ratio (6.1) than sample H1conv (3.6) (Table 2).
Assuming that magnesium and lithium are in the hectorite struc-
ture (in the layers or in the interlamellar space), the higher Mg/Li
ratio of H1 could be attributed to the presence of higher amounts of
magnesium ions in the interlamellar space. Higher BET areas were
observed for H1ac and H2 than for H1 and H1conv (Table 2). In the
case of H1ac, this could be related to some disorder in stacking,
occurred during its preparation, which resulted in a loss of defini-
tion of the 001 reflection. The higher BET area of sample H2 could be
explained by the almost totally absence of silica, which may block
the pores of the other samples.
3
. Results and discussion
3
.1. Characterisation of hectorites
XRD patterns of all samples showed the presence of reflections
◦
characteristic of clays minerals and a broad band around 22.8 cor-
responding to amorphous silica resulting in a worse definition of
the 004 reflection (Fig. 1). The silica band was more appreciable in
samples H1, H1conv and H1ac than in H2. Therefore, residual silica
was lower in H2.
3.2. Catalytic activity
H1 and H2 had similar crystallinity, as deduced from their
similar crystallite size, calculated from the 060 reflection of XRD
patterns (Table 2). In contrast, crystallite sizes of H1ac and H1conv
Hectorites (H1, H1ac, H1conv, and H2) and the silica used as
reagent showed negligible conversion values for the styrene oxide

3
4
I. Vicente et al. / Applied Catalysis A: General 408 (2011) 31–37
Table 2
Characterisation data of hectorite samples.
2
Sample
Si/Mg ratio
Mg/Li ratio
d (0 0 1) (nm)
Crystallite size (0 6 0) (nm)
CEC (meq/100 g)
BET area (m /g)
H1
2.7
–
2.3
1.9
6.1
–
3.6
5.5
13.3
13.2
12.6
14.2
9.3
7.4
7.9
9.9
64
–
53
75
162
219
165
199
H1ac
H1conv
H2
hydrogenation, styrene oxide isomerisation, and styrene oxide
ring-opening reactions after 6 h of reaction. The presence of weak
(Fig. 4). The first peak had low intensity for both samples. Since dif-
ferent NiO–support interactions were observed for each precursor,
their corresponding catalysts may have nickel sites with differ-
ent characteristics, and consequently different activity should be
expected. In a previous study, Ni/saponite catalysts were tested for
this reaction [31]. TPR curves with similar shape were obtained for
their precursors. On the whole, we observed that, the more active
the catalyst, the higher the intensity of the third reduction peak.
This was related to NiO sites with high interaction with the support.
Therefore, the higher conversion of catalyst Nis5-H1conv could be
attributed to the presence of higher content of metal nickel sites
with higher influence of support that results in a higher intensity
of the third peak of its precursor (NiOs5-H1conv) when compared
to that of NiO5s-H1.
High selectivity to 2-PEA was explained in nickel catalysts
promoted by basic solutions through the formation of ␲ benzyl
complex from the adsorbed styrene oxide, that yielded a primary
alkoxide ion, which was stabilised by NaOH, and finally hydro-
genated to the primary alcohol [9]. Another possible pathway in
basic medium was the formation of the secondary alkoxide ion,
which through hydride attack gave 2-PEA [3,9,32].
The high selectivity to 2-PEA of these Ni/hectorite catalysts
without basic promotion could be explained by the forma-
tion of a primary alkoxide ion, stabilised by acid sites, which
yielded the primary alcohol by hydrogenation. Although conver-
sion was negligible for H1, H1ac, H1conv and H2, the subsequent
calcination–reduction steps performed to obtain the catalysts may
result in some modifications of their acidic properties. Actually, the
presence of Brønsted acid sites in Nis5-H1conv was confirmed since
phenylacetaldehyde (PA) was detected at shorter reaction times.
This compound was quickly hydrogenated to 2-PEA in the first two
reuses. However, PA remained longer during the third run although
after 2 h the only product observed was 2-PEA. For catalyst Nis5-
H1conv, the higher accessibility to the acid sites, because of the
lower crystallinity of the initial hectorite, could be also respon-
sible for the deactivation of the catalyst due to the formation of
condensation products from PA [13–15].
acid sites may explain these results. Nis5-SiO catalyst yielded 25%
2
of conversion and 100% of selectivity to 2-PEA for the hydrogena-
tion of styrene oxide at 1 h of reaction but deactivated from the first
reuse.
Catalysts Nis5-H1 and Nis5-H1conv, which were obtained by
impregnating with 5 wt% of nickel content hectorites H1 (synthe-
sised with microwaves) and H1conv (synthesised by conventional
heating), showed total selectivity to 2-PEA (Table 3). Catalyst Nis5-
H1conv showed higher conversion (88%) than catalyst Nis5-H1
(
40%). However, the catalyst prepared from the hectorite synthe-
sised by conventional heating suffered higher deactivation since
after 3 runs for 1 h of reaction conversion decreased to 60%. On the
other hand, for Nis5-H1 no deactivation after 3 runs was observed
(
3
Table 3). Both catalysts maintained total selectivity to 2-PEA after
runs.
TEM images of both catalysts (Fig. 3) showed the presence of
dispersed nanoparticles with sizes around 8–15 nm (Fig. 3a).
In order to understand the differences observed in the catalytic
behaviour of these catalysts, the reducibility of their precursors was
studied by TPR. The two thermograms showed mainly three peaks
Catalyst Nis5-H2, prepared by impregnating with 5 wt% of nickel
content hectorite H2, which had lower silica content than H1,
showed total selectivity to 2-PEA and higher activity (78% conver-
sion) than Nis5-H1 (Table 3). After the third run, its conversion
dropped to 58% and 2% of PA was observed. The appearance of PA
as by-product after one hour of reaction on the third reuse (Table 3),
and the deactivation observed for Nis5-H2 can be explained by the
higher accessibility of the reactants to the acid sites because of the
higher surface area of this support (H2, Table 2).
When TPR curve of its precursor (NiOs5-H2, Fig. 4) was ana-
lysed, similar shape to that of NiOs5-H1conv, was observed. This
explains the similar catalytic behaviour obtained for both catalysts
(Table 3). Moreover, TPR of NiOs5-H2 showed higher intensity of
the first reduction peak than NiOs5-H1conv. The intensity of this
peak can be related to NiO that can be easily reduced to Ni particles
that afterwards became agglomerated because of their low inter-
action with the support. This resulted in the formation of bigger
Ni nanoparticles in Nis5-H2 (around 20–25 nm) than in Nis5-H1,
as observed by TEM (Fig. 5). Apart from these Ni particles, other Ni
particles of around 5–10 nm were detected for this catalyst (Fig. 5).
Fig. 2. FTIR spectra of samples.

I. Vicente et al. / Applied Catalysis A: General 408 (2011) 31–37
35
Table 3
Catalytic results at 1 h of reaction for samples Nis5-H1, Nis10-H1, Nis5-H1ac, Nis5-H1conv and Nis5-H2 after 1 and 3 runs.
Catalyst
1 run
3 runs
Conversion
Sel. to 2-PEA (%)
Conversion
Sel. to PA (%)
Sel. to 2-PEA (%)
Nis5-H1
40
100
66
88
78
100
100
100
100
100
40
100
54
58
58
n.d.
n.d.
4
1
2
100
100
96
99
98
Nis10-H1
Nis5-H1ac
Nis5-H1conv
Nis5-H2
Sel., selectivity; PA, phenylacetaldehyde; 2-PEA, 2-phenylethanol, n.d., non-detected.
Fig. 3. TEM images of samples Nis5-H1 and Nis5-H1conv.
One more catalytic test was performed with catalyst Nis5-H1ac,
which was prepared by impregnating with a 5 wt% of nickel content
an acidified hectorite (H1ac). This catalyst showed total selectivity
to 2-PEA and higher conversion (66%) than Nis-H1 (Table 3). TEM
image of Nis5-H1ac (Fig. 6) exhibited the presence of Ni nanopar-
ticles with sizes around 8–15 nm, similar to those observed for
Nis5-H1 (Fig. 3).
In order to explain the catalytic behaviour of Nis5-H1ac, TPR
of its catalytic precursor was performed. TPR curve of NiOs5-H1ac
(Fig. 7) showed again three peaks: the first one with low inten-
sity, and the second and third peaks the intensities of which were
higher than those of NiOs5-H1 (Fig. 7). This is in agreement with
the higher activity observed for Nis5-H1ac when compared with
Nis5-H1. However, 4% of phenylacetaldehyde was obtained after
one hour of the third reuse for catalyst Nis5-H1ac (Table 3). The
appearance of PA together with the decrease of activity observed
Fig. 4. TPR of samples NiO5-SiO2, NiOs5-H1, NiOs5-H1conv and NiOs5-H2.
Fig. 5. TEM image of sample Nis5-H2 ×500k.
Fig. 6. TEM image of sample Nis5-H1ac ×500k.

3
6
I. Vicente et al. / Applied Catalysis A: General 408 (2011) 31–37
Fig. 8. TEM image of sample Nis10-H1 ×400k.
The use of ethylene glycol to reduce Ni2+ in the exchanged hec-
torite H1 (catalyst Nie-H1/1) gave well dispersed Ni nanoparticles,
with sizes around 2–5 nm, as observed by TEM (Fig. 10a). The low
conversion (30%, Fig. 9) obtained after 2 h of reaction for this cata-
lyst may be related to the low percentage of nickel in these samples
(2%) together with the presence of some active acid sites, which
initially catalysed the formation of PA and afterwards deactivated
the catalysts by the formation of condensation products [15]. In
order to neutralise these acid sites, this catalyst was treated with
a NaOH 0.05 M solution after reduction followed by a drying step
prior to the catalytic test (Ni-H1/NaOH). Conversion and selectivity
to 2-PEA slightly increased to 40% and 15%, respectively, at 2 h of
reaction. However, catalyst deactivation was also observed.
Fig. 7. TPR curves of catalytic precursors.
may be related to the formation of condensation products catalysed
by the acid sites of the support that deactivated the catalyst.
In the TEM image of catalyst Nis10-H1, we observed Ni-
nanoparticles with two main average diameters (20 and 10 nm).
This catalyst was prepared using H1 as support but increasing the
Ni content to 10 wt%. Total conversion and total selectivity to 2-PEA
were obtained after 1 h of reaction for this catalyst. This catalytic
result maintained after 3 runs.
TPR curve of NiOs10-H1 showed higher intensity of the three
peaks than those of NiOs5-H1 (Fig. 7). The nickel nanoparticles of
0 nm observed by TEM for catalyst Nis5-H1 (Fig. 8) could be related
to the higher intensity of the first reduction peak since there was a
higher amount of NiO particles that could be more easily reduced.
This favoured the existence of more agglomerated Ni particles. The
high selectivity to 2-PEA and high conversion of this catalyst may
be related to the higher amount of Ni metal interacting strongly
with the support as confirmed by the higher intensity of the third
peak from the TPR curve of its precursor, NiOs10-H1.
Hydrazine was used to reduce Ni2+ exchanged hectorite H1 with
the aim to control the hectorite acid sites taking advantage of its
basicity. The hydrazine excess was eliminated by heating under
pure nitrogen at 473 K for 1 h until the pH of released vapours was
8 (Nie/2pH8-H1). When this catalyst was tested for the hydrogena-
tion of styrene oxide, extremely low conversion after 2 h of reaction
(8%, Fig. 9), was obtained. This could be related to an important
blockage of metal and acid sites by remaining hydrazine. In order
to improve the behaviour of this catalyst, hydrazine was eliminated,
after the reduction step, under pure nitrogen at 473 K for 3 h until
the pH of released vapours was 7 (Nie/2pH7-H1). The resulting cat-
alyst showed Ni particles with sizes around 10–15 nm (Fig. 10b). A
significant improvement of conversion (50%, Fig. 9) at 2 h of reac-
tion was observed when using this catalyst although only 30% of
selectivity to 2-PEA was obtained (Fig. 9). The high selectivity to PA
2
Fig.
9 illustrates conversion and selectivity results of
Ni/hectorite catalysts obtained by cation exchange (Nie-H1/1,
Ni-H1/NaOH, Nie-H1/2pH8 and Nie-H1/2pH7). Activity of the
silica reagent submitted to the same preparation procedure than
the exchanged catalysts was also tested. No adsorption of nickel
by silica was observed and conversion was negligible.
Fig. 9. Catalytic results of samples Nie-H1/1, Nie-H1/1NaOH, Nie-H1/2pH8and Nie-H1/2pH7.

I. Vicente et al. / Applied Catalysis A: General 408 (2011) 31–37
37
Fig. 10. TEM images of samples (a) Nie/1-H1 ×500k and (b) Nie/2pH7-H1 ×500k.
(
70%, Fig. 9) and the catalyst deactivation could suppose that not
[2] K. Bauer, D. Garbe, H. Surburg, Common Fragrance, Flavor Mate-
rials Preparation, Properties and Uses, Wiley-VCH, Holzminden,
only the hydrazine that blocked the metal sites was eliminated, but
also the hydrazine that blocked the acid ones.
2001.
[
3] C.V. Rode, M.M. Telkara, R.V. Chaudhar, in: A. Corma, F.V. Melo, S. Mendioroz,
J.L.G. Fierro (Eds.), Studies in Surface Science and Catalysis, Elsevier, Amster-
dam, 2000, pp. 533–538.
4
. Conclusions
[
[
4] E.L. Eliel, D.W. Delmonte, J. Org. Chem. 21 (1956) 596–597.
5] S. Krishnamurthy, R.M. Schubert, H.C. Brown, JACS 95 (1973) 8486–8487.
The use of microwaved-synthesised hectorites as supports of
[6] R. Sreekumar, R. Padmakumar, P. Rugmini, Tetrahedron Lett. 39 (1998)
695–2696.
2
nickel nanoparticles for the catalytic hydrogenation of styrene
oxide resulted in high active, high selective to 2-phenylethanol and
high resistant to deactivation catalytic systems.
[
[
7] T.F. Wood, N.J. Clifton, US Patent 2524096 (1950).
8] C.A. Gibson, L.F. Theiling, US Patent 4,064,186 (1977).
[9] S. Mitsui, S. Imaizumi, M. Hisashige, Y. Sugi, Tetrahedron 29 (1973)
4093–4097.
The activity of Ni/hectorite, obtained from impregnation of hec-
torites with nickel nitrate and later calcination–reduction, mainly
depended on the use of microwave during hectorite synthesis, the
nickel content and the hectorite purity. These results were related
to the presence of different amounts and accessibility of acid sites,
and also to metal nickel sites with different activity since different
NiO–hectorite interactions were observed by TPR for their pre-
cursors. Ni/hectorites, prepared by Ni2 cation exchange of one
hectorite synthesised with microwaves and later calcined–reduced,
showed lower conversion, lower selectivity to 2-phenylethanol and
higher selectivity to phenylacetaldehyde than impregnated cata-
lysts. Conversion and selectivity to 2-PEA increased when higher
amount of hydrazine, which was used as reducing agent, was
removed. However, the hydrazine elimination was not selective
for the metal sites. The phenylacetaldehyde formation and the cat-
alyst deactivation could be explained by the accessibility to the acid
sites.
[
[
10] V.G. Yadav, S.B. Chandalia, Org. Proc. Res. Dev. 2 (1998) 294–297.
11] O. Bergadà, P. Salagre, Y. Cesteros, F. Medina, J.E. Sueiras, Appl. Catal. A: Gen.
331 (2007) 19–25.
[12] O. Bergadà, P. Salagre, Y. Cesteros, F. Medina, J.E. Sueiras, Contr. Catal. Lett. 122
2008) 259–266.
(
[
13] O. Bergadà, E. Boix, P. Salagre, Y. Cesteros, F. Medina, J.E. Sueiras, Appl. Catal. A:
Gen. 368 (2009) 163–169.
[14] M.D. González, Y. Cesteros, P. Salagre, F. Medina, J.E. Sueiras, Micropor. Mesopor.
Mater. 118 (2009) 341–347.
+
[
15] I. Salla, O. Bergadà, P. Salagre, F. Medina, J.E. Sueiras, T. Montanari, J. Catal. 232
2005) 239–245.
[16] X. Liu, C. Breen, Macromol. Rapid Commun. 26 (2005) 1081–1086.
(
[
[
17] S. Xue, T.J. Pinnavaia, Micropor. Mesopor. Mater. 107 (2008) 134–140.
18] M. Casagrande, L. Storaro, M. Lenarda, S. Rossini, Catal. Commun. 6 (2005)
568–572.
[19] R.S. Varma, Tetrahedron 58 (2002) 1235–1255.
[
[
20] J.T. Kloprogge, E. Komarneni, S.J. Amonette, Clays Clay Miner. 47 (1999)
29–554.
21] I. Vicente, P. Salagre, Y. Cesteros, F. Guirado, F. Medina, J.E. Sueiras, Appl. Clay
Sci. 43 (2009) 103–107.
5
[22] I. Vicente, P. Salagre, Y. Cesteros, Phys. Proc. 8 (2010) 88–93.
23] R.B. De Lima, V. Paganin, T. Iwasita, W. Vielstich, Electrochim. Acta 49 (2003)
5–91.
[24] Z. Li, C. Han, J. Shen, J. Mater. Sci. 41 (2006) 3473–3480.
[
8
Acknowledgments
[
[
[
25] R.W. Cheary, A.A. Coelho, J. Appl. Cryst. 25 (1992) 109–121.
26] A.A. Coelho, TOPAS v3.0, Bruker AXS, 2005.
27] A.R. Stokes, A.J.C. Wilson, Proc. Camb. Philos. Soc. 38 (1942) 313–322.
The authors are grateful for the financial support of the
Ministerio de Educación y Ciencia of Spain and FEDER funds
[28] F. Bergaya, M. Vayer, Appl. Clay Sci. 12 (1997) 275–280.
[29] J.T. Kloprogge, R.L. Frost, Vib. Spectrosc. 23 (2000) 119–127.
(
CTQ2008-04433/PPQ) and for the FPU Grant (AP2006-00835) also
[
30] J.D. Rusell, A.R. Fraser, in: M.J. Wilson (Ed.), Clay Mineralogy: Spectro-
scopic and Chemical Determinative Methods, Chapman & Hall, London, 1994,
pp. 11–67.
financed by the Ministerio de Educación y Ciencia of Spain.
References
[31] I. Vicente, P. Salagre, Y. Cesteros, Appl. Clay Sci. 53 (2011) 212–219.
32] C.V. Rode, M.M. Telkar, R. Jaganathan, R.V. Chaudhari, J. Mol. Catal. A: Chem.
00 (2003) 279–290.
[
2
[1] B.D. Mookherjee, R.A. Wilson, R.E. Kirk, Encyclopedia of Chemical Technology,
vol. 4, fourth ed., Jonh Wiley, New York, 1996.