Selective hydrogenation of benzaldehyde to benzyl alcohol over Au/Al 2O3

Article abstract of DOI:10.1016/j.catcom.2011.09.017

We provide the first evidence of exclusive benzyl alcohol production in the continuous gas phase hydrogenation of benzaldehyde (T = 393 K; P = 1 atm) over Au/Al₂O₃, where activity was largely maintained for up to 15 h on-stream. XRD and DRS UV-vis have established the presence of metallic Au post temperature programmed reduction to 473 K, with a surface area weighted mean Au particle size (from TEM analysis) of 7.9 nm. Under the same reaction conditions, Ni/Al₂O₃ and Pd/Al₂O₃ were non-selective, generating toluene as the principal product via hydrogenolysis of benzaldehyde and exhibiting an appreciable temporal loss of activity. The results establish Au/Al₂O₃ as a promising catalyst for the production of benzyl alcohol.

Full text of DOI:10.1016/j.catcom.2011.09.017

Catalysis Communications 16 (2011) 159–164
Contents lists available at SciVerse ScienceDirect
Catalysis Communications
journal homepage: www.elsevier.com/locate/catcom
Short Communication
Selective hydrogenation of benzaldehyde to benzyl alcohol over Au/Al O
2
3
1
Noémie Perret, Fernando Cárdenas-Lizana , Mark A. Keane ⁎
Chemical Engineering, School of Engineering and Physical Sciences, Heriot-Watt University, Edinburgh EH14 4AS, Scotland
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 5 August 2011
Received in revised form 14 September 2011
Accepted 14 September 2011
Available online 29 September 2011
We provide the first evidence of exclusive benzyl alcohol production in the continuous gas phase hydrogena-
tion of benzaldehyde (T=393 K; P=1 atm) over Au/Al , where activity was largely maintained for up to
5 h on-stream. XRD and DRS UV–vis have established the presence of metallic Au post temperature pro-
grammed reduction to 473 K, with a surface area weighted mean Au particle size (from TEM analysis) of
.9 nm. Under the same reaction conditions, Ni/Al and Pd/Al were non-selective, generating toluene
as the principal product via hydrogenolysis of benzaldehyde and exhibiting an appreciable temporal loss of
2 3
O
1
7
2
O
3
2 3
O
Keywords:
activity. The results establish Au/Al
2
O
3
as a promising catalyst for the production of benzyl alcohol.
© 2011 Elsevier B.V. All rights reserved.
Benzaldehyde
gas phase hydrogenation
benzyl alcohol
Au/alumina
1
. Introduction
hydrogenolysis to generate toluene and benzene [7]. Moreover, catalyst
deactivation from coke deposition has been a common feature of the
catalysts tested to date [9]. In gas phase operation, activity has been re-
lated to the strength of reactant interaction while selectivity is governed
by the competitive adsorption of benzaldehyde and benzyl alcohol [14].
Where benzaldehyde is strongly held on the surface, e.g. in the presence
of strong basic sites, benzene formation is preferred [15]. The use of
Benzyl alcohol is a commercially important chemical (with a
projected production of 60 kt in 2011 [1]), employed as a solvent
for inks, paints, lacquers and as precursor for the manufacture of a
range of esters in the cosmetics and flavoring industries [2]. The stan-
dard industrial production route has involved the hydrolysis of benzyl
chloride in alkaline media but the associated production of dibenzyl
ether with chlorine release represent serious drawbacks that have
now limited the applicability of this approach [3]. Two alternative
catalytic routes, generally in liquid phase, are: (i) oxidation of toluene,
typically over vanadia [4]; (ii) disproportionation of benzaldehyde
2 2 2
reducible oxides (TiO , ZrO or CeO ) favours interaction via the
hydroxyl oxygen with the resultant generation of toluene, whereas
benzyl alcohol interaction with irreducible supports is limited [11].
Moreover, a high surface capacity for hydrogen uptake (including spill-
over) is associated with increased activity and selectivity to toluene [12].
In this report, we consider, for the first time, the use of supported
Au to promote the selective hydrogenation of benzaldehyde to benzyl
alcohol. Nevertheless, it should be noted that the reverse reaction
(
Cannizzaro reaction) over irreducible basic oxides such as MgO [5].
The oxidation route generates significant quantities of benzoic acid
and benzaldehyde, requiring additional downstream separation/
purification, while the use of additives (such as sodium bromide) to
increase yield and/or avoid formation of by-products results in reactor
corrosion [6]. The Cannizzaro reaction has not been studied to any sig-
nificant extent and the low rates achieved (0.2 molalcohol molMgO h )
have prevented commercial implementation. The catalytic hydrogena-
tion of benzaldehyde represents an alternative, which has now been
promoted using transition metals (Pt, Pd, Ru and Ni) supported on
(oxidation of benzyl alcohol) has been studied over Au (on TiO
2
and
Al ) at elevated pressures (up to 147 atm) with a higher selectivity
2 3
O
(close to 100%) to benzaldehyde than that obtained with Pt and Pd
catalysts [16]. Catalysis by Au has attracted appreciable research
activity as demonstrated in the review by Hutchings et al. [17] but
the work to date has largely dealt with oxidation reactions involving
CO, alcohol, aldehyde, and diol reactants [18]. By comparison, Au pro-
moted hydrogenation has not been considered to the same extent,
focusing largely on the conversion of ketones, aldehydes and nitroar-
enes [19]. Although supported Au delivers lower hydrogenation rates
than conventional transition metal catalysts (e.g. Pd, Pt and Ni),
enhanced performance in terms of selectivity and stability has been
reported, notably in the hydrogenation of aldehydes to alcohols (e.g.
crotonaldehyde, citral, acrolein) [20]. Moreover, there is some evi-
dence [21] that edge sites on Au nanoparticles favour C=O adsorp-
tion/activation and hydrogenation. In this report, we link critical
-
1
oxides (Al
2 3
O , ZrO
2 2 2 2
, CeO , TiO and SiO ), largely in the liquid phase
(T=295–405 K; P=1–60 atm) [7-10] but with limited gas phase [11-
1
3] application. However, C=O hydrogenation selectivity is challeng-
ing and has not exceeded 80% over Group VIII metals due to a catalysed
⁎
Corresponding author. Tel.: +44 131 451 4719.
E-mail address: M.A.Keane@hw.ac.uk (M.A. Keane).
Present address: Group of Catalytic Reaction Engineering, Ecole Polytechnique Fédérale
1
de Lausanne (GGRC-ISIC-EPFL), Lausanne CH-1015, Switzerland.
2 3
Au/Al O characterisation results with benzaldehyde hydrogenation
1566-7367/$ – see front matter © 2011 Elsevier B.V. All rights reserved.
doi:10.1016/j.catcom.2011.09.017

160
N. Perret et al. / Catalysis Communications 16 (2011) 159–164
data in continuous flow gas phase operation and compare perfor-
mance with Ni/Al and Pd/Al
(Oxford Instruments) operated at an accelerating voltage of 200 kV
using Gatan DigitalMicrograph 3.4 for data acquisition/manipulation.
The sample for analysis was prepared by dispersion in acetone and
deposited on a holey carbon/Cu grid (300 Mesh). The mean Au parti-
cle size is given as a surface area-weighted mean (dTEM) according to:
2
O
3
2 3
O .
2
. Experimental
2
.1. Catalyst Preparation and Activation
3
∑i nidi
dTEM
¼
ð2Þ
Au/Al
2
O
3
(1.1% w/w) was prepared by standard impregnation of
solu-
∑i nid2i
Al
2
O
3
(5 g, Puralox, Condea Vista Co.) with an aqueous HAuCl
4
-
3
-1
tion (Aldrich, 7.3×10 M). The slurry was heated at 2 K min to
53 K under vigorously stirring (ca. 600 rpm) and maintained in a
where n
i
is the number of particles of diameter d
Characterisation details for the Pd/Al
been published previously [22].
i
and Σ n
i
=650.
3
2 3 2 3
O and Ni/Al O catalysts have
-1
He purge. The solid residue was dried in He at 2 K min to 383 K
for 3 h, sieved into a batch of 75 μm average particle diameter and
stored under He at 278 K (in the dark). The Au loading was deter-
mined by inductively coupled plasma-optical emission spectrometry
2
.3. Catalysis Procedure
(
ICP-OES, Vista-PRO, Varian Inc.). For comparative purposes,
Pd/Al and Ni/Al prepared by impregnation with Pd(NO
and Ni(NO [22], were employed as reference catalysts. Prior to
use in catalysis, Au/Al , Pd/Al and Ni/Al were activated in
at 2 K min to a final temperature in the range
73–723 K. The samples were passivated in 1% v/v O /He at room
Reactions were carried out under atmospheric pressure at
2
O
3
2
O
3
3 2
)
T=393 K, in situ immediately after activation, in a fixed bed vertical
glass tubular reactor (60 cm×15 mm i.d.). The catalytic reactor and
operating conditions to ensure negligible heat/mass transport limita-
tions, have been fully described elsewhere [23] but some features,
pertinent to this study, are given below. A layer of borosilicate glass
beads served as preheating zone, ensuring that the organic reactant
was vaporized and reached reaction temperature before contacting
the catalyst. Isothermal conditions (± 1 K) were ensured by diluting
the catalyst bed with ground glass (75 μm). The reaction temperature
was continuously monitored using a thermocouple inserted in a ther-
mowell within the catalyst bed. Benzaldehyde (or benzyl alcohol) as a
solution in ethanol was delivered at a fixed calibrated flow rate to the
reactor via a glass/teflon air-tight syringe and teflon line using a mi-
croprocessor controlled infusion pump (Model 100 kd Scientific). A
co-current flow of benzaldehyde (or benzyl alcohol) and ultra pure
3 2
)
2
O
3
2
O
3
2 3
O
3
-1
-1
6
4
0 cm min
H
2
2
temperature for ex situ analysis.
2
.2. Catalyst Characterization
Temperature programmed reduction (TPR) and BET surface area
analysis were conducted using the commercial CHEM-BET 3000
Quantachrome) unit. The Au/Al sample was loaded into a U-
shaped Quartz cell (10 cm×3.76 mm i.d.) and heated in
(
2 3
O
3
-1
1
7 cm min (Brooks mass flow controlled) 5% v/v H
2
/N
min to 473± 1 K. The effluent gas passed through a liquid N
and changes in H consumption were monitored by TCD with data
2
at 2 K
-1
2
trap
2
4
-1
H
2
(b1% v/v organic in H
with an inlet reactant molar flow (F)=4.8×10 mol h . The H
2
) was maintained at a GHSV=2×10 h
acquisition/manipulation using the TPR Win™ software. The reduced
sample was maintained at the final temperature in a constant flow of
H /N until the TCD signal returned to baseline, swept with
2 2
-5
-1
2
con-
tent was up to 3000 times in excess of the stoichiometric require-
ment, the flow rate of which was monitored using a Humonics
3
-1
6
5 cm min flow N
sample was then subjected to H
10 μl) titration procedure. Hydrogen pulse introduction was repeat-
2
for 1.5 h and cooled to room temperature. The
(
Model 520) digital flowmeter. The molar metal to inlet molar organ-
2
chemisorption using a pulse
-3
-3
ic feed rate (n/F) ratio spanned the range 5×10 –88×10 h. In a se-
ries of blank tests, passage of reactant in a stream of H through the
empty reactor or over the Al support did not result in any detect-
able conversion. A flow of benzaldehyde in a stream of He over
Au/Al generated traces quantities of benzyl alcohol via dispropor-
(
2
ed until the signal area was constant, indicating surface saturation. In
a series of blank tests, chemisorption measurements on the alumina
support did not result in any hydrogen uptake. BET area was recorded
with a 30% v/v N
At least 3 cycles of N
employed to determine total surface area using the standard single
point method. BET surface area and H uptake values were reproduc-
ible to within± 5%; the values quoted represent the mean. UV–vis
spectra of the HAuCl aqueous solution and Au/Al were obtained
2 3
O
2 3
O
2 2
/He flow using pure N (99.9%) as internal standard.
tionation [5] where the amount produced was less than 3% of that
formed by selective hydrogenation under the same reaction condi-
tions. The reactor effluent was frozen in a liquid nitrogen trap for sub-
sequent analysis, which was made using a Perkin-Elmer Auto System
XL gas chromatograph equipped with a programmed split/splitless
2
adsorption-desorption in the flow mode were
2
4
2 3
O
injector and
0 m×0.33 mm i.d., 0.20 μm film thickness capillary column (J&W
Scientific). Benzaldehyde (Fluka, ≥ 98% w/w purity), benzyl alcohol
Riedel-de Haën, ≥99% w/w purity) and ethanol (Sigma Aldrich, ≥
9% v/v) were used without further purification. Fractional benzalde-
hyde conversion (x) is defined by
a flame ionization detector, employing a DB-1
using, respectively, a Spectronic Unicam HEλIOS β UV–vis Spectro-
photometer and a Perkin-Elmer Lambda 35 UV–vis Spectrometer
5
(
4
with BaSO powder as reference). Powder X-ray diffractograms
(
9
were recorded on a Bruker/Siemens D500 incident X-ray diffractom-
eter with Cu Kα radiation. The sample was scanned at a rate of
0
fied using the JCPDS-ICDD reference standards, i.e. Al (10–0425)
and Au (04–0784). Metal particle size (dhkl) was estimated using
the Scherrer equation:
-
1
.02° step over the range 20°≤2θ≤80°. Diffractograms were identi-
2 3
O
½
benzaldehydeꢀ −½benzaldehydeꢀ
in
out
x ¼
ð3Þ
½
benzaldehydeꢀ
in
where the subscripts “out” and “in” refer to the outlet and inlet gas
streams, respectively. Selectivity in terms of benzyl alcohol as the
target product is given by
K×λ
β × cosθ
dhkl
¼
ð1Þ
where K=0.9, λ is the incident radiation wavelength (1.5056 Å), β is
the peak width at half the maximum intensity and θ represents the
diffraction angle corresponding to the main plane associated with
metallic Au (2θ=38.1°). Au particle morphology and size were also
determined by transmission electron microscopy analysis; JEOL JEM
½
benzyl alcoholꢀ
out
Sbenzyl alcohol% ¼
×100
ð4Þ
½
benzaldehydeꢀ −½benzaldehydeꢀ
in
out
Repeated reactions with the same batch of catalyst delivered a
product composition that was reproducibility to within± 3%.
2
011 HRTEM unit with a UTW energy dispersive X-ray detector

N. Perret et al. / Catalysis Communications 16 (2011) 159–164
161
B
415 K
A
D
2
0
40
60
80
2θ
A
3
85 nm
5
45 nm
C
4
00
600
800
4
00
600
800
Wavelength (nm)
Wavelength (nm)
20 nm
3+
Au
Au
473 K
B
3
00
350
400
450
500
T (K)
Fig. 1. (A) UV–vis spectrum of HAuCl
spectrum and (D) XRD pattern for Au/Al
2 3
on JCPDS-ICDD reference data: (♦) γ-Al O
4
aqueous solution, (B) TPR profile, (C) UV–vis
. Note: XRD peak assignments are based
(10–0425); (w) Au ( 04–0784).
2 3
O
3
. Results and Discussion
3
.1. Catalyst Characterization
The UV–vis spectrum recorded for the Au precursor solution is
presented in Fig. 1A where the band at ca. 385 nm can be attributed
3
+
to Au [24]. The TPR of the Au/Al
2 3
O precursor generated the profile
shown in Fig. 1B, which is characterised by a single peak with
5 nm
3
+
T
max =415 K. A single step reduction of cationic gold (Au ) to the
zero valent state has been recorded elsewhere for Au/Al prepared
2 3
O
by deposition-precipitation at 436 K [25] and 490 K [26]. Moreover,
hydrogen consumption during TPR (Table 1) matched that required
for the reduction of the precursor to the metallic form. It should be
noted that both cationic (Au and Au ) [27] and metallic [28] gold
have been proposed as the active species in hydrogenation reactions.
Fig. 2. Representative (A) medium and (B) high magnification TEM images (with asso-
ciated diffractogram pattern) of Au/Al
2 3
O .
+
3+
XRD profile (Fig. 1D) presents peaks at 2θ=45.7° and 66.7° that cor-
respond, respectively, to the (400) and (440) planes of cubic γ-Al
The BET surface area for Au/Al
2
O
3
(Table 1) was lower than that of the
2 3
O ,
2
-1
starting Al
2
O
3
support (172 m g ), suggesting partial pore filling by
where the broadness of the peaks is indicative of short range order. In
addition, peaks at 2θ=38.1°, 44.3°, 64.6° and 77.5° can be assigned to
the (111), (200), (220) and (311) planes of metallic Au. Standard X-
the metal component [11]. The UV–vis spectrum (Fig. 1C) of the acti-
vated catalyst exhibits an absorption band with maximum intensity
(Amax) at ca. 545 nm characteristic of zero valent gold [25]. The wave-
length corresponding to maximum band intensity is dependent on Au
dispersion with a shift to lower wavelengths with decreasing Au par-
ticle size. The value quoted in this work is close that reported in the
0
.3
0.16
0.12
A
0.2
literature [25] for highly dispersed Au/Al
O
2 3
(Amax =540 nm;
0
0
.1
.0
d=4 nm) suggesting the presence of Au nano-particles. The UV
response is consistent with a supported Au0 phase post-TPR. The
0
1 2 3 4 5 6
0
0
0
.08
.04
.00
t (h)
Table 1
Characteristics of the Au/Al
2 3
O catalyst.
Au loading (% w/w)
1.1
2
-1
BET area (m
TPR Tmax (K)
TPR H
consumption (μmol g-1
g
)
161
415
78
B
2
)
UV–vis Amax (nm)
545
8.5
1-20
7.9
0
1
2
3
4
a
dhkl (nm)
2
b
10 n/F (h)
Au particle size range (nm)
b
dTEM (nm)
Fig. 3. (A) Variation of benzaldehyde fractional conversion (x) with time-on-stream
O (n/F=3.5×10 h); (B) Pseudo-first order kinetic plot for the hydroge-
2 3
nation of benzaldehyde to benzyl alcohol (P=1 atm; T=393 K).
a
-2
from XRD analysis: see Eq.(1).
from TEM analysis.
over Au/Al
b

162
N. Perret et al. / Catalysis Communications 16 (2011) 159–164
Benzene
(
II)
3 H2
CH CH O
OCH2CH3
O
3
2
O
(VIII)
(VI)
3
H2
2
C H OH
2
5
Benzaldehyde diethyl acetal
Benzaldehyde
Hexahydro-benzaldehyde
H
2
(
IX)
(I)
H2
H
(IV)
O
5
2
H
2
C
OCH CH3
2
OH
(X)
(III)
C H OH
H2
2
5
Benzyl ethyl ether
Toluene
Benzyl alcohol
(V)
3 H2
(VII) 3 H2
HO
Cyclohexylmethanol
Methylcyclohexane
Fig. 4. Possible reaction pathways associated with the hydrogenation of benzaldehyde to the target benzyl alcohol ( ), isolated by-products (➔) and reaction products reported in
the literature ( ).
ray line broadening analysis based on the Scherrer formula (Eq. (1))
yields a mean Au particle size of 8.5 nm (Table 1). Representative
(dTEM, Table 1) of 7.9 nm which is in good agreement with that
obtained from XRD analysis.
(
2 3
A) medium and (B) high magnification TEM images of Au/Al O
are given in Fig. 2, which reveal discrete Au particles at the nano-
scale that exhibit a quasi-spherical morphology. The diffractogram
pattern for an isolated Au particle is included as an inset in Fig. 2B
where the associated d-spacings (0.20/0.23) are consistent with the
3.2. Catalytic Activity/Selectivity
The temporal dependence of benzaldehyde fractional hydrogena-
tion (x) over Au/Al
(for 6 h on-stream) conversion is in evidence. Benzaldehyde hydroge-
nation over Au/Al generated benzyl alcohol as the sole product, i.e.
2 3
O is shown in Fig. 3A where a time invariant
(
200) and (111) planes of metallic gold (0.204/0.235 nm). Although
the nature of hydrogen/Au interactions has yet to be conclusively
established, the consensus that emerges from the literature [29]
suggests that chemisorption occurs predominantly at step, edge and
corner sites, which are more prevalent on smaller particles. More-
over, a well dispersed Au phase at the nano-scale (b10 nm) is
required in order to optimize the catalytic response. A total count of
2 3
O
100% selective in promoting exclusive carbonyl group reduction to
the alcohol. Reaction selectivity is challenging in benzaldehyde hy-
drogenation as a range of intermediates and by-products are possible
as shown in Fig. 4, which includes all the reaction pathways that have
been proposed in the literature [5,7,10,12,13,30,31]; step I represents
the target reduction of the carbonyl function. In both liquid and gas
6
50 Au particles delivered a surface area-weighted metal diameter

N. Perret et al. / Catalysis Communications 16 (2011) 159–164
163
Table 2
Hydrogen chemisorption, pseudo-first order specific rate constants (k), initial reaction selectivities (S
0
) and those obtained after 15 h on-stream (S15 h) for the conversion of benz-
2 3 2 3 2 3
aldehyde and benzyl alcohol over Au/Al O , Pd/Al O and Ni/Al O .
Catalyst
H
2
uptake
Benzaldehyde feed
Benzyl alcohol feed
-
1
(
μmol g )
k
(
S
0
(%)
S
0
(%)
S
15h (%)
S
15h (%)
k
S
0
(%)
0
S (%)
mmol m m- 2etal h-1
)
benzyl alcohol
toluene
benzyl alcohol
toluene
(mmol mmetal h-1
)
-2
toluene
benzene
Au/Al
Pd/Al
2
O
3
b1
24
4
0.5
2.6
3.1
100
10
0
0
90
100
100
95
-
0
5
-
b0.01
1.7
1.4
0
100
100
100
0
0
2
O
3
a
a
Ni/Al
2
O
3
a
no detectable activity after 15 h on-stream.
phase operation, catalytic hydrogenolysis with the formation of tolu-
ene and benzene is promoted [5,13]. It has been proposed [5] that
benzene is formed directly from benzaldehyde (step II) while toluene
can result from the subsequent conversion of the alcohol (step III)
and hydrogenolysis of benzaldehyde (step IV). Saadi and co-workers
2 3
alcohol over Au/Al O was maintained over extended time on-stream
(see Table 2). The formation of benzyl alcohol (from benzaldehyde)
has been proposed to occur via a nucleophilic mechanism where the
carbonyl function is activated at the metal/support interface [11]. Ad-
sorption of benzaldehyde where both the aromatic ring and carbonyl
group are co-planar to the surface facilitates C=O polarization, ren-
dering the oxygen susceptible to nucleophilic attack with proton
transfer to generate the alcohol [35]. The nature of the metal/support
interface and the degree of metal dispersion are considered to be crit-
ical in determining benzyl alcohol formation [11]. Initially, toluene
[
12] isolated methylcyclohexane as a result of the further reduction of
toluene (step V) during the gas phase hydrogenation of benzaldehyde
over Ni/Al . In liquid phase operation over Ru/C, hydrogenation of the
2 3
O
aromatic ring (steps VI and VII) generated hexahydro-benzaldehyde [30]
and cyclohexylmethanol [10] from benzaldehyde and benzyl alcohol,
respectively. The nature of the solvent can influence both activity and se-
lectivity [13] where the use of ethanol has resulted in benzaldehyde
diethyl acetal formation by reaction with benzaldehyde (step VIII) and
benzyl ethyl ether, formed from both benzaldehyde (step IX) and benzyl
alcohol (step X) [7,31]. In our exclusive production of benzyl alcohol from
was the principal product resulting from reaction over Pd/Al
2 3
O
with a low selectivity (10%) to benzyl alcohol while Ni/Al generat-
2 3
O
ed toluene as the only product. These selectivity trends find general
agreement with the literature [7,9,12,23]. Initial selectivities to benzyl
alcohol of 20-40% [9] and 10-20% [12] have been reported for
benzaldehyde over Au/Al
2
O
3
, etherification of the reactant and/or product
2 3 2 3
Pd/Al O (in liquid phase) and Ni/Al O (in gas phase, T=383–
with the solvent (ethanol) did not occur and there was no detectable
hydrogenolysis to generate toluene and/or benzene.
413 K), respectively. Toluene production from benzaldehyde is
favoured by a strong interaction between the carbonyl oxygen and
the surface with hydrogenolytic cleavage of C=O by dissociated hy-
drogen (Fig. 4, step IV) [14]. In order to probe the possible reaction
pathway, the conversion of benzyl alcohol under the same reaction
conditions was considered and the results are included in Table 2.
The rate constants generated for the alcohol feed were consistently
lower than those obtained for the reaction of benzaldehyde and this
can be attributed to the electrophilic character of C=O which renders
this functionality more reactive relative to O-H [13]. Transformation
of benzyl alcohol to toluene (Fig. 4, step III) has been proposed to pro-
ceed via adsorption/activation on the support (formation of a benzo-
Benzaldehyde reaction kinetics have been reported for both liquid
[8,9] and gas phase [11,12] operation. The applicability of a pseudo-
first order kinetic treatment can be tested using the relationship [32]
ꢀ
ꢁ
−1
n
F
lnð1−x0Þ ¼ k
ð5Þ
0
where x represents the initial fractional conversion and n/F has the
physical meaning of contact time. The linear relationship between
-
1
ln(1-x
first order behaviour and the associated specific rate constant (k) is
given in Table 2. In order to explore the potential of the Au/Al cat-
alyst, we employed Ni/Al and Pd/Al as benchmark catalysts.
0
)
and n/F, shown in Fig. 3B, confirms adherence to pseudo-
ate on Al
2
O
3
) and reaction with hydrogen from the metal site [35].
from benzaldehyde
2
O
3
The rate of toluene production over Ni/Al O
2 3
O
2 3
O
2 3
was significantly higher (by a factor of 2) than that obtained from
benzyl alcohol and path IV (Fig. 4) can be taken to be the predomi-
nant route to toluene. The rate of conversion of benzyl alcohol over
Au/Al O was two orders of magnitude lower than benzaldehyde hy-
2 3
drogenation; trace quantities of benzene were detected. The low
The activities given in Table 2 are in good agreement with values
quoted in the literature [9,12,33] where both Pd and Ni catalysts
delivered higher (by a factor up to 6) hydrogenation rates relative
to Au. A lower hydrogenation activity for supported Au can be attrib-
uted to the higher energy barrier for H
uptake (at ambient temperature) on Au/Al
lower than that measured for Pd/Al and Ni/Al
terms of sustainable processing, the initial reaction rate is not the
only consideration and catalyst lifetime must also be addressed. This
feature was investigated for reaction over a prolonged period (15 h)
and the results are presented in Fig. 5, where temporal conversion
2
dissociation [34]. Hydrogen
was significantly
(Table 2). In
2 3
O
1.0
0.8
0.6
2
O
3
2
O
3
is related to the initial value. While Au/Al
tial level of conversion (x15h/x =ca. 0.9), Ni/Al
immediate decline in activity to deliver negligible conversion after
h on-stream while Pd/Al showed a continual decrease that was
particularly marked at extended times (N 8 h on-stream). A temporal
catalyst deactivation during benzaldehyde hydrogenation has been
observed previously, attributed to coke deposition and associated
with toluene formation [9].
2
O
3
largely retained the ini-
0
2 3
O
exhibited a drastic
0
0
0
.4
.2
.0
3
2 3
O
0
4
8
12
16
t (h)
Initial reaction selectivities (S
0
) are presented in Table 2. In com-
or Ni/Al did not result
mon with Au/Al , reaction over Pd/Al
2
O
3
2
O
3
2 3
O
Fig. 5. Ratio of benzaldehyde fractional conversion with time on-stream (x) to the
initial value (x ) for reaction over Au/Al (■), Pd/Al (▲) and Ni/Al (●)
in any aromatic ring hydrogenation, hydrogenolysis to form benzene
or reactant etherification/acetalisation. Reaction exclusivity to benzyl
0
2
O
3
2
O
3
2 3
O
(n/F=3.5×10- h; P=1 atm; T=393 K).
2

164
N. Perret et al. / Catalysis Communications 16 (2011) 159–164
[
[
2] B. Nair, International Journal of Toxicology 20 (2001) 23–50.
3] V. van Brunt, J.S. Kanel, in: S. Kulprathipanja (Ed.), Reactive Separation Processes,
Taylor & Francis, London, 2002, pp. 51–92.
reactivity of the alcohol on Au/Al
clusivity with respect to the benzaldehyde→benzyl alcohol step. In
the case of Pd/Al , benzyl alcohol conversion generated toluene as
the sole product. The temporal decline in activity exhibited by
Pd/Al during benzaldehyde conversion was accompanied by a
2 3
O must contribute to reaction ex-
2
O
3
[4] A. Bottino, G. Capannelli, A. Comite, R. di Felice, Catalysis Today 99 (2005)
71–177.
[
1
5] D. Haffad, U. Kameswari, M.M. Bettahar, A. Chambellan, J.C. Lavalley, Journal of
Catalysis 172 (1997) 85–92.
2 3
O
switch in selectivity from toluene as the predominant product to ben-
zyl alcohol. This response indicates a temporal inhibition of steps III
and IV, limiting hydrogenolysis, while reaction via step I proceeds,
resulting in the preferential production of benzyl alcohol. Ultimately,
hydrogenation step (I) is limited with a residual conversion recorded
after 15 h on-stream. There was no detectable toluene formation in
[6] C.-C. Guo, Q. Liu, X.-T. Wang, H.-Y. Hu, Applied Catalysis A: General 282 (2005)
55–59.
[
7] D. Procházková, P. Zámostný, M. Bejblová, L. Červený, J. Čejka, Applied Catalysis A:
General 332 (2007) 56–64.
[8] D. Divakar, D. Manikandan, G. Kalidoss, T. Sivakumar, Catalysis Letters 125 (2008)
277–282.
[
9] F. Pinna, F. Menegazzo, M. Signoretto, P. Canton, G. Fagherazzi, N. Pernicone,
Applied Catalysis A: General 219 (2001) 195–200.
10] P. Kluson, L. Červený, Journal of Molecular Catalysis A: Chemistry 108 (1996)
any reaction over Au/Al
in behaviour from both Pd/Al
2
O
3
, which represents a marked divergence
and Ni/Al . We attribute this re-
[
[
[
O
3
2 3
O
107–112.
2
11] A. Saadi, Z. Rassoul, M.M. Bettahar, Journal of Molecular Catalysis A: Chemistry
sponse to fundamental differences in surface reactive hydrogen and
arene reactant activation at the metal/support interface.
164 (2000) 205–216.
12] A. Saadi, R. Merabti, Z. Rassoul, M.M. Bettahar, Journal of Molecular Catalysis A:
Chemistry 253 (2006) 79–85.
[
[
13] M.A. Vannice, D. Poondi, Journal of Catalysis 169 (1997) 166–175.
14] R. Merabti, K. Bachari, D. Halliche, Z. Rassoul, A. Saadi, Reaction Kinetics, Mecha-
nisms and Catalysis 101 (2010) 195–208.
4
. Conclusion
We have established, for the first time, the viability of benzalde-
hyde hydrogenation over Au/Al O as a sustainable alternative route
2 3
to benzyl alcohol. The reaction has been conducted under relatively
mild reaction conditions (393 K, 1 atm) in continuous flow gas
phase operation where 100% selectivity with respect to benzyl alco-
hol was achieved and maintained for 15 h on-stream. Temperature
[15] M.A. Peralta, T. Sooknoi, T. Danuthai, D.E. Resasco, Journal of Molecular Catalysis
A: Chemistry 312 (2009) 78–86.
[
16] P. Haider, B. Kimmerle, F. Krumeich, W. Kleist, J.-D. Grunwaldt, A. Baiker, Catalysis
Letters 125 (2008) 169–176.
[17] G.J. Hutchings, M. Brust, H. Schmidbaur, Chemical Society Reviews 37 (2008)
1759–1765.
[
[
[
18] M. Besson, P. Gallezot, Catalysis Today 57 (2000) 127–141.
19] L. McEwan, M. Julius, S. Roberts, J.C.Q. Fletcher, Gold Bulletin 43 (2010) 298–306.
20] P. Claus, Applied Catalysis A: General 291 (2005) 222–229.
2 3
programmed reduction (TPR) of Au/Al O to 473 K generated a sup-
ported Au metal phase, as confirmed by XRD and UV–vis analyses,
characterized by quasi-spherical particles with a surface weighted
mean metal diameter=7.9 nm. While Pd/Al O and Ni/Al O deliv-
2 3 2 3
[21] P. Claus, H. Hofmeister, C. Mohr, Gold Bulletin 37 (2004) 181–186.
[
22] F. Cárdenas-Lizana, S. Gómez-Quero, M.A. Keane, Applied Catalysis A: General
34 (2008) 199–206.
23] M.A. Keane, Journal of Molecular Catalysis A: Chemistry 118 (1997) 261–269.
3
[
ered higher activity (under the same reaction conditions), both
exhibited severe deactivation with time-on-stream and promoted
hydrogenolysis to toluene. We attribute the exclusive selectivity
achieved with Au/Al O to a polarization of the C=O bond rendering
2 3
the oxygen susceptible to nucleophilic hydrogen attack to generate
the alcohol, which does not undergo further reaction.
[24] M. Harada, H. Einaga, Langmuir 23 (2007) 6536–6543.
[
25] A.C. Gluhoi, X. Tang, P. Marginean, B.E. Nieuwenhuys, Topics in Catalysis 39
2006) 101–110.
(
[
26] C.K. Costello, J. Guzman, J.H. Yang, Y.M. Wang, M.C. Kung, B.C. Gates, H.H. Kung,
The Journal of Physical Chemistry. B 108 (2004) 12529–12536.
[
[
27] X. Zhang, H. Shi, B.-Q. Xu, Journal of Catalysis 279 (2011) 75–87.
28] P. Claus, A. Brückner, C. Mohr, H. Hofmeister, Journal of the American Chemical
Society 122 (2000) 11430–11439.
[
29] G. Bond, C. Louis, D.T. Thompson, in: G.J. Hutchings (Ed.), Catalysis by Gold, Cata-
lytic Science Series, Vol.6, Imperial College Press, London, 2006, p. 150.
30] L. Červený, Z. Bĕlohlav, M.N.H. Hamed, Research on Chemical Intermediates 22
Acknowledgements
[
(
1996) 15–22.
The authors acknowledge the contribution of Dr. S. Gómez-Quero
to this project and EPSRC support for free access to the TEM facility at
the University of St Andrews.
[31] M. Arai, A. Obata, Y. Nishiyama, Journal of Catalysis 166 (1997) 115–117.
[32] M.A. Keane, Applied Catalysis A: General 271 (2004) 109–118.
[
33] T. Osaki, T. Hamada, Y. Tai, Reaction Kinetics and Catalysis Letters 78 (2003)
17–223.
34] B. Hammer, J.K. Nørskov, Nature 376 (1995) 238–240.
2
[
References
[35] K. Lanasri, A. Saddi, K. Bachari, D. Halliche, O. Cherifi, Studies in Surface Science
and Catalysis 174 (2008) 1279–1282.
[1] SIDS Initial Assessment report, OECD SIDS, Benzoates, CAS N° 65-85-0, 532-32-1,
582-25-2, 100-51-6, UNEP publications, 2001, p. 9.