Selective hydrogenation of nitrocyclohexane to cyclohexanone oxime by alumina-supported gold cluster catalysts

Article abstract of DOI:10.1016/j.molcata.2011.05.018

Metal oxides (Al₂O₃, SiO₂, MgO)-supported Au cluster catalysts prepared by colloid deposition method and well established Au/TiO₂ prepared by deposition-precipitation method were tested for the selective reduction of nitrocyclohexane into cyclohexanone oxime. The activity and selectivity depended strongly on the size of Au and support material. Au/Al₂O₃ with smaller Au particle size (2.5 nm) was the most effective catalyst, exhibiting a high cyclohexanone oxime yield (86%) under mild condition (0.6 MPa H₂, 100 °C). To study the origin of support and size effects, in situ FTIR experiments for OH/D ₂ exchange reaction and nitrobenzene adsorption combined with poisoning experiments using pyridine and acetic acid were carried out. It is suggested that cooperation of coordinatively unsaturated Au atoms and the acid-base pair site (Al^δ+-O^δ- site) plays an important role in H₂ dissociation step. Al^δ+-O ^δ- site also acts as the adsorption site of the nitro group.

Full text of DOI:10.1016/j.molcata.2011.05.018

Journal of Molecular Catalysis A: Chemical 345 (2011) 54–59
Contents lists available at ScienceDirect
Journal of Molecular Catalysis A: Chemical
journal homepage: www.elsevier.com/locate/molcata
Selective hydrogenation of nitrocyclohexane to cyclohexanone oxime
by alumina-supported gold cluster catalysts
a,∗
b
c
b
Ken-ichi Shimizu , Takumi Yamamoto , Yutaka Tai , Atsushi Satsuma
a
Catalysis Research Center, Hokkaido University, N-21, W-10, Sapporo 001-0021, Japan
Department of Molecular Design and Engineering, Graduate School of Engineering, Nagoya University, Nagoya 464-8603, Japan
Materials Research Institute for Sustainable Development, Chubu Research Base of National Institute of Advanced Industrial Science and Technology (AIST), Nagoya 463-8560, Japan
b
c
a r t i c l e i n f o
a b s t r a c t
Article history:
Metal oxides (Al O , SiO , MgO)-supported Au cluster catalysts prepared by colloid deposition method
2
3
2
Received 8 December 2010
Received in revised form 12 May 2011
Accepted 21 May 2011
and well established Au/TiO2 prepared by deposition–precipitation method were tested for the selective
reduction of nitrocyclohexane into cyclohexanone oxime. The activity and selectivity depended strongly
on the size of Au and support material. Au/Al2O3 with smaller Au particle size (2.5 nm) was the most
Available online 27 May 2011
effective catalyst, exhibiting a high cyclohexanone oxime yield (86%) under mild condition (0.6 MPa H2,
◦
1
00 C). To study the origin of support and size effects, in situ FTIR experiments for OH/D2 exchange
Keywords:
Selective hydrogenation
Gold catalysts
reaction and nitrobenzene adsorption combined with poisoning experiments using pyridine and acetic
acid were carried out. It is suggested that cooperation of coordinatively unsaturated Au atoms and the
acid–base pair site (Al␦ –O site) plays an important role in H2 dissociation step. Al –O site also acts
+
␦−
␦+
␦−
Cyclohexanone oxime
as the adsorption site of the nitro group.
©
2011 Elsevier B.V. All rights reserved.
1
. Introduction
Adopting this method, we have recently investigated the support
effect of Au cluster catalysts for the selective hydrogenation of nitro
group in the presence of other reducible functional groups and
found that small Au clusters supported on a non-reducible support,
Al O , gave the highest activity and selectivity [9]. Based on the fun-
Since Haruta discovered a remarkable activity of supported
gold nanoparticles (NPs) in CO oxidation, many publications have
been devoted to clarify the factors controlling the activity of gold
catalysts [1–6]. Among them the most relevant appears to be
particle size, the nature of the support and preparation method.
For many catalytic systems, including oxidation and hydrogena-
tion reactions, there is a general tendency that Au NPs on inert
oxides, such as SiO2 and Al O , exhibit lower activity for various
2
3
damental studies, we proposed that cooperation of the acid–base
pair site on Al O3 and the coordinatively unsaturated Au atoms
2
on Au clusters is responsible for the rate-limiting H dissociation
2
+
−
to yield a H /H pair at metal/support interface. Theoretical and
experimental studies established that dissociative chemisorption
2
3
reactions than those on reducible semiconductor oxides, such as
TiO2 [1,3,7], and it is widely believed that this tendency is due
to the oxygen vacancies at metal support interface of Au/TiO2.
The low activity of Au NPs on inert oxides should be partly
due to a difficulty in controlling the size and oxidation state of
gold using the conventional deposition–precipitation method. For
of H does not occur on clean gold extended surfaces but can occur
2
on low coordinated Au atoms at corners or edges of Au NPs [10–13].
Based on these backgrounds, it is of fundamental importance to
expand the applicability of our hypothesis to other Au-catalyzed
hydrogenation reactions.
The production of cyclohexanone oxime is an important step
in the production of nylon-6. The methods presently employed in
the manufacture of cyclohexanone oxime cannot avoid the pro-
duction of ammonium sulfate and the use of hazardous chemicals
such as oleum, halides and nitrogen oxides. There have been many
attempts to change the process into environmentally benign ones
[14–18]. Considering recent advances in the selective nitration of
cyclohexane to nitrocyclohexane [19,20], one step hydrogenation
of nitrocyclohexane into cyclohexanone oxime can be one of the
alternative methods [16,17]. However, previous examples, salts of
Cu(II) and Ag(I) as homogeneous catalysts [17] and a supported
Pd catalyst [16], have drawbacks such as high pressures (35 bar)
and the formation of environmentally unfriendly inorganic wastes.
example, when Al O -supported gold catalysts are prepared by the
2
3
deposition–precipitation method, small Au NPs (2–3 nm) as major
species, larger Au particles (4.3–20 nm) [5,6] and slightly oxidized
gold species [5] coexist on the support, and such non-uniform struc-
tural features can lead to lower catalytic activity. Recently, one of
the authors [8] has reported that the colloid deposition method
allows dispersion of metallic gold NPs with a narrow size distri-
bution on various supports, including inert oxides such as SiO2.
∗ Corresponding author.
E-mail address: kshimizu@cat.hokudai.ac.jp (K.-i. Shimizu).
1
381-1169/$ – see front matter © 2011 Elsevier B.V. All rights reserved.
doi:10.1016/j.molcata.2011.05.018

K.-i. Shimizu et al. / Journal of Molecular Catalysis A: Chemical 345 (2011) 54–59
55
◦
Recently, Serna et al. [18] have developed a TiOx-decorated Pt NPs
catalyst with Na-doping, which is effective for the selective hydro-
genation of nitrocyclohexane to cyclohexanone oxime under mild
Pt/Al O -1.3 sample at 25 C using the pulse-adsorption of CO in a
flow of He.
2
3
hydrogenation conditions (4 bar of H ). Design of alternative cata-
2.2. In situ FTIR
2
lysts without the additives and platinum-group metals, such as gold
catalysts, is a fundamentally more challenging target. However, a
well-established Au/TiO2 catalyst was shown to be ineffective for
this reaction [18].
In situ FTIR spectra were recorded on a JASCO FT/IR-620
equipped with a quartz IR cell connected to a conventional flow
reaction system. The sample was pressed into a 20 mg of self-
supporting wafer and mounted into the quartz IR cell with CaF2
windows. Spectra were measured accumulating 5–20 scans at a
Herein, we report that small Au clusters on Al O , prepared by
2
3
the colloid deposition method, catalyze the selective hydrogena-
tion of nitrocyclohexane into cyclohexanone oxime. To establish a
design concept of Au-based selective hydrogenation catalysts, we
show structural studies that address the influence of the particle
size and support material on catalytic efficiencies.
−
1
resolution of 4 cm . A reference spectrum of the catalyst wafer in
He taken at measurement temperature was subtracted from each
spectrum. Prior to each experiment the catalyst disk was heated
3
−1
◦
in He flow (100 cm min ) at 300 C for 0.5 h, followed by cooling
to the desired temperature under He flow. For the introduction of
organic compounds to the IR disc, the liquid compound was injected
◦
under the He flow preheated at 150 C which was fed to the in situ
2
. Experimental
IR cell. Then, the IR disk was purged with He for 600 s.
2.1. Catalyst preparation
2.3. Typical procedures for the catalytic test
The TiO -supported Au catalyst synthesized by
deposition–precipitation procedure (Au = 1.5 wt%, average Au
a
Commercially available organic compounds were used with-
out further purification. As-received or as-prepared catalysts were
used in catalytic experiments without any pre-treatments. Cat-
alytic experiments were carried out in a 30 cm3 autoclave with
a glass tube inside equipped with magnetic stirring. For hydro-
genation of nitrocyclohexane, 0.5 mmol of nitrocyclohexane (from
Tokyo Chemical Industry Co., Ltd. with purities above 95%), EtOH
2
particle size = 3.6 ± 0.28 nm) was purchased from the World Gold
Council and was named Au/Ti
SBET) of 224 m g
Catapal B Alumina purchased from Sasol) at 600 C for 3 h.
Au NPs supported on various supports (Al O , SiO , MgO) were
-3.6. ␥-Al O with surface area
2 3
WGC
2
−1
was prepared by calcination of ␥-AlOOH
(
(
◦
2
3
2
3
prepared by the colloid deposition method [8]. The method we
used to prepare Au NPs is the well-known two phase reduction
method developed by Schiffrin and co-workers [21]. In a typical
(1.5 cm ) as solvent, and catalyst (Au = 0.25 mol%) were placed into
the autoclave. After being sealed, the reactors were flushed with H2
and then pressurized at 0.6 MPa, and then heated to the required
−
◦
preparation of Au NPs, AuCl4 ions were extracted from the water
temperature (typically 100 C). Conversion and yields of products
into the toluene phase by excess tetraoctylammonium bromide
were determined by GC (GC-14B, Shimadzu) with DB-1 capillary
column (Shimadzu) using n-dodecane as an internal standard. The
products were identified by gas chromatography/mass spectrome-
try (GCMS-QP5000, Shimazu) equipped with the same column and
in the same conditions as GC and also by comparison with commer-
cially pure products. For hydrogenation of nitrobenzene, 3.0 mmol
(
TOAB). After separating the toluene phase, the protecting agent
−1
dodecanethiol (DDT, Au/DDT = 1:1 mol mol ) was added to it at
4
◦
0 C under vigorous stirring. The obtained solution was then left
under stirring for 30 min. A following rapid injection of an aqueous
−
1
solution of NaBH (Wako, 95% purity, Au/NaBH = 1:10 mol mol ),
4
4
3
led to formation of a dark orange-brown solution, indicating the
formation of the gold sol. Au cations are reduced at the boundary of
of nitrobenzene in EtOH (1.5 cm ) with Au catalyst (Au = 0.1 mol%)
◦
were heated at 110 C for 2 h.
the water and toluene phases. This method is advantageous in that
+
unreacted NaBH and other water-soluble byproducts such as Na ,
3. Results and discussion
4
−
Cl and borates can be separated from Au NPs quite easily. Trans-
mission electron microscopy (TEM) analysis showed that the mean
particle diameter of Au NPs thus prepared was 2.3 ± 0.41 nm. The
support was then added to the colloidal gold solution under stirring
and kept in contact until total adsorption (1 wt% of gold on the sup-
port) occurred, indicated by decoloration of the solution. The solids
were collected by filtration followed by washing the solids with
toluene to remove all the soluble species. The resulting composites
3.1. Catalytic properties
Time-conversion profiles for the nitrocyclohexane hydrogena-
tion with Au/Al-2.5, Au/Ti-3.6, and Pt/Al-1.3 are compared in
Fig. 1. Recently, Serna et al. [18] showed that hydrogenation
of nitrocyclohexane by H2 over Pt/Al O3 and Pd/C, as conven-
tional hydrogenation catalysts, results in non-selective reduction
to cyclohexylamine rather than selective reduction to cyclohex-
2
◦
were dried at room temperature and calcined at 300 C for 4 h under
air for the combustion of thiols [8]. Al O -supported Au NPs cata-
anone oxime. We tested this reaction with a conventional Pt/Al O3
2
3
2
lysts are designated as Au/Al-x, where x is the mean size of Au NPs
nm) estimated in our previous study. For Al O , gold colloid with
catalyst (Pt/Al-1.3). As shown in Fig. 1C, the undesirable byprod-
uct (cyclohexylamine, 3) was selectively produced by Pt/Al-1.3.
After 3 h, where the conversion reached 100%, the selectivity of
cyclohexanone oxime (2) was only 1%. Cyclohexanone (4), ini-
tially appeared as an unstable product, disappeared after 1 h, and
then the selectivity of dicyclohexylamine (6) increased. Based
on kinetic studies, Serna et al. [18] proposed a general reaction
scheme for the metal-catalyzed hydrogenation of nitrocyclohex-
ane. Assuming that this model is applicable in our system, a
tentative reaction scheme is shown in Scheme 1, which will be
verified hereafter. The results in Fig. 1C can be explained by this
scheme, in which the reaction of cyclohexylamine (3) and cyclo-
(
2
3
larger mean diameters (5.9 ± 0.49 nm) prepared through annealing
◦
the mixture of as-prepared Au NPs and TOAB at 165 C was used,
and the prepared catalyst was named Au/Al-6.0. The sample named
Au/Al-30, composed of large gold particles (30 nm), was prepared
◦
by calcining the Au/Al-2.5 sample at 1000 C for 3 h. Preparation
methods and structural characteristics of the catalysts are listed in
Table 1.
Al O -supported Pt catalyst (Pt/Al-1.3, Pt = 1 wt%) was pre-
2
3
pared by impregnating ␥-Al O3 with an aqueous HNO3 solution
2
◦
of Pt(NH ) (NO ) , followed by drying at 80 C for 12 h, calcining
at 500 C for 2 h, and reducing at 300 C for 0.5 h in H . The average
3
2
3 2
◦
◦
hexanone (4) under H yields dicyclohexylamine (6). As shown in
2
2
particle size of Pt (1.3 nm) was estimated with the CO uptake of the
Fig. 1B, a well-established Au nanoparticle catalyst, TiO -supported
2

5
6
K.-i. Shimizu et al. / Journal of Molecular Catalysis A: Chemical 345 (2011) 54–59
Table 1
List of the catalysts.
Catalysts-xa
Supports
DNP (nm)b
M (wt%)
Preparation method
Tcal ( C)c
◦
Au/Al-2.5
Au/Al-6.0
Au/Al-30
Au/Si-1.9
Au/Mg-3.0
Au/Ti-3.6d
Pt/Al-1.3
␥-Al2O3
␥-Al2O3
␥-Al2O3
SiO2
MgO
TiO2 (anatase)
␥-Al2O3
2.3
5.9
2.3
2.3
2.3
–
1
1
1
2.5
1
Colloid deposition
Colloid deposition
Colloid deposition
Colloid deposition
Colloid deposition
Deposition–precipitation
Impregnation
300
300
1000
300
300
–
1.5
1
–
500
a
Average particle size (nm) of metallic Au species estimated from the EXAFS except for Au/Al-6.0 (TEM), Au/Al-30 (XRD), Au/TiWGC-3.6 (from WGC) and Pt/Al-1.3 (CO
adsorption).
b
Mean diameter of gold NPs in gold colloid used for catalyst preparation.
Calcination temperature.
Standard catalyst supplied from WGC.
c
d
Au (Au/Ti-3.6) was not suitable catalyst. After 12 h, where the con-
version reached 100%, the selectivity of cyclohexanone oxime (2)
was only 9%. Initially, cyclohexylamine (3), cyclohexanone (4), and
cyclohexyl-cyclohexylidene amine (5) were main products. How-
ever, with an increase in the reaction time, the selectivities to
these products and mass balance decreased. This indicates that
these compound are unstable in the reaction condition and will be
converted to GC-undetectable side products. In contrast, cyclohex-
anone oxime was selectively produced by the Au/Al-2.5 catalyst.
After 12 h, the yield of cyclohexanone oxime reached 83% and small
amount (6.8% yield) of cyclohexanone (4) was observed as byprod-
uct.
(5), and dicyclohexylamine (6) appeared as by-products. The selec-
tivity pattern is consistent with Scheme 1, where cyclohexylamine
(3) and cyclohexanone (4), produced by hydrogenation and hydrol-
ysis of cyclohexanone oxime (2), undergo condensation to yield
cyclohexyl-cyclohexylidene amine (5), which will be hydrogenated
to give dicyclohexylamine (6). Fig. 2B shows the conversion and
◦
selectivities at 100 C after 12 h as a function of the initial H pres-
2
sure. No reaction occurred under the normal pressure. Increase
in the H2 pressure from 0.1 MPa to 0.6 MPa resulted in a drastic
increase in the conversion (from 0% to 100%), which suggests that
H activation is the rate-limiting step of the present reaction. Above
2
0.6 MPa, the selectivity to cyclohexanone oxime (2) decreased, and
the selectivity to cyclohexanone (4) was maximum at 1.1 MPa. At
higher pressure, the selectivity to cyclohexanone (4) decreased,
and cyclohexyl-cyclohexylidene amine (5) and dicyclohexylamine
(6) appeared. This selectivity pattern is again consistent with
Scheme 1. From these results in Figs. 1 and 2, we could conclude
that Scheme 1 is a general reaction scheme for the hydrogenation
of nitrocyclohexane. As for the catalytic performance, Au/Al-2.5
Fig. 2 shows the effect of reaction conditions on the cat-
alytic properties of Au/Al-2.5 for nitrocyclohexane hydrogenation.
Fig. 2A shows the conversion and selectivities at the reaction
time of 12 h as a function of the reaction temperature. Up to
◦
1
00 C, where the conversion reached 100%, the selectivities to
cyclohexanone oxime (2) and cyclohexanone (4) increased with
temperature, and they decreased at higher temperature. Simulta-
neously, cyclohexylamine (3), cyclohexyl-cyclohexylidene amine
◦
shows, under the optimized reaction condition (T = 100 C, t = 12 h,
A 100
B 100
80
C100
80
8
6
4
2
0
0
0
0
0
60
60
40
40
20
20
0
0
0
5
10
15
0
5
10
15
0
1
2
3
Time / h
Time / h
Time / h
Fig. 1. Conversion of nitrocyclohexane (1, ꢀ) and yields of cyclohexanone oxime (2, ꢀ), cyclohexylamine (3, ꢁ), cyclohexanone (4, ꢁ), cyclohexyl-cyclohexylidene amine (5,
◦
×
), and dicyclohexylamine (6, ᭹) vs. time for the hydrogenation of nitrocyclohexane at 100 C with (A) Au/Al-2.5, (B) Au/Ti-3.6, and (C) Pt/Al-1.3.
NH2
NO2
H2
NOH
H2
3
H
N
H2
N
H2
H O
O
1
2
5
6
2
4
Scheme 1. Possible reaction scheme for the hydrogenation of nitrocyclohexane with H2.

K.-i. Shimizu et al. / Journal of Molecular Catalysis A: Chemical 345 (2011) 54–59
57
A 100
B100
100
Au/Al-2.5
1
8
0
0.8
0.6
0.4
0.2
0
8
6
4
2
0
0
0
0
80
60
40
20
0
60
40
20
0
Au/Ti-3.5
2.6
Au/Si-1.9
2.8
Au/Mg-3.0
.2 2.4
0
8
0
90
100 110 120
0
1
2
2
3
o
T/ C
P(H2)/ MPa
Electronegativity of support metal oxide
Fig. 2. Effects of (A) temperature and (B) hydrogen pressure on the conversion of
nitrocyclohexane (1, ꢀ) and yields of cyclohexanone oxime (2, ꢀ), cyclohexylamine
Fig. 4. (ꢀ) Conversion of nitrocyclohexane and (ꢀ) cyclohexanone oxime selectivity
◦
(
3, ꢁ), cyclohexanone (4, ꢁ), cyclohexyl-cyclohexylidene amine (5, ×), and dicy-
for nitrocyclohexane hydrogenation at 100 C and (᭹) initial rates of OH/D2 isotope
clohexylamine (6, ᭹) for the hydrogenation of nitrocyclohexane with Au/Al-2.5.
exchange as a function of electronegativity of support oxide.
◦
Conditions: (A) t = 6 h, P(H2) = 0.6 MPa; (B) T = 100 C, t = 6 h.
as a parameter of acidity (or electrophilicity)of metal oxides [9]. The
result shows that the support with strong basic character (MgO)
P(H ) = 0.6 MPa), the cyclohexanone oxime yield of 83% with 6.8%
2
yield of cyclohexanone (4) as byproduct. The cyclohexanone oxime
yield is comparable to that of Na-doped Pt/TiO2 catalyst [18]. It
should be noted that Au/Al-2.5, as an additive-free non-platinum
metal catalyst, achieved selective hydrogenation of nitrocyclohex-
ane to cyclohexanone oxime.
and that with acidic character (SiO ) resulted in lower activity and
2
selectivity than an amphoteric oxide, Al O . This result suggests
2
3
that both acidic and basic sites of the support material are necessary
for this reaction.
3
.3. Roles of Au cluster and Al O
2 3
3.2. Size- and support-dependent activity of Au NPs
Previously, Claus et al. reported the isotopic exchange of OH
For a series of Au/Al catalysts with the same Au loading (1 wt%)
groups of oxide supports to OD groups in D2 for Ag/SiO2 cata-
lyst [22]. They showed that the reaction begins with the cleavage
of D2. In our previous study [9], we carried out in situ IR exper-
iments for the isotopic exchange of M−OH to M−OD under D2
on metal oxides (MOx)-supported Au catalysts, and the rate of
OH/D2 exchange was used to compare the relative rate of H2 dis-
sociation on various Au catalysts. As shown in Fig. 5, the initial
rates of OD formation (ꢀAOD/ꢀt) on Au/Al–2.5, estimated from
the slope of the curve, was significantly higher than that for ␥-
Al2O3, which indicates that the Au cluster plays a significant role
in the cleavage of a H-H bond. To obtain mechanistic reasons
of the size- and support-dependent activity of gold catalyst, the
OD formation rate reported in our previous study [9] is plotted
in Figs. 3 and 4. For Au/Al catalysts with the same Au loading
but with different mean Au particle sizes, the rate of OD forma-
tion decreased with increase in the Au particle size. This trend is
but with different gold particle size, the conversion and the selectiv-
ity to cyclohexanone oxime are plotted as a function of the average
particle size in Fig. 3. Clearly, the Au clusters with the smallest size
(
2.5 nm) gave highest conversion. The increase of Au particle size
from 2.5 nm to 6.0 nm results in a drastic decreases in the con-
version (from 100% to 4.5%) and the selectivity to cyclohexanone
oxime (from 83% to 0%). It is clearly shown that the present reaction
requires small Au clusters.
The activity of the gold clusters having similar mean size
1.9–3.6 nm) depends strongly on the acid–base characteristics of
support material. In Fig. 4, conversion of nitrocyclohexane and yield
of cyclohexanone oxime are plotted as a function of the electroneg-
ativity of metal oxide which has been used, to a first approximation,
(
100
0.8
8
6
4
2
0
0
0
0
0
0
0
0
.6
.4
.2
1
00
50
0
1
5
10
50
0
100
200
300
Time / s
400
500
600
Mean Au particle size/ nm
◦
Fig. 3. (ꢀ) Conversion of nitrocyclohexane and (ꢀ) cyclohexanone oxime selectivity
Fig. 5. Area of OD band as a function of the time of D2 exposure at 150 C to
(ꢀ) untreated Au/Al-2.5, (ꢁ) pyridine-adsorbed Au/Al-2.5, (ꢁ) CH3COOH-adsorbed
Au/Al-2.5, and (᭹) ␥-Al2O3.
◦
for nitrocyclohexane hydrogenation at 100 C and (᭹) initial rates of OH/D2 isotope
exchange as a function of average particle size of Au in Au/Al.

5
8
K.-i. Shimizu et al. / Journal of Molecular Catalysis A: Chemical 345 (2011) 54–59
Table 2
Effect of additives (20 mol%) on Au/Al-2.5 catalyzed hydrogenation of nitrocyclohexane and nitrobenzene.b
a
ꢀAOD/ꢀtc (cm−1
s
−1
)
[PhNO2ad]d (cm−1
)
1 Conv. (%)
2 Sel. (%)
Yield (%)
Yield (%)
Additive
2
3
4
5
6
–
0.73
0.14
0.12
18.1
6.0
2.6
100
77
26
83
54
65
83
42
17
0
2.4
0
6.8
5.2
5.0
0
0
0
0
2.4
0
94
31
53
Pyridine
CH3COOH
a
◦
◦
Nitrocyclohexane (0.5 mmol), EtOH (1.5 mL), Au/Al-2.5 (0.25 mol%), t = 6 h, P(H2) = 0.6 MPa, T = 100 C.
b
c
Nitrobenzene (3 mmol) in EtOH (1.5 mL) with Au/Al-2.5 (0.1 mol%): t = 2 h, P(H2) = 2.0 MPa, T = 110 C. Yield of aniline was determined by GC.
The initial rates of OD formation under D2 flow (from Fig. 5).
d
◦
−1
The relative amount of adsorbed nitrobenzene at 100 C estimated from integrated intensity of the ꢁs(NO2) band (1350 cm ) of the adsorbed nitrobenzene in Fig. 6.
consistent with the changes in the activity and selectivity for
the selective hydrogenation (Fig. 3). Taking into account the fact
that dissociative chemisorption of H2 on the extended surface of
polycrystalline or single crystal gold surface is thermodynamically
unfavorable, coordinatively unsaturated sites on small Au clusters
are responsible for the H2 dissociation as a possible rate-limiting
step for the title reaction. As for the support effect, the OH/D2
0
.1
exchange reaction rate for acid–base bifunctional support (Al O )
2
3
is higher than those for basic (MgO) and acidic (SiO ) supports, and
2
this trend is consistent with the changes in the activity and selec-
tivity for the selective hydrogenation (Fig. 4). This indicates that the
surface acid–base pair sites on the support are also required for the
a
b
c
H dissociation step.
2
To further investigate the role of the Al O3 surface, the effects
2
of basic or acidic additives to the reaction mixture were exam-
ined. Table 2 shows the catalytic results for the Au/Al–2.5 catalyzed
hydrogenation of nitrocyclohexane in the absence or presence of
1600
1400
Wavenumber / cm-1
2
0 mol% of pyridine or acetic acid. The addition of these molecules
◦
decreased the conversion and the selectivity to cyclohexanone
oxime. The yield of cyclohexanone oxime in the presence of pyri-
dine and acetic acid was ca. 1/2 and 1/5 of the value in the
additive-free condition. The additive effects were also studied for
the hydrogenation of nitrobenzene as a test reaction. The addition
of pyridine and acetic acid decreased the aniline yield. As shown in
Fig. 6. IR spectra of the adsorbed nitrobenzene at 100 C on (a) untreated Au/Al-2.5,
(
b) pyridine-adsorbed Au/Al-2.5, and (c) CH3COOH-adsorbed Au/Al-2.5.
nucleophilic H␦ species on Au atom transfer to the polar nitro
−
group.
For the selective hydrogenation of nitrocyclohexane on the
Na-doped Pt/TiO2 catalyst, Serna et al. [18] proposed that the
preferential adsorption of the nitro group on Pt–TiO2 interface is
the origin of the selective production of cyclohexanone oxime.
Fig. 5, the rate of OH/D exchange reaction after the pre-adsorption
2
of pyridine (0.6 mmol/g) was ca. 1/5 of the value in the additive-
free condition. The pre-adsorption of acetic acid (0.6 mmol/g) also
resulted in a decrease in the rate of OH/D exchange reaction by a
To study a possible role of acid–base pair sites on Al O3 in pref-
2
2
factor of 1/6. These results suggest that Lewis acid–base pair sites
erential adsorption of the nitro group, in situ FTIR experiments
of nitrobenzene adsorption on Au/Al–2.5 was carried out. Note
that the IR spectrum of adsorbed nitrocyclohexane was compli-
cated, and hence nitrobenzene was used as probe molecule. When
also play an important role in H dissociation and subsequent for-
2
mation of acidic Al–OH group. Most probable model to account
for the above results is that the dissociation of H2 occurs at gold
cluster–support interface [23], where coordinatively unsaturated
◦
nitrobenzene was adsorbed on Au/Al–2.5 at 100 C (Fig. 6), IR bands
␦
+
␦−
Au atoms and the acid–base pair site (Al –O site) are located
close to each other. Recently, Garcia and co-workers showed direct
IR evidences on heterolytic dissociation of H2 on ceria-supported
due to the ꢁas(NO ) and ꢁs(NO ) of the adsorbed nitrobenzene
2
2
−
1
were observed at 1525 and 1350 cm , respectively. In our previ-
ous study, the ꢁas(NO ) band with lower frequency than gas phase
2
◦
−1
Au NPs at 150 C, leading to a proton bonded to a ceria oxygen and
PhNO (1552 cm ) has been assigned to nitrobenzene ad-species
2
a hydride bonded to gold [24]. For the chemoselective hydrogena-
tion of polar bonds (C O or C N) with homogeneous metal–ligand
bifunctional catalysts, it is widely accepted that the reaction begins
interacting with the Al O surface through the nitro group [9]. The
2
3
pre-adsorption of pyridine and acetic acid decreased the inten-
sity of these bands; the relative amounts of adsorbed nitrobenzene
on Au/Al–2.5, estimated from the integrated area of the band at
+
with heterolytic cleavage of H2 to yield H in an OH or NH lig-
−
+
−
−1
and and H in metal hydrides, and the H /H pair preferentially
transfers to the polar bonds [25,26]. On the other hand, it is estab-
1350 cm , were 1/3 and 1/7 of the value in the additive-free con-
dition. Taking into account the fact that the adsorption of acetic acid
lished that H addition to metal particles on solid acids, such as
on Al O3 results in the formation of the acetate ion the acid–base
2
2
Pt/SO42 –ZrO , Pt/Al O , and Rh/Al O , results in the formation
−
pair site (Al –O site) [28], the result indicates the preferential
␦+
␦−
2
2
3
2
3
␦+
␦−
of acidic OH groups on the support via hydrogen spillover [27].
Adopting these models, a cooperative mechanism catalyzed by low
adsorption of the nitro group on Al –O site. This can be addi-
tional reason why alumina is the most effective support material
for the gold catalyzed selective hydrogenation of nitro-compounds.
coordinated Au atoms and acid–base pair site of Al O3 for the
2
hydrogenation of a nitro group may be proposed as follows. The
␦−
dissociation of H at gold–support interface yields an H atom on
4. Conclusion
2
the low coordinated Au atom and an H atom which releases an elec-
tron to Lewis acid site of the support. The latter species becomes
a proton stabilized at the oxygen atom (Lewis base) nearby the
We demonstrated that highly selective reduction of nitrocy-
clohexane into cyclohexanone oxime is effectively catalyzed by
␥-Al O -supported Au cluster catalyst, prepared by colloid depo-
␦
+
Lewis acid site. The electrophilic H species on the support and
2
3

K.-i. Shimizu et al. / Journal of Molecular Catalysis A: Chemical 345 (2011) 54–59
59
sition method. The activity depended strongly on the size and
support oxides, and small Au clusters (2.5 nm) on the acid–base
[6] T. Ishida, N. Kinoshita, H. Okatsu, T. Akita, T. Takei, M. Haruta, Angew. Chem.
Int. Ed. 47 (2008) 9265–9268.
[
7] A. Corma, P. Serna, H. Garcia, J. Am. Chem. Soc. 129 (2007) 6358–6359.
bifunctional support (Al O ) gave highest activity and selectiv-
2
3
[8] Y. Tai, M. Watanabe, K. Kaneko, S. Tanemura, T. Miki, J. Murakami, K. Tajiri, Adv.
Mater. 13 (2001) 1611–1614.
ity. We propose a possible mechanism that account for these
observations; at gold cluster–support interface, cooperation of
coordinatively unsaturated Au atoms and the acid–base pair site of
[
9] K. Shimizu, Y. Miyamoto, T. Kawasaki, T. Tanji, Y. Tai, A. Satsuma, J. Phys. Chem.
C 113 (2009) 17803–17810.
[
10] M. Okumura, Y. Kitagawa, M. Haruta, K. Yamaguchi, Appl. Catal. A 291 (2005)
37–44.
[11] A. Corma, M. Boronat, S. Gonzalez, F. Illas, Chem. Commun. (2007) 3371–
3373.
12] E. Bus, J.T. Miller, J.A. van Bokhoven, J. Phys. Chem. B 109 (2005) 14581–14587.
[13] J.A. van Bokhoven, J.T. Miller, J. Phys. Chem. C 111 (2007) 9245–9249.
␦
+
␦−
alumina (Al –O site) plays an important role in the dissociation
of H2 as a possible rate-limiting step. The preferential adsorption
␦
+
␦−
of the nitro group on Al –O site can be additional reason of high
activity and selectivity.
[
[
[
[
[
14] T. Tatsumi, N. Jappar, J. Catal. 161 (1996) 570–576.
15] J.A. Robertson, J. Org. Chem. 13 (1948) 395–398.
16] GB Patent 857902 (1958), to DuPont.
Acknowledgments
17] J.F. Knifton, J. Catal. 33 (1974) 289–298.
This work was supported by the Japanese Ministry of Education,
Culture, Sports, Science and Technology (Grant-in-Aids for Scien-
tific Research B 20360361 and for Young Scientists A 22686075) and
JST (Adaptable and Seamless Technology Transfer Program through
Target-driven R&D).
[18] P. Serna, M. Lopez-Haro, J.J. Calvino, A. Corma, J. Catal. 263 (2009) 328–334.
[
19] S. Sakaguchi, Y. Nishiwaki, T. Kitamura, Y. Ishii, Angew. Chem. Int. Ed. 40 (2001)
22–224.
20] K. Yamaguchi, S. Shinachi, N. Mizuno, Chem. Commun. (2004) 424–425.
2
[
[21] M. Brust, M. Walker, D. Bethell, D.J. Schiffrin, R. Whyman, J. Chem. Soc. Chem.
Commun. (1994) 801–802.
[
22] M. Bron, D. Teschner, A. Knop-Gericke, F.C. Jentoft, J. Krohnert, J. Hohmeyer, C.
Volckmar, B. Steinhauer, R. Schlögl, P. Claus, Phys. Chem. Chem. Phys. 9 (2007)
3559–3569.
References
[
23] T. Fujitani, I. Nakamura, T. Akita, M. Okumura, M. Haruta, Angew. Chem. Int. Ed.
4
8 (2009) 9515–9518.
24] R. Juárez, S.F. Parker, P. Concepción, A. Corma, H. García, Chem. Sci. 1 (2010)
31–738.
[
[
[
[
[
[
[
1] M. Haruta, Catal. Today 36 (1997) 153–166.
[
2] A. Corma, H. Garcia, Chem. Soc. Rev. 37 (2008) 2096–2126.
3] M.M. Schubert, S. Hackenberg, A.C. van Veen, M. Muhler, V. Plzak, R.J. Behm, J.
Catal. 197 (2001) 113–122.
7
25] R. Noyori, T. Ohkuma, Angew. Chem. Int. Ed. (2001) 40–73, 40-73.
26] R.M. Bullock, Chem. Eur. J. 10 (2004) 2366–2374.
27] T. Shishido, H. Hattori, Appl. Catal. A 146 (1996) 157–164.
28] K. Shimizu, H. Kawabata, A. Satsuma, T. Hattori, J. Phys. Chem. B 103 (1999)
[4] P. Claus, A. Brückner, C. Mohr, H. Hofmeister, J. Am. Chem. Soc. 122 (2000)
1
1430–11439.
[5] C.K. Costello, J. Guzman, J.H. Yang, Y.M. Wang, M.C. Kung, B.C. Gates, H.H. Kung,
J. Phys. Chem. B 108 (2004) 12529–12536.
5240–5245.