Acceptor-free dehydrogenation of secondary alcohols by heterogeneous cooperative catalysis between Ni nanoparticles and acid-base sites of alumina supports

Article abstract of DOI:10.1016/j.jcat.2013.01.005

Nickel-nanoparticle-loaded θ-Al₂O₃ (Ni/θ-Al₂O₃), prepared by H₂-reduction of NiO/θ-Al₂O₃, acts as an effective and reusable heterogeneous catalyst for acceptor-free dehydrogenation of alcohols in a liquid phase. Among various supports, amphoteric supports (such as θ-Al ₂O₃), having both acidic and basic sites, gave higher activity than acidic or basic supports. Among Ni/θ-Al₂O ₃ catalysts with different Ni particle sizes, turnover frequency (TOF) per surface Ni increases with decreasing Ni particle size. These results suggest that low-coordinated Ni⁰ sites and metal/support interfaces play important roles in the catalytic cycle. The reaction mechanism is investigated by in situ IR study combined with kinetic analysis (kinetic isotope effect), and the following mechanism is proposed: (1) reaction of an alcohol with Lewis acid (Al^δ+)-base (AlO^δ-) pair site of alumina yields an alkoxide on the Al^δ+ site and a proton on the AlO^δ- site, (2) CH dissociation of the alkoxide by Ni ⁰ site to form NiH and a ketone, and (3) protolysis of NiH by a neighboring proton to release H₂ gas. The proposed mechanism provides fundamental reasons for the higher activity of Ni on the acid-base bifunctional support (Al₂O₃) than on basic and acidic ones.

Full text of DOI:10.1016/j.jcat.2013.01.005

Journal of Catalysis 300 (2013) 242–250
Contents lists available at SciVerse ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
Acceptor-free dehydrogenation of secondary alcohols by heterogeneous
cooperative catalysis between Ni nanoparticles and acid–base sites
of alumina supports
a
a
b
Ken-ichi Shimizu a,b, , Kenichi Kon , Katsuya Shimura , Siddiki S.M.A. Hakim
⇑
a Catalysis Research Center, Hokkaido University, N-21, W-10, Sapporo 001-0021, Japan
b Elements Strategy Initiative for Catalysts and Batteries, Kyoto University, Katsura, Kyoto 615-8520, Japan
a r t i c l e i n f o
a b s t r a c t
Article history:
Nickel-nanoparticle-loaded h-Al2O3 (Ni/h-Al2O3), prepared by H2-reduction of NiO/h-Al2O3, acts as an
effective and reusable heterogeneous catalyst for acceptor-free dehydrogenation of alcohols in a liquid
phase. Among various supports, amphoteric supports (such as h-Al2O3), having both acidic and basic
sites, gave higher activity than acidic or basic supports. Among Ni/h-Al2O3 catalysts with different Ni
particle sizes, turnover frequency (TOF) per surface Ni increases with decreasing Ni particle size. These
results suggest that low-coordinated Ni0 sites and metal/support interfaces play important roles in the
catalytic cycle. The reaction mechanism is investigated by in situ IR study combined with kinetic anal-
ysis (kinetic isotope effect), and the following mechanism is proposed: (1) reaction of an alcohol with
Lewis acid (Ald+)–base (AlAOdꢀ) pair site of alumina yields an alkoxide on the Ald+ site and a proton on
the AlAOdꢀ site, (2) CAH dissociation of the alkoxide by Ni0 site to form NiAH and a ketone, and (3)
protolysis of NiAH by a neighboring proton to release H2 gas. The proposed mechanism provides fun-
damental reasons for the higher activity of Ni on the acid–base bifunctional support (Al2O3) than on
basic and acidic ones.
Received 30 November 2012
Revised 8 January 2013
Accepted 9 January 2013
Keywords:
Alcohols
Alumina
Cooperative catalysis
Dehydrogenation
Nickel
Ó 2013 Elsevier Inc. All rights reserved.
1. Introduction
mental and economical acceptability, much effort has been de-
voted to developing heterogeneous catalysts based on noble
Replacement of noble metal catalysts with cheap and abundant
metals is a major goal of sustainable chemistry. Nickel (Ni), as a
cheap element, is an ideal candidate if high activity and selectivity
can be achieved. Despite the fact that Ni-containing solids, such as
sponge Ni (Raney Ni), have been used as promoters or catalysts for
hydrogenation of various compounds, advances in the field of het-
erogeneous non-noble metal-catalyzed organic synthesis are still
in their infancy. Recently, Yus et al. [1] demonstrated that stoichi-
ometric or substoichiometric amounts (20–100 mol%) of Ni nano-
particles (NPs) promote various types of hydrogen-transfer
reactions, which are typically implemented by noble metal-based
catalysts. Our ongoing interest in supported metal NP catalysis
for dehydrogenation/hydrogenation reactions [2,3], combined with
the pioneering studies [1], motivated us to investigate possibilities
of supported Ni NPs as cheaper alternatives to the well-established
noble metal catalysts in hydrogen-transfer-type reactions [4].
Oxidation of alcohols to carbonyl compounds is one of the most
important topics in synthetic catalysis [5–8]. To achieve environ-
metals (Pd [9], Ru [10,11], Pt [12], and Au [13,14]) for aerobic oxi-
dation of alcohols. Non-noble metal-based heterogeneous catalysts
also catalyze the aerobic oxidation of alcohols [5–7,15–20], but
their efficiencies in terms of activity and substrate scope are gener-
ally lower than for the noble metal catalysts. For example, Ni-
based catalysts [15–20] reported so far are effective only for acti-
vated (aromatic) alcohols, but not for aliphatic alcohols, and their
turnover numbers (TONs) are low (TON = 0.2–1.5). From the view-
point of atom efficiency and safety of the reaction, oxidant-free
catalytic dehydrogenation of alcohols to carbonyl compounds
and molecular hydrogen is ideal. This process is also of current
interest as a potential method for hydrogen production from bio-
mass products [21,22]. Several systems for the dehydrogenative
oxidation of alcohols using homogeneous Rh [23], Ru [24–29],
and Ir [30–32] catalysts are reported, but most of them suffer from
the requirement for acid or base additives, difficult catalyst synthe-
sis and manipulation, difficulties in catalyst reuse, and high price.
To overcome these problems, heterogeneous Ru [33–35], Ag
[2,36], and Au [37–39] catalysts have recently been developed,
but these systems require the use of noble metals. There are a
few examples based on non-noble metals (Ni [40–42], Co [43],
and Cu [44,45]), but they suffer from drawbacks such as high cat-
⇑
Corresponding author at: Catalysis Research Center, Hokkaido University, N-21,
W-10, Sapporo 001-0021, Japan. Fax: +81 11 706 9163.
E-mail address: kshimizu@cat.hokudai.ac.jp (K.-i. Shimizu).
0021-9517/$ - see front matter Ó 2013 Elsevier Inc. All rights reserved.
http://dx.doi.org/10.1016/j.jcat.2013.01.005

K.-i. Shimizu et al. / Journal of Catalysis 300 (2013) 242–250
243
alyst loading, low yields of carbonyl compounds [40–42], limited
scope [40–42,45], or high reaction temperature (200–350 °C)
[42,43].
(M = Ni, Co, Cu, Ru, Pd, Rh, Ag), with metal loading of 5 wt.% were
prepared by an impregnation method in a manner similar to that
for Ni/h-Al2O3 using an aqueous solution of metal nitrates (for Ni,
Co, Cu), RuCl3, or an aqueous HNO3 solution of Rh(NO3)3, Pd(NO3)2,
or Pt(NH3)2(NO3)2. Au/c-Al2O3 (Au = 1 wt.%) was prepared by the
deposition–precipitation method with HAuCl4. Raney Ni (B113W,
As for a catalyst design concept, it is notable that excellent orga-
nometallic catalysts for acceptor-free oxidation of alcohols are
rationally designed by a concept of cooperative catalysis, where
the metal center acts as a CAH activation site while a basic site
in a ligand abstracts proton from a OH group [28–32,34]. For exam-
ple, the Ru catalyst developed by Milstein possesses a cooperative
basic site in the b position to the metal, which facilitates binding of
the alcohol substrate to give an alkoxide intermediate [28]. Know-
ing that amphoteric metal oxides, such as alumina, have acid–base
sites on the surface, we hypothesized that the alumina surface
might show a similar cooperative effect. We report here that Ni-
loaded alumina acts as an effective heterogeneous catalyst for
the acceptor-free alcohol dehydrogenation of aliphatic secondary
alcohols to ketones. Studies on structure–activity relationships will
clarify that Ni0 species and acid–base sites on the support are
indispensable for this catalytic system. Spectroscopic and kinetic
studies will suggest a cooperative reaction mechanism of Ni0 spe-
cies and the acid–base sites.
Ni > 90%) was supplied by Evonik Industries.
2.3. Characterization
Transmission electron microscopy (TEM) measurements were
carried out using a JEOL JEM-2100F TEM operated at 200 kV. XRD
patterns of the powdered catalysts were recorded with a Rigaku
MiniFlex II/AP diffractometer with Cu Ka radiation. Temperature-
programmed reduction under H2 (H2 TPR) was carried out with
BELCAT (BELL Japan Inc.). The sample, NiO/h-Al2O3 (Ni load-
ing = 5 wt.%, 20 mg), mounted in a quartz tube, was heated with
a temperature ramp-rate of 5 °C minꢀ1 in a flow of 5% H2/Ar
(20 cm3 minꢀ1). The effluent gas was passed through a trap con-
taining MS4A to remove water and then through the thermal con-
ductivity detector. The amount of H2 consumed during the
experiment was detected by a thermal conductivity detector. Tem-
perature-programmed desorption (TPD) experiments were carried
out using BELCAT. Prior to each experiment, catalyst (0.1 g) was
heated in a flow of He (20 cm3 minꢀ1) at 500 °C for 0.5 h, followed
by cooling to 40 °C under He flow. Then, the catalyst was exposed
to a flow of 100% CO2 or 5% NH3/He for 10 min. After being purged
in He for 0.5 h, the catalyst was heated linearly at 10 °C minꢀ1 to
900 °C in a flow of He, and CO2 (m/e = 44) and NH3 (m/e = 16 and
17) in the outlet gas were analyzed by the mass spectrometer (BEL-
MASS). The BET surface area was obtained by measuring N2
adsorption at ꢀ196 °C by BELCAT. The results are summarized in
Table 1.
2. Experimental
2.1. General
Commercially available organic compounds (from Tokyo Chem-
ical Industry or Kishida Chemical) were used without further puri-
fication. 2-Propanol-O-d, (CH3)2CHOD with 98 atom% D, and 2-
propanol-2-d1, (CH3)2CDOH with 99.4 atom% D, were purchased
from Aldrich. The GC (Shimadzu GC-14B) and GCMS (Shimadzu
GCMS-QP2010) analyses were carried out using N2 and He as the
carrier gases, respectively, with an Ultra ALLOY capillary column
UA+-5 (Frontier Laboratories Ltd.).
2.4. Typical procedures of catalytic reactions
2.2. Catalyst preparation
Typically, Ni/h-Al2O3 (Ni loading = 5 wt.%) prereduced in H2 at
500 °C for 0.5 h was used as a standard catalyst. After the prereduc-
tion, we carried out catalytic tests using a batch-type reactor with-
out exposing the catalyst to air. A mixture of o-xylene (1 g), alcohol
(1 mmol), and n-dodecane (0.5 mmol) was injected into the prere-
duced catalyst inside a reactor (cylindrical glass tube) through a
septum inlet, followed by filling with N2. Then, the resulting mix-
ture was stirred under reflux conditions; the heating temperature
was 155 °C, and the reaction temperature was ca. 144 °C.
c
and
-Al2O3 (with surface area 124 m2 gꢀ1), h-Al2O3 (112 m2 gꢀ1),
-Al2O3 (17 m2 gꢀ1) were prepared by calcination of
-AlOOH
a
c
(Catapal B Alumina purchased from Sasol) for 3 h at 900, 1000, and
1200 °C, respectively. MgO (JRC-MGO-3, 19 m2 gꢀ1), TiO2 (JRC-
TIO4, 50 m2 gꢀ1), and CeO2 (JRC-CEO3, 81.4 m2 gꢀ1) were supplied
from the Catalysis Society of Japan. Hydroxides of Zr, Sn, Y, and La
were prepared by hydrolysis of zirconium oxynitrate 2-hydrate,
SnCl4ꢁ6H2O, or metal nitrates in distilled water by gradually adding
an aqueous NH4OH solution (1.0 mol dmꢀ3), followed by filtration
of precipitate, washing with distilled water three times, and drying
at 100 °C for 12 h. Ca(OH)2 and Nb2O5ꢁnH2O (supplied by CBMM)
were commercially supplied. CaO, Y2O3, La2O3, ZrO2, Nb2O5, and
SnO2 were prepared by calcination of these hydroxides at 500 °C
for 3 h.
Table 1
List of metal oxide-supported Ni (5wt%) catalysts.
a
Catalysts
SBET
Acid
Base
Acid
Base
(m2 gꢀ1
)
amount
amount
density
density
b
c
d
d
(mmol gꢀ1
)
(mmol gꢀ1
)
(
l
mol mꢀ2
)
(l )
mol mꢀ2
NiO/h-Al2O3 (Ni loading = 5 wt.%) was prepared by an impreg-
nation method; a mixture of h-Al2O3 (18 g) and an aqueous solu-
tion of Ni(NO3)2ꢁ6H2O (0.162 M, 100 cm3) were evaporated at
50 °C, followed by drying at 90 °C for 12 h, and calcination in air
at 300 °C for 1 h. Before each catalytic experiment, Ni/h-Al2O3
was prepared by in situ prereduction in NiO/h-Al2O3 in a glass (Pyr-
ex or quartz) tube under a flow of H2 (20 cm3 minꢀ1) at a reduction
temperature (TH2) of 500 °C for 0.5 h. Ni/h-Al2O3 with different Ni
loadings (5, 10, 20, 50 wt.%) and 5 wt.% Ni loaded on various sup-
Ni/CaO
8.0
15
15
12
69
70
74
45
40
25
<0.001
0.008
0.029
0.023
0.207
0.211
0.287
0.083
0.150
0.029
0.008
0.287
0.176
0.180
0.182
0.171
0.484
0.348
0.234
0.022
0.021
0.015
0.002
0.234
0
22
12
12
14
7.0
5.0
3.2
0.49
0.53
0.60
0.48
1.9
Ni/MgO
Ni/La2O3
Ni/Y2O3
Ni/CeO2
Ni/h-Al2O3
Ni/ZrO2
Ni/TiO2
0.53
1.9
1.9
3.0
3.0
3.9
1.8
3.7
1.2
1.9
2.4
Ni/Nb2O5
Ni/SnO2
Ni/a-Al2O3 4.2
ports (a-Al2O3, La2O3, Y2O3, CeO2, ZrO2, TiO2, Nb2O5, SnO2) were
Ni/c
-Al2O3 122
prepared in the same manner as Ni/h-Al2O3. CaO-, and MgO-sup-
ported Ni (5 wt.%) were prepared by the impregnation method, fol-
lowed by drying at 90 °C for 12 h, and by in situ prereduction under
a
b
c
Specific surface area of the catalyst determined by N2 adsorption at ꢀ196 °C.
Amount of surface acidic site determined by TPD of NH3.
Amount of surface basic site determined by TPD of CO2.
d
H2 at 500 °C.
c
-Al2O3-supported metal catalysts, M/
c-Al2O3
Density of surface acidic or basic site (from the surface area and TPD results).

244
K.-i. Shimizu et al. / Journal of Catalysis 300 (2013) 242–250
Table 3
Conversion and yields of products were determined by GC using n-
dodecane as an internal standard. Progress of the reaction was
monitored by GC analysis of aliquots (ca. 0.05 g). The initial rate
of reaction was measured under the conditions where the conver-
sion was below 40%. The products were identified by GC–MS
equipped with the same column as GC and by comparison with
commercially pure products. GC analysis of the gas-phase product
(H2) was carried out by the mass spectrometer (BELMASS).
Preparation conditions, Ni particle size, and catalytic activity of Ni/h-Al2O3 catalysts.
c
d
Ni loading
(wt.%)
TH2
(°C)
Db
(nm)
DV
V0
TOFe
(nm)
(mol molNi hꢀ1
)
(hꢀ1
)
5
10
20
50
50
500
500
500
500
800
6.1
6.5
8.8
12.2
15.6
7.6
9.0
9.6
13.5
16.3
53
35
25
6
346
267
203
67
5
68
a
b
c
Temperatures of prereduction in H2.
Mean diameter of Ni particles estimated by TEM.
Volume–area mean diameter of Ni particles estimated by TEM.
Initꢀia1l formation rate of cyclododecanone 2a per total Ni atoms (mol of 2a
2.5. In situ IR
d
In situ IR spectra were recorded on a JASCO FT/IR-620 equipped
with a quartz IR cell connected to a conventional flow reaction sys-
tem. The sample was pressed into a 50-mg self-supporting wafer
and mounted into the quartz IR cell with CaF2 windows. Spectra
were measured by accumulating 10 scans at a resolution of
4 cmꢀ1. A reference spectrum of the catalyst wafer in He taken at
the measurement temperature was subtracted from each spec-
trum. Prior to each experiment, the catalyst disk was heated in
10% H2/He flow (100 cm3 minꢀ1) at 500 °C for 0.5 h, followed by
cooling to 150 °C in H2/He 60 °C and to 60 °C in He.
molNi h ).
e
TOF per surface Ni atoms calculated from V0 and Dv.
for a series of metal oxide (MOx)-supported Ni catalysts with the
same loading (5 wt.%). Alumina with different crystal phases was
tested as a support of Ni (entries 2–4). The alumina with spinel
structure (h-Al2O3 and
corundum structure (
c
-Al2O3) gave higher yields than that with
a
-Al2O3), and h-Al2O3 gave the highest yield
(45%). Among various support materials (entries 2, 8–18), Ni/
h-Al2O3 showed the highest yield. Use of large amounts (5 mol%)
of conventional Ni catalysts, Raney Ni, and Ni powder (entries 6,
7) resulted in a lower yield of 2a than Ni/h-Al2O3. NiO-loaded h-
Al2O3 (NiO/h-Al2O3) and h-Al2O3 were inactive (entries 1, 5). To
optimize the metal content in Ni/h-Al2O3, catalysts with different
Ni content (5, 10, 20, 50 wt.%) were tested under the reflux condi-
tion, and the initial formation rate of 2a was measured under the
condition where the conversion of 1a was below 40% (Table 3). It
was found that the metal content of 5 wt.% gives the highest activ-
ity. The effect of the prereduction temperature (300–600 °C) of 5
wt.% Ni/h-Al2O3 was also tested, and the results showed that a
reduction temperature of 500 or 600 °C was optimum (Table S1
in the supporting information). Among various metal (Ni, Co, Cu,
3. Results and discussion
3.1. Catalyst optimization
Table 2 shows the influence of various catalyst parameters on
the catalytic activity for dehydrogenation of cyclododecanol 1a to
cyclododecanone 2a under the same reaction conditions (under
N2 in reflux conditions for 1 h). First, we carried out catalytic tests
Table 2
Dehydrogenation of cyclododecanol.a
Ru, Rh, Pd, Ag, Au)-loaded
c-Al2O3 catalysts (Table 1, entries 2,
OH
O
cat.(1 mol%)
17–23), Ni/ -Al2O3 showed the highest yield of 2a.
c
H2
+
o
-xylene (1 g)
144 oC, 1 h
1a
2a
1.0 mmol
3.2. Catalytic performance of Ni/h-Al2O3
b
Entry
Catalyst
Conv. (%)
Yield (%)
V0 (mol molNi hꢀ1
)
The catalyst optimization results show that the 5 wt.% Ni/h-
Al2O3 catalyst prereduced at 500 °C is the best catalyst. With this
catalyst, we studied reusability and the scope of the present sys-
tem (Table 4). The dehydrogenation of cyclooctanol by Ni/h-Al2O3
(1 mol%) for 3 h resulted in a >99% yield of cyclooctanone (entry
1). After 0.5 h of the reaction, the yields of gas-phase H2 (27%)
and cyclooctanone (30%) were nearly identical to the alcohol con-
version (30%), indicating that H2 was generated quantitatively. The
reaction was completely terminated by removal of the catalyst
from the reaction mixture after 0.5 h; further heating of the filtrate
for 2.5 h under reflux conditions did not increase the yield. ICP
analysis of the filtrate confirmed that the content of Ni in the solu-
tion was below the detection limit. These results confirm that the
reaction is attributable to the heterogeneous catalysis of
Ni/h-Al2O3. After the reaction, the catalyst was easily separated
from the reaction mixture by centrifugation. The filtered catalyst
was washed with acetone, followed by drying in air at 90 °C for
1 h, and by reducing in H2 at 500 °C for 0.5 h. As shown in Table 4
(entry 1), the recovered catalyst was reused at least four times [46].
In the low-catalyst-loading (Ni = 0.02 mol%) and solvent-free con-
ditions at 180 °C, the catalyst showed 81% yield of the product after
47 h, corresponding to the TON of 4050 and turnover frequency
(TOF) of 86 hꢀ1 (Scheme 1). The TON is three orders of magnitude
higher than those of previous Ni-based catalysts (TON = 0.2–1.5)
[15–20] for the liquid-phase aerobic oxidation of alcohols.
1c
2
3
NiO/h-Al2O3
Ni/h-Al2O3
0
0
0
45
43
7
45
43
7
64
61
7
Ni/c
-Al2O3
4
Ni/a-Al2O3
5d
6e
7e
8
h-Al2O3
0
0
–
Ni powder
Raney Ni
Ni/CaO
Ni/MgO
Ni/La2O3
Ni/Y2O3
Ni/CeO2
Ni/ZrO2
Ni/TiO2
4.3
35
1.1
1.6
2.9
15
35
16
8.0
1.6
2.0
30
39
4.6
3.1
2.5
3.0
12
1.7
32
1.1
1.6
2.9
15
35
16
7.0
1.6
1.3
30
34
4.6
2.8
2.5
3.0
8.6
0.33
6.0
1.1
1.6
2.9
15
35
16
7.0
1.6
1.3
30
45
4.6
2.7
2.5
3.0
8.6
9
10
11
12
13
14
15
18
17
18
19
20
21
22
23c
Ni/Nb2O5
Ni/SnO2
Co/c-Al2O3
Cu/
Ru/
Rh/
c
c
c
-Al2O3
-Al2O3
-Al2O3
Pd/
Ag/
Au/
c
c
c
-Al2O3
-Al2O3
-Al2O3
a
Conversion of 1a and yield of 2a were determined by GC.
b
Initial ꢀf1ormation rate of cyclododecanone 2a per total Ni atoms (mol of
2a molNi h ).
c
Catalysts were not prereduced.
h-Al2O3 = 50 mg.
5 mol% of Ni.
d
e

K.-i. Shimizu et al. / Journal of Catalysis 300 (2013) 242–250
245
Table 4
Dehydrogenation of various alcohols by Ni/h-Al2O3.
O
OH
cat.(Ni: 0.01 mmol)
+ H2
R2
R1
R1
R2
o
-xylene (1 g), 144 oC
1 mmol
Entry
Alcohol
t (h)
Conv. (%)
Yield (%)
1
2
3
4
5
6
7
8
9
Cyclooctaol
Cycloheptanol
Cyclohexanol
Cyclopentadecanol
Cyclododecanol
exo-Norborneol
4-Octanol
3
99, 99,a, 94b, 94,c, 95d
99, 98a, 94b, 94c, 92d
24
72
24
8
24
24
12
24
100
94
86
91
96
96
86
97
75
74
86
91
93
91
82
97
2-Octanol
5-Nonanol
a
b
c
First reuse.
Second reuse.
Third reuse.
Fourth reuse.
d
(TH2 = 300, 400, 500, 600 °C), and the fraction of metallic Ni0 in
each catalyst (Table S1 in the supporting information) was esti-
mated by H2 TPR (Fig. S1). Fig. 1 shows the formation rate of cycl-
ododecanone 2a per total moles of Ni as a function of the fraction
of metallic Ni0. It is clear that the activity increases with the frac-
tion of metallic Ni0, which indicates that the metallic Ni0 is the ac-
tive species.
OH
O
cat.(Ni: 0.004 mmol)
o
neat, 180 C, 47 h
20 mmol
81% conv.
TON = 4050
TOF = 86 h
81% yield
-1
Next, we discuss the effect of Ni metal particle size. As shown in
Table 2, a series of Ni/h-Al2O3 catalysts with different particle sizes
were prepared by changing the Ni loading and reduction tempera-
ture, and the initial rate for the formation rate of 2a was deter-
mined. The XRD patterns of these Ni/h-Al2O3 catalysts are shown
in Fig. S2. A sharp line at 51.9° due to Ni metal was observed in
the patterns of the high-Ni-loading (20, 50 wt.%) catalysts. A signif-
icantly broad and weak line due to Ni metal was observed in the
patterns of the 5 and 10 wt.% Ni catalysts. When this is combined
with the H2 TPR result, it can be concluded that dominant Ni spe-
cies in these catalysts are Ni metal particles. TEM analysis of Ni/h-
Al2O3 catalysts (Ni = 5, 10, 20, 50 wt.%) showed spherical or semi-
spherical Ni particles. The particle size distributions and the mean
diameter of Ni particles (D) are shown in Fig. S3 and Table 3,
Scheme 1. A gram scale dehydrogenation of cyclooctanol by Ni/h-Al2O3 under neat
conditions under N2 flow.
The scope of the Ni/h-Al2O3-catalyzed dehydrogenation of vari-
ous alcohols is shown in Table 4. In the presence of Ni/h-Al2O3
(1 mol%), various aliphatic secondary alcohols were selectively
converted to the corresponding ketones with good to high yield
(74–93%). Because H2 escapes from the reaction mixture into the
gas phase, the overall endothermic dehydrogenation reaction is
driven to the product side, leading to a high conversion of alcohols
to ketones. For the reaction of a primary alcohol (1-octanol), yields
of the corresponding aldehydes were low (2%). It should be noted
that the previous examples of Ni-based heterogeneous catalysts
for the aerobic alcohol oxidation were effective for activated (aro-
matic) alcohols but not for less activated aliphatic alcohols [15–
20]. Considering that previous Ni catalysts for the acceptor-free
dehydrogenation of alcohols suffer from high reaction temperature
(200–350 °C) [42], low yields of ketones (<70%) [41,42], and lim-
ited scope [41,42], our result represents the first successful exam-
ple of Ni-catalyzed dehydrogenation of various alcohols. The
catalytic activity of Ni/h-Al2O3 for dehydrogenation of alcohols is
higher than that of Ag/Al2O3, reported in our previous study [2].
For the dehydrogenation of cyclododecanol, as an example, Ni/h-
Al2O3 shows a TOF of 11 hꢀ1, which is five times higher than that
of Ag/Al2O3 (TOF = 2.0 hꢀ1) [2].
80
500 oC
60
600 oC
40
20
400 oC
0.6
NiO/
θ
-Al2O3
300 oC
0.4
3.3. Structure–activity relationship
0
0.2
0.8
1
Adopting the dehydrogenation of cyclododecanol 1a to cyclod-
odecanone 2a by Ni/h-Al2O3 as a model reaction, we studied the
structure–activity relationship. First, we studied the effect of Ni
oxidation states on the activity. NiO/h-Al2O3, as an inactive precur-
sor, was reduced in a flow of H2 at various reduction temperatures
Ni0/(Ni0+NiO)
Fig. 1. Formation rate of cyclododecanone per total mole of Ni versus the fraction of
metallic Ni0 (from H2 TPR) for 5 wt.% Ni/h-Al2O3 catalysts reduced under H2 at
various temperatures.

246
K.-i. Shimizu et al. / Journal of Catalysis 300 (2013) 242–250
respectively. It is known that the mean particle diameter estimated
from the number of surface atoms on the assumption of spherical
particles agrees with a volume–area mean diameter (DV) measured
by TEM [47]. Adopting this model, the number of the surface Ni
atoms was estimated using volume–area mean diameter. Using
the number of surface Ni atoms and the initial rates, the TOF per
surface Ni atom was estimated (Table 3). As shown in Fig. 2, the
TOF per surface Ni atom decreases with the mean diameter of Ni
particles. Considering that the fraction of Ni atoms at coordinative-
ly unsaturated sites and metal/support perimeter sites decreases
with the particle size, the results suggest that the present catalytic
system requires Ni0 species located at coordinatively unsaturated
sites and metal/support perimeter sites.
CaO
4
3
2
1
20
10
Y2O3
La2O3
MgO
CeO2
ZrO2
θ-Al2O3
5 SnO2
Nb O
TiO2
2
0
60
0
CeO2
If the metal/support perimeter sites play an important role, the
activity should depend strongly on the type of support material.
Here, we discuss the effect of the support. As listed in Table 1,
we prepared a series of metal oxide-supported Ni catalysts with
the same Ni content (5 wt.%) and the same reduction temperature
(500 °C). Basic (MgO, CaO), amphoteric (h-Al2O3, Y2O3, La2O3, CeO2,
ZrO2), and relatively acidic (TiO2, Nb2O5, SnO2) supports were se-
lected according to well-known classifications of the acid–base
character of metal oxides [48]. The initial rate of the reaction is
plotted in Fig. 3 as a function of the electronegativity of the support
metal oxide, which is generally used as a parameter of acidity of
metal oxides [49]. The rate of dehydrogenation of 2-octanol 1b to
2-octanone 2b by Ni/h-Al2O3 (Table S2) was also determined and
shown in Fig. 3. Clearly, volcano-type relationships between the
catalytic activity of the support and the electronegativity of the
support are obtained. Among the supports with basic to ampho-
teric nature (CaOAh-Al2O3), the activity increases with the electro-
negativity (acidity) of the support oxide, while acidic support
result in lower activity than amphoteric supports. As listed in
Table 1, densities of surface acidic and basic site on Ni-loaded cat-
alyst were estimated based on the surface area of the catalyst and
the number of desorbed NH3 and CO2 during NH3 TPD and CO2 TPD
experiments. In Fig. 3, the densities of acidic and basic site as
experimental indexes of the acid–base character of the catalyst
are plotted versus the support electronegativity. For the supports
with basic to amphoteric nature (CaO ꢂ h-Al2O3), the increase in
electronegativity generally results in a decrease in the density of
the basic sites and an increase in the acid site density. The supports
with electronegativity higher than 2.5 (TiO2 ꢂ SnO2) show negligi-
ble value of basic site density. Comparison of the trends observed
40
20
θ-Al2O3
ZrO2
Y2O3
La2O3
TiO2
Nb2O5
CaO
MgO
SnO2
0
θ-Al2O3
60
40
20
0
CeO2
Y2O3
ZrO2
TiO2
Nb2O5
La2O3
SnO2
2.8
MgO
CaO
2
2.2
2.4
2.6
Electronegativity of support metal oxide
Fig. 3. Reaction rates per total mole of Ni for the dehydrogenation reactions of (s)
cyclododecanol and (}) 2-octanol and the surface densities of basic (N) and acidic
( ) sites as a function of the electronegativity of the support metal oxide.
r
in acid–base character with those found in the catalytic activity al-
lows us to conclude that the co-presence of acidic and basic sites
on the surface of support oxides is indispensable for the Ni-
catalyzed dehydrogenation of alcohols.
300
200
100
0
To verify this hypothesis, we studied the poisoning effects of
acidic or basic additives in the reaction mixture on the activity of
Ni/h-Al2O3 for the dehydrogenation of 1a. As shown in Fig. 4, the
presence of catalytic amounts of acidic (acetic acid) and basic (pyr-
idine) additives decreased the reaction rate. It is established that
adsorption of acetic acid on Al2O3 results in the formation of the
acetate ion at the acid–base pair site (Ald+AOdꢀ site) [50], while
pyridine adsorbs on the surface Lewis acid site (exposed Al cation)
of Al2O3 [51]. Therefore, the result of the poisoning experiment
indicates that the acid–base pair site and the surface Lewis acid
site are necessary for this reaction.
5
10
15
20
3.4. Reaction mechanism
Mean diameter of Ni / nm
Fig. 2. Formation rate of cyclododecanone per total mole of Ni as a function of
average Ni particle size in Ni/h-Al2O3 (from Table 3).
To study the reaction mechanism, we studied in situ IR experi-
ments at 60 °C (Fig. 5). Fig. 5 shows the spectra of adsorbed species

K.-i. Shimizu et al. / Journal of Catalysis 300 (2013) 242–250
247
60
40
20
0
60
40
20
0
0
0.2
0.4
0.6
0.8
0
0.2
0.4
0.6
0.8
1
1
[Acetic acid / M]
[Pyridine / M]
Fig. 4. Formation rate of cyclododecanone over Ni/h-Al2O3 per total moles of Ni as a function of the concentration of pyridine and acetic acid.
formed from the adsorption of 2-propanol-O-d, (CH3)2CHOD, on h-
Al2O3 preheated in He at 500 °C for 0.5 h. In the initial spectrum
(t = 60 s), bands due to 2-propoxide groups [2,52] are predomi-
nant: 2970 cmꢀ1 (methyl asymmetric CAH stretching),
2940 cmꢀ1 (methyl asymmetric CAH stretching), 2874 cmꢀ1 (cou-
should account for the experimental result: the reaction of 2-
propanol with AlAOdꢀ group (basic site on alumina) results in
the formation of 2-propoxide coordinated to the unsaturated Ald+
site (Lewis acid site) and an AlAOD group.
The same IR experiment was conducted using Ni/h-Al2O3 pre-
reduced in H2 at 500 °C for 0.5 h (Fig. 6A). The bands due to
(CH3)2CHAOAl and AlAOD species were also observed for Ni/h-
Al2O3. It is noteworthy that band positions for h-Al2O3 and Ni/
h-Al2O3 are exactly the same, indicating that 2-propoxide is
not adsorbed on the Ni site but on the support surface. Fig. 6B
shows the height of the bands due to 2-propoxide, adsorbed ace-
tone, and AlAOD species as a function of time in a flow of He.
After fast formation of 2-propoxide species (t = 60 s), intensity
of the bands due to the 2-propoxide species gradually decreased,
and a new band at 1700 cmꢀ1 due to C@O stretching of adsorbed
acetone appeared and its intensity increased with time. The
same IR results were observed for the reaction of 2-propanol,
(CH3)2CHOH, on Ni/h-Al2O3 (Fig. 7A). These results suggest that
2-propoxide undergoes CAH cleavage, resulting in the formation
of acetone:
pling of
2722 cmꢀ1
deformation), 1376 cmꢀ1 (methyl symmetric deformation),
1328 cmꢀ1 -CAH deformation), and 1166 and 1130 cmꢀ1
a
-CAH stretching and methyl symmetric CAH stretching),
(a
-CAH stretching), 1465 cmꢀ1 (methyl asymmetric
(a
(CAO/CAC coupled stretching). Broad and multiple bands in the
range 2300–2750 cmꢀ1 due to AlAOD groups are also observed.
The broad feature due to the OD stretching vibration indicates
the acidic (protonic) nature of the AlAOD groups. The band due
to OH deformation of undissociated 2-propanol (1280 cmꢀ1) is
hardly observed. These results indicate the presence of aluminum
alkoxide, (CH3)2CHAOAl, and acidic AlAOD groups on the alumina
surface and the absence of non-dissociatively adsorbed 2-propanol.
Intensities of the bands due to (CH3)2CH-OAl and AlAOD species
gradually decreased with time in a flow of He, but these bands still
remained after 600 s, which indicates that these species were rel-
atively stable. The following surface reaction,
R2CH ꢀ OAl þ Ni0 ! ðCH3Þ C@O þ Ni ꢀ H þ Al
ð2Þ
2
ðCH3Þ CH ꢀ OD þ Aldþ þ Al ꢀ Odꢀ ! ðCH3Þ CH ꢀ OAl þ Al ꢀ OD
2
2
Note that 2-propoxide species on Ni-free alumina did not
convert to acetone (Fig. 5). Thus, metallic Ni is necessary in
the CAH dissociation step. The kinetic curve in Fig. 6B clearly
shows that the formation rate of 2-propoxide is higher than its
consumption rate, which indicates that step (1) is faster than
step (2).
The intensity of the band due to AlAOD groups (2300–
2750 cmꢀ1) decreased with time, indicating the consumption of
the acidic AlAOD groups. The following pathway should account
for this result: protolysis of NiAHdꢀ species with the protonic Dd+
species in an adjacent AlAOD group would release dihydrogen
(HD) accompanied by the regeneration of the Ni0 site and the basic
AlAOdꢀ site of alumina:
ð1Þ
0.1
60 s
Ni ꢀ H þ Al ꢀ OD ! HD þ Ni0 þ Al ꢀ Odꢀ
ð3Þ
300 s
Fig. 7B shows IR spectra of adsorbed species formed by the reac-
tion of Ni/h-Al2O3 with 2-propanol-2-d1 (a-deuterized 2-propanol,
(CH3)2CDOH). Bands at 2970 cmꢀ1(methyl asymmetric CAH
stretching), 2940 cmꢀ1 (methyl asymmetric CAH stretching),
600 s
3000
2500
2000
1500
2870 cmꢀ1 (methyl symmetric CAH stretching), 2104 cmꢀ1
(a-
Wavenumber / cm-1
CAD stretching), 1465 cmꢀ1 (methyl asymmetric deformation),
1376 cmꢀ1 (methyl symmetric deformation), and 1190 cmꢀ1
Fig. 5. IR spectra of adsorbed species on h-Al2O3 as a function of time in a flow of He
at 60 °C. At t = 0 s, (CH3)2CHOD (0.5 mmol gꢀ1) was introduced into the sample.
(CAO/CAC coupled stretching) were assigned to
a-deuterized 2-

248
K.-i. Shimizu et al. / Journal of Catalysis 300 (2013) 242–250
A
0.4 B
Al-OiPr
0.1
0.3
0.2
0.1
0
acetonead
60 s
Al-OD
300 s
600 s
0
200
400
600
3000
2500
2000
1500
Wavenumber / cm-1
t
/ s
Fig. 6. (A) IR spectra of adsorbed species on Ni/h-Al2O3 and (B) height of the bands due to (s) 2-propoxide on the support (1130 cmꢀ1), (d) adsorbed acetone (1700 cmꢀ1),
and (4) Al–OD group (2570 cmꢀ1) as a function of time in a flow of He at 60 °C. At t = 0 s, (CH3)2CHOD (0.5 mmol gꢀ1) was introduced into the sample.
B
A
0.1
0.1
60 s
60 s
300 s
300 s
600 s
600 s
3000
2500
2000
1500
3000
2500
2000
1500
Wavenumber / cm-1
Wavenumber / cm-1
Fig. 7. IR spectra of adsorbed species on Ni/h-Al2O3 under He flowing at 60 °C. At t = 0 s, 0.5 mmol gꢀ1 of (A) (CH3)2CHOH or (B) (CH3)2CDOH was introduced into the sample.
kH
/
k
D = 2.0
0
-0.2
-0.4
-0.6
-0.8
-1
H
O
H
O
OH
Al
OH
Al
Al
Al
O
O
OH
H
H
O
0
100
200
300
400
O
Al
O
Al
Al
t
/ s
Al
O
O
Fig. 8. First-order plots for the decomposition of the surface 2-propoxide species on
Ni/h-Al2O3 at 60 °C. Rate constants estimated from the changes in the area of the IR
band in the range 1080–1210 cmꢀ1 for (CH3)2CHOH (s) and 1100–1250 cmꢀ1 for
(CH3)2CDOH (N) were 3.4 ꢃ 10ꢀ3 sꢀ1 and 1.7 ꢃ 10ꢀ3 sꢀ1, respectively.
H2
Fig. 9. A possible mechanism for the dehydrogenation of alcohols by Ni/h-Al2O3.

K.-i. Shimizu et al. / Journal of Catalysis 300 (2013) 242–250
249
propoxide groups on alumina, (CH3)2CDAOAl [2]. A weak band at
metal-free and ligand-free catalysts for hydrogen-transfer-type
reactions.
2174 cmꢀ1 may be assigned to
a-CAD stretching of a different type
of 2-propoxide. The band at 2104 cmꢀ1 due to the CAD bond of
(CH3)2CDAOAl species decreased with time, and a new band due
to adsorbed acetone (1700 cmꢀ1) increased. This is direct evidence
of Ni-catalyzed CAD bond dissociation of the (CH3)2CDAOAl spe-
cies (Eq. (2)).
Acknowledgments
This work was supported by a ‘‘Grant for Advanced Industrial
Technology Development’’ in 2011 from the New Energy and
Industrial Technology Development Organization (NEDO) and a
MEXT program ‘‘Elements Strategy Initiative to Form Core Re-
search Center’’ (since 2012), Japan. The authors thank Naomichi
Imaiida (Nagoya University) for his contributions in the prelimin-
ary experiments.
Kinetic analysis of the conversion of the 2-propoxide groups
was undertaken using time-resolved IR data (Fig. 7). The relative
amounts of (CH3)2CHAOAl and (CH3)2CDAOAl species were esti-
mated from the changes in the area of the IR band in the ranges
1080–1210 cmꢀ1 and 1100–1250 cmꢀ1
, respectively, because
these bands had the least overlap with other bands. As shown
in Fig. 8, the first-order plots of the 2-propoxide group during
its conversion to acetone gave fairly good straight lines. The
first-order rate constants of 2-propoxide consumption were esti-
mated to be 3.4 ꢃ 10ꢀ3 and 1.7 ꢃ 10ꢀ3 sꢀ1 for (CH3)2CHAOAl
and (CH3)2CDAOAl species, respectively. The kinetic isotope effect
(kH/kD) at 60 °C was estimated to be 2.0, corresponding to the dif-
ference in the activation energies of 1.9 kJ molꢀ1. This value is
lower than the zero-point energy difference (3.7 kJ molꢀ1) in
Appendix A. Supplementary material
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.jcat.2013.01.005.
References
[1] F. Alonso, P. Riente, M. Yus, Acc. Chem. Res. 44 (2011) 379–391.
[2] K. Shimizu, K. Sugino, K. Sawabe, A. Satsuma, Chem. Eur. J. 15 (2009) 2341–
2351.
[3] K. Shimizu, K. Sawabe, A. Satsuma, Catal. Sci. Technol. 1 (2011) 331–341.
[4] K. Shimura, K. Shimizu, Green Chem. 14 (2012) 2983–2985.
[5] C. Parmeggiani, F. Cardona, Green Chem. 14 (2012) 547–564.
[6] T. Mallat, A. Baiker, Chem. Rev. 104 (2004) 3037–3058.
[7] T. Matsumoto, M. Ueno, N. Wang, S. Kobayashi, Chem. Asian J. 3 (2008) 196–
214.
[8] N. Dimitratos, J.A. Lopez-Sanchez, G.J. Hutchings, Chem. Sci. 3 (2012) 20–44.
[9] K. Mori, K. Yamaguchi, T. Hara, T. Mizugaki, K. Ebitani, K. Kaneda, J. Am. Chem.
Soc. 124 (2002) 11572–11573.
[10] K. Yamaguchi, N. Mizuno, Angew. Chem. Int. Ed. 41 (2002) 4538–4542.
[11] S. Mori, M. Takubo, K. Makida, T. Yanase, S. Aoyagi, T. Maegawa, Y. Monguchi,
H. Sajiki, Chem. Commun. 45 (2009) 5159–5161.
[12] Y.M.A. Yamada, T. Arakawa, H. Hocke, Y. Uozumi, Angew. Chem. Int. Ed. 46
(2007) 704–706.
[13] A. Abad, A. Corma, H. Garcia, Chem. Eur. J. 14 (2008) 212–222.
[14] H. Tsunoyama, N. Ichikuni, H. Sakurai, T. Tsukuda, J. Am. Chem. Soc. 131 (2009)
7086–7093.
[15] B.M. Choudary, M.L. Kantam, A. Rahman, C.V. Reddy, K.K. Rao, Angew. Chem.
Int. Ed. 40 (2001) 763–766.
[16] C. Li, H. Kawada, X. Sun, H. Xu, Y. Yoneyama, N. Tsubaki, ChemCatChem 3
(2011) 684–689.
[17] V.R. Choudhary, P.A. Chaudhari, V.S. Narkhede, Catal. Commun. 4 (2003) 171–
175.
[18] H. Ji, T. Wang, M. Zhang, Q. Chen, X. Gao, React. Kinet. Catal. Lett. 90 (2007)
251–257.
[19] T. Kawabata, Y. Shinozuka, Y. Ohishi, T. Shishido, K. Takaki, K. Takehira, J. Mol.
Catal. A 236 (2005) 206–215.
[20] H. Ji, T. Wang, M. Zhang, Y. She, L. Wang, Appl. Catal. A 282 (2005) 25–30.
[21] R.J. Angelici, ACS Catal. 1 (2011) 772–776.
[22] T.C. Johnson, D.J. Morris, M. Wills, Chem. Soc. Rev. 39 (2010) 81–88.
[23] S. Shinoda, T. Kojima, Y. Saito, J. Mol. Catal. 18 (1983) 99–104.
[24] G.B.W.L. Ligthart, R.H. Meijer, M.P.J. Donners, J. Meuldijk, J.A.J.M. Vekemans,
L.A. Hulshof, Tetrahedron Lett. 44 (2003) 1507–1509.
[25] G.R.A. Adair, J.M.J. Williams, Tetrahedron Lett. 46 (2005) 8233–8235.
[26] H. Junge, M. Beller, Tetrahedron Lett. 46 (2005) 1031–1034.
[27] S. Shahane, C. Fischmeister, C. Bruneau, Catal. Sci. Technol. 2 (2012) 1425–
1428.
[28] J. Zhang, M. Gandelman, L.J.W. Shimon, H. Rozenberg, D. Milstein,
Organometallics 23 (2004) 4026–4033.
[29] M. Nielsen, A. Kammer, D. Cozzula, H. Junge, S. Gladiali, M. Beller, Angew.
Chem. Int. Ed. 50 (2011) 9593–9597.
a
-CAH (2722 cmꢀ1) and
a
-CAD (2104 cmꢀ1) in the IR spectra.
This indicates that the rate-determining step of 2-propoxide
conversion to acetone is the cleavage of the
other kinetically relevant steps may also exist.
a-CAH bond, but
Based on the mechanistic experiments on the dehydrogenation
of alcohols by Ni/h-Al2O3, a reaction mechanism is presented in
Fig. 9. Initially, a basic AlAOdꢀ group of alumina abstracts a proton
from alcohol to afford an alkoxide group on an adjacent Ald+ site
(Lewis acid site) and a proton on the AlAOdꢀ site (AlAOHd+). An
alkoxide at the interface between the Ni NP and alumina under-
goes hydride abstraction to yield hydride species on Ni (NiAH)
and acetone as the rate-determining step. Then, the NiAH species
undergoes protolysis by a neighboring AlAOHd+ site to release H2
gas, accompanied by the regeneration of the Ni0 site and the basic
AlAOdꢀ site of alumina. On the basis of the proposed mechanism,
the strong effect of support acid–base nature on the Ni-catalyzed
alcohol dehydrogenation (Fig. 3) can be explained as follows. The
abstraction of protons in alcohol by the metal oxide surface re-
quires basic sites. Thus, on the Ni-loaded acidic oxides (Nb2O5,
SnO2), dissociative adsorption of 2-propanol is limited by the weak
nucleophilic character of the metal oxide surface. On the other
hand, the protonic (electrophilic) nature of the OH groups on the
basic oxides (CaO, MgO) is weaker than that on the amphoteric
oxide (alumina), and the protolysis of NiAHdꢀ species with an adja-
cent OH group is slow.
4. Conclusions
Alumina-supported Ni metal NP was found to act as a heteroge-
neous and reusable catalyst for the acceptor-free oxidation of sec-
ondary alcohols. Among various supports, amphoteric supports
(such as alumina) having both acidic and basic sites gave higher
activity than acidic and basic supports. The TOF per surface Ni in-
creases with decreasing Ni particle size. These results suggest that
low-coordinated Ni0 sites and the metal/support interface play
important roles in the catalytic cycle. On the basis of in situ IR
and kinetic experiments, the following mechanism is proposed:
[30] K.-I. Fujita, N. Tanino, R. Yamaguchi, Org. Lett. 9 (2007) 109–111.
[31] K.-I. Fujita, T. Yoshida, Y. Imori, R. Yamaguchi, Org. Lett. 13 (2011) 2278–2281.
[32] S. Musa, I. Shaposhnikov, S. Cohen, D. Gelman, Angew. Chem. Int. Ed. 50 (2011)
3533–3537.
[33] M. Yamashita, K. Kawamura, M. Suzuki, Y. Saito, Bull. Chem. Soc. Jpn. 64 (1991)
272–278.
[34] J.H. Choi, N. Kim, Y.J. Shin, J.H. Park, J. Park, Tetrahedron Lett. 45 (2004) 4607–
4610.
[35] W.H. Kim, I.S. Park, J. Park, Org. Lett. 8 (2006) 2543–2545.
[36] T. Mitsudome, Y. Mikami, H. Funai, T. Mizugaki, K. Jitsukawa, K. Kaneda,
Angew. Chem. Int. Ed. 47 (2008) 138–141.
[37] W. Fang, Q. Zhang, J. Chen, W. Deng, Y. Wang, Chem. Commun. 46 (2010)
1547–1549.
[38] Q. Zhang, W. Deng, Y. Wang, Chem. Commun. 47 (2011) 9275–9292.
[39] W. Fang, J. Chen, Q. Zhang, W. Deng, Y. Wang, Chem. Eur. J. 17 (2011) 1247–
1256.
(1) reaction of an alcohol with the Lewis acid (Ald+)–base (AlAOdꢀ
)
pair site of alumina yields an alkoxide on the Ald+ site and a proton
on the AlAOdꢀ site, (2) CAH dissociation of the alkoxide by Ni0 site
to form NiAH and ketone, (3) protolysis of NiAH by a neighboring
proton to release H2 gas, accompanied by the regenerations of the
Ni0 site and the basic AlAOdꢀ site of alumina. Fundamental infor-
mation in this study will be useful as a design concept for noble

250
K.-i. Shimizu et al. / Journal of Catalysis 300 (2013) 242–250
[40] M. Noda, S. Shinoda, Y. Saito, Bull. Chem. Soc. Jpn. 61 (1988) 961–965.
[41] M. Yamashita, F. Dai, M. Suzuki, Y. Saito, Bull. Chem. Soc. Jpn. 64 (1991) 628–
634.
[42] Y. Uemichi, T. Sakai, T. Kanazuka, Chem. Lett. 18 (1989) 777–780.
[43] Y. Uemichi, K. Shouji, M. Sugioka, T. Kanazuka, Bull. Chem. Soc. Jpn. 98 (1995)
385–387.
[44] T. Mitsudome, Y. Mikami, K. Ebata, T. Mizugaki, K. Jitsukawa, K. Kaneda, Chem.
Commun. 44 (2008) 4804–4806.
[45] R.K. Marella, C.K.P. Neeli, S.R.R. Kamaraju, D.R. Burri, Catal. Sci. Technol. 2
(2012) 1833–1838.
Yields of cyclooctanone for the first and second reuse experiments were 95%
and 5%, respectively. It is found that the reactivation treatment is necessary.
[47] K. Kunimori, T. Uchijima, M. Yamada, H. Matsumoto, T. Hattori, Y. Murakami,
Appl. Catal. 4 (1982) 67–81.
[48] G. Busca, Phys. Chem. Chem. Phys. 1 (1999) 723–736.
[49] N.C. Jeong, J.-S. Lee, E.L. Tae, Y.J. Lee, K.B. Yoon, Angew. Chem. Int. Ed. 120
(2008) 10282–10286.
[50] K. Shimizu, H. Kawabata, A. Satsuma, T. Hattori, J. Phys. Chem. B 103 (1999)
5240–5245.
[51] C. Morterra, G. Magnacca, Catal. Today 27 (1996) 497–532.
[52] Z. Martinez-Ramirez, J.A. Gonzalez-Calderon, A. Almendarez-Camarillo, J.C.
Fierro-Gonzalez, Surf. Sci. 606 (2012) 1167–1172.
[46] Catalyst recycle experiment without reactivation treatment was also carried
out for the reaction of cyclooctanol (t = 6 h) under the conditions of Table 3.