Support effects of metal oxides on gold-catalyzed one-pot N-alkylation of amine with alcohol

Article abstract of DOI:10.1016/j.apcata.2011.11.017

Gold nanoparticles supported on metal oxides catalyzed the N-alkylation of aniline with benzyl alcohol to produce secondary amine under the equimolar amounts of substrates without additives under mild conditions. Selectivity to secondary amine was changed by the kinds of supports. Gold on ZrO₂ exhibited the highest catalytic performance and achieved 94% selectivity to secondary amine. Surface hydroxyl groups of ZrO₂ played important roles in the deprotonation of alcohol in the alcohol dehydrogenation step, as the adsorption sites of aniline, and as a proton source in the hydrogen transfer step.

Full text of DOI:10.1016/j.apcata.2011.11.017

Applied Catalysis A: General 413–414 (2012) 261–266
Contents lists available at SciVerse ScienceDirect
Applied Catalysis A: General
journal homepage: www.elsevier.com/locate/apcata
Support effects of metal oxides on gold-catalyzed one-pot N-alkylation of amine
with alcohol
Tamao Ishidaa,d,∗, Rena Takamurab,d, Takashi Takeib,d, Tomoki Akitac,d, Masatake Harutab,d,∗∗
a Department of Chemistry, Graduate School of Sciences, Kyushu University, 6-10-1 Hakozaki, Higashi-ku, Fukuoka 812-8581, Japan
b Department of Applied Chemistry, Graduate School of Urban Environmental Sciences, Tokyo Metropolitan University, 1-1 Minami-osawa, Hachioji, Tokyo 192-0397, Japan
c Research Institute for Ubiquitous Energy Devices, National Institute of Advanced Industrial Science and Technology (AIST), 1-8-31 Midorigaoka, Ikeda, Osaka 563-8577, Japan
d Japan Science and Technology Agency (JST), CREST, 4-1-8 Honcho, Kawaguchi, Saitama 332-0012, Japan
a r t i c l e i n f o
a b s t r a c t
Article history:
Gold nanoparticles supported on metal oxides catalyzed the N-alkylation of aniline with benzyl alcohol
to produce secondary amine under the equimolar amounts of substrates without additives under mild
conditions. Selectivity to secondary amine was changed by the kinds of supports. Gold on ZrO2 exhibited
the highest catalytic performance and achieved 94% selectivity to secondary amine. Surface hydroxyl
groups of ZrO2 played important roles in the deprotonation of alcohol in the alcohol dehydrogenation
step, as the adsorption sites of aniline, and as a proton source in the hydrogen transfer step.
© 2011 Elsevier B.V. All rights reserved.
Received 12 October 2011
Received in revised form
11 November 2011
Accepted 13 November 2011
Available online 20 November 2011
Keywords:
Gold nanoparticles
Supported gold catalyst
N-Alkylation
Dehydrogenation
1. Introduction
form imine via hemiaminal intermediate, and (iii) the hydrogen
transfer to produce secondary amine (Scheme 1). The hydrogen,
Secondary amines are one of the important chemicals or inter-
mediates for fine chemical syntheses. They are generally produced
by stoichiometric reactions using alkyl halides with primary amines
or the reductive amination of aldehydes or ketones with primary
amines via the formation of imine [1]. However, the former process
has drawbacks that a lot of inorganic salts and undesired different
amines are produced as by-products. As to the reductive amination,
alcohols are more beneficial than carbonyl compounds because
carbonyl compounds are generally produced by the oxidation of
alcohols.
generated by the step (i), can be used for the step (iii), so that neither
2 nor H2 is required for the alcohol oxidation or hydrogenation of
O
imine. In addition, water is the only by-product. The N-alkylation
has been mainly studied in the presence of homogeneous metal cat-
alysts [3–7]. More recently, heterogeneous metal catalysts, such as
Ru [8], Ag [9,10], Pd [11], and Cu [12] have been reported to be active
for the N-alkylation. However, high catalytic activity or selectivity
to secondary amine has generally been obtained by using excess
amount of alcohols or amines or in the presence of acidic or basic
additives.
The N-alkylation of primary amines with alcohols to produce
secondary amines has been recently focused as a new efficient
method for the production of secondary amines [2]. It consists three
steps in one-pot; (i) the dehydrogenation of alcohol to aldehyde, (ii)
the dehydrated condensation of aldehyde with primary amine to
Gold catalysts, that had long been considered to be catalyt-
ically inert, have gained prominence since Au exhibits unique
size- and support-dependent catalytic activity and selectivity [13].
We reported that Au clusters supported on inert porous coordi-
nation polymers could catalyze the N-alkylation of amine with
alcohol under equimolar amounts of substrates [14]. Although
the size effect of Au particles was pronounced especially for Au
particles deposited on inert supports, base was required to pro-
mote the first dehydrogenation step, even when Au could be
deposited as clusters smaller than 2 nm. Later, Cao and co-workers
reported that Au clusters deposited on TiO2 exhibited the high-
est catalytic activity and selectivity to secondary amine among the
metal oxide-supported Au catalysts under the equimolar amount
of substrates without additives [15]. However, the reaction was
∗
Corresponding author at: Department of Chemistry, Graduate School of Sci-
ences, Kyushu University, 6-10-1 Hakozaki, Higashi-ku, Fukuoka 812-8581, Japan.
Tel.: +81 92 642 7528; fax: +81 92 642 7528.
∗∗
Corresponding author at: Department of Applied Chemistry, Graduate School of
Urban Environmental Sciences, Tokyo Metropolitan University, 1-1 Minami-osawa,
Hachioji, Tokyo 192-0397, Japan. Tel.: +81 42 677 2852; fax: +81 42 677 2852.
E-mail addresses: tishida@chem.kyushu-univ.jp (T. Ishida),
haruta-masatake@center.tmu.ac.jp (M. Haruta).
0926-860X/$ – see front matter © 2011 Elsevier B.V. All rights reserved.
doi:10.1016/j.apcata.2011.11.017

262
T. Ishida et al. / Applied Catalysis A: General 413–414 (2012) 261–266
Scheme 1. Au-Catalyzed N-Alkylation of aniline with benzyl alcohol to produce secondary amine.
performed under pressurized N2. On the other hand, Peng et al.
synthesized N-benzylaniline from nitrobenzene with benzyl alco-
hol over Au/Fe2O3 in the presence of large excess amount of alcohol
[16]. These suggested that the catalytic performance of Au could be
significantly changed by the kind of supports, however, the role of
supports is not fully investigated.
In this work, we have screened a variety of metal oxide-
supported Au catalysts in the N-alkylation of aniline with benzyl
alcohol under the most atom-economical reaction conditions;
using equimolar amounts of substrates without additives under
atmospheric pressure of N2 and investigated the origin of support
effects. We also examined the N-alkylation with aliphatic alcohol.
yields of N-alkylation were measured by gas chromatography (GC)
by using a SHIMADZU GC-14B with a G-Column G-205 (1.2 mm
i.d., 40 m, Chemicals Evaluation and Research Institute Japan) using
dodecane as an internal standard. Qualitative analysis was per-
formed with a GC-MS (SHIMADZU PARVUM2 and GC-2010 with
a Shinwa Chemical ULBON HR-1 capillary column, 0.25 mm i.d.,
30 m).
2.3. Catalyst preparation
Gold on ZrO2 was prepared by the grinding method [17]. A
mixture of metal oxide support (3.0 g), Me2Au(acac) (50 mg for
1 wt% Au loading), and acetone (10 g) was ground by ball milling
(350 rpm) in air at room temperature for 30 min. The mixture was
filtered, and calcined in air at 300 ◦C for 4 h. Gold on WO3 was pre-
pared by mixing WO3 powder with Me2Au(acac) in an agate mortar
at room temperature for 20 min followed by calcination in air at
300 ◦C for 4 h [18]. Gold on NiO and Fe2O3 were prepared by the
co-precipitation [19]. Gold on CeO2, Al2O3, TiO2, and Co3O4 were
prepared by the deposition-precipitation (DP) method according to
the literature [20]. Gold on MgO was prepared by DP in the presence
of Mg citrate according to the literature [21].
2. Experimental
2.1. Materials
High purity microparticulate CeO2 and ZrO2 (RC-100) were
supplied by Dai-ichi Kigenso Kagaku Kogyo Co. Ltd. Alumina,
(Al2O3, NEOBEAD 300), and TiO2 (P-25) were supplied by Mizusawa
Industrial Chemicals, Ltd. and Nippon Aerosil Co., Ltd., respec-
tively. Proprietary WO3 was used as a support. Reagent grade,
Ni(NO3)2·6H2O, Co(NO3)2·6H2O, and Fe(NO3)2·6H2O were pur-
chased from Wako Pure Chemical Industries, Ltd. and used as
received. Cobalt oxide (Co3O4) was prepared by the neutraliza-
tion method from aqueous Co(NO3)2·6H2O solution with aqueous
Na2CO3 solution followed by calcination in air at 300 ◦C for
4 h. Tetrachloroauric acid tetrahydrate (HAuCl4·4H2O) was pur-
chased from Kishida chemical. Dimethyl gold(III)acetylacetonate,
Me2Au(acac), was purchased from Trichemical Laboratories Inc
and used as received. Reagent grade, aniline, benzyl alcohol, N-
benzylidenbenzylamine, dodecane, and toluene were purchased
and used without further purification.
2.4. Catalytic tests
A mixture of benzyl alcohol (52 ␮L, 0.5 mmol), aniline (46 ␮L,
0.5 mmol), Au catalyst (Au 1.5 mol% to substrate), and toluene
(3.0 mL) was stirred under N2 atmosphere at 110 ◦C for 22 h. After
the reaction, the mixture was extracted with Et2O and filtered.
The filtrate was analyzed by GC using dodecane as an internal
standard. In principle, there are two selectivities based on the
conversions of benzyl alcohol and of aniline. To simplify the com-
parison of the support effects, the selectivity to secondary amine
was calculated based on the total yields by the following equation:
selectivity = yield of sec. amine/total yield of products (sec. amine,
imine, aldehyde, and benzyl benzoate).
2.2. Instruments
Ball milling was carried out by using Nagao System Planet M2-3
with ZrO2 pot and ZrO2 balls (ꢀ = 2 and 5 mm). Transmission elec-
tron microscopic (TEM) and high-angle annular dark-field scanning
TEM (HAADF-STEM) observations were carried out by using a
JEOL JEM-3000F operating at 300 kV. Diffuse reflectance infra-red
Fourier transform (DRIFT) spectra were obtained by using a JASCO
FT/IR-620. Diffuse reflectance accessory (Spectra Tech, Collector)
can be evacuated and heated. Catalyst samples were pretreated at
200 ◦C for 0.5 h under vacuum (<10−3 Pa). After the catalyst was
cooled to room temperature, DRIFT spectra were obtained before
the introduction of compounds. Then, benzyl alcohol, aniline, and
benzaldehyde were adsorbed on the catalyst surfaces at room tem-
perature under vacuum for 30 min and DRIFT spectra were obtained
after the adsorption under vacuum. The conversions and product
3. Results
3.1. Support screening
Acid–base properties of metal oxides can be correlated to the
electronegativity of metal element or metal ions (ꢁi) in metal
oxide support and an increase in electronegativity increased acid-
ity of metal oxide surfaces [23]. ꢁi is expressed as ꢁi = (1 + 2z)ꢁ0,
where ꢁ0 is the Pauling’s electronegativity of metal element and
z is the valence of metal ions [23]. Table 1 shows the results of
the N-alkylation of aniline (1) with benzyl alcohol (2) over Au NPs
supported on metal oxides. Gold on basic (entries 1–3) and neu-
tral metal oxides (entries 4–8) showed relatively higher alcohol

T. Ishida et al. / Applied Catalysis A: General 413–414 (2012) 261–266
263
Table 1
N-Alkylation of aniline (1) with benzyl alcohol (2) over Au catalysts.a
Entry
Catalyst
ꢁib
Conv. 1 (%)c
Conv. 2 (%)c
Yield (%)c
Selec. 5 (%)d
3
4
5
6
1
2
3
4
5
6
7
8
9
Au/MgO
Au/NiO
Au/CeO2
Au/Al2O3
Au/Co3O4
Au/ZrO2
Au/Fe2O3
Au/TiO2
Au/WO3
6
9
9.9
10.5
11.3
12.6
12.8
13.5
22.1
44
29
33
59
90
79
42
81
39
68
43
70
68
19
39
15
17
51
88
31
40
72
15
4
0
0
17
0
0
0
0
0
0
6
39
16
14
0
60
0
10
56
5
20
0
0
0
0
5
0
13
10
8
0
46
0
d
>99
82
42
87
34
d
9
19
a
Reaction conditions: aniline (0.5 mmol), benzyl alcohol (0.5 mmol), Au catalyst (Au 1.5 mol%), toluene (3.0 mL), 110 ◦C, 1 atm of N2, 22 h.
Electronegativity of metal ion (ꢁi) in metal oxide supports is expressed as ꢁi = (1 + 2z)ꢁ0, where ꢁ0 is the Pauling’s electronegativity of metal element and z is the valence
b
of metal ions [23].
c
Conversions and yields were calculated by GC using dodecane as an internal standard. Selectivity was calculated based on the total yields of all the products detected by
GC.
d
Au 5 mol%.
Scheme 2. Formation of benzyl benzoate over Au catalysts.
conversions than Au on acidic WO3 (entry 9) except for Au/NiO
(entry 2) and Au/Fe2O3 (entry 7). Among the catalysts, Au/ZrO2 gave
both appreciable high catalytic activity and the highest selectiv-
ity to secondary amine (5) (entry 6). While Au/NiO (entry 2) and
Au/WO3 (entry 9) gave relatively low alcohol conversions, they
exhibited better selectivity to sec. amine than others. We have
reported that Au/NiO was highly active for aerobic oxidation of
methanol in the presence of amine and preferentially oxidized alco-
hol to give formamide [22], whereas it showed relatively lower
catalytic activity but better selectivity to sec. amine than others in
N-alkylation. This result implies that the catalytic nature of Au/NiO
for alcohol transformation was altered by the presence of O2. Gold
on CeO2 gave not only sec. amine but also benzaldehyde (3) and
benzyl benzoate (6) as by-products, lowering the selectivity to sec.
amine (entry 3). Benzyl benzoate was formed from benzaldehyde
and benzyl alcohol via the formation of hemiacetal followed by the
dehydrogenation (Scheme 2). Gold on Co3O4 and Au/Fe2O3 gave
only imine (4) and did not promote the hydrogen transfer even
when the amount of Au was increased (entries 5 and 7). Cao and
co-workers reported that Au/TiO2 exhibited high catalytic activity
and selectivity in the N-alkylation [15]. However, Au/TiO2 did not
show high selectivity to sec. amine under our conditions (entry 8)
probably due to the relatively larger Au NPs (2.8 nm) in our catalysts
than 1.8 nm in Au/TiO2 prepared by Cao and co-workers.
In the N-alkylation, Shimizu et al. demonstrated that the yield
of secondary amine could be correlated to the electronegativity of
metal element in metal oxides, showing that volcano type relation-
ships with Al2O3 support at the top for Ag clusters catalysts [9].
Fig. 1a and b show the relationships between the electronegativ-
ity of metal ions (ꢁi) and the conversion of benzyl alcohol and the
selectivity to sec. amine, respectively. It has been reported in the
reaction mechanism of Au NPs-catalyzed alcohol oxidation that the
deprotonation of alcohols to form alkoxide takes place at the basic
sites of metal oxide surfaces [24]. Therefore, basic oxides have been
recognized as efficient supports. It is reasonable that basic and neu-
tral oxides gave higher alcohol conversion than acidic WO3 (Fig. 1a).
On the contrary, the relationships between acid–base properties
and the selectivity to sec. amine were not observed under our con-
ditions (Fig. 1b).
3.2. Dehydrogenation of alcohol over gold catalysts
Catalytic performance of Au catalysts for the dehydrogenation
of benzyl alcohol was studied (Table 2). The reaction proceeded
in the absence of aniline and the formation of molecular hydro-
gen was confirmed. The catalytic activity order was similar to the
one for the N-alkylation. However, the conversion of benzyl alcohol
decreased to 46% (entry 3) as compared to 82% in the N-alkylation
over Au/ZrO2. It is likely that imine acted as a hydrogen acceptor
to remove hydrogen or hydride from the Au surfaces in a similar
manner to O2 in the catalytic cycle of oxidative dehydrogenation
over Au catalysts [24]. Again, Au/NiO showed poor catalytic activ-
ity in the absence of O2 (entry 1). The catalytic activity of Au/CeO2
for the dehydrogenation was not affected by the absence of aniline
(entry 2).
Fang et al. demonstrated that Au NPs supported on hydrotalcite
possessing both acidity and basicity exhibited the best catalytic per-
formance among the other supports such as acidic or basic metal
oxides for the dehydrogenation of alcohol in the absence of hydro-
gen acceptor [25]. They proposed that H+ would be supplied from
acidic sites of hydrotalcite. However, taking into account that Au
NPs supported on basic CeO2 catalyzed the dehydrogenation of
alcohol, CeO2 possesses an ability to supply H+ although there are
no acidic sites on CeO2. These results implied that hydrogen trans-
fer efficiency from imine to sec. amine did not directly correlate to
the surface acidity of metal oxide supports.
Table 2
Dehydrogenation of benzyl alcohol (2).a
Entry
Catalyst
Conv. (%)b
Yield 3 (%)b
Yield 6 (%)b
1
2
3
5
Au/NiO
Au/CeO2
Au/ZrO2
Au/WO3
16
66
46
0
14
32
35
0
0
31
0
0
a
Reaction conditions: benzyl alcohol (0.5 mmol), toluene (3.0 mL), Au 1.5 mol%,
110 ◦C, 22 h.
b
Calculated by GC using dodecane as an internal standard.

264
T. Ishida et al. / Applied Catalysis A: General 413–414 (2012) 261–266
Table 3
Optimization of reaction conditions for Au/ZrO2.
Entry
Conditionsa
Conv. 1 (%)b
Conv. 2 (%)b
Yield (%)b
Selec. 5 (%)c
3
4
5
6
1
2
3
4
A
B
C
D
79
91
–
82
>99
42
0
0
0
0
31
45
8
46
45
30
80
0
0
0
0
60
50
78
94
d
86
99
4
a
Reaction conditions: [A] aniline (0.5 mmol), benzyl alcohol (0.5 mmol), Au catalyst (Au 1.5 mol%), toluene (3.0 mL), 110 ◦C, under 1 atm of N2, 22 h. [B] Au 2 mol%, 2-propanol
(0.5 mmol) was added. [C] Conditions A but aniline (1.0 mmol). [D] Conditions A but Au 4 mol%.
b
Conversions and yields were calculated by GC using dodecane as an internal standard.
Selectivity was calculated based on total yields of all products detected by GC.
Conversion was not calculated because 2 eq. amount of 1 was used.
c
d
3.3. Optimization of reaction conditions
H2 generated by the dehydrogenation of alcohol [11]. Accordingly,
2-propanol was added as a hydrogen donor but not effective to
improve the selectivity to sec. amine (entry 2). It was probably
due to the reaction inhibition of benzaldehyde with aniline on Au
surfaces. When an excess amount of aniline (2 eq.) was used, the
hydrogen transfer efficiency was improved to give 78% selectivity
to sec. amine whereas the dehydrogenation of alcohol was inhib-
ited (entry 3). An increase in the gold content appreciably improved
the selectivity to sec. amine and 4 mol% of Au achieved the selectiv-
ity of 94% under equimolar amount of alcohol and aniline without
additives (entry 4).
We further searched better reaction conditions by using
Au/ZrO2 (Table 3). To use excess amount of alcohol or amine has
usually been adopted in order to achieve higher selectivity to sec.
amines [8–12]. The fact, that the conversion of benzyl alcohol
exceeded the conversion of aniline, suggested that side reactions
of alcohol might occur such as disproportionation to benzaldehyde
and toluene [25] and toluene formation by the hydrogenolysis with
3.4. N-Alkylation of aliphatic substrate
We examined N-alkylation by using aliphatic amine and alco-
hol. The combination of butyl amine with 1-hexanol did not give the
corresponding sec. amine but butylidene butyl amine via the dehy-
drogenative homo-coupling of butyl amine. When reactive benzyl
alcohol was used, the coupling of butyl amine was prevented and
the corresponding sec. amine was obtained in 26% yield. In addition,
1-hexanol could be converted and reacted with aniline to give sec.
amine in 15% yield. Although the selectivity to sec. amine was not
satisfied at this stage, this result proved that Au/ZrO2 could pro-
mote the dehydrogenation of aliphatic alcohol and followed the
N-alkylation.
3.5. Adsorption behavior of metal oxide supported gold catalysts
In order to investigate the support effects on the secondary
amine selectivity, we studied the adsorption behavior of benzyl
alcohol, aniline, and benzaldehyde onto the catalysts by DRIFT
spectra (Figs. 2–4). Benzyl alcohol was adsorbed on Au/CeO2 and
Au/ZrO2, but not on Au/WO3 judging from the adsorption peaks at
3060, 3024, 2954, and 2910 cm−1 which were ascribed to the C–H
stretching (Fig. 2). These results were in a good agreement with the
surface acidity of metal oxides and the catalytic activity order for
alcohol conversion.
On the aniline adsorption experiments, new adsorption peaks
were not observed for Au/CeO2 but for Au/ZrO2 and Au/WO3 (Fig. 3).
In Fig. 3b and c, the peaks at 3076 and 3020 cm−1 were assigned as
C–H stretching. The peaks at 1605 and 1496 cm−1 were assigned
as C C ring stretching [26–28].Vijayaraj et al. reported that N–H
stretching peaks of adsorbed aniline were higher wavenumber shift
(3228 and 3320 cm−1 from pure aniline at 3290 and 3355 cm−1
,
respectively) [28]. They also mentioned that the peak between
3500 and 3800 cm−1 was observed due to the interaction of polar-
ized NH2 with surface oxygen, leading to the formation of surface
OH groups [28]. The other report mentioned that a broad O–H
stretching band between 3600 and 3700 cm−1 indicated the OH/␲
interaction [26]. In Fig. 3, broad N–H stretching peaks at 3220–3330
and 3590–3630 cm−1 were observedfor Au/ZrO2 and ZrO2, suggest-
ing the adsorption of aniline on their surfaces.
Fig. 1. Relationships between the electronegativity of metal ions (ꢁi) in metal oxide
supports and the conversion of benzyl alcohol (a) and the selectivity to secondary
amine (b).

T. Ishida et al. / Applied Catalysis A: General 413–414 (2012) 261–266
265
Tanaka and Ogasawara reported that the peak was observed
at 2570 cm−1 (NH3 deformation) when aniline was adsorbed on
+
acidic sites [26]. They also proposed that the lower wavenum-
ber shift of N–H deformation band from 1630 cm−1 (pure aniline)
to 1575 cm−1 was due to aniline adsorbed on Lewis acid sites
or by hydrogen bonds [26]. In contrast, we did not observed
peaks at around 2570 cm−1 for all the catalysts. Instead, the N–H
deformation band was considered to be overlapped at 1605 cm−1
,
suggesting that aniline was not adsorbed on strong acid sites but
weakly adsorbed on Au/ZrO2 and on ZrO2 (Fig. 3b and c). These
results imply that the presence of strong acid sites is not neces-
sary since aniline can be weakly adsorbed on the Lewis acid sites
or by hydrogen bonds. In addition, the adsorption peaks of aniline
were also observed for ZrO2 without Au (Fig. 3c), indicating that
adsorption ability was come from metal oxide properties. Aniline
was adsorbed also on Au/WO3 judging from an adsorption peak
at 1492 cm−1 (C C ring stretching) although the peaks were very
weak (Fig. 3d).
Gold on CeO2 and Au/ZrO2 adsorbed benzaldehyde on their
surfaces (Fig. 4). Due to the lack of aniline adsorption ability
for Au/CeO2, benzaldehyde reacted with benzyl alcohol to form
hemiacetal followed by the dehydrogenation to produce benzyl
benzoate as a by-product (Scheme 2). Among these catalysts, only
Au/ZrO2, which could adsorb benzyl alcohol, aniline, benzaldehyde
on the surfaces, showed the highest catalytic activity and the selec-
tivity, indicating that the adsorption of benzaldehyde on Au/ZrO2
enabled the rapid condensation of aniline with aldehyde into imine
or hemiaminal intermediate on the Au surfaces.
Fig. 2. Difference DRIFT spectra of benzyl alcohol adsorbed on Au/CeO2 (a), Au/ZrO2
(b), and Au/WO3 (c).
4. Discussion
4.1. Support effects on the N-alkylation
To elucidate the reaction pathway for the N-alkylation, hydro-
gen transfer reaction from benzyl alcohol to imine over Au/ZrO2
was examined. The yield of sec. amine decreased to 17% as com-
pared to 46% for the N-alkylation (Table 1, entry 6), meaning that
imine was only slightly adsorbed on Au/ZrO2 surfaces. This was in
opposite to the case of polymer supported Au clusters [14]. In the
case of metal oxide supported Au catalysts, imine could not be eas-
ily converted to sec. amine once imine was formed in the reaction
solution without catalysts or diffused into the solution from the
catalyst surfaces. As Cao and co-workers [15] and Shimizu et al. [9]
have mentioned for the N-alkylation over Au/TiO2 and Ag/Al2O3,
respectively, it is likely that the hydrogen transfer would occur via
hemiaminal intermediates rather than imine on Au/ZrO2 surface.
Therefore, the aniline adsorption ability of metal oxides at the Lewis
acidic sites or by OH/␲-interactions would play an important role.
Fig. 3. Difference DRIFT spectra of aniline adsorbed on Au/CeO2 (a), Au/ZrO2 (b),
ZrO2 (c), and Au/WO3 (d).
4.2. Reaction pathways for N-alkylation
A plausible reaction pathway is shown in Scheme 3. Deproto-
nation of alcohol was promoted at the basic sites of ZrO2 to from
alkoxide on the Au surfaces. Gold catalyzes the ␤-hydride elimi-
nation to produce benzaldehyde. Aldehyde and aniline, which are
adsorbed on the ZrO2 surfaces beforehand, react to form hemiami-
nal. Before hemiaminal is converted to imine, hydride on the Au
NPs and proton on the surface OH groups on ZrO2 move to produce
sec. amine. Acidic OH groups might contribute as proton sources
and as aniline adsorption sites. However, taking into account that
Au NPs supported on basic CeO2 could also catalyze the dehydro-
genation of alcohol, protons could be supplied not only from acidic
sites but also other surface OH groups. Although the acidic sites of
ZrO2 would be involved in the catalytic reaction, acidity of metal
oxides was not a major factor to determine the hydrogen transfer
Fig. 4. Difference DRIFT spectra of benzaldehyde adsorbed on Au/CeO2 (a) and
Au/ZrO2 (b).

266
T. Ishida et al. / Applied Catalysis A: General 413–414 (2012) 261–266
Scheme 3. A possible reaction pathway for the N-alkylation of aniline with benzyl alcohol over Au/ZrO2.
efficiency. Furthermore, the addition of 2-propanol as a hydrogen
donor lowered the selectivity to sec. amine over Au/ZrO2 (Table 3,
entry 2). It is indicated that excess amount of alcohols inhibited the
adsorption of aniline onto ZrO2 surfaces, resulting the imine for-
mation in the reaction solution. Therefore, it can be concluded that
aniline adsorption ability of oxide supports plays an important role
to achieve high hydrogen transfer efficiency.
References
[1] R.N. Salvatore, C.H. Yoon, K.W. Jung, Tetrahedron 57 (2001) 7785–7811.
[2] A.J.A. Watson, J.M.J. Williams, Science 329 (2010) 635–636.
[3] Y. Watanabe, Y. Tsuji, H. Ige, Y. Ohsugi, T. Ohta, J. Org. Chem. 49 (1984) 3359.
[4] K.-I. Fujita, Y. Enoki, R. Yamaguchi, Tetrahedron 64 (2008) 1943–1954.
[5] B. Blank, S. Michlik, R. Kempe, Chem. Eur. J. 15 (2009) 379–3799.
[6] A. Tillack, D. Hollmann, D. Michalik, M. Beller, Tetrahedron Lett. 47 (2006)
8881–8885.
[7] M.H.S.A. Hamid, C.L. Allen, G.W. Lamb, A.C. Maxwell, H.C. Maytum, A.J.A. Wat-
son, J.M.J. Williams, J. Am. Chem. Soc. 131 (2009) 1766–1774.
[8] J.W. Kim, K. Yamaguchi, N. Mizuno, J. Catal. 263 (2009) 205–208.
[9] K. Shimizu, M. Nishimura, A. Satsuma, ChemCatChem 1 (2009) 497–503.
[10] X. Cui, Y. Zhang, F. Shi, Y. Deng, Chem. Eur. J. 17 (2011) 1021–1028.
[11] A. Corma, T. Ródenas, M.J. Sabater, Chem. Eur. J. 16 (2010) 254–260.
[12] J. He, K. Yamaguchi, N. Mizuno, Chem. Lett. 39 (2010) 1182–1183.
[13] M. Haruta, Chem. Rec. 3 (2003) 75–87.
[14] T. Ishida, N. Kawakita, T. Akita, M. Haruta, Gold Bull. 42 (2009) 267–274.
[15] L. He, X.-B. Lou, J. Ni, Y.-M. Liu, Y. Cao, H.-Y. He, K.-N. Fan, Chem. Eur. J. 16 (2010)
13965–13969.
[16] Q. Peng, Y. Zhang, F. Shi, Y. Deng, Chem. Commun. 47 (2011) 6476–6478.
[17] T. Ishida, N. Kinoshita, H. Okatsu, T. Akita, T. Takei, M. Haruta, Angew. Chem.
Int. Ed. 47 (2008) 9265–9268.
[18] T. Ishida, M. Nagaoka, T. Akita, M. Haruta, Chem. Eur. J. 14 (2008) 8456–8460.
[19] M. Haruta, N. Yamada, T. Kobayashi, S. Iijima, J. Catal. 115 (1989) 301–309.
[20] M. Haruta, S. Tsubota, T. Kobayashi, H. Kageyama, M.H. Genet, B. Delmon, J.
Catal. 144 (1993) 175–192.
[21] D.A.H. Canningham, W. Vogel, H. Kageyama, S. Tsubota, M. Haruta, J. Catal. 177
(1998) 1–10.
5. Conclusions
We have demonstrated that Au/ZrO2 works as an efficient cat-
alyst for the N-alkylation of aniline with benzyl alcohol to yield
secondary amine using equimolar amount of substrates without
additives under N2 at an atmospheric pressure. Although Au/CeO2
showed the highest catalytic activity for the dehydrogenation of
alcohol in the absence of aniline, the hydrogen transfer efficiency
was lower than that of Au/ZrO2 due to the lack of aniline adsorption
ability. It was revealed that not only basic and acidic properties of
metal oxide surfaces but also aniline adsorption ability by hydro-
gen bonding were important in the N-alkylation and that aniline
adsorption might influence on the hydrogen transfer efficiency to
give secondary amine in high yields.
[22] T. Ishida, M. Haruta, ChemSusChem 2 (2009) 538–541.
[23] K. Tanaka, A. Ozaki, J. Catal. 8 (1967) 1–7.
Acknowledgements
[24] A. Abad, A. Corma, H. García, Chem. Eur. J. 14 (2008) 212–222.
[25] W. Fang, J. Chen, Q. Zhang, W. Deng, Y. Wang, Chem. Eur. J. 17 (2011) 1247–1256.
[26] M. Tanaka, S. Ogasawara, J. Catal. 25 (1972) 111–117.
[27] R. Sokoll, H. Hobert, J. Catal. 125 (1990) 285–291.
[28] M. Vijayaraj, B. Murugan, S. Umbarkar, S.G. Hegde, C.S. Gopinath, J. Mol. Catal.
A: Chem. 231 (2005) 169–180.
We thank Mr. Y. Misaki of Tokyo Metropolitan University for
helping us with TEM observations. This work was financially sup-
ported by JST-CREST, a Grant-in-Aid for Young Scientists (B) (no.
21750160) from the Ministry of Education, Culture, Sports, Science,
and Technology, Japan.