Influence of the support and the size of gold clusters on catalytic activity for glucose oxidation

Article abstract of DOI:10.1002/anie.200802845

Not all that glitters... The activity of supported gold nanoparticles depends on the method used for their preparation. Gold clusters of about 2 nm in diameter were deposited on nonreducible metal oxides and carbon materials by solid grinding of a volatile organogold complex in a ball mill and subsequent calcination (see scheme). Au/ZrO₂ and Au/Al₂O₃ prepared in this way were extremely efficient catalysts for the aerobic oxidation of glucose. (Chemical Equation Presented).

Full text of DOI:10.1002/anie.200802845

Angewandte
Chemie
DOI: 10.1002/anie.200802845
Gold Nanoparticles
Influence of the Support and the Size of Gold Clusters on Catalytic
Activity for Glucose Oxidation**
Tamao Ishida, Naoto Kinoshita, Hiroko Okatsu, Tomoki Akita, Takashi Takei, and
Masatake Haruta*
Dedicated to the Catalysis Society of Japan on the occasion of its 50th anniversary
Gold was neglected as a catalytic metal in the history of
heterogeneous catalysis for more than 100 years. However,
when gold is deposited as nanoparticles (NPs) on basic
transition-metal oxides, it exhibits unique catalytic perfor-
mance, for example, the catalysis of the oxidation of carbon
The catalytic transformation of biomass-derived natural
resources into valuable compounds is of great importance for
sustainable developments. In particular, the aerobic oxida-
tion of glucose to gluconic acid over transition-metal catalysts,
[
6]
including Pd, Pt, and Au, has been studied during the last two
[
1]
[4,7–12]
monoxide at a temperature as low as ꢀ708C.
decades.
Gold catalysts for this transformation have
[
4,9–12]
As conventional impregnation does not lead to the high
dispersion of small gold NPs, several methods for catalyst
been studied extensively,
as Rossi and co-workers
reported that gold NPs supported on activated carbon (AC)
showed high catalytic activity and high selectivity in the
[
1]
preparation have been developed. One of the most fre-
quently used techniques is the deposition–precipitation (DP)
[8]
oxidation of glucose to gluconic acid in 2002. The reaction
[
9a]
method, in which Au(OH) embryos generated from HAuCl4
conditions, the reaction mechanism on Au/AC
colloids,
and Au
3
[
10]
[9b]
by treatment with NaOH are precipitated and deposited
exclusively on the surfaces of basic metal oxides; these
deposits are then transformed into gold NPs with diameters
the effect of the size of the gold particles,
reproducibility and the durability of Au catalysts, and the
preparation of the catalyst have been investigated inten-
sively. However, these approaches are limited to unsupported
Au colloids and Au NPs supported on AC, TiO , and Al O .
[
11]
[
4]
[
1,2]
smaller than 5 nm by calcination.
preparation methods, such as DP urea, in which urea is used
Modified catalyst-
2
2
3
[
3]
instead of NaOH as the precipitating agent, and the
Recently, we developed a new preparation method, solid
grinding (SG), by using a volatile organogold complex,
[
4]
incipient wetness method have also been used. However,
these methods are not applicable to supports that have a low
point of zero charge, such as acidic metal oxides (e.g. SiO2-
Al O ), carbon materials, and organic polymers. Accordingly,
[Me Au(acac)] (acac = acetylacetonate), and succeeded in
2
depositing Au clusters smaller than 2 nm in diameter onto
[
13]
porous coordination polymers. We deposited Au clusters
onto several kinds of metal oxides and carbon supports by this
SG method. Herein we report on the effects of the support
and the size of the particles on glucose oxidation over various
supported Au catalysts, which were prepared by two methods,
DP and SG.
2
3
gold colloids prepared beforehand were immobilized by
mixing with carbon to produce carbon-supported gold
[
5]
catalysts. Therefore, other new versatile methods are
required for the direct deposition of Au clusters from Au
precursor compounds onto various kinds of support materials.
The size of Au particles depended strongly on the
preparation method. The deposition–precipitation method
[
*] Dr. T. Ishida, N. Kinoshita, H. Okatsu, Prof. Dr. T. Takei,
Prof. Dr. M. Haruta
Graduate School of Urban Environmental Sciences
Tokyo Metropolitan University (TMU)
could stabilize Au NPs on ZrO (with a mean diameter of
2
3
.7 nm), TiO (2.9 nm), and CeO (4.0 nm). However, large
2 2
Au NPs were formed simultaneously with small Au NPs on
Al O , so that a mean diameter of 4.3 nm was observed.
2
3
1-1 Minami-osawa, Hachioji, Tokyo 192-0397 (Japan)
In the SG method, [Me Au(acac)] and the carbon
Fax: (+81)42-677-2851
2
E-mail: haruta-masatake@center.tmu.ac.jp
supports AC, which is microporous, and nanoporous carbon
NPC), which has 2 nm mesopores and hollow cores, were
(
Dr. T. Ishida, N. Kinoshita, H. Okatsu, Dr. T. Akita, Prof. Dr. T. Takei,
Prof. Dr. M. Haruta
Japan Science and Technology Agency (JST), CREST
ground in an agate mortar and in a ball mill (300 rpm),
respectively, in air at room temperature for 20 to 30 min and
then subjected to calcination at 3008C for 4 h. Solid grinding
gave Au clusters and small NPs deposited on the carbon
supports, especially on NPC. The mean diameters of the Au
particles on NPC were calculated by HAADF STEM (high-
angle annular dark-field scanning transmission electron
microscopy) and TEM to be 1.9 and 2.6 nm, respectively
4-1-8 Hon-cho, Kawaguchi, Saitama 332-0012 (Japan)
Dr. T. Akita
Research Institute for Ubiquitous Energy Devices, National Institute
of Advanced Industrial Science and Technology (AIST)
1-8-31 Midorigaoka, Ikeda, Osaka 563-8577 (Japan)
[
**] We thank Y. Misaki of TMU for TEM observation and T. Hara of Fuso
Chemical Co. for HPLC analysis. This research was supported
financially by JST, CREST, and the Gold Research Opportunities
Worldwide (GROW) program of the World Gold Council.
(Figure 1a,b).
During the optimization of the grinding conditions, we
found that ball milling (350 rpm for 1 h) was favorable for
metal oxides in contrast to organic polymers and carbon
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.200802845.
Angew. Chem. Int. Ed. 2008, 47, 9265 –9268
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
9265

Communications
III
ure 2a). Some fractions of the Au precursors adsorbed on
TiO and CeO , however, might be reduced by the support at
2
2
3
+
the M cation radical or oxygen radical sites on the surfaces
during ball milling in air (steps 2 and 3 in Figure 2b). This
hypothesis would explain the observed color change from
0
white to mauve. Once Au seed crystals had formed on the
III
supports, they enhanced the reduction of Au on the surface
of the gold seed crystals and thus caused the growth of large
Au NPs. Although the size of the Au particles formed by SG
III
and DP was similar for CeO , a smaller amount of Au
2
precursors could be adsorbed onto CeO by SG than the DP.
2
The catalytic performance of the supported gold nano-
particles was first tested for the oxidation of CO in the gas
phase (Figure 3). Solid grinding was advantageous over DP
for the preparation of Au/Al O and Au/ZrO owing to the
2
3
2
smaller size of the Au particles. However, the nature of the
support affected the catalytic activity more strongly. Lower
catalytic activity was observed with nonreducible Al O than
2
3
with reducible metal oxides, such as ZrO , TiO , and CeO ,
2
2
2
which can form oxygen-vacancy sites at the perimeter
interfaces of Au particles. Gold NPs exhibited no catalytic
activity at all at temperatures up to 1208C when they were
deposited on carbon materials and polymers. Similarly, no
catalytic activity was observed for gold NPs on carbon
supports and polymers in the gas-phase oxidation of H2.
Figure 1. HAADF STEM images of a) Au/NPC (SG) and c) Au/Al O
2
3
(
SG), and the size distribution of Au particles on b) NPC and d) Al O .
2 3
Orange and green bars indicate data obtained from HAADF STEM and
[
14]
TEM images, respectively.
materials. It is likely that harder metal-oxide
particles needed stronger mixing than soft
carbon materials. Uniform grinding partially
failed in a ball mill, in which large Au NPs were
formed simultaneously with Au clusters in the
completely solid state. Therefore, acetone was
used for uniform milling to prevent the aggrega-
tion of Au particles. The mean diameter of Au
particles on Al O was calculated to be 2.6 nm by
2
3
TEM and HAADF STEM (Figure 1c,d).
Guzman and Gates reported the reaction of
[
Me Au(acac)] with Al O upon mixing as slurry
2 2 3
[
15]
in hexane. Mixing had to be carried out under a
completely inert atmosphere owing to the insta-
bility of [Me Au(acac)], and the reaction took one
2
day under these conditions. In contrast, milling
enables the reaction time to be shortened by
enhancing the reaction in air. [Me Au(acac)] was
2
not reduced on nonconductive Al O , ZrO , and
2
3
2
carbon materials during grinding.
Solid grinding, both in the completely solid
state and in an acetone slurry, was found to be an
efficient technique for the preparation of Au
clusters on nonconductive Al O , ZrO , and
2
3
2
carbon supports, but unsuitable for semiconduc-
tive supports, in particular, TiO . Guzman and
2
Gates also investigated the mechanism of the
ligand exchange of [Me Au(acac)] with the sur-
2
face hydroxy groups of Al O (step 1 in Fig-
2
3
[
15]
III
ure 2a). In our experiments, the Au precursors
adsorbed on Al O as a result of ball milling were
2
3
stable in air for 1 h. They were then reduced to
form Au clusters by calcination (step 2 in Fig-
Figure 2. Probable pathways for the reaction of [Me Au(acac)] with a) Al O and
b) TiO to form Au particles.
2
2
2
3
0
9
266 www.angewandte.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2008, 47, 9265 –9268

Angewandte
Chemie
Comotti et al. reported that the catalytic activity of Au
NPs for glucose oxidation was detectable only when Au was
smaller than 10 nm and that it was inversely proportional to
[9b]
the diameter of the Au particles in the range of 3–6 nm.
This result indicates that the catalytic activity per surface Au
atom is independent of the diameter of the Au particles. In
contrast, metal-oxide-supported Au catalysts in this study
showed an increase in TOF with a decrease in the diameter of
the Au particles. Although the size dependencies of the TOF
values in our study and that of Commoti et al. were different,
the size of the Au particles appears to be the most important
parameter. Glucose oxidation in the liquid phase may occur at
the Au surfaces rather than at the perimeter interfaces
between Au particles and the supports, in contrast to the gas-
Figure 3. Comparison of the catalytic activities of metal-oxide-sup-
ported Au NPs prepared by SG and DP. T1/2 is the temperature for
5
0% conversion in the gas-phase oxidation of CO. Reaction condi-
[
1]
phase CO oxidation.
The rate of glucose oxidation is expressed by Equa-
tion (1):.
tions: CO (1 vol%) in air, Au catalyst (150 mg), space velocity:
ꢀ
1
ꢀ1
2
0000 mLgcat h .
Next, the oxidation of glucose was studied by bubbling
oxygen through a 5 wt% aqueous solution of glucose at a flow
a
b
0
0
b
r ¼ k½glucoseꢁ ½O ꢁ ¼ k ½glucoseꢁ ðk ¼ k½O ꢁ Þ
ð1Þ
2
2
ꢀ1
rate of 120 mLmin at 508C and pH 9.0 in the presence of Au
catalysts under ambient pressure. In contrast to CO oxidation,
the catalytic activity observed for glucose oxidation over Au
catalysts was influenced more significantly by the size of the
Au particles than by the nature of the support, as indicated by
the turnover frequency (TOF) per surface Au atom
The reaction was estimated to be 0.4 order with respect to
glucose for Au/Al and 1.0 order for Au/NPC in the
presence of excess oxygen. Beltrame et al. reported zero-
and first-order dependence on glucose and oxygen, respec-
O
3
2
[14]
[
10]
tively, for unsupported Au colloids.
However, at lower
(
Figure 4). The selectivity for the formation of gluconic acid
glucose concentrations, when oxygen was present in excess
relative to glucose, the reaction order with respect to glucose
was always very high at above 98%.
[
10]
became close to 0.4, the value estimated for the reaction
with our Au/Al O catalyst.
2
3
We derived Arrhenius plots for the temperature range
[
14]
from 40 to 708C. The catalytic activity of Au/Al O was
2
3
comparable to that of Au/ZrO (DP) and slightly higher than
2
that of Au/carbon at 708C. However, Au/Al O maintained a
2
3
higher catalytic activity than that of Au/ZrO (DP) at lower
2
temperature. The apparent activation energies were calcu-
ꢀ
1
lated to be 27, 53, 72, and 96 kJmol for Au supported on
Al O (SG), ZrO (DP), NPC (SG), and AC (SG), respec-
2
3
2
tively. The apparent activation energy for glucose oxidation
over unsupported Au colloids was reported to be
ꢀ
1 [10]
4
7 kJmol . The activation energies differed according to
the nature of the support and tended to be lower for Au on
metal oxides. The catalytic activity of Au/carbon was
depressed appreciably at temperatures below 508C. It can
be assumed that carbon materials participate in only weak or
Figure 4. Turnover frequencies for glucose oxidation (glucose reacted
(
mol) per number of surface Au atoms (mol) per second) versus the
diameter of the Au particles. Filled and open symbols indicate samples
prepared by SG and DP, respectively. Reaction conditions: 5 wt%
[9b]
no metal–support interactions
and adsorb glucose to a
ꢀ1
glucose solution (175 mL), Au catalyst (30 mg), O (120 mLmin ),
certain extent owing to the large specific surface area. These
features may explain the stronger dependence of the reaction
rate on glucose concentration and the reaction temperature
with carbon supports than with metal-oxide supports. In
contrast, metal-oxide supports might affect Au clusters
electronically; such effects may lead to the weaker rate
dependencies than those observed with Au/carbon and
unsupported Au colloids.
In summary, we have explored simple but efficient
methods for the preparation of small Au NPs on a variety
of support materials. In particular, the solid grinding (SG) of a
volatile organogold complex with a nonreducible support can
lead to the formation of Au clusters that exhibit excellent
2
5
08C, pH 9.0.
The highest catalytic activity was observed for
Au/ZrO2 (SG), with
a
turnover frequency of
ꢀ
1
ꢀ1
4
5 molglucose molsurface Au
s
at 508C and pH 9.0. The highest
ꢀ
1
TOF per surface Au atom reported to date was 42 s at 508C
[
9d]
and pH 9.5.
As the reaction rate of glucose oxidation
[
8]
increases with an increase in pH, the TOF of Au/ZrO (SG)
2
ꢀ
1
reached 56 s at 508C and pH 9.5. To the best of our
knowledge, this value is the highest reported to date for
glucose oxidation catalyzed by gold NPs.
Angew. Chem. Int. Ed. 2008, 47, 9265 –9268
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 9267

Communications
catalytic performance in the oxidation of glucose. We have
reached the following conclusions:
[
1] a) M. Haruta, Chem. Rec. 2003, 3, 75 – 87; b) M. Haruta, J. New
Mater. Electrochem. Syst. 2004, 7, 163 – 172.
2] Catalysis by Gold (Eds.: G. C. Bond, C. Louis, D. T. Thompson),
Imperial College Press, London, 2006.
1
) The catalytic mechanism of Au seems to be completely
different for gas-phase and liquid-phase oxidation reac-
[
tions. As far as the gas-phase oxidation of H and CO and
2
[3] a) R. Zanella, L. Delannoy, C. Louis, Appl. Catal. A 2005, 291,
62 – 72; b) R. Zanella, S. Giorgio, C. R. Henry, C. Louis, J. Phys.
Chem. B 2002, 106, 7634 – 7642.
the liquid-phase oxidation of glucose are concerned, the
appropriate selection of the support material is more
important for the gas-phase reactions, whereas the control
of the size of Au particles is more critical in the liquid
phase.
[
4] a) C. Baatz, N. Thielecke, U. Prꢀsse, Appl. Catal. B 2007, 70,
53 – 660; b) C. Baatz, U. Prꢀsse, J. Catal. 2007, 249, 34 – 40.
6
[
[
5] L. Prati, G. Martra, Gold Bull. 1999, 32, 96 – 101.
6] J. N. Chheda, G. W. Huber, J. A. Dumesic, Angew. Chem. 2007,
2
3
) For glucose oxidation, the influence of the support and the
size of Au NPs was discussed comprehensively for a
119, 7298 – 7318; Angew. Chem. Int. Ed. 2007, 46, 7164 – 7183.
[7] a) A. Abbadi, H. van Bekkum, J. Mol. Catal. A 1995, 97, 111 –
118; b) S. Hermans, M. Devillers, Appl. Catal. A 2002, 235, 253 –
variety of support materials. Gold on ZrO (SG) exhibited
2
264; c) M. Wenkin, P. Ruiz, B. Delmon, M. Devillers, J. Mol.
extremely high catalytic activity with TOF values per
ꢀ1
ꢀ1
Catal. A 2002, 180, 141 – 159.
surface Au atom of 45 s at 508C, pH 9.0 and 56 s at
08C, pH 9.5.
[
[
8] S. Biella, L. Prati, M. Rossi, J. Catal. 2002, 206, 242 – 247.
9] a) M. Comotti, C. Della Pina, E. Falletta, M. Rossi, Adv. Synth.
Catal. 2006, 348, 313 – 316; b) M. Comotti, C. Della Pina, R.
Matarrese, M. Rossi, Angew. Chem. 2004, 116, 5936 – 5939;
Angew. Chem. Int. Ed. 2004, 43, 5812 – 5815; c) M. Comotti, C.
Della Pina, R. Matarrese, M. Rossi, A. Siani, Appl. Catal. A
5
) The most suitable method for Au deposition depends on
the support.
2
005, 291, 204 – 209; d) M. Comotti, C. Della Pina, E. Falletta,
M. Rossi, J. Catal. 2006, 244, 122 – 125.
10] P. Beltrame, M. Comotti, C. Della Pina, M. Rossi, Appl. Catal. A
006, 297, 1 – 7.
11] a) N. Thielecke, M. Aytemir, U. Prꢀsse, Catal. Today 2007, 121,
Experimental Section
[
[
Solid grinding: A metal oxide (3.0 g), [Me Au(acac)] (50 mg for
2
2
1
wt% loading of Au), and acetone (10 g for Al O ) were ground by
2 3
ball milling (350 rpm) at room temperature for 1 h. The product was
reduced by calcination in air at 3008C for 4 h. Grinding of carbon
materials was performed without solvents. The grinding conditions
were 300 rpm for 300 min in a ball mill (NPC) and for 30 min in an
agate mortar (AC), respectively.
115 – 120; b) N. Thielecke, K.-D. Vorlop, U. Prꢀsse, Catal. Today
2007, 122, 266 – 269; c) A. Mirescu, H. Berndt, A. Martin, U.
Prꢀsse, Appl. Catal. A 2007, 317, 204 – 209.
[
[
12] a) Y. ꢁnal, S. Schimpf, P. Claus, J. Catal. 2004, 223, 122 – 133.
13] T. Ishida, M. Nagaoka, T. Akita, M. Haruta, Chem. Eur. J. 2008,
1
4, 8459 – 8460.
Received: June 16, 2008
Published online: October 10, 2008
[
[
14] See the Supporting Information.
15] J. Guzman, B. C. Gates, Langmuir 2003, 19, 3897 – 3903.
Keywords: gold · heterogeneous catalysis · nanoparticles ·
.
oxidation · solid grinding
9
268 www.angewandte.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2008, 47, 9265 –9268