Journal of New Materials for Electrochemical Systems 7, 163-172 (2004)
© J. New. Mat. Electrochem. Systems
Nanoparticulate Gold Catalysts for Low-Temperature CO Oxidation
M. Haruta
Research Institute for Green Technology, National Institute of Advanced Industrial Science and Technology (AIST)
16-1 Onogawa, Tsukuba 305-8569, Japan
(Received November 27, 2003; received revised form December 29, 2003)
Abstract: Gold can be deposited as nanoparticles on a variety of support materials by coprecipitation, deposition-precipitation of Au(OH)3 ,
grafting of organo-gold complexes such as dimethyl-Au(III)-acetylacetonate, mixing of colloidal Au particles, and vacuum deposition. The unique
and practically useful catalytic performance of Au emerges by the strong contact with the support, the selection of support materials, and the size
control of Au particles. The oxidation of CO can take place even at a temperature as low as 200 K, owing to the moderate adsorption of CO on the
edges and corners of Au nanoparticles and to the activation of O2 , probably at the perimeter interface with the supports. However, an exception is
that acidic supports such as Al2O3-SiO2 and activated carbon do not lead to low temperature CO oxidation, which imposes an essential scientific
question related to polymer electrolyte fuel cells. A comprehensive mechanism for low-temperature CO oxidation is presented that can account for
the whole catalytic behavior of supported gold catalysts : an increase in TOF with a decrease in the diameter of Au particles, large difference in
the enhancing effect of moisture among support metal oxides, zeroth order kinetics at concentrations of CO and O2 above 0.1 vol%, and low
apparent activation energies.
Key words: gold, nanoparticles, CO oxidation, catalyst preparation, structure sensitivity
over frequencies, TOFs) are almost independent of the size and
shape of Au particles [13]. In contrast, other reactions, typically
CO oxidation and propylene epoxidation are markedly structure
sensitive. Both the rate and selectivity are defined by the three
major factors: contact structure of Au particles (the length of
perimeter interface), selection of the support, and the size of Au
particles. This means that catalyst preparation is crucially
important for the catalysis by Au nanoparticles. Accordingly, the
first attempt in this paper is to make a comprehensive overview of
the techniques currently available for preparing nanoparticulate
Au catalysts.
1. INTRODUCTION
Gold was regarded to be catalytically inert, however, it is now
well known that Au is about 1000 times more active than Pt in the
electrochemical and catalytic oxidation of CO under basic but not
acidic encironments[1-7]. Gold can also promote many other
reactions when stabilized in the form of nanoparticles attached to
metal oxide and activated carbon supports [8-12]. Since Au/Fe2O3
supported on zeolite-coated paper honeycomb was commercially
used for an odor eater in modern Japanese rest rooms in 1992,
R&D activity on the catalysis of Au has been rapidly growing in
the world, expanding the frontiers of sciences and applications of
the noblest metal of Au [9].
The oxidation of CO is one of the simplest reactions, while it
covers a wide range of applications from gas masks, gas sensors,
indoor air quality control to hydrogen purification for polymer
electrolyte fuel cells. The reaction attracts renewed interest both
in fundamental [14] and applied research of catalysis and
electrochemistry [1-3,7,15]. As summarized in Table 1, supported
Au catalysts are superior to both the other supported noble metal
Over supported nanoparticulate Au catalysts, H2 oxidation and
hydrogenation reactions of unsaturated hydrocarbons are structure
insensitive and the rates per surface exposed metal atom (turn
*To whom correspondence should be addressed: Fax:+81-29-861-8458;
E-mail: [email protected]
163
164
M. Haruta / J. New. Mat. Electrochem. Systems 7, 163-172 (2004)
catalysts and base metal oxides in room temperature catalytic
activity and moisture enhancement. In this context, the second
attempt in this paper is to deliver answers to the general questions
why the combination of inactive components for the adsorption of
CO and O2, Au with Al2O3 or TiO2, for example, can bring
surprisingly high catalytic activity for CO oxidation [16] and why
moisture enhances the catalytic activity [17].
calcination of HAuCl4 crystallites dispersed on the support
surfaces. Table 2 lists four categories of techniques that can
deposit Au nanoparticles with diameters below 10 nm [22]. A
suitable technique should be chosen depending on the kind of
support materials such as basic or acidic metal oxides,
carbonaceous materials, or single crystals.
The discovery of highly active Au catalysts was led by a simple
assumption based on a volcano-like relation between the
catalytic activity of metal oxides for the oxidation and the heat
of metal oxide formation per one oxygen atom which
corresponds to the metal-oxygen (M-O) bonding strength
[11,18]. It had been expected that the combination of metals
located at the opposite sides of the volcano plots, Ag or Au
which has weaker M-O bonding than PtO2 with Mn, Fe, Co, Ni
which have stronger bonding, might bring about complex
oxides as active as PtO2 which has optimum M-O bonding
strength. In fact, some gold composite oxides (later they were
found to be Au nanoparitcles supported on transition metal
oxides) were much more active than PtO2 in CO oxidation.
Jaksic has intensively studied the nature, causes, and
consequences of volcano plots of transition metals for hydrogen
electrode reactions in an attempt to explore optimum synergetic
intermetallic systems [19-21]. He has also extended his work to
metal-oxide electrocatalytic systems for simultaneous anodic
hydrogen and CO oxidation [15]. These achievements will
surely flourish in the near future in providing novel electrode
catalysts in fuel cell applications, as in the case of
nanoparticulate Au catalysts. This article, devoted to the 70th
anniversary of Professor Jaksic, therefore, will also convey a
Table 2. Preparation techniques for nanoparticulate Au catalysts.
Table 1. Comparison of performances of three classes of catalysts for CO
oxidation at room temperature.
Catalyst
Activity
Moisture
Nanoparticulate gold
Excellent
Activated
Other noble metals
Poor
Activated
Base metal oxides
Good
Deactivated
comprehensive overview of the requirements and the working
mechanism of Au catalysts.
2. PREPARATION OF NANOPARTICULATE GOLD
CATALYSTS
It is difficult to deposit Au as nanoparticles on metal oxides by the
impregnation method (IMP) mainly because Au has lower melting
point (Au: 1336 K, Pd: 1823 K, Pt: 2042 K) and much lower
affinity to oxygen than Pd and Pt. Another reason is that chloride
ion markedly enhances the coagulation of Au particles during
Categories
Preparation of mixed
precursors of Au and
the metal component
of supports
Strong interaction of
Au precursors with
support materials
Mixing colloidal Au
with support materials
Model catalysts using
single crystal
Preparation techniques
Support materials
Ref.
Coprecipitation
(hydroxides or
carbonates) CP
Be(OH)2, TiO2*, Mn2O3,
Fe2O3, Co3O4, NiO, ZnO,
In2O3, SnO2
4,
23-26
co-sputtering (oxides)
in the presence of
O2 CS
amorphous alloy
(metals) AA
Co3O4
27
ZrO2
28
deposition-precipitation Mg(OH)2*, Al2O3, TiO2,
(aqueous HAuCl4
Fe2O3, Co3O4, NiO, ZnO,
solution) DP
ZrO2, CeO2, Ti-SiO2
29-33
liquid phase grafting
(organogold complex in
organic solvents) LG
TiO2, MnOx, Fe2O3
34, 35
gas phase grafting
(organogold complex)
GG
all kinds, including
SiO2, Al2O3-SiO2,
and activated C
36, 37
colloid mixing CM
TiO2, activated C
12, 38,
39
40-46
vacuum deposition VD Defects are the sites for
(at low temperature)
deposition. MgO, Si, TiO2
The addition of Mg citrate during or after coprecipitation or deposition-precipitation is
indispensable for depositing Au as nanoparticles.
The first category is characterized by the preparation of wellmixed precursors, for example, hydroxide and/or carbonate,
oxide, or metal mixtures of Au with the metal component of the
support by coprecipitation (CP)[4,23-26], co-sputtering by Ar
containing O2 (CS)[27], or amorphous alloying (AA)[28],
respectively. These precursor mixtures are calcined in air at
temperatures above 550K. Since Au tends to become metallic
while the support metal forms crystalline oxide phase, Au is
gradually excluded to the surfaces of the support to form metallic
Au particles strongly attached to the crystalline metal oxides such
as α-Fe2O3, Co3O4, and ZrO2.
Among the three techniques, coprecipitation is the most useful
and simplest way of preparation. An aqueous solution of
HAuCl4•4H2O and water soluble metal salts, most preferably
nitrates, such as Fe(NO3)2•9H2O is poured into an aqueous
alkaline solution under agitation in a few minutes. After aging for
1 hour, the precipitate is washed with water for more than 5 times
until the pH of the supernatant reaches a steady value around 7
and then filtrated. This procedure makes the coprecipitation
Nanoparticulate Gold Catalysts for Low-Temperature ... / J. New. Mat. Electrochem. Systems 7, 163-172 (2004)
method different from the impregnation method in that the
catalyst precursor is almost free from Na and Cl ions (several tens
ppm). The removal of Cl ions is indispensable for preventing Au
nanoparticles from coagulation. The hydroxide and/or carbonate
mixture is dried overnight and calcined in air to obtain powder
catalysts. The following conditions are important to obtain
homogeneous dispersion of Au nanoparticles.
165
20nm
1) Concentration of metal salt solution : 0.1~0.4M/l.
2) Neutralizer : Na2CO3 or K2CO3
Since gold hydroxide, Au(OH)3, is amphoteric, its solubility
increases due to the formation of Au(OH)4- anion when pH is too
high. Therefore, precipitation is the most efficient in the pH range
of 7-10. When metal salt solution is added to Na2CO3 or K2CO3
solution within a few minutes, precipitate can be formed at a
relatively constant pH in the range of 8-9, resulting in the
homogeneous distribution of Au nanoparticles and thus better
reproducibility of catalytic performance. When NaOH solution is
used, the pH widely changes during coprecipitation. The use of
NH4OH solution usually results in the formation of Au particles
larger than 10 nm in diameter.
3) Temperature : 320-360K for precipitation and 550-670K for
calcination
Aqueous solutions for coprecipitation should be warmed to a
temperature in the range of 320-360K in order to promote the
exchange of chloride in AuCl4- ion with OH to transform into
Au(Cl)4-n(OH)n- . To prepare Au/Co3O4, coprecipitation should be
carried out in the temperature range of 270-300K so as to depress
the reduction of AuCl4- ion with the oxidation of Co2+ ion to Co3+
ion. The catalytic activity for low-temperature CO oxidation of
coprecipitates reaches a maximum when they are calcined at a
temperature around 570K, where hydroxide species of Au mostly
changes to metallic Au particles [25]. Above 670K the
coagulation of Au nanoparticles becomes accelerated.
Figure 1. TEM photograh for Au/α-Fe2O3 prepared by coprecipitation
followed by calcinationi at 673K. Au/Fe=1/19 (atomic ratio).
adjusted at a fixed point in the range of 6 to 10, and is selected
primarily based on the isoelectric points (IEP) of the metal oxide
supports shown in Figure 2. As for the neutralizer, NaOH or KOH
is preferable to Na2CO3 or K2CO3. This is probably because
hydroxides can adjust the pH of HAuCl4 solution with smaller
amount than carbonates and accordingly bring weaker ionic
strength of the solution. As in the case of coprecipitation, NH4OH
or urea is, in principle, not recommended because of the
formation of larger Au particles with wider size distribution. It has
recently been reported that aging for 16 hours leads to a high
loading of about 8 wt% of small Au nanoparticles over TiO2 when
homogeneous precipitation using urea as a neutralizer is applied
[32]. Another example is Au/θ-Al2O3 prepared by ion exchange
with AuCl4- in acidic aqueous solution, followed by washing and
by neutralization with ammonia aqueous solution [33].
Figure 1 shows a TEM photograph for Au/α-Fe2O3 prepared by
coprecipitation. Gold nanoparticles are homogeneously dispersed
on α-Fe2O3 particles, with a standard deviation of the diameter of
about 30%. The applicability of coprecipitation is limited to metal
hydroxides or carbonates that can be coprecipitated with
Au(OH)3. Actually, Au can be supported in the form of welldispersed nanoparticles on α-Fe2O3, Co3O4, NiO, and ZnO while
not on TiO2, Cr2O3, MnOx, and CdO [24]. In the case of TiO2, the
addition of Mg citrate during or after coprecipitation is necessary
to obtain good dispersion of Au nanoparticles [29], while in the
case of Mn2O3 precipitation in aqueous LiCO3 solution leads to
better catalytic activity for selective CO oxidation in H2
stream[26].
The second category is based on the deposition or adsorption of
Au compounds, for example, Au hydroxide by depositionprecipitation (DP)[29-33] or organogold complex by liquid phase
grafting (LG) [34,35] and by gas phase grafting (GG)[36,37]. The
DP method is the easiest to handle and is used now for producing
commercial Au catalysts. The pH of aqueous HAuCl4 solution is
Figure 2. Isoelectric points for various metal oxides.
166
Careful control of the concentration around 10-3M/l, pH in the
range of 6 to 10, and temperature in the range of 320K to 360K of
the aqueous HAuCl4 solution enables selective deposition of
Au(OH)3 only on the surfaces of support metal oxides without
precipitation in the liquid phase. It is easier to deposit Au(OH)3 on
metal oxide supports if their specific surface areas are larger than
10 m2/g. Because the precursor can be washed before drying, Na
and Cl ions are removed to a level of a few tens ppm, as in the
case of coprecipitation. The only drawback of DP is that it is not
applicable to metal oxides, the IEPs of which are below 5 (see
Fig. 2), and to activated carbon. Gold hydroxide cannot be
deposited on SiO2 (IEP=2), SiO2-Al2O3 (IEP=1), or WO3 (IEP=1).
Gold deposited on the hydroxides of alkaline earth metals exhibits
the highest catalytic activity for CO oxidation at 200 K only when
Au is small clusters with diameters below 1.0 nm [31].
Magnesium citrate is indispensable to prepare such tiny Au
clusters. When BeO or MgO powder is dispersed in an aqueous
solution of HAuCl4 at 343 K, the oxide is transformed into
hydroxide and is partly dissolved to change the pH of the solution
toward 10. At this pH and at 343 K, an aqueous solution of
magnesium citrate is added in an amount of about 6 times as the
molar of Au. After aging at 343 K for 1h, the solid precursor is
washed for several times, dried overnight, and is calcined in air at
a temperature ranging from 523 K to 553 K. When the precursor
is calcined at temperatures above 553 K, the hydroxide support is
transformed into crystalline metal oxide, being accompanied by
the formation of large Au particles. This causes the loss of
catalytic activity for CO oxidation even at 473 K.
The role of magnesium citrate is schematically drawn in Figure 3.
Since the concentration of dissociated free citrate anion is
automatically controlled at pH=10 by the dissociation
equilibrium, the reduction of AuCln(OH)4-n- (“n” depends on pH)
into metallic Au can be avoided. In contrast, the reduction of Au
anions proceeds in Na citrate solution due to its complete
dissociation. Because citrate anion is trivalent anion, it can adsorb
M. Haruta / J. New. Mat. Electrochem. Systems 7, 163-172 (2004)
strongly on metal oxide surfaces, probably surrounding the Au
hydroxide precipitates. Through DTA and TG it has been proved
that during calcination the thermal decomposition of Au
hydroxide takes place earlier than the desorption and combustion
of citrate anion. This suggests that citrate anion might act as a
barrier which prevents Au clusters and particles from coagulation.
Liquid phase grafting (LG) using organo-gold complexes such as
phenylphoshine-gold nitrate in organic solvent can produce Au/
MnOx catalyst active for CO oxidation [34]. It has also the same
limitation as that of DP in its applicability to the type of metal
oxide supports [35]. Furthermore, it needs freshly prepared metal
hydroxides [34] probably because the concentration of surface
OH groups in organic solvent should be high enough for the
interaction with the organo-gold complex. In contrast, gas phase
grafting (GG) using dimethyl-Au(III)-aceytylacetonate is unique
in that it can deposit Au nanoparticles on almost all kinds of
supports, for example, SiO2 and SiO2-Al2O3 , and activated
carbon [36,37]. Before introducing the vapor of organic gold
complex, metal oxide support is usually evacuated at 523 K
followed by oxygen treatment at the same temperature to remove
hydrocarbon contaminants over the surfaces. This pretreatment
improves the reproducibility of Au deposition. Dimethy-Au(III)acetylacetonate has a planar structure and its adsorption structure
was estimated by DFT (Density Functional Theory) calculations
to have a chemical interaction with OH groups of metal oxide
surfaces at its oxygen atoms in the acetylacetonate ring. The
horizontal adsorption might lead to the better contact of Au
nanoparticles with the support and the homogeneous distribution
of the gold complex over the metal oxide surfaces,.
The third category is to use monodispersed Au colloids stabilized
by organic ligands or polymer compounds [12,38-39]. Among the
former six techniques, only GG is effective for depositing Au
nanoparticles on activated carbon, however, the sizes of Au
particles are relatively large and around 10 nm. Precise control of
dispersion and size (within 5 nm in diameter) of Au nanoparticles
can be accomplished by dipping the support in Au sols stabilized
with polyvinyl pyrrolidone or tetrakis(hydroxymethyl)
phosphonium chloride [12].
The last category is the preparation of model catalysts using
single crystals of MgO, TiO2 (rutile), FeO, Fe3O4, and Si as a
support. Size selected Au anion clusters can be deposited with
homogeneous dispersion at temperatures below 273 K [40-46],
and then they are annealed at higher temperatures to stabilize
them. Surface defects or specific surface cages are suggested as
sites for stabilizing the Au clusters.
Figure 3. Schematic illustration for the role of Mg citrate in depositing Au on
rare earth metal hydroxides.
In future applications gold catalysts may be used in combination
with other transition metal catalysts and base metal oxide
catalysts, especially in electrochemical systems [47] and
environmental pollutant abatement [48,49]. Even a mechanical
mixture of Au/Fe2O3 catalyst with Pt black led to improved
performance in the electrochemical oxidation of MeOH and CO
[50]. It has been demonstrated by us that the most prosperous way
is to integrate a few different supported precious metal catalysts in
167
Nanoparticulate Gold Catalysts for Low-Temperature ... / J. New. Mat. Electrochem. Systems 7, 163-172 (2004)
catalysts prepared by DP, photochemical deposition (PD), and
IMP methods [6]. The DP method yields hemispherical metal
particles with their flat planes strongly attached to the TiO2
support, often by epitaxial contact, Au (111) to anatase TiO2 (112)
as shown in Figure 5 and rutile TiO2 (110), while PD and IMP
methods yield spherical particles, which are simply loaded on the
TiO2 support and, therefore, are much larger, particularly in the
case of Au. Over Pt/TiO2, the reaction of CO with O2 takes place
mainly on the Pt surfaces with or without decoration of TiO2 and
the metal oxide support itself is less involved in the reaction. This
can explain why different methods of preparation do not make
any appreciable difference in the TOF of Pt catalysts. In contrast,
the TOF of Au/TiO2 markedly depends on the methods of
preparation and changes by four orders of magnitude. The TOF of
strongly attached hemispherical Au particles exceeds that of Pt by
one order of magnitude. The dramatic difference in TOFs
suggests that the contact structure is the most critical factor in
supported Au catalysts.
nm scale, as schematically shown in Figure 4. The idea is to
combine synergetically the different performances of each
precious metal by choosing the most suitable support respectively
for a target reaction and then by integrating them into a single
catalytic system. An integrated catalyst composed of Pd or Pt/
SnO2, Au/Fe2O3, and Ir/La2O3 exhibited enhanced catalytic
activities not only for dioxin decomposition but also for the
oxidation of H2 and (CH3)3N [49].
Dioxin Decomposition below 423 K
O
Cl
CO, HCL
HCHO etc.
Cl
O
CO2
Au
Pt
Ir
Fe2O3
SnO2
‫ޓ‬Carrier
The sharp contrast between Pt and Au metal catalysts in CO
oxidation suggests that the reactions might take place at the
perimeter interfaces around the Au particles. To confirm this
hypothesis, Vannice prepared an inversely supported catalyst,
namely, TiO2 layers deposited on a Au substrate, and observed
appreciable catalytic activity [51]. Another type of Au/TiO2
catalyst was prepared by mechanically mixing a colloidal solution
La2O3
Figure 4. Schematic representation of a multi-functional catalyst integrating
three supported precious metal catalysts.
A mixture of SnO2 and Fe2O3 was prepared by coprecipitation
from an aqueous solution of SnCl4 and Fe(NO3)2, followed by
calcination in air at 673K. Palladium is first deposited by DP in an
aqueous solution of Pd(CH3COO)2 at pH=7, where Pd2+ cation
can approach negatively charged SnO2, but is not allowed to
approach Fe2O3 which is positively charged. After washing and
calcination at 673K, the sample composed of PdO/SnO2 and
Fe2O3 was aged in an aqueous solution of HAuCl4 at pH7.5. Gold
hydroxide Au(OH)3 is selectively deposited from AuCl4-n(OH)n anion on positively charged Fe2O3, without precipitation on
negatively charged SnO2. After washing and calcination, the
Figure 5. TEM for Au/TiO2 prepared by deposition-precipitation method,
sample composed of PdO/SnO2 and Au/Fe2O3 is dispersed in an
aqueous solution of Na2CO3, to which an aqueous solution of
IrCl4 and La(NO3) is poured to obtain finally a mixture of Pd/
Table 3. CO oxidation over Pt/TiO2 and Au/TiO2 prepared by different methods [6].
through
washing,
SnO2+Au/Fe2O3+Ir/La2O3
calcination, and reduction in H2 stream.
Metal
3. STRUCTURE SENSITIVITY OF CO
OXIDATION
The catalytic activity of supported Au catalysts for the
low-temperature oxidation of CO strongly depends on
preparation methods and conditions. This is because the
catalytic activity is defined by the contact structure of
Au particles with support, the type of support, and the
size of Au particles.
3.1. Contact Structure of Gold Particles
Table 3 lists TOFs and apparent activation energies for
CO oxidation at 300K over Pt/TiO2 and Au/TiO2
Preparation
methods
Metal
loading (wt
%)
Diameter of
Au (nm)
T1/2*
(K)
Rate at 300K
(mol s−1
g-cat−1)
TOF at 300K
(s−1)
Ea (kJ/
mol)
DP
1.0
1.3 ± 0.3
334
1.4 x 10−7
2.7 x 10−3
49
IMP
1.0
1.4 ± 0.3
339
1.9 x 10−7
3.8 x 10−3
60
PD
0.9
2.4 ± 0.6
363
2.4 x 10−8
9.2 x 10−3
53
DP
0.7
3.1 ± 0.7
282
6.9 x 10−7
3.4 x 10−2
19
DP
1.8
2.7 ± 0.6
253
5.5 x 10−6
1.2 x 10−1
18
1.0
Pt
Au
IMP
PD
1.0
10 <
4.6 ± 1.5
481
477
−10
1.7 x 10
−10
1.5 x 10
ሪ
58
−6
9.6 x 10
* T1/2 : temperature for 50% conversion of 1 vol.% CO in air under a space velocity of 2 x 104 h−1 ml/g-cat.
Preparation methods : DP deposition-precipitation, IMP Impregnation, PD photochemical deposition.
56
168
M. Haruta / J. New. Mat. Electrochem. Systems 7, 163-172 (2004)
of 5nm Au particles with TiO2 powder and by calcination in air at
different temperatures [52]. Calcination at 873K promotes the
coagulation of Au particles in forming larger particles with
diameters above 10nm, but at the same time, with stronger
contact (observed by TEM), leading to much higher catalytic
activity than calcination at 573K.
3.2. Type of Metal Oxide Support
For CO oxidation, many oxides except for strongly acidic
materials such as Al2O3-SiO2 and activated carbon can be used as
a support to induce catalytic activity at temperatures below 300K.
For Pd and Pt, semiconductive metal oxides lead to enhanced
catalytic activities but at temperatures above 300K (see T1/2 in
Table 3). Semiconductive metal oxides such as TiO2, Fe2O3, and
NiO provide more stable Au catalysts than insulating metal
oxides such as Al2O3 and SiO2. The difference also appears in the
enhancing effect of moisture: the concentration of H2O required
by Al2O3 and SiO2 supports is at least 100 ppm higher than that by
TiO2, for CO oxidation to proceed at room temperature [53,54].
Alkaline earth metal hydroxides such as Be(OH)2 and Mg(OH)2
lead to the highest activity at a temperature as low as 196K
[23,31]. In contrast, when an acidic material such as Al2O3-SiO2,
WO3, or activated carbon is used as a support, gold exhibits poor
activity; even at temperatures above 473K the conversions are far
below 100% [37].
As clearly seen in Figure 6, low-temperature CO oxidation under
acidic environment has not yet been accomplished by any
catalysts. This reaction is very important in relation to polymer
electrolyte fuel cells, which are operated at relatively low
temperatures around 373K [55]. In order to use methanol directly
as a fuel, the anode should also be active for the electrochemical
oxidation of CO, for which, however, the present Pt electrode is
not only inactive but also deactivated. Low-temperature CO
oxidation under an acidic environment is an essential scientific
issue in opening a new stage of fuel cell development.
3.3. Size of the Au Particles
Reaction Temperature (K)
Figure 7 plots the TOFs of CO oxidation over Pt/SiO2 and Au/
TiO2 as a function of the mean diameter of metal particles. A
sharp increase in the TOF is observed with a decrease in the
diameter from 4 nm. In contrast, the Pt group of metals shows a
Acidity/Basicity of the Support
Figure 6. Temperature for CO oxidation (50% conversion) as a function of
acidity and basicity of support materials. Reaction conditions : CO 1 vol% in
air, SV=2 X 104h-1/ml•g-cat.
decreasing or steady TOF [7]. The rates over Au/TiO2 were about
one order of magnitude greater when measured as the temperature
was lowered from 353K than when measured as the temperature
was raised from 203K [56]. This difference is assumed to occur
from the accumulation of carbonate species on the surfaces of the
support at low temperatures resulting in the loss of the activating
power of the perimeter interfaces for O2. Therefore, the steady
state rate over Au/TiO2, which was deactivated during experiments
at lower temperatures, is regarded to be close to the rate of CO
reaction with O2 over the surfaces of the Au particles with the
least contribution of O2 activation at the perimeter interfaces. The
one order of magnitude difference in the rate between fresh
(obtained by high temperature measurements) and deactivated Au/
TiO2 can be ascribed to the contribution of the TiO2 support.
Accordingly, the increase in the TOF can be explained if the
adsorption sites for CO are edge, corner, or step sites, and the
reaction zone is the periphery around the Au particles, the
fractions of which increase in hemispherical Au particles with a
decrease in their size [57].
Figure 7. Turnover frequency of CO oxidation over Au/TiO2 and Pt/SiO2 as a
function of the mean diameter of metals.
4. REACTION PATHWAYS IN CO OXIDATION
Figure 8 shows Arrhenius plots for CO oxidation over noble metal
catalysts. A unique feature of Au catalysts is that apparent
activation energies (Ea) are very low. At temperatures below
300K, the value of Ea is 20 to 40 kJ/mol and is nearly zero at
temperatures above 300K. In contrast, Pt group metals have a
value of Ea ranging from 50 kJ/mol (see Table 3) to 170 kJ/mol
[58] and are more active than Au but only at temperatures above
500K. At room temperature, Au is more active by 1-4 orders of
magnitude.
The rate of CO oxidation over Au/TiO2, Au/Fe2O3, Au/Co3O4 is
independent of the concentration of CO and is only slightly
dependent on the concentration of O2 (on the order of 0 to 0.25) in
the range of concentrationdown to 0.1 vol% [5]. This
independence suggests that CO and O2 are adsorbed on the
catalyst surfaces nearly to saturation and that the reaction of the
two adsorbed species is the rate-determining step.
Nanoparticulate Gold Catalysts for Low-Temperature ... / J. New. Mat. Electrochem. Systems 7, 163-172 (2004)
%CVCN[UV6GORGTCVWTG㧔K㧕
TOF㧔s -1 㧕
Figure 8. Arrhenius plots for CO oxidation over noble metal catalysts. The
data for Ru and Pt are taken from ref [58].
In FT-IR spectra for CO adsorption at 90K over Au/TiO2 calcined
in air at different temperatures [59], the most active sample
(calcined at 573K, with Au particles having a mean diameter of
2.4 nm) exhibited the largest peak intensities at 2110 to 2120
cm-1. This peak is attributed to the linear adsorption of CO on the
metallic Au sites [60]. When the diameter of the Au particles
became larger than 10 nm (sample calcined at 873K), the peak
almost disappeared , indicating that CO adsorption took place on
the steps, edges, and corners of Au particles but not on the smooth
surfaces. The results agree well to the fractions of edges and
corners calculated by Mavrikakis and his coworkers as a functiona
of the diameter of Au nanoparticles [57].
No direct experimental evidence has yet been presented showing
where oxygen is activated to react with CO adsorbed on the Au
surfaces and whether oxygen molecule is dissociatively or nondissociatively adsorbed. A TAP (temporal analysis of products)
study of O2 adsorption and the reaction of O2 with CO [61,62],
18O isotope experiments [59,61,62], and ESR measurements
2
[63,64] indicate that molecularly adsorbed O2, most likely O2- at
the perimeter interface, is involved in the oxidation of CO.
Figure 9 shows FT-IR for the introduction of C16O at 300K to Au/
TiO2 preadsorbed with 18O2 [59]. The fact that C16O2 was formed
in a quantity comparable to that of C16O18O means that the
GE
PC
DT
QU
D#
300 K
C18O16O
5 mbar 18O2
+
10 mbar C16O
C18O2
oxygen species (16O) contained in the surface layer of the TiO2
support was also involved in CO oxidation at room temperature.
Although metallic Au particles appears to be indispensable, a
question arises as to why the periphery of Au particles can
activate O2 molecules at low temperatures. As firstly proposed by
Bond and Thompson [65] and later discussed by Kung [17], it is
probable that the perimeter interfaces contain oxidic Au species,
most probably Au(OH)3 or Au(OH), under usual conditions
wherever H2O is present at concentrations above 1 ppm. These
hydroxides might be stabilized and reversibly formed and
decomposed by the aid of the metal oxide supports.
1000/ T㧔K-1 㧕
C16O2
169
㧟min. later
㧞min. later
C18O
9CXGPWODGT㧔cm-1㧕
Figure 9. FI-IR spectra for the introduction of C16O at 300K to Au/TiO2
preadsorbed with 18O2.
An argument proposed by Goodman and his coworkers is that the
non-metallic nature of Au clusters leads to high catalytic activity
[40,66]. This conclusion is questionable except for the case of Au/
Mg(OH)2, which exhibits CO oxidation activity only when 13
atoms Au clusters having icosahedral structure are present [31].
The transition of the electronic state was measured for one
specific Au cluster of a defined diameter by scanning tunneling
spectroscopy, whereas the catalytic activity was measured for an
entire Au/TiO2 model catalyst specimen with a mean diameter of
all Au clusters. A maximum in the catalytic activity with respect
to the mean diameter of the Au clusters was observed where the
transition of the electronic state of the specific Au cluster
occurred from metallic to non-metallic. This result can be more
reasonably explained by assuming that metallic Au surfaces are
necessary for CO adsorption and that the peripheries act as
reaction zones with O2. A maximum total surface area or number
of step sites [57] of metallic Au clusters and the longest perimeter
interface is obtained at around a thickness of 2 atoms, where the
transition to non-metallic state begins [67].
There are other arguments as well regarding the active species of
Au, especially in the case of Au/Fe2O3. These include oxidized Au
species, Au+ [68], Aud+ [69], or metal oxide support surfaces with
modified reducibility by the interaction with Au nanoparticles
[70,71]. It is unlikely that oxidic Au species are the major
catalytically active phases because the most active supported Au
catalysts are prepared by calcination in air at 573K, where Au
precursors (hydroxides or organo complexes) are mostly
transformed into metallic particles. A certain fraction of Au
species remain as atomically dispersed species in the matrix of the
support, as proved by EXAFS [25,69,72,73], XPS [70],
Mössbauer [71,74], and IR of CO adsorbed [59,68]. However, no
correlation between the amount of oxidic Au species and catalytic
activity has yet been presented. A reason why the samples
primarily consisting of oxidic Au exhibited high catalytic activity
might be that the oxidic Au species were transformed into
metallic particles during reaction or storage of the sample after
preparation. Recently it has been reported that oxidic Au is
responsible for water gas shift reaction over Au/CeO2 catalyst
[75]. However, the catalytic activity is much less than that of a
commercial Cu/ZnO/Al2O3 catalyst, which was opposite in the
case of Au/CeO2 properly prepared in our laboratory [76].
170
M. Haruta / J. New. Mat. Electrochem. Systems 7, 163-172 (2004)
The above arguments are summarized in Table 4 in terms of
accountability of the proposed major active sites for the known
characteristic behavior of supported Au catalysts. A recent
overview [77], classifies the proposed active sites into four types:
electron rich Au clusters (eight or more atoms), non-metallic Au
nanoparticles (diameter below 3 nm, thickness dropped to 2
atomic layer), jagged step edges at the sides of Au nanoparticles,
and perimeter sites around Au nanoparticles. The former three
hypotheses simply describes some probable changes in the nature
of Au nanoparticles and clusters, which can explain only a part of
the reaction parhways (for example, adsorption of CO and O2) in
CO oxidation. They are included into one category, Au surfaces,
in Table 4, because combined effect of theses three effects may
dominate in the catalysis by Au nanopareticles. What cannot be
explained well in this hypothesis is the effect of support,
especially the remarkable difference in moisture enhancement
among TiO2, Al2O3, SiO2 [54].
Table 4. Accountability of major active sites for the characteristic behaviour
of nanoparticulate Au catalysts.
Active sites
Shape &
size
effect
Support
Effect
0th order
kinetics
Small
Ea
H2O
effect
Oxidic Au
–
+
–
+
–
Au
surfaces
+
–
¢
+
¢
Support
surfaces
+
+
–
–
–
Perimeter
interfaces
+
+
+
+
+
Oxidic Au species, either in the form of isolated Au species
incorporated into the matrix of the support metal oxides or in the
form of Au hydroxide or oxide aggregates on the support surfaces,
may exhibit some interesting features in their catalysis, however,
they cannot explain why higher catalytic activities are obtained by
Au catalysts mainly composed of metallic Au particles. Because
oxidic Au sites are assumed to behave like transition metal oxides,
they should be deactivated by moisture and the rate should be
sensitive to the concentration of at least one of the reactants.
Au nor supports (Al2O3, SiO2, TiO2) adsorb CO and O2 can be
explained in terms of the size effect of Au particles and the
reactivity of the perimeter interfaces. Gold surfaces can adsorb
CO when the diameter of Au particles is smaller than 10nm due to
an increase in step, edge, and corner sites [59] and the active Au/
TiO2 catalysts can adsorb O2 much more than Au and TiO2 [5]
indicating that new sites evoked by the strong contact of Au
particles with the support, most probably perimeter interfaces, are
the sites for O2 adsorption. It may happen that the adsorbed O2
can move to the step, edge, and corner sites of Au nanoparticles,
however, more probably, it reacts with CO adsorbed on the Au
surfaces at the perimeter interfaces.
Because the supported Au catalysts should not be evacuated or
reduced by H2 otherwise severly deactivated [79], the decoration
model proposed in connection with so called “strong metal
support interaction” is not applicable to supported Au catalysts.
This is one of characteristic feature of Au which is different from
the noble metals of the group VIII in the periodic table. It was
reported that the formation of intermetallic alloy compounds
Pt3Sn induced the catalytic activity for CO oxidation at room
temperature when Pt/SnO2 catalysts were reduced by H2 [80].
There is a coincidence with the prediction and experimental
results reported for H2 electrode catalysts by Jaksic [15].
Based on the above arguments, the most probable pathway for CO
oxidation over supported Au catalysts can be depicted as shown in
Figure 10.
5. FUTURE PROSPECTS
In order to make recent attempts to combine experimental work
using real and model catalysts with theoretical calculations more
synergetic [81-85], the effect of moisture should be taken into
account. Most work in surface science is done in an ultrahigh
vacuum, while measurements of the activity of real catalysts are
conducted in a fixed bed flow reactor using reactant gas
containing moisture at least 1 ppm, usually 10 ppm. The catalyst
surfaces are covered with OH groups and water molecules at
room temperature. For CO oxidation, which does not produce
H2O and proceeds at temperatures below 373K, moisture
The support surfaces, which we first assumed to be the active sites
[4], have also similar difficulties to those of oxidic Au. Although
they can explain the shape and size effect, they may not bring
about such a low apparent activation energy as 20 kJ/mol.
The perimeter interfaces appear to be the most reasonable to
account for the whole characteristic behaviour of nanoparticulate
Au catalysts. This hypothesis may also involve the ideas proposed
for the increase in side steps on Au surfaces [78] as the sites for
CO adsorption and the ideas for support surfaces as the sites for
O2 adsorption. The reason why supported Au catalysts properly
prepared are surprisingly active for CO oxidation while neither
Figure 10. A schematic representation of CO oxidation pathways over
supported Au catalysts.
Nanoparticulate Gold Catalysts for Low-Temperature ... / J. New. Mat. Electrochem. Systems 7, 163-172 (2004)
markedly changes the catalytic activity of the metal oxides and
the Au catalysts. Under dry conditions with an H2O concentration
of 80 ppb, CO oxidation can occur over Co3O4 without Au even at
210K [86,87], while supported Au catalysts prefer moisture
[53,54].
It can be speculated from our preliminary investigation that H2O
enhances the dissociation of O2 at the perimeter interface around
Au particles [88] and also stabilizes the cationic Au species. The
copresence of zerovalent and cationic Au species may stabilize
the atomic structure at the contact interfaces which can repeatedly
adsorb O2 and liberate active oxygen species.
2Au + H2O + O2
or
[2Au]-CO + H2O + O2
O* + 2(OH)- + 2Au+ (or 2/3Au3+)
CO2 + 2(OH)- + 2Au+(or 2/3Au3+)
It has recently been reported that Au nanoparticles of 4.8nm in
diameter synthesized by block copolymer micelle encapsulation
exhibited good catalytic activity in electro-oxidation of CO in
0.5M KOH solution [89,90]. This is the first example, to our
knowledge, that shows the outbirth or noticeable enhancement in
the electrocatalytic activity of nanoparticulate Au. Based on the
recent publication on the magic number of Au clusters for the
stability and chemical reactivity [91,92], size-controlled
nanoparticulate Au dispersed on electrically conductive polymer
may open a new possibility in electrode catalysts.
6. CONCLUSIONS
1) There are at least eight techniques to prepare highly dispersed
gold catalysts. Their catalytic performances for CO oxidation
markedly depend on the preparation methods and conditions
because three factors, strong contact between Au particles and the
support, selection of the support, and the size of the Au particles,
define the catalytic properties of Au nanoparticles.
2) Perimeter interfaces provide the sites for the reaction of CO
with O2. Metallic Au surfaces, especially steps, edges, and
corners, are necessary as the sites for CO adsorption and the
support surfaces around Au particles for O2 adsorption.
3) Very low apparent activation energies and enhancement by
moisture present advantages to Au catalysts in environmental
applications, which often prefer low-temperature operation.
REFERENCES
[1] H. Kita, H. Nakajima, K. Hayashi, J. Electroanal. Chem., 190,
141 (1985).
[2] L. D. Burke, D. T. Buckley, J. A. Morrissey, Analyst, 119, 841
(1994).
[3] L. D. Burke, P. F. Nugent, Gold. Bull., 30, 43 (1997) and 31,
39 (1998).
[4] M. Haruta, N. Yamada, T. Kobayashi, S. Iijima, J. Catal , 115,
301 (1989).
[5] M. Haruta, S. Tsubota, T. Kobayashi, H. Kageyama, M. J.
Genet, B. Delmon, J. Catal., 144, 175 (1993).
[6] G. R. Bamwenda, S. Tsubota, T. Nakamura, M. Haruta, Catal.
Lett , 44, 83 (1997).
171
[7] M. Haruta, in “Catalysis and Electrocatalysis at Nanoparticle
Surfaces”, Ed., A. Wieckowski, E. R. Savinova, C. G.
Vayenas, Marcel Dekker, New York, USA, 2003, p645.
[8] G. C. Bond, D. T. Thompson, Catal. Rev. – Sci. Eng., 41, 319
(1999).
[9] M. Haruta, M. Daté, Appl. Catal. A: General, 222, 427 (2001).
[10]M. Haruta, CATTECH, 6, 102 (2002).
[11] M. Haruta, Chem. Record, 3, 75 (2003).
[12] L. Prati, M. Rossi, in “Green Chemistry: Challenging
Perspectives”, Ed., P. Tundo, P. Anastas, Oxford Univ. Press,
Oxford, Great Britain, 2000, p.183.
[13] M. Okumura, S. Nakamura, S. Tsubota, T. Nakamura, M.
Azuma, M. Haruta, Catal. Lett., 51, 53 (1998).
[14] G. Ertl, Chem. Record, 1, 33 (2001).
[15] S. G. Neophytides, S. H. Zafeiratos, M. M. Jaksic, J.
Electrochem. Soc., 150 (10), E512 (2003).
[16] U. Diebold, Surf. Sci. Report, 48, 534 (2003).
[17] H. H. Kung, M. C. Kung, C. K. Costello, J. Catal., 216, 425
(2003).
[18] M. Haruta, H. Sano, Stud. Surf. Sci. Catal, 16, 225 (1982).
[19] M. M. Jaksic, J. New Mat. Electrochem. Systems, 3, 153
(2000).
[20] M. M. Jaksic, C. M. Lacnjevac, B. N. Grgur, N. V. Krstajic, J.
New Mat. Electrochem. Systems, 3, 169 (2000).
[21] M. M. Jaksic, Electrochim. Acta, 45, 4085 (2000).
[22] M. Haruta, S. Tsubota, M. Okumura, in “Advances in
Catalyst Preparation (in Japanese)”, Ed., Y. Ono, Assoc. for
the Promotion of Catal. Preparation Chem., Tokyo, Japan,
2000, p39.
[23] M. Haruta, T. Kobayashi, S. Tsubota, Y. Nakahara, Chem.
Express, 3, 159 (1998).
[24] M. Haruta, H. Kageyama, N. Kamijo, T. Kobayashi, F.
Delannay, Stud. Surf. Sci. Catal., 44, 33 (1988).
[25] H. Kageyama, N. Kamijo, T. Kobayashi, M. Haruta, Physica,
B158, 183 (1989).
[26] R. M. Torres Sanchez, A. Ueda, K. Tanaka, M. Haruta, J.
Catal., 168, 125 (1997).
[27] T. Kobayashi, M. Haruta, S. Tsubota, H. Sano, Sensors and
Actuators, B1, 222 (1990).
[28] M. Shibata, N. Kuwata, T. Matsumoto, H. Kimura, Chem.
Lett., 1605 (1985).
[29] D. A. H. Cunningham, W. Vogel, R. M. Torres Sanchez, K.
Tanaka, M. Haruta, J. Catal., 183, 24 (1999).
[30] S. Tsubota, M. Haruta, T. Kobayashi, A. Ueda, Y. Nakahara,
Stud. Surf. Sci. Catal., 63, 695 (1991).
[31] D. A. H. Cunningham, W. Vogel, H. Kageyama, S. Tsubota,
M. Haruta, J. Catal., 177, 1 (1998).
[32] R. Zanella, S. Giorgio, C. R. Henry, C. Louis, J. Phys. Chem.
B., 106, 7634 (2002).
[33] Q. Xu, K. C. C. Kharas, A. K. Datye, Catal. Lett., 85, 229
(2003).
[34] Y. Yuan, A. P. Kozlova, K. Asakura, H. Wan, K. Tsai, Y.
Iwasawa, J. Catal., 170, 191 (1997).
[35] M. Okumura, M. Haruta, Chem. Lett., 396 (2000).
172
[36] M. Okumura, K. Tanaka A. Ueda, M. Haruta, Solid State
Ionics, 95, 143 (1997).
[37] M. Okumura, S. Tsubota, M. Haruta, J. Mol. Catal. A., 199,
73 (2003).
[38] J-D. Grunwaldt, C. Kiener, C. Wögerbauer, A. Baiker, J.
Catal., 181, 223 (1999).
[39] C. L. Bianchi, S. Biella, A. Gervasini, L. Prati, M. Rossi,
Catal. Lett., 85, 91 (2003).
[40] M. Valden, S. Pak, X. Lai, D. W. Goodman, Catal. Lett., 56, 7
(1998).
[41] U. Heiz, W. –D. Schneider, J. Phys. D; Appl. Phys., 33, R85
(2000).
[42] W. T. Wallace, R. L. Whetten, J. Phys.Chem. B., 104, 10964
(2000).
[43] K. Kishi, M. Daté, M. Haruta, Surf. Sci., 486, L475 (2001).
[44] F. Cosandey, T. E. Madey, Surf. Rev. Lett., 8, 73 (2001).
[45] Sh. K. Shaikhutdinov, R. Meyer, M. Naschitzki, M. Bäumer,
H. –J. Freund, Catal. Lett., 86, 211 (2003).
[46] E. Wahlström, N. Lopez, R. Schaub, P. Thostrup, A. Rønnau,
C. Africh, E. Lægsgaard, J. K. Nørskov, F. Besenbacher, J.
Phys. Chem., in press.
[47] C. E. Larsson, T. H. Baum, R. L. Jackson, J. Electrochem.
Soc., 134, 266 (1987).
[48] A. Ueda, M. Haruta, Appl. Catal. B; Environ., 18, 115
(1998).
[49] M. Okumura, T. Akita, M. Haruta, X. Wang, O. Kajikawa, O.
Okada, Appl. Catal. B: Environomental, 41, 43 (2003).
[50] P. C. Biswas, Y. Nodasaka, M. Enyo, M. Haruta, J.
Electronanl. Chem. 381, 167 (1995).
[51] Z. M. Liu, M. A. Vannice, Catal. Lett., 43, 51 (1997).
[52] S. Tsubota, T. Nakamura, K. Tanaka, M. Haruta, Catal. Lett.,
56, 131 (1998).
[53] M. Daté, M. Haruta, J. Catal., 201, 221 (2001).
[54] M. Daté, M. Okumura, S. Tsubota, M. Haruta, in
“Proceedings of Gold 2003”, Vancouver, Canada, Sept. 28Oct. 1, 2003.
[55] J. Lipkowski, P. N. Jr. Ross, “Electrocatalysis, Frontiers in
Electrochemistry”, VCH, New York, USA, 1997, Vol. 5.
[56] Y. Iizuka, T. Tode, T. Takao, K. Yatsu, T. Takeuchi, S.
Tsubota, M. Haruta, J. Catal., 187, 50 (1999).
[57] M. Mavrikakis, P. Stoltze, J. K. Nørskov, Catal. Lett., 64, 101
(2000).
[58] C. H. F. Peden, in “Surface Science of Catalysis – In Situ
Probes and Reaction Kinetics”, ACS Symp. Ser. 482, Ed., D.
J. Dwyer, F. M. Hoffmann, Amer. Chem. Soc., Washington
DC, USA, 1992, p. 143.
[59] F. Boccuzzi, A. Chiorino, M. Manzoli, P. Lu, T. Akita, S.
Ichikawa, M. Haruta, J. Catal., 202, 256 (2001).
[60] C. Ruggiero, P. Hollins, Surf. Sci., 377-379, 583 (1997).
[61] M. Olea, M. Kunitake, T. Shido, Y. Iwasawa, Phys. Chem.
Chem. Phys., 3, 627 (2001).
[62] M. M. Schubert, S. Hackenberg, A. C. van Veen, M. Muhler,
V. Plzak, R. J. Behm, J. Catal., 197, 113 (2001).
[63] H. Liu, A. I. Kozlov, A. P. Kozlova, T. Shido, K. Asakura, Y.
Iwasawa, J. Catal., 185, 252 (1999).
M. Haruta / J. New. Mat. Electrochem. Systems 7, 163-172 (2004)
[64] M. Okumura, J. M. Coronado, J. Soria, M. Haruta, J. C.
Conesa, J. Catal., 203, 168 (2001).
[65] G. C. Bond, D. T. Thompson, Gold Bull., 33, 41 (2000).
[66] M. Valden, X. Lai, D. W. Goodman, Science, 281, 1647
(1998).
[67] Y. Maeda, M. Okumura, S. Tsubota, M. Kohyama, M.
Haruta, Appl. Scurf. Sci.,222, 409 (2003).
[68] S. Minicò, S. Scirè, C. Crisafulli, A. M. Visco, S. Galvagno,
Catal. Lett., 47, 273 (1997).
[69] Z. Hao, L. An, H. Wang, T. Hu, React. Kinet. Catal. Lett., 70,
153 (2000).
[70] D. Horváth, L. Toth, L. Guczi, Catal. Lett., 67, 117 (2000).
[71] F. E. Wagner, S. Galvagno, C. Milone, A. M. Visco, L.
Stievano, S. Calogero, J. Chem. Soc. Faraday Trans., 93,
3403 (1997).
[72] H. Kageyama, K. Kadono, K. Fukumi, T. Akai, N. Kamijo,
M. Haruta, J. Phys. IV France 7, C2, 935 (1997).
[73] S. Tsubota, D. A. H. Cunningham, Y. Bando, M. Haruta,
Stud. Surf. Sci. Catal., 91, 227 (1995).
[74] Y. Kobayashi, S. Nasu, S. Tsubota, M. Haruta, Hyperfine
Interactions, 126, 95 (2000).
[75] Q. Fu, H. Saltsburg, M. Flytzani-Stephanopoulos, Science,
301, 935 (2003).
[76] H.Sakurai, S. Tsubota, M. Haruta, Shokubai (Catalysts &
Catalysis), 44, 416 (2002).
[77] A. Cho, Science, 299, 1684 (2003).
[78] Y. Xu. M. Mavrikakis, J. Phys. Chem. B. 107, 9298 (2003).
[79] D. A. H. Cunningham, S. Tsubota, N. Kamijo, M. Haruta,
Res. Chem. Interm., 19, 1 (1993).
[80] S. D. Gardner, G. B. Hoflund, M. R. Davidson, D. R.
Schryer, J. Catal., 115, 132 (1989).
[81] R. Schlögel, CATTECH, 5, 146 (2001).
[82] C. J. Zhang, P. Hu, J. Am. Chem. Soc., 122, 2134 (2000).
[83] J. K. Nørskov, Nature, 414, 405 (2001).
[84] J. R. Hahn, W. Ho, Phys. Rev. Lett. 87. 166102 (2001).
[85] L. M. Molina, B. Hammer, Phys, Rev. lett., 90 (20), 206102
(2003).
[86] D. A. H. Cunningham, T. Kobayashi, N. Kamijo, M. Haruta,
Catal. Lett., 25, 257 (1994).
[87] M. Haruta, M. Yoshizaki, D.A.H. Cunningham, T. Iwasaki,
Ultraclean Technology (in Japanese), 8, 1 (1996).
[88] F. Boccuzzi, A. Chiorino, S. Tsubota, M. Haruta, in “Sensors
and Microsystems”, Ed, C. Di Natale, A. D’Amico, World
Sci.Publ., Singapore, 1996, p. 116.
[89] M. M. Maye, Y. Lou, C. –J. Zhong, Langmuir, 16, 7520
(2000).
[90] T. F. Jaramillo, S.-H. Baeck, B. R. Cuenya, E. W. McFarland,
J. Am. Chem. Soc., 125, 7148 (2003).
[91] H.-G. Boyen, G. Kästle, F. Weigl, B. Kaslowski, C. Dietrich,
P. Ziemann, J. P. Spatz, S. Riethmuller, C. Hartmann, M.
Moller, G. Schmid, M. G. Garnier, P. Oelhafen, Science, 297,
1533 (2002).
[92] J. Li, X. Li, H.-J. Zhai, L.-S. Wang, Science, 299, 864
(2003).