CO oxidation over titanate nanotube supported Au: Deactivation due to bicarbonate

Article abstract of DOI:10.1016/j.jcat.2008.11.009

New experimental findings, which offer further information concerning the deactivation mechanism of CO oxidation over titania nanotube (TN) supported gold, are reported. Contrary to earlier reports by others, we have observed that the formation of bicarbo

Full text of DOI:10.1016/j.jcat.2008.11.009

Journal of Catalysis 261 (2009) 94–100
Contents lists available at ScienceDirect
Journal of Catalysis
www.elsevier.com/locate/jcat
CO oxidation over titanate nanotube supported Au:
Deactivation due to bicarbonate
T.A. Ntho a, J.A. Anderson b, M.S. Scurrell a,∗
a
Molecular Sciences Institute, School of Chemistry, University of the Witwatersrand Johannesburg, South Africa
Chemistry Department, University of Aberdeen, Scotland, UK
b
a r t i c l e i n f o
a b s t r a c t
Article history:
New experimental findings, which offer further information concerning the deactivation mechanism of CO
oxidation over titania nanotube (TN) supported gold, are reported. Contrary to earlier reports by others,
we have observed that the formation of bicarbonate species during CO oxidation on the Au/TN catalyst
is a competing reaction that leads to catalyst deactivation. We regard the formation of bicarbonate
Received 24 February 2008
Revised 3 November 2008
Accepted 4 November 2008
Available online 4 December 2008
species not as an intermediate step but rather a product of the CO2 produced during CO oxidation.
−1
The bicarbonate species detected at 1290 cm
was very stable and only desorbed at relatively high
Keywords:
Gold
Titania
Nanotubes
Deactivation
CO oxidation
Bicarbonate
Infrared
TPD
temperatures. In the presence of water the bicarbonate species (1290 cm−1) did not form under reaction
conditions. CO2-TPD results implied that water possibly reacted with and removed the already formed
bicarbonate species and prevented its further formation and in the process promoted the reaction and
prevented deactivation. The possible reactions of how water (OH) might remove the gold associated
bicarbonate and hence reactivate and promote the oxidation of CO are discussed. The mechanisms
for the formation of the observed bicarbonate may account for the deactivation of the Au/TN catalyst
during CO oxidation; this reaction would consume the OH groups on Au which have been reported as
essential in the mechanism of CO oxidation. The above hypothesis of CO2 being able to chemisorb on
gold nanoparticles is not unique. Based on IR studies [N.M. Schubert, A. Venugopal, M.J. Kalich, V. Plazk,
R.J. Behm, J. Catal. 222 (2004) 32; B. Schumacher, Y. Denkwitz, V. Plzak, M. Kinne, R.J. Behm, J. Catal. 224
(2004) 449], evidence was provided which pointed to the fact that CO and CO2 competed for adsorption
on the same site on Au surfaces. To the best of our knowledge, this is the first time that DRIFTS and
CO2-TPD have been combined to show that CO2 is adsorbed on the gold-nanoparticles in the form of
bicarbonate to the detriment of the oxidation of CO.
CO2
© 2008 Elsevier Inc. All rights reserved.
1. Introduction
the interface, which is governed by the metal–support interaction
[8–12], (ii) the importance of particle geometry which dictates the
presence of low coordinated atoms on the nanoparticles [13,14]
and (iii) quantum size effects due to the low dimensionality of the
gold particles [15,16].
Much of the interest in gold catalysis in the last decade stems
from the pioneering work by Haruta et al. [1]. In particular, exten-
sive work has been done on low temperature CO oxidation over
metal oxide supported gold catalysts [2–7]. Despite the high initial
activity for CO oxidation, gold catalysts suffer rapid deactivation
within the first few hours of time-on-line. Although catalyst de-
activation is an important phenomenon in industrial applications,
the deactivation of gold catalysts has not always received the at-
tention it deserves as evidenced by the lack of deactivation kinetic
data.
Generally, there is discord in the results that are reported in
the literature regarding the rate of (de)activation and means of
reactivation [17]. Some of the possible reasons cited in the litera-
ture for deactivation include growth of metal particles, blockage of
(TiO2) interfacial sites by water, build up of ‘carbon species,’ and
the depletion of surface hydroxyl groups ([17] and refs. therein).
Deactivation due to the reduction of Aux+ ions by carbon monox-
ide has also been reported [18].
Adsorbed carbonate species are formed at most metal oxide
surfaces because of chemisorption of carbon dioxide [19]. The exact
speciation and coordination of these species is crucial to under-
stand fully, due to the important role they supposedly play in
CO oxidation. Originally, Haruta et al. [2] proposed a mechanism
involving the formation and decomposition of carbonates as the
There are conflicting arguments in the literature regarding the
mechanism of CO oxidation catalyzed by supported gold. The most
popular mechanistic models reported involve: (i) the importance of
Corresponding author.
E-mail address: michael.scurrell@wits.ac.za (M.S. Scurrell).
*
0021-9517/$ – see front matter © 2008 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2008.11.009

T.A. Ntho et al. / Journal of Catalysis 261 (2009) 94–100
95
◦
lowed operation under different atmospheres up to 600 C. Spectra
rate-limiting step in CO oxidation catalyzed by gold. Daniells et al.
[20] later proposed that the active site was an ensemble of Au3+
and Au0, with hydroxyls from the cationic gold and the support
promoting bicarbonate formation and their subsequent decompo-
sition to produce CO2. However, Gupta and Tripathi [21] argued
that both carbonates and bicarbonates were not possible interme-
diates in the mechanism. Formates [8] and bicarbonates [10] have
also been reported as possible reaction intermediates during CO
oxidation.
−1
were collected at a resolution of 4 cm
on a Perkin–Elmer 1720
FTIR spectrometer. Gases were led into the cell via a computer-
controlled gas blender and exit gases were analyzed by a Baltzers
◦
Prisma QMS. All samples were heat treated at 300 C in air
(20 ml/min) for 30 min prior to DRIFTS studies. A typical gas blend
was 2% CO + 10% air, balanced in N2. Water (100 μL) was injected
with a syringe where it was used.
In this paper, new experimental findings, which offer further in-
formation concerning the deactivation mechanism of CO oxidation
over titania nanotube (TN) supported gold, are reported. Contrary
to earlier reports [22], we have observed that the formation of bi-
carbonate species during CO oxidation on the Au/TN catalyst is a
competing reaction that leads to catalyst deactivation.
2.5. Temperature-programmed desorption (TPD)
All TPD experiments were carried out in a flow apparatus with
helium as carrier gas (50 ml min−1) using a quartz U-tube reactor
(6 mm i.d.). For detection of desorbed CO2 a Baltzers Prisma QMS
was used. Prior to adsorption, samples (220 mg) were heat treated
◦
at 300 C in a flow of 5% O2 balanced in He. In a typical TPD exper-
2. Experimental
iment, adsorbents (e.g. 2% CO+5% O2 balanced in He) were passed
−1
over the catalyst bed at 343 K at a flow rate of 100 ml min
for
2.1. Synthesis of titanate nanotube support
30 min. Then the reactor was cooled down to room temperature
for the start of CO2-TPD. All samples were heated up to 873 K at
The synthesis procedure followed to generate the titanate and
the structure of such materials have been described elsewhere
[23,24]. Commercial TiO2 (P25) powder (25 g) was added to a
◦
the rate of 8 C min−1. A water bubbler-saturator was used where
water was a co-adsorbent.
−1
200 ml solution of 2.8 × 10
M KOH in a 1 L stainless steel au-
2.6. Catalytic activity measurement
◦
toclave. The mixture was heated for 24 h at 120 C (stirring at
500 rpm). The resulting material was cooled down to room tem-
perature, then allowed to age for 2 days in the base solution. It was
then repeatedly centrifuged with deionized water until the con-
ductivity was below 100 μS cm−1. The material was finally dried
◦
The catalyst sample (Au/TN) was calcined at 300 C in 10% O2
(balance He) for 30 min prior to testing. The catalytic performance
of the catalyst for CO oxidation was tested in a continuous flow
quartz fixed-bed reactor (6 mm i.d.). 100 mg of solid catalyst (pow-
der particle size about 150 μm) was loaded into the reactor. The
gas mixtures consisted of 5% CO and 10% O2 balanced in He and
the total flow was generally 40 ml min−1. The course of the reac-
tion was monitored quantitatively by use of a gas chromatograph
facility equipped with a thermal conductivity detector. All mea-
◦
in air at 120 C for at least 12 h. The potassium content may be
expected to lie in the approximate rage 18–32 mass% K. From here
onwards the resultant material shall be referred to simply as the
titanate nanotube (TN).
2.2. Preparation of gold-supported catalysts
◦
surements were carried out at 70 C and atmospheric pressure.
Deposition–precipitation was the method used to load gold
onto the titanate nanotube support [17]. The support was slur-
3. Results
ried in distilled water (600 ml) under vigorous stirring. Required
3.1. Characterization of the support and catalyst
−2
amount of diluted HAuCl4 solution (10
M) was added slowly to
the support slurry with continuous stirring. Due to the basicity
of the titanate support, the pH was already at ∼9; so the ad-
dition of base was not necessary. The precipitated solution was
aged for 15 h. A solution of NaBH4 was prepared in ice water
and added in the required amount for reduction of Au (III) to
Au (0). The solid was washed with hot water (total volume ap-
proximately 3000 cm3) and dried at room temperature. No further
pretreatments were applied unless otherwise stated. In our expe-
rience with using borohydride reduction, the residual quantities of
sodium and boron (and chlorine) on the surface are negligible pro-
vided that washing of the solid is carried out as described. The
resulting catalyst powder of Au/TN contained 0.48 wt% Au, as de-
termined by fire assay with atomic absorption finish (Performance
Laboratories, South Africa).
The BET surface areas of the both the support (TN) and cat-
alyst (Au/TN) were approximately 148 m2/g with a pore volume
of 0.26 cm3/g. The gold loading on the support was 0.48 wt%
(intended loading was 1 wt%). Titanate nanotubes in aqueous sus-
pension have a relatively high negative zeta-potential over a wide
range of pH [23]. The fact that we used a negatively charged gold
precursor (HAuCl4) explained why our resultant gold loading was
relatively low. The morphologies of the titanate nanotube (TN) and
the catalyst Au/TN were studied by the use of HRTEM and the cor-
responding images are shown in Fig. 1. The average particle size of
gold on the catalyst Au/TN was found to be 4.5 nm. Careful anal-
ysis of the TN material suggested that it was in the form of very
thin sheets, which could roll up into two or more layered sheets
of titanium dioxide [25]. The chemical composition of the TN sam-
ple was examined by EDS (Fig. 2), which revealed the presence of
potassium.
2.3. Sample characterization
Nitrogen adsorption–desorption isotherms were measured on a
Micromeritics TRISTAR 3000 analyzer. The samples were degassed
under vacuum for several hours before the nitrogen adsorption
measurements. The specific surface areas were determined by the
Brunauer–Emmett–Teller (BET) method.
3.2. CO oxidation
The primary aim of this paper is to interrogate the reasons why
the Au/TN catalyst system deactivates during time-on-line of CO
oxidation. In order to study deactivation properly, diffusion/mass
transfer limitations must be avoided. This is best done by keeping
catalytic conversions low where the rate of the reaction proceeds
under kinetic control. Fig. 3 illustrates the deactivation observed
during a typical CO oxidation run over the Au/TN catalyst. It is
2.4. DRIFT spectroscopic measurements
Samples (50 mg) were loaded into a DRIFTS cell with CaF2
windows; the cell was equipped with a heating system that al-

96
T.A. Ntho et al. / Journal of Catalysis 261 (2009) 94–100
Fig. 1. The TEM images of the titanate nanotube support (right) and the catalyst Au/TN (left).
Fig. 2. EDS spectrum of the titanate nanotube support confirming the presence of
potassium in the material.
Fig. 3. Deactivation of the Au/TN catalyst during time-on-line during the oxidation
of CO at 70 C.
◦
interesting to note that there was a short induction period, after
which the catalyst gained maximum activity. However, after that,
it could not sustain that activity as it rapidly deactivated. Data
recorded during the first 1 h on stream are not regarded as sig-
nificant for our present purposes, since the catalyst is essentially
developing catalytic activity during this period.
Fig. 4 illustrates the effect of introducing water into the feed
stream during CO oxidation. In this experiment, a DRIFTS cell was
used as a reactor and the exit gases were analyzed using a quadru-
ple mass spectrometer. Water promoted the reaction and the cor-
responding DRIFTS spectra for these particular experiments are
shown in Fig. 8. For operational reasons data were not recorded
in a completely continuous fashion during the period when water
was being removed from the reaction system.
ture of CO + O2 (Fig. 5a) and CO2 (Fig. 5b) were flowed for 30 min
◦
at 70 C. Under both adsorption conditions (CO + O2 and CO2) the
support material (TN) showed the presence of basic surface sites
◦
from which desorption took place at temperatures above 300 C.
◦
However, below 300 C there was no CO2 desorbed for the case
where CO + O2 mixture was passed over the TN support. In the
case where CO2 was adsorbed, two low temperature desorption
◦
◦
peaks were observed at 99 C and 199 C.
3.3.2. Au/TN catalyst
Fig. 6 shows the desorption profiles of CO2 from the catalyst
Au/TN. The desorption profiles were obtained after a mixture of
◦
CO + O2 had been passed over the catalyst for 30 min at 70 C
3.3. CO2-TPD
under dry and wet conditions. Under dry conditions, there were
◦
CO2 desorption peak species at 94, 153, and 229 C. The latter
two peaks were absent when the adsorption was done under
wet conditions. Similarly, there were CO2 desorption peaks at 89
and 102 C under the wet adsorption/desorption experiment.
3.3.1. Titanate nanotube (TN)
The CO2-TPD profiles illustrated in Fig. 5 show the desorption of
carbon dioxide from the pure support material (TN); after a mix-
◦

T.A. Ntho et al. / Journal of Catalysis 261 (2009) 94–100
97
Fig. 4. Effect of water (100 μL) on the catalytic activity of the Au/TN catalyst measured in a DRIFTS cell: (a) promotion of the oxidation of CO reaction by added water,
(b) graph showing the quantification of the CO2 produced.
◦
Fig. 7. CO2-TPD for the Au/TN catalyst after adsorption of CO2 at 70 C.
Fig. 5. CO2-TPD spectra for the titanate nanotube (TN) support: (a) CO + O2 co-
adsorbed, (b) CO2 adsorbed.
Fig. 7 shows the typical CO2-TPD spectrum of CO2 after a 1%
CO2 (balance N2) gas was passed over the catalyst for 30 min at
◦
70 C. The distinct feature from this figure was the CO2 desorption
◦
peak at 162 C, which for all intents and purposes, was due to the
◦
same CO2 species that desorbed at 153 C when CO + O2 mixture
was passed over the dry catalyst. These species of CO2 seemed to
be associated with the presence of Au on the support as it was not
observed on the pure support when either CO + O2 or CO2 were
adsorbed on the pure support.
3.4. DRIFT spectroscopic measurements
In order to characterize the adsorption state of carbon diox-
ide on both the pure support (TN) and the catalyst (Au/TN), DRIFT
spectroscopic studies were carried out.
In principle, IR-active vibration frequencies are highly sensitive
to the structure, protonation states, and speciation of carbonate
(adsorbed CO2) [26]. However, in practice, there are many inherent
difficulties in assigning the molecular structure of carbonate from
its vibrational frequencies. Accurate definition of the important car-
bonate asymmetric C–O stretching region (∼1450–1600 cm−1) is
difficult when DRIFTS measurements are done on samples where
Fig. 6. CO2-TPD spectra for the catalyst (Au/TN): (a) CO + O2 co-adsorbed under dry
conditions, (b) CO + O2 co-adsorbed under wet conditions (∼2% vol/vol H2O).
H2O is one of the co-reactants/adsorbents. Adsorbed water might

98
T.A. Ntho et al. / Journal of Catalysis 261 (2009) 94–100
−1
Fig. 8. FTIR spectra scanned before and after the addition of water (100 μL) into the feed stream: (a) spectra showing the build up of the peak at 1290 cm
−1
in the absence
of water in the feed, (b) plot of CO2 yield vs the intensity of the peak at 1290 cm
.
Fig. 9. FTIR spectrum scanned on the titanate nanotube (TN) support under 2% CO +
Fig. 10. FTIR spectrum scanned on the Au/TN catalyst under 2% CO + 5% O2 (bal-
◦
◦
5% O2 (bal. He) flow at 70 C.
ance He) flow at 70 C: (a) 100 min after addition of water (100 μL), (b) 28 min of
reaction under dry conditions.
contribute H–O–H bending intensity at locations different from
that of bulk water [26].
been assigned as an OH deformation of bicarbonate species [27].
To confirm whether these bicarbonate species needed the presence
of gold to form or they could form on the support under reaction
conditions without gold, a control experiment was performed and
Fig. 8 shows the DRIFT spectra collected over the catalyst Au/TN
during a typical CO + O2 + H2O co-adsorption experiment. The ex-
periment was done under the flow of the adsorbates (CO + O2)
◦
◦
the result is illustrated in Fig. 9.
at 70 C. All the spectra were recorded at 70 C. Spectra were
recorded at 13 and 28 min during time-on-line; after which 100
μL of water was introduced into the feed and subsequently spectra
were recorded as a function of time. The quantification of the CO2
produced in this particular experiment is shown in Fig. 4.
−1
Indeed the absence of the peak at 1290 cm
(Fig. 9) con-
firmed that gold was a necessary ingredient to form the bicar-
−1
bonate species (see Fig. 10). The vibration band at 1329 cm
in
Figs. 9 and 10, due to unidentate carbonate species [27], was not
affected by the presence of water; while that of the bicarbonate
(1290 cm−1) was almost completely removed by water (Fig. 10a).
What was concluded from Figs. 4 and 8 is that as previously
reviewed [17], water had a promoting effect on the Au/TN cata-
lyst system. Fig. 4 showed a doubling increase in amount of CO2
produced when H2O was introduced and the same promoting ef-
fect was shown in Fig. 8, characterized by the increased intensity
4. Discussion
of the V3 vibration of linear, non-adsorbed (i.e. gas-phase) CO2 at
The high activity of gold catalysts for the oxidation of CO is well
documented [17]. However, activity alone is not the only criterion
that must be fulfilled for a suitable catalyst for CO oxidation [28].
Long term stability is also a crucial feature.
Fig. 3 illustrated the deactivation of the Au/TN catalyst while
Fig. 4 showed the beneficial effect of water in the oxidation of CO.
Also in Fig. 4 it was shown that as long as there was moisture in
−1
2361 cm
.
In addition, in Figs. 4 and 8, it is shown that as soon as the
amount of water decreases in the feed/surface the catalysts suf-
fered rapid deactivation. Fig. 8 illustrates that the deactivation
coincided with a rapid build up of carbonaceous species with a vi-
bration band at 1290 cm−1. This carbonate species has previously

T.A. Ntho et al. / Journal of Catalysis 261 (2009) 94–100
99
the feed, deactivation was inhibited. The observed deactivation in
Fig. 3 was correlated to the growth of the bicarbonate band [1290
cm−1]. As shown in Fig. 4, it took about one hour for water to
effect the promotion of CO oxidation and that could be related
to how long water takes to decompose the bicarbonate species
(Scheme 1). Comparison of Figs. 5a and 6a shows that the TPD
data for the catalyst do not show any signs of the desorption of
co-adsorbed (CO + O2) or carbonate decomposition products from
the support alone. Rather the TPD data are dominated by the des-
orption of CO2, formed from, for example CO +O2 reaction on gold
sites.
the formation of CO2 did not go through the bicarbonate as an
intermediate during the oxidation of CO but rather, CO2 was a re-
actant in the formation of the bicarbonate. There are two obvious
pathways by which CO2 can react with gold (Au–OH) to form the
gold-associated bicarbonate species (responsible for deactivation):
(i) The first possibility is the reaction of CO2 (g) with Au–OH in a
Rideal-Eley type mechanism,
Also the CO2-TPD data (Figs. 6 and 7a) showed desorption
◦
◦
peaks (153 C and 162 C, respectively) due to some CO2 species
whose formation was directly linked to the presence of Au, as they
were not formed on the pure support (Fig. 5) under reaction con-
ditions. We deduce that the CO2 species seen from the DRIFTS
(ii) The second possibility could be the reaction of CO2 adsorbed
◦
data (1290 cm−1) and CO2-TPD data (153 C) are the same species
on Lewis acid sites (Ti4+
[Au(p)],
) surrounding the gold particles
which lead to the deactivation of the catalysts. This conclusion was
reached because the formation of these CO2 species coincided with
the presence of Au on the support. The gradual accumulation of
the CO2 species (bicarbonate) leads to the blockage/occupation of
active sites on the catalysts.
To understand the mechanism of deactivation one needs to first
look at the proposed reaction mechanisms. There has been much
debate concerning the nature of the active site for catalysts [17].
More than a decade ago, a reaction mechanism was proposed for
Au–TiO2 involving the effect of water in the CO oxidation reaction
[29]. The mechanism involved the dissociative adsorption of water
on the TiO2 surface, followed by the formation of formate species
(Au–COOH); which subsequently decomposes to CO2 and H2. How-
ever, this mechanism has recently been disputed [30]. Schubert
et al. [22] proposed that water transforms the carbonate species
formed during the oxidation of CO into less thermally stable bicar-
bonate species. They suggested that in the presence of water these
bicarbonate species are possibly reaction intermediates.
The above mechanisms for the formation of the observed bicarbon-
ate may account for the deactivation of the Au/TN catalyst during
CO oxidation; this reaction would consume the OH groups on Au
which have been reported as essential in the mechanism of CO ox-
idation [17]. The above hypothesis of CO2 being able to chemisorb
on gold nanoparticles is not unique. Based on IR studies [22,33],
evidence was provided which pointed to the fact that CO and CO2
competed for adsorption on the same site on Au surfaces. However,
we are not proposing that the adsorption of carbon dioxide on gold
is so strong that each product CO2 molecule undergoes adsorption.
A simple calculation based on our experimental conditions shows
that this is unlikely to occur. For the feed gas, 40 ml/min gas, con-
taining 5% CO with an average conversion of 10 over the first 10 h
will result in 3.2 × 1020 molecules CO converted per hour. 0.48%
Au on 100 mg catalyst with an average clusters size of 4.5 nm
with on 5% active gold atoms on the gold titania interface will
result in 7.6 × 1016 active gold sites. In other words more than
4000 molecules of CO will be converted per active gold site per
hour and thus more than 4000 molecules of CO2 will be formed
per hour per active gold site (assuming the entire interface active
gold has the same performance). Nevertheless CO2-mediated deac-
tivation appears to be occurring. To the best of our knowledge, this
is the first time that DRIFTS and CO2-TPD have been combined to
show that CO2 is adsorbed on the gold-nanoparticles or support in
the form of bicarbonate to the detriment of the oxidation of CO.
From our work, what we found out is that the formation of bi-
carbonate species is not an intermediate step but rather a product
of the CO2 produced during CO oxidation. The bicarbonate species
−1
detected at 1290 cm
(Figs. 8 and 10a) was very stable and only
◦
desorbed at 153 C (Fig. 6). The thermal stability of the bands due
to bicarbonate has been also reported elsewhere [31]. In the pres-
ence of water the bicarbonate species (1290 cm−1) did not form
(Fig. 10) under reaction conditions. Our CO2-TPD results implied
that water possibly reacted with and removed the already formed
bicarbonate species and prevented its further formation and in
the process promoted the reaction and prevented deactivation. The
possible reactions of how water (OH) might remove the gold as-
sociated bicarbonate and hence reactivate and promote the oxida-
tion of CO are outlined in Scheme 1. Schumacher et al. [32] have
also proposed that water reacts with the poisoning carbonate-like
species (possibly via CO3 +H2O → CO2 +2OH) preventing catalysts
deactivation.
Acknowledgments
Au–CO3H–Ti + 2HO → Au–OH + Ti–CO3 + H2O (Reactivation),
Au–O–Au + H2O → 2Au–OH,
Au–OH + CO → Au–H + CO2(g) (Reaction),
Au–O + Au–H → Au–OH + Au.
(1.1)
(1.2)
(1.3)
(1.4)
The authors are grateful to project AuTEK (Mintek and Anglo-
gold Ashanti) for financial support. The work was further facilitated
by the award of a grant under the Royal Society (London) and Na-
tional Research Foundation (South Africa) Science, Engineering and
Technology Programme.
Scheme 1. Possible reactions of how water might reactivate and promote the oxida-
tion of CO on Au/TN catalyst system.
References
[1] M. Haruta, T. Kobayashi, H. Sano, N. Yamanda, Chem. Lett. 2 (1987) 405.
[2] M. Haruta, S. Tsubota, T. Kobayashi, H. Kayeyama, M.J. Genet, B. Delmon, J.
Catal. 144 (1993) 175.
[3] F. Boccuzzi, A. Chiorino, S. Tsubota, M. Haruta, J. Phys. Chem. 100 (1996) 3625.
[4] M. Daté, M. Okumora, S. Tsubota, M. Haruta, Chem. Int. Ed. 43 (2004) 2129.
Fig. 7 shows a very interesting result in that it shows the for-
mation of the gold-associated CO2 species (bicarbonate) by simply
passing 1% CO2 (balance N2) over the catalyst. These species did
not form on the pure support (Fig. 5). This result suggested that

100
T.A. Ntho et al. / Journal of Catalysis 261 (2009) 94–100
[5] M.S. Chen, D.W. Goodman, Science 306 (2004) 252.
[19] J.C. Lavalley, Catal. Today 27 (1996) 377.
[6] P. Konova, A. Naydenov, C. Venkov, D. Mehandjiev, D. Andreeva, T. Tabakova, J.
Mol. Catal. A 213 (2004) 235;
M. Manzoli, A. Chiorino, F. Boccuzzi, Appl. Catal. B 52 (2004) 259.
[7] M.A. Sachez-Castillo, C. Couto, W.B. Kim, J.A. Dumesic, Angew. Chem. Int. Ed. 43
(2004) 1140.
[8] G.C. Bond, D.T. Thompson, Gold Bull. 33 (2000) 41.
[9] M. Haruta, Gold Bull. 37 (2004) 27.
[10] K. Costello, M.C. Kung, H.S. Oh, Y. Wang, H.H. Kung, Appl. Catal. A 232 (2002)
159.
[11] L.M. Molina, B. Hammer, Appl. Catal. A 291 (2005) 21.
[12] Z.P. Lin, X.Q. Gong, J. Kohanoff, C. Sanchez, P. Hu, Phys. Rev. Lett. 91 (2003)
266102.
[13] N. Lopez, T.V.W. Janssens, B.S. Clausen, Y. Xu, M. Mavrikakis, T. Bligaard, J.K.
Narskow, J. Catal. 223 (2004) 232.
[14] C. Lemise, R. Meyer, S. Shaikhutdinov, H.J. Freund, Angew. Chem. Int. Ed. 43
(2004) 118.
[15] M.S. Chem, D.W. Goodman, Science 306 (2004) 252.
[16] B. Yoon, H. Hakkinen, U. Landman, A.S. Worz, J.M. Antonetti, S. Abbet, K. Judai,
Science 307 (2005) 403.
[20] S.T. Daniells, M. Makkee, J.A. Moulijn, Catal. Lett. 100 (2005) 39.
[21] N.M. Gupta, A.K. Tripathi, Gold Bull. 34 (4) (2001) 120.
[22] N.M. Schubert, A. Venugopal, M.J. Kalich, V. Plazk, R.J. Behm, J. Catal. 222
(2004) 32.
[23] D.V. Bavykin, A.A. Lapkin, P.K. Plucinki, L. Torrente-Murciano, J.M. Friedrich, F.C.
Walsh, Top. Catal. 39 (2006) 151.
[24] L.C. Sikhwivhulu, N.J. Coville, T. Ntho, M.S. Scurrell, Catal. Lett. 123 (2008)
193.
[25] T. Kasuga, M. Hiramatsu, A. Hoson, T. Sekino, K. Niihara, Langmuir 14 (12)
(1998) 3160.
[26] J.R. Bargar, J.D. Kubicki, R. Reitmeyer, J.A. Davis, Geochim. Cosmochim. Acta 69
(2005) 1527.
[27] H. Knozinger, Adv. Catal. 25 (1976) 235.
[28] F. Moreau, G.C. Bond, Catal. Today 114 (2006) 362.
[29] D. Andreeva, V. Idakiev, T. Tabakova, A. Andreev, R. Giovanoli, Appl. Catal. 134
(1996) 275.
[30] M.A. Debeila, R.P.K. Wells, J.A. Anderson, J. Catal. 239 (2006) 162.
[31] M. Maciejewski, P. Fabrizioli, J. Grunwaldt, O.S. Becker, A. Baiker, Phys. Chem.
Chem. Phys. 3 (2001) 3846.
[17] G.C. Bond, C. Louis, D.T. Thompson, Catalysis by Gold, Imperial College Press,
London, 2006.
[18] A.M. Visco, F. Neri, G. Neri, A. Donato, C. Milone, S. Galvagno, Chem. Phys. 1
(1999) 2869.
[32] B. Schumacher, Y. Denkwitz, V. Plzak, M. Kinne, R.J. Behm, J. Catal. 224 (2004)
449.
[33] M.A. Bollinger, M.A. Vannice, Appl. Catal. B 8 (1996) 417.