Role of support in the oxidation of acetylene over gold catalysts

Article abstract of DOI:10.1016/j.jcat.2009.10.011

The influence of the carrier (CeO₂, TiO₂ or ZrO₂) on the reaction of acetylene oxidation over a 2% gold catalyst has been investigated in the presence or in the absence of CO. XPS reveals that gold is metallic in every case and the percentage of gold on the surface is always higher than that in the bulk confirming the wide dispersion as evaluated by TEM. There is a significant effect of the support on the activity in the absence of CO, but the presence of CO affects the reactivity differently according to the support. In a mixture of CO and C₂H₂, the oxidation of CO is systematically inhibited by C₂H₂ whereas the behaviour is completely different for the oxidation of C₂H₂. CO has no effect upon Au/CeO₂, acts as an inhibitor with Au/TiO₂ and as a promoter with Au/ZrO₂. Kinetic study reveals that the order towards C₂H₂ remains unchanged whereas the order towards O₂ clearly depends upon the support. These effects on activity and kinetics may result from various factors discussed in this paper.

Full text of DOI:10.1016/j.jcat.2009.10.011

Journal of Catalysis 269 (2010) 26–32
Contents lists available at ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
Role of support in the oxidation of acetylene over gold catalysts
*
Yassine Azizi, Corinne Petit, Véronique Pitchon
Laboratoire des Matériaux, Surfaces et Procédés pour la Catalyse, UMR 7515 du CNRS ECPM, 25, rue de Becquerel, 67087 Strasbourg, France
a r t i c l e i n f o
a b s t r a c t
Article history:
The influence of the carrier (CeO2, TiO2 or ZrO2) on the reaction of acetylene oxidation over a 2% gold cat-
alyst has been investigated in the presence or in the absence of CO. XPS reveals that gold is metallic in
every case and the percentage of gold on the surface is always higher than that in the bulk confirming
the wide dispersion as evaluated by TEM. There is a significant effect of the support on the activity in
the absence of CO, but the presence of CO affects the reactivity differently according to the support. In
a mixture of CO and C2H2, the oxidation of CO is systematically inhibited by C2H2 whereas the behaviour
is completely different for the oxidation of C2H2. CO has no effect upon Au/CeO2, acts as an inhibitor with
Au/TiO2 and as a promoter with Au/ZrO2. Kinetic study reveals that the order towards C2H2 remains
unchanged whereas the order towards O2 clearly depends upon the support. These effects on activity
and kinetics may result from various factors discussed in this paper.
Received 23 July 2009
Revised 9 October 2009
Accepted 9 October 2009
Available online 25 November 2009
Keywords:
Acetylene oxidation
Gold
Au catalyst
Automotive exhaust
Kinetics
Ó 2009 Elsevier Inc. All rights reserved.
1. Introduction
of CO in the reaction as had been postulated previously by other
groups employing a steady-state regime [11,12]. Of note is the
Acetylene oxidation is a reaction widely studied as a model for
the processes occurring in the combustion of hydrocarbon (HC)
molecules [1–3]. The chemistry yielding products of this reaction,
such as *CHCO, *CHO, C2H* and *CH2 radicals, is of great importance
to the elucidation of mechanisms implied in the flame combustion
of HC. The chemistry of acetylene oxidation has been the subject of
an extensive range of publications, over several decades [4,5]. New
developments in this area have occurred due to studies in hetero-
geneous catalysis dedicated to automotive exhaust gas depollu-
importance of ceria, which through an oxygen interacting with
acetylene adsorbed on the noble metal (Pt, Rh) exerts a strong
influence upon this direct bi-functional reaction.
The work of Johnson and Krieger [13] on acetylene oxidation on
silver nitrate revealed that the heats of combustion of acetylene and
carbon monoxide are not very far apart (310 and 240 kcal/mol,
respectively), and concluded that if any carbon monoxide is formed,
it must be a short-lived intermediate. Also proposed was that the
reactivity is a function of the partial pressure of C2H2 and required
a critical concentration of acetylene below which the reaction
would not occur. This suggests that the reaction initiates at a high
coverage of acetylene, a pre-requisite necessary for the formation
of surface combinations of acetylene, such as diacetylene and tri-
acetylene, species more easily oxidised than acetylene itself. For
these dimers or trimers on silver nitrate, oxygen would have no di-
rect influence on the rate, implying a zero order towards O2.
A fundamental study of acetylene oxidation on the (1 1 1) plat-
inum surface was undertaken by Burnett et al. [14] which revealed
that the adsorption of acetylene occurs through a strong interac-
tion. Indeed, during fuel combustion within an engine,
a
proportion of the HC remains unburned and is found in the exhaust
along with other incomplete combustion products such as CO, NO
and H2 or other oxygenated compounds such as formaldehyde.
These unburned residues are either intact or have been partially
transformed by pyrolysis processes to produce short chain HC
(C1–C2) or aromatic compounds (C6–C9). Among the unburned HC
are found methane, ethylene and acetylene, this latter being refrac-
tory to catalytic oxidation in the post-combustion device [6–8].
Acetylene is characterised by its very strong adsorption on the sur-
face of noble metals which is due to its very high electron-donor
character. There are few reports on the kinetics of acetylene oxida-
tion in the literature despite a prevailing interest in the study of
this reaction. Harmsen et al. [9,10], using a mixture of acetylene
and carbon monoxide on Pt/Rh/CeO2/d-Al2O3, revealed that carbon
monoxide is an intermediate product of the reaction. The experi-
ments, run under transient conditions, confirmed the involvement
tion between the unsaturated
p bonding and the Pt surface. A sim-
ilar type of interaction was observed between the
p–r bonding of
ethylene and propylene on Pt (1 1 1) [15].
To our knowledge, there are as yet no fundamental studies of
acetylene oxidation on gold catalysts with the exception of our
own in which gold was proposed as a potential catalyst noble me-
tal for use in the oxidation of hydrocarbon at low temperature, the
so-called ‘‘low light-off depollution” of exhaust gas [16]. However,
studying the oxidation of acetylene is particularly important in this
process because of its electron-donor character. Oxidation of
* Corresponding author.
E-mail address: pitchon@chimie.u-strasbg.fr (V. Pitchon).
0021-9517/$ - see front matter Ó 2009 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2009.10.011

Y. Azizi et al. / Journal of Catalysis 269 (2010) 26–32
27
acetylene in a group of gasses is difficult, despite its inhibition of
the oxidation of other compounds, particularly carbon monoxide
whose oxidation is fundamental for the viability of the process of
low light-off oxidation. We have therefore studied the oxidation
of acetylene alone, or in the presence of CO. Acetylene behaves dif-
ferently towards oxygen in the presence of CO and equally in the
presence of ceria, as described by Harmsen et al. [10]. According
to the authors ceria is involved in the reaction mechanism, a direct
bi-functional interaction between acetylene adsorbed on the noble
metal and oxygen from ceria. This experiment reveals the impor-
tant role played by the support in the reaction and we therefore ex-
tended the present study of reactions on gold on different supports
to assess their possible effect, or influence upon this noble metal.
(1486.6 eV). The aluminium anode was operated at an accelerating
voltage of 15 kV, 15 mA, 20 Volts. Base pressure in the analysis
chamber was maintained within the range of 5 ꢁ 10ꢀ9 mbar.
2.4. Total oxidation of CO and/or acetylene
Both reactions were performed using the same apparatus. Flow
regulation was achieved using Tylan flow controllers. The temper-
ature was monitored and controlled by a Eurotherm system. The
experiment was repeated several times in order to measure the
reproducibility. Before each test, the catalyst was pre-treated at
300 °C in air for 2 h.
In the case of CO oxidation, the reaction was performed at
atmospheric pressure using a fixed bed quartz micro-reactor
packed with 50 mg of catalyst diluted in cordierite and a gas mix-
ture of 1.5% CO and 5% O2 in He with a total flow rate of
75 cm3 minꢀ1 (GHSV ꢂ 35,000 hꢀ1). Analysis for both CO and CO2
was performed by a Rosemount Infrared analyser.
2. Experimental
2.1. Catalyst preparation
For acetylene oxidation, the reaction was performed at atmo-
spheric pressure with 0.15 g of catalyst and a gas mixture contain-
ing 0.2% of acetylene and 5% of oxygen in helium with a flow rate
adjusted to 75 cm3 minꢀ1. The analysis of C2H2 was performed by
FID gas (DB-Wax column, 2 m. long and of 2.5 mm inner diameter),
CO2 production was monitored continuously using the on-line IR
Rosemount analyser as for CO oxidation. The experiment was re-
peated twice in order to measure the reproducibility. Before each
test, the catalyst was treated ex situ at 300 °C in air for 4 h.
2.1.1. Supports
The simple or mixed supports TiO2, ZrO2 and CeO2 were pre-
pared by the precipitation of metallic hydroxides. The starting re-
agents (Strem chemicals), respectively, isopropoxide titanium
(IV) (Strem chemicals), n-propoxide zirconia (IV) and cerium ni-
trate, were dissolved in ethanol and then precipitated by an ammo-
nia solution of concentration 4 mol Lꢀ1. The solid obtained was
filtered, washed several times and dried at 100 °C overnight and fi-
nally calcined at 750 °C for TiO2 and ZrO2 and at 450 °C for CeO2.
2.5. Kinetic study
2.1.2. Gold impregnation
The method of preparation is based upon the direct anionic ex-
change (DAE) of the gold species with the hydroxyl groups of the
support as reported in a previous publication [17]. An aqueous solu-
tion of HAuCl4 of concentration 10ꢀ4 mol Lꢀ1 (pH 3.5) was prepared
in order that a final Au loading of 2 wt% be obtained. The solution
was heated to 70 °C and the support (TiO2, ZrO2 or CeO2), was
added. The resulting slurry was filtered. A washing procedure with
ammonia was then applied in order to remove the chlorine residues
from the solution, from the gold chloro-hydroxy complexes and/or
the adsorbed chlorides on the support. The catalyst was submerged
for 1 h in an ammonia solution of concentration 4 mol Lꢀ1. The solid
was then filtered and washed with water, dried overnight in an
oven at 120 °C and calcined in air at 300 °C for 4 h.
In order to determine the kinetics of the reaction, a study of the
variationofO2 orC2H2 pressurewasundertaken. Thekineticorderto-
wards O2 or C2H2 was determined at low conversion in order to avoid
any diffusion phenomenon. The reaction rate can be expressed as
m
n
r ¼ k ꢁ ½PC
ꢃ
ꢁ ½PO
ꢃ
ðmol minꢀ1 gcata
Þ
2H2
2
where m and n represent the order towards acetylene and oxygen,
respectively.
The partial pressure of O2 was varied while maintaining the par-
tial pressure of C2H2 constant and vice versa. When possible, i.e.
when the conversion remained low over a broader range of tem-
perature, the orders were measured at different temperatures.
The level of concentration of acetylene was varied between 0.05%
and 0.4% and the oxygen concentration was varied according to
the required O2/HC ratio. The experimental conditions for the
kinetics measurements are reported in Table 1.
2.2. Chemical analysis, BET and microscopy
The chemical analysis of Au and Cl was performed by induc-
tively coupled plasma atom emission spectroscopy at the CNRS
Centre of Chemical Analysis (Vernaison, France). The detection lim-
it was 150 ppm for Cl. The analyses were carried out following the
calcination of the catalyst. The Au weight loading is expressed as
the weight of Au (g) per weight of calcined catalyst (g).
The surface area (by BET method) is determined by the adsorp-
tion and desorption of nitrogen at the temperature of liquid nitro-
gen (77 K). The specific surface area measurements were
performed in an automated device (Tristar 3000) following several
hours of degassing at 250 °C. The average masses used for these
measurements varied between 100 and 200 mg.
3. Results
3.1. Effect of support on the size and on the chemical state of gold
3.1.1. Chemical analysis
The BET surface along with the final gold concentration is re-
ported in Table 2, the chlorine concentration was always inferior
Table 1
Experimental conditions for the kinetics measurements of apparent orders of
reaction.
The catalysts were characterised by Transmission Electron
Microscopy (TEM) Topcon EM002B to determine the gold particle
size distribution.
Order/O2
Order/C2H2
Mass catalyst (mg)
[O2] (%)
50 or 200a
0.5, 1, 2 and 3
0.2
50 or 200a
1
0.05, 0.1, 0.2 and 0.4
75
0.5, 1.0, 2.0 and 4.0
[C2H2] (%)
Flow rate (cm3/min)
Partial pressure (mbar)
75
2.3. Xps
5, 10, 20 and 30
a
As the conversions were lower on Au/ZrO2 a larger mass of catalyst was used in
order to work in the same range of temperature as for Au/CeO2 and Au/ZrO2.
X-ray photoelectron spectra (XPS) were acquired on a Multi-lab
2000 spectrometer (Thermo VH Scientific), using Al K
a radiation

28
Y. Azizi et al. / Journal of Catalysis 269 (2010) 26–32
Table 2
Gold weight as percentage, surface concentration, BET surface and average particle
size of the catalysts.
Catalyst
Au (%)a
Surface Au (%)b
SBET (m2 gꢀ1
)
Particle size (nm)
(c)
Au/CeO2
Au/TiO2
Au/ZrO2
1.93
1.56
1.32
1.91
7.56
2.1
47
8
9
2–3
3–4
Bimodal. 3 and 15
a
Determined by chemical analysis.
Calculated from XPS data.
b
to detection limits by atomic absorption, i.e. less than 150 ppm.
The method of preparation consists of an exchange of the hydroxyl
group or the chloride ion of the gold precursor HAuCl4 which is
partially hydroxylated while put in aqueous solution. For CeO2,
the amount of gold deposited is close to the theoretical value of
2%. This is explained by the fact that the exchange between the
precursor and CeO2 is very easy. Although the surfaces for TiO2
and ZrO2 were relatively small, deposition was possible though
the gold content was lower than an amount of the solution equiv-
alent to yielding 2% gold. This loss was nevertheless comparatively
modest for both Au/TiO2 (22% loss) and Au/ZrO2 (34% loss). The dif-
ficulty to deposit the total amount of gold arises from the fact that
each support has a different type of interaction with HAuCl4 in
solution [18]. However, the amount of gold deposited remains high
in comparison with that obtained using the deposition–precipita-
tion method [19,20]. The technique of preparation by anionic ex-
change, originally developed for alumina, has proved very
successful with a whole range of supports including the one used
in the present study [21]. This is also underlined by the average
particle size obtained, as measured by electronic microscopy (Ta-
ble 2). On TiO2 and CeO2, the particles are small, are homoge-
neously dispersed and have an average size within the range of
3–4 nm and 2–3 nm, respectively. The gold particles examined
on Au/ZrO2 are larger, formed by the agglomeration of smaller gold
crystals and with a broader size distribution. Since it was impossi-
ble to obtain a number of particles sufficient to build a truly repre-
sentative histogram, only the size range (between 3 and 15 nm) is
given in Table 2.
(b)
(a)
75
80
85
90
95
Binding Energy/ eV
Fig. 1. X-ray photoelectron spectra of the energy level 4f of gold: (a) Au/ZrO2, (b)
Au/TiO2 and (c) Au/CeO2.
3.1.2. Chemical state: XPS
The de-convolution of XPS spectra following the ex situ calcina-
tion allowed the detection, assuming their presence, of multiple
oxidation states, Au0 and Au3+ [22]. In the present study, the detec-
tion of a sole doublet indicated the presence of gold in a metallic
state, exclusively [23,24]. For the three catalysts, as described in
Fig. 1, the values of binding energy ( 0.15 eV) are Au/TiO2: [Au
4f7/2 83.4 eV, Au 4f5/2 87.1 eV]; Au/ZrO2 [Au 4f7/2 83.3 eV, Au 4f5/2
87.0 eV] and Au/CeO2, [Au 4f7/2 83.3 eV, Au 4f5/2 87.0 eV]).
In order to calculate the surface ratio, the spectra were also de-
convoluted in the region corresponding to the binding energy of
the elements comprising the support. Fig. 2a presents the spectrum
corresponding to level 3d of zirconium which is characterised by a
doublet located at 181.9 0.15 and 184.2 0.15 eV which accord-
ing to the literature corresponds to Zr4+ [25]. In Fig. 2b, the spec-
trum of the 2p energy level of Ti is shown with a doublet at
458.3 0.15 and 464 0.15 eV characterising Ti4+ [26]. The inter-
pretation of the cerium spectrum is more complex but has been
extensively described in the literature. As shown in Fig. 2c, the
main peaks are at the binding energies which conform to the val-
ues reported in the literature [27,28]. Evaluating these values led
177
180
183
186
189
Energie de liaison, eV
Fig. 2a. X-ray photoelectron spectrum of the energy level 3d of zirconium in Au/
ZrO2.
to the conclusion that there was a non-negligible amount of Ce3+
,
representing 13.2%, while the remainder was Ce4+
.
452
456
460
464
468
472
Following this de-convolution, a quantitative evaluation of the
percentage of gold on the surface could be made and compared
with the bulk values as obtained by chemical analysis, which are
Energie de liaison, eV
Fig. 2b. X-ray photoelectron spectrum of the energy level 2p of titanium in Au/TiO2.

Y. Azizi et al. / Journal of Catalysis 269 (2010) 26–32
29
100
U’’’
V’’’
U
U’ U’’
80
60
40
20
V
V’
V’’
Au/TiO2
Au/ZrO2
Au/CeO2
860
870
880
890
900
910
920
930
Energie de liaison, eV
0
0
50
100
150
200
250
300
Fig. 2c. X-ray photoelectron spectrum of the energy level 3d of cerium in Au/CeO2.
Temperature (°C)
Fig. 3. Conversion of CO vs. T °C for Au/TiO2, Au/CeO2 and Au/ZrO2.
reported in Table 2. In the case of Au/CeO2 and Au/ZrO2, the values
of the surface concentration are close to those in the bulk (2.1 in-
stead of 1.32 pour Au/ZrO2 and 1.91 compared with 1.93 for Au/
CeO2). Conversely, for Au/TiO2, the corresponding value is almost
5 times higher on the surface than on the bulk (7.56 compared
with 1.56), which confirms the very broad dispersion of gold on
titania as observed by TEM. The quantitative analysis of XPS data
is not straightforward as XPS is not strictly surface specific. In this
case, the difference in the surface values for gold concentration on
Au/TiO2 and Au/CeO2 may result from a difference of porosity,
although particle shape effect could not be excluded.
Au/TiO2
Au/ZrO2
Au/CeO2
3.2. CO oxidation
0
50
100
150
200
250
300
350
400
450
The catalysts were tested and their performance was compared
Temperature (°C)
for the CO oxidation reaction. The rates of the reaction expressed in
ꢀ1
mol CO min
g
Au
ꢀ1 at 50 °C are displayed in Table 3 while the re-
Fig. 4. Conversion of C2H2 vs. T °C for Au/TiO2, Au/CeO2 and Au/ZrO2.
sults of conversion vs. temperature are presented in Fig. 3.
Au supported on both TiO2 and CeO2 proves very active with a
conversion of c.a. 40% at room temperature and with total conver-
sion reaching at c.a. 75 °C. Therefore the rates of these reactions
must be similar. The behaviour of Au/ZrO2, however, is different
with the catalyst being less active. The 100% conversion was at-
tained at c.a. 300 °C. As the amount of gold is comparable for the
three catalysts, this difference in reactivity results principally from
a difference in particle size though a support effect should not be
excluded since such effects are known to be among the key param-
eters in the achievement of an excellent activity from gold cata-
lysts in the oxidation of CO. This effect has been recently
rationalised in terms of support reducibility and nature of the Au
support interaction [29].
it has been reported as a transient primary product of reactions
during experiments on Pt and on Pt–Rh [9,11,14]. In our study,
the subsequent conversion of CO to CO2 must be a very fast stea-
dy-state event. Moreover, the marked ability of gold to catalyse
the CO reaction may prevent CO observation in the gas phase,
although its formation as an adsorbed intermediate should per-
haps not be ruled out.
The most active catalyst is Au/CeO2 with a T50 value of 170 °C
and contrary to the oxidation of CO this catalyst is markedly more
active than Au/TiO2, while Au/ZrO2 remains the less active for the
oxidation reaction. As particle size for the first two catalysts are
equivalent it is clear that this cannot be the only parameter influ-
encing the activity. Indeed, as the reaction occurs at a temperature
higher than that of CO oxidation, other factors such as oxygen
mobility and the nature of the adsorbed molecular oxygen species
may be involved [30]. Another important factor is the fraction of
gold surface which is active. Recently, Hutchings et al. have re-
ported that active gold species in CO reaction have a peculiar bi-
layer cluster which contains only c.a. 10 atoms and are therefore
not detectable by conventional TEM [31]. The oxygen storage abil-
ity of ceria is well-known and this, allied to the ease with which
the surface oxygen can be reduced, may be an influence upon the
reaction, in contrast to the non-reducible TiO2 and ZrO2. Indeed,
the reducibilities of the Au-supported catalysts were checked by
TPR (5% H2, not shown). For TiO2 and ZrO2 no hydrogen consump-
tion was obtained upto 400 °C, whereas c.a. 42% of Ce4+ is reduced
to Ce3+ between 130 and 400 °C.
3.3. Acetylene oxidation
The conversion of acetylene vs. temperature for Au supported on
CeO2, TiO2 or ZrO2 is presented in Fig. 4. The activity varies according
to the nature of the support in the following order Au/CeO2 > Au/
TiO2 > Au/ZrO2. The rates of reaction at 210 °C are reported in Table
3, along with the temperature of 50% conversion (T50).
The only observed product of reaction was CO2, with H2O not
being measured. The presence of CO was never detected though
Table 3
Rate in CO oxidation at 50 °C (alone) and rate in C2H2 oxidation at 210 °C (alone) and
T50 for Au/CeO2, Au/TiO2 and Au/ZrO2.
Catalyst
CO
r (mol h
C2H2
C2H2
T50 (°C)
The rates of reaction expressed in Table 3 are c.a. 100 times low-
er than those obtained for CO oxidation even though the activities
were compared at different temperatures. To understand further
the kinetics phenomenon implied in both reactions, a study of
the simultaneous oxidation of CO and acetylene was undertaken.
ꢀ1
ꢀ1
ꢀ1
g
)
r (mol hꢀ1 gAu
)
Au
Au/CeO2
Au/TiO2
Au/ZrO2
1.91
1.84
0.08
6.15 ꢁ 10ꢀ2
5.33 ꢁ 10ꢀ3
1.77 ꢁ 10ꢀ3
170
265
340

30
Y. Azizi et al. / Journal of Catalysis 269 (2010) 26–32
Table 4
3.3.1. Effect of CO addition in the feed
Kinetics data for acetylene oxidation: orders of reaction towards C2H2 and O2
(without CO).
The catalytic behaviour of each catalyst in the C2H2 oxidation
was studied by adding 1% CO to the feed while maintaining fixed
the flow rate and an equivalent mass of catalyst, i.e. 75 cm3 minꢀ1
and 0.15 g of catalyst. The conversions in the presence of CO, as
represented in Fig. 5a–c, were systematically compared with those
obtained in the absence of CO in the feed.
T (°C)
Order C2H2
Order O2
Au/CeO2
Au/ZrO2
Au/TiO2
100
130
270
300
300
330
360
0.2
0.2
ꢀ0.3
ꢀ0.2
0.0
0.0
0.0
0.5
0.5
0.8
0.7
0.8
In the case of Au/CeO2, acetylene reacts in a similar way across
the temperature domain in the presence or absence of CO (Fig. 5a).
The situation is somehow different with Au/TiO2; the presence of
CO delays slightly the reaction of acetylene combustion, as proven
by an increase in the T50 value of c.a. 80 °C, when compared with
the value observed when acetylene is the sole gas in the combus-
tion mixture (Fig. 5b). The case of Au/ZrO2 is also different from
the two previous ones. Indeed, the value of the T50 is decreased
by c.a. 60 °C, revealing a positive effect of CO (Fig. 5c). For all the
three catalysts, the CO oxidation reaction starts only after the
C2H2 oxidation. This delay in reactivity could signify that CO is ad-
sorbed onto the same sites as C2H2 only when these sites are va-
cated by the acetylene molecules or from the latter’s reaction
products. Here again, a support effect in the competitive reaction
between CO and C2H2 is decisive; therefore, this support effect
was studied in a great detail by a kinetics analysis in order to find
a possible explanation for its role of the support in the reaction.
0.0
0.0
3.3.2. Kinetics study
The activity measurements were completed by a kinetic study
undertaken by varying the partial pressure of acetylene at constant
oxygen pressure and vice versa in order to determine the apparent
order of reaction for Au supported on CeO2, ZrO2 and TiO2, respec-
tively. Each order was measured at different temperatures in the
kinetic regime; the orders for both O2 and C2H2 are not tempera-
ture dependent in these sets of experimental condition. In a num-
ber of cases a variation of 0.1 in unit was observed depending
upon the temperature, but was considered as being experimentally
insignificant, proving that mass transfer is not limiting. The orders
of reaction towards O2 or C2H2 are displayed in Table 4.
On all the catalysts, it must be stressed that acetylene and O2
may adsorb on different types of site. For Au/CeO2, at low temper-
ature, i.e. below 130 °C, and therefore at low conversion, there is no
variation in the conversion of acetylene with an increase of PO . In
100
2
this range, the order towards acetylene is slightly positive while a
zero order towards oxygen is obtained, suggesting that the rate
does not depend upon oxygen pressure. The situation is rather dif-
ferent for the other two catalysts and particularly for oxygen. The
order towards acetylene remains close to zero. Nevertheless, here
the variations are significant, in the case of ZrO2, the order towards
acetylene is slightly negative with a value of ꢀ0.2 suggesting that
during the course of the reaction the catalyst surface is covered by
strongly adsorbed acetylene. The order towards oxygen is positive
and equal to 0.5 which means that the rate is only dependent on
oxygen partial pressure and might suggest a dissociation of oxygen
on the surface. Finally, for Au/TiO2, a zero order towards C2H2 is ob-
served which means that the surface is covered by acetylene and
the order towards O2 is equal to 0.8 proving that the rate of reac-
tion is a function of the oxygen pressure and not of acetylene in
this set of experimental conditions.
(a)
80
60
40
20
0
C
C
C
C
22H2 (miixture)
OO (mixxture)
22H2
O
0
100
200
300
400
500
Temperature (°C)
100
80
60
40
20
0
(b)
C2H2 (mixture)
CO (mixture)
CO
4. Discussion
C2H2
The oxidation reaction of CO is systematically inhibited in the
presence of acetylene. However, this is not the case with acetylene
which behaves differently in the presence of CO according to the
support used. The behaviour could be summarised by the following
observations:
0
100
200
300
400
500
Temperature (°C)
100
80
60
40
20
0
- On Au/CeO2 there is no influence of CO in the C2H2 oxidation;
- the C2H2 oxidation is inhibited by the presence of CO on Au/
TiO2;
- the C2H2 oxidation is enhanced by the presence of CO on Au/
ZrO2.
(c)
CO (mixture)
C2H2 (mixture)
C2H2
This offers an indication of the mechanism of oxidation – the
adsorption of acetylene and CO most probably occurs on the same
active site. The adsorption is competitive in favour of acetylene
according to CO. The acetylene is strongly adsorbed to the surface
at room temperature and occupies all adsorption sites which do
not allow an adsorption of CO. However, it does provoke an in-
crease in the temperature required for acetylene conversion with
CO
0
100
200
300
400
500
Temperature (°C)
Fig. 5. Influence of CO on C2H2 conversion: (a) Au/CeO2, (b) Au/TiO2, (c) Au/ZrO2.

Y. Azizi et al. / Journal of Catalysis 269 (2010) 26–32
31
an increase of around 50 °C over a whole range of temperatures
compared with the conversion of acetylene alone. Once the oxida-
tion of acetylene is complete the oxidation of CO can begin. The de-
lay in the oxidation of CO indicates that CO could adsorb and react
only on those sites vacated by acetylene or its oxidation products.
Acetylene is characterised by its very strong adsorption on the
surface of noble metals [15], due to its high capacity as an electron
donor, this characteristic is verified by this experiment where inhi-
bition of CO oxidation is systematically observed in the presence of
C2H2. This strong adsorption may result from the acidic character
of the acetylene molecule as conferred by the sp hybridisation
which gives 50% of s-character to the molecule. A particular type
of affinity of acetylene for gold has been postulated by DFT calcu-
lations when comparing adsorption capacity of either acetylene or
ethylene by Segura et al. [32]. Our experience in acetylene hydro-
genation, indicates that this observation is valid in the presence of
both oxygen and hydrogen [33].
It is known that the reaction order of a surface-catalysed reac-
tion is strongly affected by the degree of coverage achieved by a
reactant. The variation of order found in the oxidation of the three
catalysts of this study may be the result of a difference in the
adsorption strength. A significant result is that the orders of reac-
tion in oxygen and acetylene depend upon the type of support. The
order of oxygen is never negative, while the order in HC is zero or
slightly positive for Au/TiO2 and Au/CeO2, respectively, and is
slightly negative for Au/ZrO2 indicating that, in this latter case,
HC is strongly bonded on the surface and therefore could reflect
a competitive adsorption with other reactant or other intermedi-
ate. It reacts with difficulty and could account for the compara-
tively reduced activity for this catalyst. However, only a small
variation of the HC order is observed. Although significant, (exper-
imental error being of 0.1), it does not seem that the difference in
the behaviour of the three catalysts results from a difference in the
acetylene adsorption and/or reaction in adsorbed phase. The differ-
ence in the orders of reaction towards oxygen is much more signif-
icant and indicates that the mode of adsorption and/or reaction
depends only on oxygen.
There is a particular interest in comparing the behaviour in the
presence of H2 where strongly adsorbed acetylene forms an acety-
lene residue and a hydrogen atom. This process depends mainly
upon the support whereas in the presence of hydrogen, adsorbed
acetylene tends to form carbonic species [34] which is specific of
dispersed gold [33]. The results proposed here underline a similar-
ity between the reactions of acetylene with either H2 or O2 and im-
plies a peculiar ability of gold to undergo a carbon–carbon rupture.
However, the role of oxygen remains predominant in the oxidation
reaction, the question being how and in which form?
A first effect observed is the role of the support which is evi-
denced by the superiority of CeO2 over the two other supports
which may result from the fact that ceria generally serves as a res-
ervoir for oxygen in the oxidation reaction, though this effect is
usually observed above 300 °C [35,36]. For Harmsen et al. [10],
the role of ceria is determinant for the mechanism they have pro-
posed which suggests that the oxidation occurs via a bi-functional
reaction between adsorbed acetylene and oxygen from the ceria.
Although realistic on a Pt–Rh catalyst, this mechanism cannot be
the sole possibility in our case, for while the reaction is promoted
with ceria, it is not necessarily the only form of active oxygen as
both Au/ZrO2 and Au/TiO2, although less active than ceria, never-
theless exhibit a relatively high activity when compared with other
metals such as Pt and Rh [8].
a higher reactivity; but what our results show is that the size is
not the only factor of importance but intriguingly it results directly
from the support itself, modifying the kinetics parameter. In partic-
ular, when the order towards oxygen changes this is an indication
that there is a different mechanism according to the support.
Although this work does not completely explain this effect, several
assumptions can be made; the electronic factor could be elimi-
nated since as in all cases gold is purely metallic, no ionic gold
was detected, even following an oxidation in air. The support effect
may result from a structuring influence of the support on the par-
ticle, either by epitaxy or by the creation of different particle
shapes which could subsequently produce geometrical effects. In-
deed previous work enabled us to discuss this type of effect on gold
particles supported on alumina, which in this case were preferen-
tially adopting a cubo-octahedric shape (packing) leading to an
interesting reactivity towards oxygen. The chemisorption of oxy-
gen does not occur on massive gold, but there are several refer-
ences in the literature which report that chemisorption and/or
dissociation of oxygen become feasible on small gold particles
[38–40], the location of oxygen from the support vacancies, on
the particle, at the interface being dependent on the support itself.
The recent work of Hutchings et al. emphasises the role of shape
and morphology in the case of gold catalysts. They have shown
that a fraction only of gold surface which is active, and that this ac-
tive fraction is mainly under the form of bilayer clusters that are
c.a. 0.5 nm in diameter with a number of atoms not exceeding
10, and therefore not detectable by classical microscopy [31]. Then
our work supports this idea, the support playing an essential role
in the formation of these very active particles.
5. Conclusions
The present study on acetylene oxidation has revealed several
features characteristic of gold catalysts. First of all, the oxidation
of acetylene which usually is a difficult reaction on metal becomes
quite feasible on gold due to the very strong interaction between
acetylene and gold nanoparticles. This confirms the extreme versa-
tility of gold catalysis for oxidation reactions as observed for other
contaminants such as CO and volatile organic compounds. The
activity is a function of the nature of the support with the following
order of activity: Au/CeO2 > Au/TiO2 > Au/ZrO2. The oxidation of CO
is systematically inhibited by the presence of C2H2; this indicates
that the adsorption competition is strongly in favour of C2H2. How-
ever, in the case of the reaction of C2H2 oxidation, the situation is
different: CO has no effect on Au/CeO2, the reaction is poisoned
for Au/TiO2 and promoted for Au/ZrO2. The reason lies in a different
kinetic behaviour, mainly for the mode of adsorption and/or reac-
tion of oxygen proven by a variation in the kinetic order, as there is
no difference in the order of reaction towards C2H2 for the three
catalysts. The effect of oxygen is not clearly elucidated but a
change of mechanism resulting from the kinetics is postulated
which might be the result of a different structuring of the gold par-
ticles according to the nature of the support.
References
[1] J. Peeters, Bull. Soc. Chim. Belg. 106 (1997) 337.
[2] J.M. Levy, Combust. Flame 46 (1982) 7.
[3] T. Koike, K. Morinaga, Bull. Chem. Soc. Japan 60 (1987) 4229.
[4] J.M. Hay, G.W. Norrish, Proc. Roy. Soc. Ser. 288 (1412) (1965) 17.
[5] J.M. Hay, D. Lyon, Proc. Roy. Soc. Ser. A 317 (1258) (1970) 1.
[6] I. Amon-Meziere, F. Castagna, M. Prigent, SEA SP-1094 (1995) 249.
[7] J.M. Bart, A. Pentenero, M.F. Prigent, ACS Symposium Ser. 495 (1992) 42.
[8] B. Mabilon, D. Durand, Ph. Courty, Stud. Surf. Sci. Catal 96 (1995) 181.
[9] J.M.A. Harmsen, J.H.B.J. Hoebink, J.C. Schouten, Chem. Eng. Sci. 56 (2001) 2019.
[10] J.M.A. Harmsen, J.H.B.J. Hoebink, J.C. Schouten, Topics Catal. 16/17 (1–4) (2001)
397.
It is not the first time that the role of oxygen is evoked particu-
larly in oxidation reactions over gold nanoparticles [37]. Generally,
particle size is involved because for small particles gold loses its
metallic character due to an increase of the d-electron density
and also to a high degree of surface de-saturation which leads to
[11] C.E. Megiris, P. Berlowitz, J.B. Butt, H.H. Kung, Surf. Sci. 159 (1985) 184.

32
Y. Azizi et al. / Journal of Catalysis 269 (2010) 26–32
[12] R.L. Palmer, J. Vac. Sci. Technol. 12 (1975) 1403.
[27] P. Burroughs, A. Hammett, A.F. Orchard, G. Thorntom, J. Chem. Soc., Dalton
Trans. 17 (1976) 1686.
[28] A. Fujimori, Phys. Rev. B 27 (1983) 3992.
[29] Y. Azizi, Ph.D. Dissertation, University of Strasbourg, 2009.
[30] M.M. Schubert, S. Hackenberg, A.C. van Veen, M. Muhler, V. Plzak, R.J. Behm, J.
Catal. 197 (2001) 113.
[13] W.T.M. Johnson, K.X. Krieger, J. Am. Chem. Soc. 76 (1954) 4782.
[14] D.J. Burnett, A.M. Gabelnick, D.A. Fischer, Anderson L. Marsh, J.L. Gland, J. Catal.
230 (2005) 282.
[15] A.M. Gabelnick, A.T. Capitano, S.M. Kane, J.L. Gland, D.A. Fischer, J. Am. Chem.
Soc. 122 (2000) 143.
[16] S. Ivanova, C. Petit, V. Pitchon, Gold Bull. 39 (1) (2006) 3–8.
[17] S. Ivanova, C. Petit, V. Pitchon, Appl. Catal. A 267 (2004) 191–201.
[18] Y. Azizi, V. Pitchon, C. Petit J. Mol. Catal., submitted for publication.
[19] X. Zhang, H. Wang, B.Q. Xu, J. Phys. B 109 (2005) 9678.
[20] F. Menegazzo, F. Pinna, M. Signoretto, V. Trevisan, F. Boccuzzi, A. Chiorino, M.
Manzoli, Appl. Catal. A 356 (2009) 31.
[21] S. Ivanova, V. Pitchon, C. Petit, J. Mol. Catal. A: Chem. 256 (2006) 278.
[22] A.M. Visco, F. Neri, G. Neri, A. Donato, C. Milone, S. Galvagno, Phys. Chem.
Chem. Phys. 11 (1999) 2869.
[31] A.A. Herzing, C.J. Kiely, A.F. Carley, P. Landon, G.J. Hutchings, Science 321
(2008) 1331.
[32] Y. Segura, N. Lopez, J. Perez-Ramirez, J. Catal. 247 (2007) 383.
[33] Y. Azizi, C. Petit, V. Pitchon, J. Catal. 256 (2008) 338.
[34] J.U. Ried, S.J. Thomson, G. Webb, J. Catal. 30 (1973) 372.
[35] A.F. Diwell, R.R. Ramaram, H.A. Shaw, T.J. Truex, Stud. Surf. Sci. Catal. 71 (1991)
139.
[36] A. Trovarelli, Catal. Rev. 38 (1996) 439.
[37] B. Grzybowska-Swierkosz, Catal. Today 112 (2006) 3–7.
[38] T. Fukushima, S. Galvagno, G. Parravano, J. Catal. 57 (1979) 177.
[39] H. Berndt, I. Pitsch, S. Evert, K. Struve, M.-M. Pohl, J. Radnik, A. Martin, Appl.
Catal. A 244 (2003) 169.
[23] S. Minicò, S. Scirè, C. Crisafulli, S. Galvagno, Appl. Catal. B 34 (2001) 277.
[24] P. Konova, A. Naydenov, T. Tabakova, D. Mehandjiev, Catal. Commun. 5 (2004)
537.
[25] N. Nakayama, T. Hayashi, Composites A 38 (2007) 1996.
[26] W.M. Mullins, B.L. Averbach, Surf. Sci. 206 (1988) 29.
[40] M. Kotobuki, R. Leppelt, D.A. Hansgen, D. Widmann, R.J. Behm, J. Catal. 264
(2009) 67.