Vital roles of hydroxyl groups and gold oxidation states in Au/ZrO 2 catalysts for 1,3-butadiene hydrogenation

Article abstract of DOI:10.1016/j.jcat.2011.01.002

Hydrogenation of 1,3-butadiene over Au/ZrO₂ catalysts with different gold loadings and calcination temperatures was reported. The catalysts were characterized in depth to understand their structure-property relationship. Gold oxidation states a

Full text of DOI:10.1016/j.jcat.2011.01.002

Journal of Catalysis 279 (2011) 75–87
Contents lists available at ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
Vital roles of hydroxyl groups and gold oxidation states in Au/ZrO2 catalysts
for 1,3-butadiene hydrogenation
Xin Zhang a,b, Hui Shi a,1, Bo-Qing Xu a,
⇑
a Innovative Catalysis Program, Key Lab for Organic Optoelectronics & Molecular Engineering, Department of Chemistry, Tsinghua University, Beijing 100084, China
b State Key Laboratory of Heavy Oil Processing, China University of Petroleum, Changping, Beijing 102249, China
a r t i c l e i n f o
a b s t r a c t
Article history:
Hydrogenation of 1,3-butadiene over Au/ZrO2 catalysts with different gold loadings and calcination tem-
peratures was reported. The catalysts were characterized in depth to understand their structure–property
relationship. Gold oxidation states and surface hydroxyl groups, which were found to be sensitive to the
gold loading, calcination temperature, and treatment with water, were shown to play vital roles in the
hydrogenation activity of Au/ZrO2. Continued activity decrease was seen when the density of surface
hydroxyl groups was lowered by elevating the pre-calcination temperature of ZrO2. Fully dehydroxylated
Au/ZrO2 was essentially inactive, but became very active after partial regeneration of the hydroxyl groups
by water treatment. Moreover, the activity of Au/ZrO2 increased with increasing Au3+/Au0 ratio. Isolated
Au3+ ions at the support surface showed up to two orders of magnitude higher activity than Au0 atoms on
Au particles. Several models are proposed to address the structural features of active sites for H2 activa-
tion in Au/ZrO2 catalysts.
Received 17 September 2010
Revised 31 December 2010
Accepted 4 January 2011
Available online 18 February 2011
Keywords:
Gold nanoparticles
Hydrogenation catalysis
ZrO2
Butadiene
Surface hydroxyls
Gold oxidation state
Active sites
Ó 2011 Elsevier Inc. All rights reserved.
1. Introduction
speculated, without direct evidence, that the hydrogenation activ-
ity of their gold catalysts arose from small gold particles having a
Supported gold catalysts have unexpectedly been found to be
active for a wide variety of reactions [1]; however, most of the
studies have been devoted to the unprecedented activity of nano-
sized gold catalysts for CO oxidation. The nature of active sites and
reaction mechanisms for CO oxidation on gold catalysts has been
extensively studied [1–6]. In contrast, little work has been done
to gain insight into the activity and nature of gold catalysts for
hydrogenation reactions [7,8], especially for the hydrogenation of
unsaturated hydrocarbons [9–13], in spite of some elegant docu-
large number of defects, or resided in even smaller particles not
witnessed by electron microscopy. Buchanan and Webb [10] stud-
ied the reaction of 1,3-butadiene with H2/D2 over Au/Al2O3 cata-
lysts with widely varied Au sizes. They claimed that the special
sites capable of activating molecular hydrogen would reside in ex-
tremely small gold particles not sensed by their electron micros-
copy. Due to the limitations of the impregnation method for
preparing gold catalysts (e.g., wide Au size distribution and possi-
ble contamination by poisonous chloride residues [1,4,12]), these
earlier attempts to understand the catalytic sites might be severely
distorted. Using improved preparation methods (e.g., deposition–
precipitation), Au nanoparticles on oxide supports were prepared
and tested as catalysts for the hydrogenation of unsaturated hydro-
carbons [12,13]. For the hydrogenation of acetylene over Au/Al2O3,
a maximum reaction rate was observed over the sample having an
average Au particle size of 3 nm [13]. For hydrogenation of 1,3-
butadiene, however, the reaction rates over oxide-supported Au
catalysts appeared almost insensitive to the Au particle size as well
as to the nature of the oxide supports [12].
mentation on gold-catalyzed hydrogenation of
a,b-unsaturated
aldehydes/ketones [14–18]. This is surprising, since catalytic
hydrogenation represents a well-studied and industrially impor-
tant category of chemical reactions on conventional transition me-
tal catalysts [19].
Supported gold catalysts prepared by the conventional impreg-
nation method were found to show very low catalytic rates for the
hydrogenation of unsaturated hydrocarbons in the 1970s [9,10].
The first report was made by Bond et al. [9], who investigated
impregnated Au/SiO2 and Au/c-Al2O3 catalysts for the hydrogena-
tion of 1-pentene, 1,3-butadiene, and 2-butyne. These authors
In contrast to the well-established knowledge of H2 chemisorp-
tion/activation over Group VIII metals (e.g., Pd, Pt), H2 chemisorp-
tion/activation on supported gold catalysts is far from clear.
Chemisorption of H2 molecules does not occur on clean ‘‘bulky’’
gold surfaces but is occasionally reported on supported small gold
nanoparticles [7,11,20,21]. Theoretical calculations of H2 activation
⇑
Corresponding author. Fax: +86 10 6279 2122.
E-mail address: bqxu@mail.tsinghua.edu.cn (B.-Q. Xu).
Present address: Department Chemie, Technische Universität München,
1
Lichtenbergstr. 4, D-85748 Garching, Germany.
0021-9517/$ - see front matter Ó 2011 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2011.01.002

76
X. Zhang et al. / Journal of Catalysis 279 (2011) 75–87
by dissociation on single-crystal Au surfaces or gold clusters with
varying numbers of Au atoms provide some information [22–24],
but these theoretical models can be far from real catalysts and
sometimes produce controversial results. It should be pointed
out that these earlier experimental and/or theoretical assessments
of H2 chemisorption and hydrogenation reactions were in most
cases done on supported metallic gold catalysts, which could have
neglected possible involvements of cationic gold species and the
underlying support surfaces in H2 activation.
Conductivity of the final filtrate was less than 10ꢁ5 S mꢁ1. The sam-
ples were dried at 393 K overnight and then calcined in flowing air
(60 ml minꢁ1) at temperature T for 5 h (ramp rate: 10 K minꢁ1).
The samples were denoted as Au/ZrO2-CP(AN)-t-T. t refers again
to the pre-calcination temperature of ZrO2 support (before gold
was loaded) and T is the catalyst calcination temperature (after
gold was loaded). Unless specified, the actual gold loadings were
between 0.63 and 0.76 wt.% by ICP-AES analysis, but the nominal
gold loading in the preparation was 1.0 wt.% for these Au/ZrO2-
CP-t-T and -AN-t-T samples. Variation in gold loading in the range
of 0.01–1.0 wt.% was intentionally made within the Au/ZrO2-CP-
673-473 samples, by adjusting the amounts of aqueous HAuCl4
during the deposition–precipitation stage; these samples are la-
beled as m%Au/ZrO2-CP-673-473 and m% denotes the actual gold
loading by weight percentage. In addition, a 0.08%Au/ZrO2-CP-
673-473 was prepared by treatment of the Au/ZrO2-CP-673-473
(1.0% nominal Au loading) with a 2% KCN solution as described pre-
viously [31].
For CO oxidation [25–28] and water–gas shift reactions [29],
cationic gold species (Au3+ and/or Au+) sometimes are found to
be superior to metallic gold (Au0), at least on several oxide sup-
ports. The first demonstration of high activity of immobilized
Au3+ ions for olefin hydrogenation was documented using MgO-
supported mononuclear Au (III) complex catalyst for ethene hydro-
genation [30]. It was clearly shown with in situ X-ray absorption
near-edge structure (XANES) that mononuclear Au3+ complexes,
but not metallic gold clusters, were responsible for the hydrogena-
tion activity [30]. We previously communicated that isolated Au3+
ions on a ZrO2 surface were highly active for the selective hydroge-
nation of 1,3-butadiene [27,31]. These cationic gold species were
also found to be active for olefin (propene) hydrogenation [32].
These results are encouraging, which identified that immobilized
surface cationic gold species have potential as new catalyst for
the hydrogenation of unsaturated hydrocarbons, even though H2
could potentially be a reducing agent to the cationic gold species.
It was reported that moisture (water vapor) in the reaction feed
or in some cases hydroxyl groups on the catalyst surface [33,34]
could influence the catalytic activity of oxide-supported gold for
CO oxidation. Nevertheless, effects of moisture and surface hydro-
xyl groups on the gold-catalyzed hydrogenation of unsaturated
hydrocarbons have not yet been explored. Here, a series of Au/
ZrO2 catalysts, with different gold loadings and calcination temper-
atures, are prepared by deposition–precipitation and investigated
for the selective hydrogenation of 1,3-butadiene. A detailed analy-
sis of the catalyst structure is provided, based on characterizations
using a variety of techniques. We attempt to gain insight into the
structure–performance relationship for 1,3-butadiene hydrogena-
tion over these Au/ZrO2 catalysts. Particular attention is paid to
the effects of moisture in the reaction feed, density of hydroxyl
groups on the support surface, and percentage of Au3+ species in
gold. Several model structures are envisaged to discuss the struc-
tural features of active sites in Au/ZrO2 catalysts.
2.2. Catalyst characterizations
The catalysts were characterized by a number of techniques
including XRD, TEM, XPS, TPR, ESR, IR, TG-DTA, N2 adsorption,
and element analysis (ICP-AES), the details of which are given in
the Supplementary material. H2-TPD measurements were per-
formed on a TPR apparatus [31] equipped with both TCD and MS
(Balzers MSC 200) detectors. Typically, the catalyst sample
(200 mg for 0.05%Au/ZrO2-CP-673-473, 100 mg for the others)
was put into a U-shaped quartz reactor (i.d. 4 mm) and pretreated
in flowing Ar at the catalyst calcination temperature T for 1 h. After
the sample was cooled to 393 K (the reaction temperature of 1,3-
butadiene hydrogenation), the reactor was switched to H2 flow
(45 ml minꢁ1) for 1 h. The reactor was then switched back to Ar
and purged for 0.5 h, followed by cooling to room temperature.
TCD and MS (m/z = 2) signals were recorded when the reactor
was heated from room temperature up to 1073 K at a rate of
10 K minꢁ1 in flowing Ar (45 ml minꢁ1).
2.3. Catalytic hydrogenation of 1,3-butadiene
Hydrogenation of 1,3-butadiene was carried out at 393 K in a
fixed-bed flow stainless reactor (i.d. 6 mm) under atmospheric pres-
sure. Unless otherwise specified, the reaction feed was a gas mixture
of 2.15 vol.% 1,3-butadiene in H2, which was introduced into the
reactor at a space velocity (GHSV) of 8100 ml hꢁ1 (g cat.)ꢁ1. The cat-
alyst (ca. 100 mg Au/ZrO2 in 80–120 mesh) was diluted with 500 mg
of quartz sand before being loaded into the reactor. Before it was
switched to the reaction feed gas, the catalyst was pretreated in
flowing Ar (30 ml minꢁ1) at 473 K for 2 h. The reaction conditions
were similar to those applied by Okumura et al. [12]. In some exper-
iments, a 0.05%Au/ZrO2-CP-673-473 catalyst was pretreated with
flowing H2 (30 ml minꢁ1) at 473 or 523 K for 2 h to investigate the
effect of hydrogen reduction on catalyst activity; the reduced cata-
lyst was then cooled to 393 K in flowing Ar (30 ml minꢁ1) to evaluate
its catalytic activity for the hydrogenation reaction. Any significant
2. Experimental
2.1. Catalyst preparation
Two kinds of zirconia were prepared, using ZrO(OH)2 hydrogel
obtained by hydrolysis of zirconyl chloride in aqueous ammonia,
as described previously [5,35]. Following our earlier sample coding,
ZrO2-CP-t denotes the sample prepared by calcination for 5 h in
flowing air of the ZrO(OH)2 hydrogel at temperature t = 673, 873,
973 and 1073 K, respectively. A ZrO(OH)2 alcogel was prepared
by extensive washing with anhydrous ethanol of the ZrO(OH)2
hydrogel [35]. A follow-up calcination for 5 h of the ZrO(OH)2
alcogel in flowing N2 at temperature t produced the ZrO2-AN-t
samples.
Zirconia-supported gold catalysts were prepared using a depo-
sition–precipitation method. Briefly, zirconia powders (2 g) were
added into an appropriate amount of aqueous HAuCl4 solution.
The solution pH was adjusted to ca. 9 by controlled addition of
0.2 M NH3ꢀH2O solution. The aqueous dispersions were stirred for
6 h, aged for 2 h, and suction filtered. Extensive washing with
deionized water was then continued until it was free of Clꢁ ions.
mass-transfer limitation was precluded in
a series of pre-
experiments at a fixed space velocity (e.g., GHSV = 8100 ml hꢁ1
(g cat.)ꢁ1) so that the conversion of 1,3-butadiene over a highly
active Au/ZrO2 catalyst was not significantly affected by changing
the flow rate of the reaction feed and the catalyst pellet size in
40–120 mesh. The reactor effluent was analyzed on line using a
GC-8A gas chromatograph equipped with a GDX-501 column and
a FID detector.
To study the effect of moisture on the hydrogenation reaction,
the reaction feed gas was set to pass through a glass vessel
(250 ml) containing ca. 50 g H2O. The temperature was varied

X. Zhang et al. / Journal of Catalysis 279 (2011) 75–87
77
between 203 and 273 K in order to change the concentration of
moisture (gaseous H2O) in the range of 20–6000 ppm. The reaction
feed in these experiments was a gas mixture composed of 1,3-
butadiene, H2, and N2 with a volumetric ratio of 1:49:97 and was
of the hydroxyl groups (HOA, formula mass 17) in the original gel
with ethoxyl groups (CH3CH2OA, formula mass 45) in ethanol:
(AOH)s + CH3CH2OH ! (AOCH2CH3)s + H2O, leading to formation
of surface ethoxyls in the alcogel [38]. Therefore, the density of sur-
face hydroxyl groups would be higher on a ZrO2 sample obtained
from the ZrO(OH)2 hydrogel (ZrO2-CP) than on the sample ob-
tained from its derived ZrO(OH)2 alcogel (ZrO2-AN) using the same
calcination procedure. In other words, the ZrO2-AN-t sample would
carry fewer hydroxyl groups than ZrO2-CP-t. As can be judged from
the TG curves, the density of hydroxyl groups for both ZrO2-AN-t
and -CP-t samples would decrease with t. The undetectable weight
loss from the TG curves in the range 1073–1273 K would suggest
that nearly fully dehydroxylated ZrO2 samples can be prepared
by calcination of ZrO(OH)2 at a temperature as high as 1073 K.
The textural properties of ZrO2-CP-t and -AN-t are summarized
in Table 1. Nitrogen adsorption measurements show that BET sur-
face areas of ZrO2-CP-t are 120, 44, 23, and 14 m2 gꢁ1 at t = 673,
873, 973, and 1073 K, respectively. These surface area data are
close to those documented earlier by Nakano et al. [39]. In compar-
ison with ZrO2-CP-t, much higher surface areas and pore volumes
were observed for ZrO2-AN-t (Table 1). XRD patterns (see Fig. S1
in the Supplemental material) show that the present ZrO2 samples
are all composed of monoclinic (M) and tetragonal (T) phases [5].
The phase composition of ZrO2-CP-t and -AN-t when t was higher
than 873 K could not be differentiated (Table 1). Although the aver-
age crystal sizes determined by the X-ray line broadening analysis
(XLBA) seemed similar for ZrO2-CP-t and -AN-t when t was less
than 1073 K, the average domain sizes of ZrO2-AN-t particles in
TEM images appeared much smaller than those of their ZrO2-CP-t
counterparts. The similar particle sizes obtained from XLBA and
TEM measurements for the ZrO2-AN-t would hint that these sam-
ples were composed of discrete ZrO2 nanocrystals. In contrast,
the difference between the XLBA and TEM sizes of the conventional
ZrO2-CP-t would indicate that the primary crystallites in these
samples agglomerated more or less into larger particles. We be-
lieve that the transformation of the conventional ZrO(OH)2 hydro-
gels to its alcogels can reduce the capillary pressure and density of
hydroxyl groups in the gel network, because the surface tension of
ethanol would be much lower than that of water (22.39 vs
72.88 dyn/cm at 293 K), which could endow the alcogel-derived
ZrO2-AN sample with greater resistance to structure collapse and
thermal sintering [38,40].
introduced at a GHSV of 23,820 ml hꢁ1 (g cat.)ꢁ1
.
3. Results
3.1. Physicochemical properties of zirconia support
Fig. 1A shows TG/DTA curves for the ZrO(OH)2 hydrogel and its
derived alcogel. Mass spectroscopic analyses of the effluent gas
during the TG/DTA process of the hydrogel are shown in Fig. 1B.
The weight loss between 323 and 450 K on the TG curve can be
generally ascribed to the desorption of weakly adsorbed water
molecules. Further weight loss at above 450 K would be associated
with the desorption of strongly adsorbed water and/or dehydration
by surface dehydroxylation [36]. The DTA curve of the ZrO(OH)2
hydrogel shows a weak but broad endothermic peak ranging from
323 to 673 K, due to the water desorption and dehydration by sur-
face dehydroxylation. The sharp exothermic peak at 719 K on the
DTA curve of the hydrogel is known as the glow exotherm associ-
ated with the formation of crystalline zirconia [37]. An additional
exothermic feature was observed at a lower temperature (634 K)
on the DTA curve of the alcogel. This new peak would arise from
the decomposition and/or combustion of surface ethoxy groups,
which agrees with the detected evolutions of H2O, CH3CHO/CO2,
and C2H4 (m/z = 18, 44, 28, and 27) as shown in Fig. 1B. The weight
loss on the TG curve for the alcogel (22.1%) was more than twice
that for the hydrogel (10.5%), which demonstrated that the prepa-
ration of ZrO(OH)2 alcogel by repeated washing of the ZrO(OH)2
hydrogel with anhydrous ethanol induced significant replacement
100
94
88
82
76
70
0.2
0
A
634 K 735 K
719 K
-0.2
-0.4
-0.6
3.2. Physicochemical properties of zirconia-supported gold catalysts
3.2.1. Textural and morphologic properties
ICP-AES analyses showed that zirconia-supported gold samples
contained ca. 0.70% (in the range 0.63–0.76%), 0.23%, 0.05%, and
0.01% gold by weight, corresponding to nominal gold loadings of
1.0%, 0.25%, 0.05%, and 0.01%, respectively. Nitrogen adsorption
measurement showed that the deposition of gold had little impact
on the texture properties of zirconia as long as the catalyst calcina-
tion temperature T remained no higher than 473 K. The BET surface
areas of the Au/ZrO2-CP-673-T catalysts decreased from 100 to
74 m2 gꢁ1 when T was increased from 573 to 773 K (Table 2).
XRD patterns of the Au/ZrO2 catalysts were found similar to those
of their corresponding ‘‘pure’’ ZrO2 supports; no diffraction associ-
ated with gold was observed.
273
473
673
873
1073 1273
Temperature (K)
m/z = 18
B
44
28
27
Fig. 2 shows some representative TEM/HRTEM images for the
Au/ZrO2-CP-t-T samples (the images for Au/ZrO2-AN-t-673
samples were given elsewhere [5]). As can be seen in Fig. 2A, our
careful search for Au particles in the HRTEM images for m%Au/
ZrO2-CP-673-473 samples at gold loadings of 0.01%, 0.05%, and
0.08% always produced negative results, which would indicate that
gold in these samples was atomically dispersed on the zirconia
support surface. This agrees with Fu et al. [29], who showed that
373 473 573 673 773 873 973 1073
Temperature (K)
Fig. 1. (A) TG/DTA curves of ZrO(OH)2 hydrogel (solid line) and its derived alcogel
(dotted line); (B) mass spectroscopic analyses of the effluent gas from the alcogel
during TG/DTA.

78
X. Zhang et al. / Journal of Catalysis 279 (2011) 75–87
Table 1
Texture properties of ZrO2-CP-t and ZrO2-AN-t samples.
Sample
BET surface area (m2/g)
Pore volume (cm3/g)
Average pore size (nm)
XLBA sizea (nm)
TEM size (nm)
ZrO2-phasea (%)
ZrO2-CP-673
ZrO2-CP-873
ZrO2-CP-973
ZrO2-CP-1073
ZrO2-AN-673
ZrO2-AN-873
ZrO2-AN-973
ZrO2-AN-1073
120
44
23
14
162
61
0.16
0.12
0.11
0.07
0.36
0.28
0.26
0.19
3.9
8.5
16.3
14.5
6.7
15.3
24.8
33.9
M6/T5
M12/T11
M22
10–30
30–100
40–200
40–200
5–10
12–17
20–25
25–40
M54/T46
M88/T12
M92/T8
M98/T2
M76/T24
M88/T12
M92/T8
M97/T3
M40
M6/T5
M12/T11
M21
36
20
M30
a
M and T represent monoclinic and tetragonal phases, respectively.
Table 2
Physicochemical properties and catalytic activity of Au/ZrO2 catalysts in selective hydrogenation of 1,3-butadiene.
Catalyst
BET surface (m2/g) dAu (nm)
OH (mmol/g) Red. temp. (K) H2 uptakea Au3+/Au0 Conv. (%) Rxn. rate (mmol sꢁ1 (g Au)ꢁ1) TOFb (sꢁ1
)
Au/ZrO2-AN-1073-673
Au/ZrO2-AN-973-673
Au/ZrO2-AN-873-673
Au/ZrO2-AN-673-673
Au/ZrO2-CP-1073-473
Au/ZrO2-CP-1073-573
Au/ZrO2-CP-1073-673
Au/ZrO2-CP-973-673
Au/ZrO2-CP-873-673
Au/ZrO2-CP-673-773
Au/ZrO2-CP-673-673
Au/ZrO2-CP-673-573
Au/ZrO2-CP-673-473
Au/ZrO2-CP-673-393
18
34
53
142
14
14
14
22
34
74
90
100
100
100
3.0–7.0 (5.4)
–
–
–
–
–
–
–
–
424
–
–
–
–
0
0
0
0
0.14
0
0
0
0
0
0
0
0
0.10
0
0
0
0
0
0
0.20
0.33
0.42
0.83
3.0
–
–
3.0–7.0 (4.7)
3.0–7.0 (4.7)
3.0–7.0 (5.4)
12.0
18.0
49.0
19.0
3.4
2.6
8.8
32.0
6.8
39.0
53.4
75.2
98.4
39.3
85.4
49.9
10.3
0.03
0.05
0.13
0.05
–
0.03
0.04
0.10
0.04 (0.11)
–
–
0.01
0.09
0.03
7–13 (10.0)
3.0–7.0 (4.0)
4.0–8.3 (6.2)
6.5–13.4 (11.7) –
5.6–10.4 (8.0) 0.75
0.23 (0.30)c
0.38
0.68
–
0.02
0.09
0.02
0.10
0.14
0.20
0.38
0.34
2.15
2.01
2.08
–
–
0
0
0.12
4.5–9.0 (6.7)
3.0–7.0 (4.1)
2.5–5.5 (3.5)
2.0–4.0 (2.9)
n.d.d
–
–
–
–
–
–
–
488
483
483
482
496
504
508
0.25
0.37
0.44
0.68
1.50
1.49
1.57
0.17 (0.17)
0.17 (0.16)
0.31 (0.25)
0.17 (0.16)
(0.42)
0.23%Au/ZrO2-CP-673-473 110
0.08%Au/ZrO2-CP-673-473 –
0.05%Au/ZrO2-CP-673-473 120
0.01%Au/ZrO2-CP-673-473 120
e
1
n.d.d
1
(0.40)
(0.41)
e
n.d.d
1
e
a
b
c
The unit is mol(H2)/mol(Au).
The TOF data were calculated on the basis of exposed Au0 atoms or Au3+ ions (data in the parentheses) in the catalyst.
Data in the parentheses is the hydroxyl concentration of 0.74% Au/ZrO2-CP-1073-673 treated with deionized water.
n.d.: undetectable in the HRTEM measurement.
d
e
Only Au3+ ions were found.
0.5% Au/Ce(La)O2 obtained by subjecting their 5% Au/Ce(La)O2 cat-
alyst to an aqueous solution of 2% NaCN contained only atomically
dispersed gold species. Nearly spherical Au particles were observed
for the Au/ZrO2-CP-673-T samples; their average size became
greater and their size distribution wider when T was increased
from 393 to 773 K. For comparison, Au particles in those
Au/ZrO2-AN-t-673 samples (Table 2) showed a relatively narrower
size distribution (3–7 nm) and a smaller average sizes (4–5 nm)
[5].
increased from 673 to 873 K and became invisible when t was
increased to 973 K or higher. The increase in t also significantly re-
duced the adsorption bands for bi- (ca. 3680 cmꢁ1) and tri-bridged
hydroxyl groups (ca. 3630 cmꢁ1). Ultimately, the surface of Au/
ZrO2-CP-1073-673 (Fig. 3d) showed basically no sign of any of
those hydroxyl groups. Interestingly, simple water treatment of
this hydroxyl-free Au/ZrO2-CP-1073-673 sample resulted in partial
regeneration of all three kinds of surface hydroxyl groups (Fig. 3e),
due to rehydration of the support oxide.
The number of hydroxyl groups was evaluated by TG measure-
ments of the Au/ZrO2-CP-t-673 samples, which were pretreated
in situ at 673 K (473 K for the water-treated Au/ZrO2-CP-
1073-673) in flowing dry air for 0.5 h. The numbers were 0.75,
0.68, 0.38, and 0.23 mmol/g for Au/ZrO2-CP-t-673 samples with
t = 673, 873, 973, and 1073 K, respectively (Table 2). We found that
the number for Au/ZrO2-CP-1073-673 increased from 0.23 to
0.30 mmol/g after it was treated with water, in accordance with
the observed regeneration of surface hydroxyls after the water
treatment in the IR measurements. These results clearly show that
the concentration of hydroxyl groups in Au/ZrO2 catalysts depends
sensitively on the support calcination temperature t.
3.2.2. Assessment of surface hydroxyl groups
The Au/ZrO2-CP-t-673 samples were characterized with IR spec-
troscopy to compare their hydroxyl groups. Fig. 3 shows the IR
spectra in the AOH stretching vibration region. Three absorption
bands at around 3770, 3680, and 3630 cmꢁ1 were observed on
Au/ZrO2-CP-673-673. It was understood that weakly adsorbed
water molecules (or coordinated molecular water) on highly hy-
drated zirconia surfaces could be eliminated easily upon evacua-
tion at 450 K, but hydroxyl groups on the same surfaces were
sustained up to ca. 873 K [36]. Because all the samples were pre-
treated in situ at 450 K in vacuum for 1 h to remove any coordi-
nated water molecules before their IR spectra were recorded, it
would be impossible to relate the broad peak at around
3680 cmꢁ1 to a presence of molecularly absorbed water. The three
peaks of Fig. 3a are then assigned to terminal and bi- and tri-
bridged hydroxyl groups, respectively [41]. The terminal hydroxyl
groups (ca. 3770 cmꢁ1) were sharply weakened when t was
3.2.3. Oxidation states of Au/ZrO2 catalysts
Fig. 4 shows the XPS spectra of Au4f in Au/ZrO2-CP-673-T
samples. Compared with the Au4f7/2 and 4f5/2 binding energies
(BEs) for both T = 673 and 773 K samples (Fig. 4d and e), the
Au XPS peaks for the T = 393, 473, and 573 K samples (Fig. 4a–

X. Zhang et al. / Journal of Catalysis 279 (2011) 75–87
79
B
A
3 nm
3 nm
C
D
Au
3 nm
3 nm
F
E
6 nm
6 nm
Fig. 2. Representative TEM/HRTEM images of Au/ZrO2-CP-t-T catalysts: (A) Au/ZrO2-CP-673-473 containing 0.01%, 0.05%, and 0.08% Au by weight; (B) 0.23%Au/ZrO2-CP-673-
473; (C) Au/ZrO2-CP-673-393; (D) Au/ZrO2-CP-673-473; (E) Au/ZrO2-CP-673-573; (F) Au/ZrO2-CP-673-673; (G) Au/ZrO2-CP-673-773; (H) Au/ZrO2-CP-873-673; (I) Au/ZrO2-
CP-973-673; (J) Au/ZrO2-CP-1073-673.
c) showed significantly higher BEs (by up to ca. 1.0 eV). The BEs
at 83.6 and 87.4 eV for both T = 673 and 773 K samples were
characteristic of metallic Au [5,26]. The higher Au4f BEs for the
samples with T < 673 K could easily be attributed to the presence
of cationic gold species (Au3+ or/and Au+). However, it is under-
stood that quantitative assessment of the gold oxidation state by
XPS characterization alone is limited because the final state ef-
fects associated with particle size could heavily disturb the XPS
features of Au [6,25,42] and the exposure to photoelectrons
under high vacuum during XPS measurement might also effect
change in the gold oxidation state [25,43]. In principle, the Auger
parameter can be used to check the electronic nature by the
initial state effects, but the high kinetic energies of the relevant
Auger electrons would make it difficult to measure supported
gold nanoparticles; an earlier attempt made by Hutchings et al.
[25] failed in obtaining valuable information even for a 5%
Au/Fe2O3 catalyst because of poor signal intensity.
We recognize that it is more reliable to titrate the oxidation
states of gold using quantitative H2-TPR, as shown in earlier liter-
ature [27,31,44,45]. Fig. 5A shows the TPR profiles for Au/ZrO2-
CP-673-T of varying gold loadings. The profiles for the samples
with T = 393, 473, and 573 K featured two hydrogen consumption
peaks, one in low (482–508 K)- and the other in high (ca. 830 K)-
temperature regions. It is known that a sufficiently high calcination

80
X. Zhang et al. / Journal of Catalysis 279 (2011) 75–87
G
H
20 nm
J
I
10nm
Fig. 2 (continued)
3680
3630
0.1
e
3770
a
b
d
c
c
d
b
a
e
92
90
88
86
84
82
80
4000 3900 3800 3700 3600 3500 3400
Wavenumber (cm-1)
Binding energy (eV)
Fig. 3. Infrared spectra of hydroxyl groups on Au/ZrO2-CP-t-673 of t = 673 (a), 873
Fig. 4. XPS spectra of Au4f in Au/ZrO2-CP-673-T samples of T = 393 (a), 473 (b), 573
(b), 973 (c), and 1073 K (d), and the water-treated Au/ZrO2-CP-1073-673 (e).
(c), 673 (d), and 773 K (e).
temperature (generally no less than 673 K) would induce complete
decomposition/reduction of the cationic species to metallic gold
particles [25–27]. This point is confirmed here by our observation
that no reduction of cationic gold could be detected for the
T = 673 K sample. Thus, the low-temperature peak would be re-
lated to a reduction of cationic gold species, while the high-
temperature one would be related to a partial (<3%) reduction of
Zr4+ to Zr3+ in the support oxide [31,45].
The calibrated hydrogen consumption data disclosed that the
stoichiometry for the hydrogen reduction of gold was H2/Au (mo-
lar) = 1.50 0.2 for the Au/ZrO2-CP-673-473 samples containing no
more than 0.08% gold. This stoichiometry clearly shows that Au3+
ions were the only gold species in the samples, which is in line
with the absence of metallic Au particles in their HRTEM images
(Fig. 2A). The residual gold species in cyanide-leached Au/Al2O3
(0.16% Au) [46] and Au/CeO2 (0.08% Au) [47] were also detected

X. Zhang et al. / Journal of Catalysis 279 (2011) 75–87
81
Fig. 5B shows the TPR profiles for Au/ZrO2-CP-1073-T samples
(gold loading 0.7%, T = 473–673 K). The water-treated Au/ZrO2-
CP-1073-673 was also subjected to TPR measurement, which is in-
cluded as profile a’. Compared to those for Au/ZrO2-CP-673-473
(Fig. 5A), the gold reduction peak shifted to a lower temperature
(ca. 424 K) and the measured relative amount of Au3+ became as
low as Au3+/Au0 = 0.10 for Au/ZrO2-CP-1073-473 (Table 2). This
comparison suggests that the fully dehydroxylated surface of
ZrO2-CP-1073 was inferior to that of -CP-673 for stabilizing the
cationic gold species (Au3+ ions). No peak that was characteristic
of Au3+ reduction was found on Au/ZrO2-CP-1073-T samples at
T = 573 and 673 K. The TPR profile for the water-treated sample ap-
peared similar to that for its untreated counterpart, indicating that
the water treatment and subsequent drying and calcination in-
duced no change in the gold species. However, it seems that hydro-
gen reduction of the support material (from Zr4+ to Zr3+) became
more pronounced in the presence of Au3+ ions, as one could judge
from the difference between profiles a and c. This would hint that
the presence of Au3+ probably could promote the reduction of Zr4+
into Zr3+ via H2 chemisorption.
A
a
b
c
d
e
f
273
473
673
873
1073
Temperature (K)
B
XPS spectra analysis of the Au/ZrO2 samples after the H2-TPR
measurement (Fig. S2) gave no signal for Zr3+ ions but only Zr4+
ions, with Zr3d BEs at 184.4 and 182.3 eV. The materials were then
subjected to ESR measurement; the results are shown in Fig. 6. The
Au/ZrO2-CP-673-473 after H2-TPR produced ESR signals at
g? ¼ 1:976 and g|| = 1.957, which are indeed distinctive for Zr3+
ions. An additional ESR signal at g = 2.003 on this reduced sample
signified formation of so-called F-centers (single electrons trapped
in oxygen vacancies), which was frequently observed on reduced
Au/ZrO2 and Au/TiO2 samples [45,48]. Surprisingly, a pair of ESR
signals attributed to Zr3+ ions was also registered on both Au/
ZrO2-CP-673-473 and -673-673 samples. Similar signals were also
recorded earlier on gold catalysts supported by mesoporous zirco-
nia [45], as well as on zirconia nanoparticles with small sizes [49].
c
b
a'
a
273
473
673
873
1073
Temperature (K)
Fig. 5. TPR profiles of Au/ZrO2-CP-673-T (A) and Au/ZrO2-CP-1073-T (B) samples.
T = 673 (a), 573 (b), 473 (c), and 393 K (d); (a0) denotes the water-treated Au/ZrO2-
CP-1073-673; (e) and (f) refer to 0.23% and 0.05%Au/ZrO2-CP-673-473 samples.
3.2.4. H2 activation on Au/ZrO2 catalysts
as Au3+ ions by in situ EXAFS/XANES, though the exact nature of
those gold cations in the cyanide-leached Au/Ce(La)O2 samples re-
mained unclear in Fu et al.’s work [29]. The lack of clear evidence of
Au+ ions in these supported gold catalysts with low gold loadings
encouraged us to exclude any involvement of Au+ ions in our Au/
ZrO2 samples. Thus, deconvolution of the XPS spectra in Fig. 4
was done based on the assumption that any gold species in Au/
ZrO2 catalysts would involve either metallic Au0 atoms or cationic
Au3+ ions. Although the BE shifts seemed too large in Fig. 4, the
deconvolution resulted in two crossed doublets that are associated,
respectively, with Au3+ ions (higher BEs) and Au0 atoms (lower
BEs) [27]. It is not surprising that the relative intensity of the dou-
blet for Au3+ declined steadily with increasing catalyst calcination
temperature T.
The reduction peak temperature, hydrogen consumption, and
quantitative Au3+/Au0 ratio derived from the TPR measurements
are summarized in Table 2. The peak temperature was lowered
from 508 to 483 K when the gold loading was increased from
0.01% to 0.70% in the Au/ZrO2-CP-673-473 samples, accompanied
by concomitant decreases in the hydrogen consumption and
Au3+/Au0 ratio. The Au3+/Au0 ratio for the Au/ZrO2-CP-673-T sam-
ples (gold loading 0.7%) declined from 0.44 to 0 when T was in-
creased from 393 to 673 K. These data are seemingly at variance
with the much earlier work of Sermon et al. [11], who claimed that
drying treatment of Au/SiO2 and Au/Al2O3 catalysts containing less
than 1.0 wt.% gold at a temperature as low as 393 K in either air or
vacuum was sufficient to reduce their gold species to metallic gold.
However, that earlier claim has not yet been confirmed, because it
was based simply on the visual sample colors.
H2-TPD experiments were performed to gain information about
H2 activation/dissociation at 393 K on the Au/ZrO2 catalysts
(Fig. 7). In order not to cause variation in the gold oxidation state,
the samples were pretreated with flowing Ar, rather than H2, at
their calcination temperature (T). Two H2-desorption peaks were
detected for 0.05%Au/ZrO2-CP-673-473 containing only isolated
Au3+ ions, and Au/ZrO2-CP-673-393 and -CP-673-473 containing
Au3+ ions as well as Au0 atoms. The peak in the low-temperature
region (ca. 590–780 K) appeared much smaller than the peak in
2.003
a
b
1.975
1.956
2.043
2.043
1.976
1.957
1.957
1.976
c
d
3100
3200
3300
3400
3500
3600
H (Gauss)
Fig. 6. ESR spectra measured at room temperature: (a) Au/ZrO2-CP-673-473 after
the H2-TPR measurement; (b) Au/ZrO2-CP-673-673; (c) Au/ZrO2-CP-673-473; and
(d) the blank (the empty tube).

82
X. Zhang et al. / Journal of Catalysis 279 (2011) 75–87
butadiene conversion levels were always no higher than that of
the blank thermal reaction (2–3%).
Although the Au/ZrO2-CP-1073-673 catalyst showed no activity
in these experiments, remarkably enhanced activity was observed
after the sample was treated with deionized water (Fig. 9). The
conversion of 1,3-butadiene over the water-treated sample in-
creased to 48.4%, which is one order of magnitude higher than that
for the blank reaction (3.4%). However, the conversion of butadiene
over this water-treated catalyst also decreased sharply when ca.
20 ppm moisture was introduced into the reaction feed. It should
be pointed out that the ‘‘moisture-deactivated’’ catalyst could not
be reactivated by switching the reaction feed off the moisture gen-
erator. Even in situ recalcination at 473 K in flowing air could only
reactivate such a ‘‘moisture-deactivated’’ catalyst up to ca. 20% of
its original activity. These results indicate that most of the active
sites on the water-treated catalyst could be blocked or ‘‘destroyed’’
by the presence of added moisture. The adsorption energies of H2O,
H2, and trans- and cis-1,3-butadiene on Au(1 1 1), according to a re-
cent DFT calculation study [50], were ꢁ4.51, ꢁ0.81, ꢁ5.71, and
ꢁ3.75 kcal/mol, respectively. This theoretical calculation would
imply that the presence of moisture can not only block the gold
sites for H2 adsorption/activation but also compete with butadiene
for the active gold sites during the hydrogenation reaction, if the
hydrogenation activity is associated only with metallic gold.
Fig. 10 shows the reaction rate changes over Au/ZrO2-CP-t-673
and -AN-t-673 catalysts with different t. The elevation of t from
673 up to 1073 K resulted in a pronounced and continuous de-
crease in the catalyst activity for both series of catalysts. As shown
in Fig. 10 and Table 2, the average butadiene consumption rate and
turnover frequency (TOF) based on exposed gold atoms within 6 h
of TOS both decreased sharply with the increment of t. Notably, the
conversion levels of butadiene over both Au/ZrO2-CP-1073-673
and -AN-1073-673 were found no higher than that of the blank
thermal reaction, implying that the metallic gold particles in both
catalysts were essentially inactive for the hydrogenation reaction.
a
b
c
d
e
f
273 373 473 573 673 773 873 973 1073
Temperature (K)
Fig. 7. H2-TPD curves after the chemisorption of H2 at 393 K over Au/ZrO2-CP-673-T
of T = 393 (b), 473 (c), and 673 K (d); (a), (e), and (f) refer to 0.05%Au/ZrO2-CP-673-
473, Au/ZrO2-CP-1073-673, and ZrO2-CP-673, respectively.
the high-temperature region (ca. 800–940 K). However, only a very
small peak was detected in the high-temperature region for Au/
ZrO2-CP-673-673 containing only Au0 atoms. Notably, hydrogen
desorption was observed neither for OH-free ZrO2-CP-1073 carry-
ing Au particles nor for Au-free ZrO2-CP-673 having surface
OH-groups. It is deduced therefore that hydrogen activation in
the absence of Au3+ ions would require some form of cooperation
between metallic Au particles and ZrO2-based hydroxyl groups.
3.3. Selective hydrogenation of 1,3-butadiene over Au/ZrO2 catalysts
3.3.1. Effect of moisture and surface hydroxyl groups
Fig. 8 shows the effect of moisture (water vapor) in the reaction
feed on the reaction rates for butadiene consumption normalized
to the gold weight (i.e., mass-specific activity of gold) at 393 K over
Au/ZrO2-CP-673-673 and -CP-1073-673 catalysts. This figure regis-
ters a detrimental effect of moisture on the hydrogenation reaction
of 1,3-butadiene for a wide range of moisture concentrations (20–
6000 ppm). The reaction rate over Au/ZrO2-CP-673-673 was low-
ered by 90% with the addition of ca. 6000 ppm moisture into the
standard ‘‘dry’’ reaction feed. On the other hand, the same change
in moisture concentration produced little effect on the reaction
rate over Au/ZrO2-CP-1073-673 catalyst (Au nanoparticles
supported on a fully dehydroxylated ZrO2), and the measured
3.3.2. Effect of gold oxidation state
Fig. 11 shows the reaction rate changes over Au/ZrO2-CP-673-T
catalysts with varied T. The rates are 0.02, 0.10, 0.14, 0.20, and
0.38 mmol sꢁ1 (g Au)ꢁ1 for T = 773, 673, 573, 473, and 393 K,
respectively. These data mean that the reaction rate decreased
with a decline in the catalyst’s Au3+/Au0 ratio when T varied be-
tween 393 and 673 K, since T elevation from 393 up to 673 K
50
40
30
0.05
0.04
0.03
0.02
0.01
0.00
After 2 h purging with
20
10
0
flowing air at 473 K
0
1
2
3
4
5
6
7
8
9 10 11 12 13 14
Reaction time (h)
Fig. 9. Time course of 1,3-butadiene conversion over as-prepared Au/ZrO2-
CP-1073-673 (h) and its water-treated counterpart sample (N). After 6 h, the
reaction over the water-treated sample was continued by adding ca. 21 ppm
10
100
1000
10000
Water vapor (ppm)
moisture into the reaction feed gas (.). The last set of data (D) shows the results
Fig. 8. Effect of water vapor (moisture) on the reaction rate by 1,3-butadiene
consumption at 393 K over Au/ZrO2-CP-t-673 of t = 673 ( ) and 1073 K (j).
with the standard ‘‘dry’’ reaction feed (without addition of moisture) after purging
the reacted catalyst with flowing air for 2 h at 473 K.
D

X. Zhang et al. / Journal of Catalysis 279 (2011) 75–87
83
a reversible variation in butadiene conversion between 98.4% and
61% when the feed GHSV was varied between 8100 and
19,200 ml hꢁ1 (g cat.)ꢁ1 (see Fig. S3). Specifically, the measured
0.12
0.10
0.08
0.06
0.04
0.02
0.00
A
overall reaction rate was 0.26, 0.38, and 0.38 mmol sꢁ1 (g Au)ꢁ1
,
respectively, when the GHSV was 8100, 13,800, and
19,200 ml hꢁ1 (g cat)ꢁ1. The actual catalytic rate for this Au/ZrO2-
CP-673-393 catalyst is therefore 0.38 mmol sꢁ1 (g Au)ꢁ1, which is
given in Table 2. It should be emphasized that the activity of Au/
ZrO2-CP-673-393 remained unchanged during the entire reaction
period measured at 393 K (14 h, Fig. S3).
Fig. 12 shows the changes in butadiene conversion over Au/
ZrO2-CP-1073-T catalysts with varied T. The conversion over the
T = 573 and 673 catalysts (2–3%) was no higher than that in the
blank thermal reaction, much lower than that over the T = 473 cat-
alyst (ca. 19%). As these two inactive catalysts contained only
metallic Au nanoparticles, while the active T = 473 catalyst con-
tained both Au3+ and Au0, these results provide additional evidence
for our finding that metallic gold particles supported on hydroxyl-
free ZrO2-CP-1073 possess no active sites for the hydrogenation
reaction.
0
1
2
3
4
5
6
TOS (h)
0.16
0.14
0.12
0.10
0.08
0.06
0.04
0.02
0.00
B
Considering that the function of Au3+ ions in the hydrogenation
catalysis could be discounted or masked by possible involvement
of zirconia-based hydroxyl groups, which would change with the
t and T temperatures of a catalyst, we further investigated a series
of m%Au/ZrO2-CP-673-473 catalysts with lower gold loadings
(m 6 0.7). The preparation of these catalysts with invariant t and
T temperatures could avoid, with the highest probability, changes
in other properties of zirconia support (including hydroxyl groups,
surface area, and crystalline phase). It is shown in Fig. 13 that the
butadiene conversion (ca. 50%) over 0.05%Au/ZrO2-CP-673-473
was higher than that (ca. 40%) over 0.23% Au/ZrO2-CP-673-473 cat-
alyst, though the gold loading in the later catalyst was about four
times higher. The measured average reaction rates (within 6 h
TOS) were 2.01, 0.34, and 0.20 mmol sꢁ1 (g Au)ꢁ1 for 0.05%,
0.23%, and 0.76%Au/ZrO2-CP-673-473 catalysts with Au3+/Au0 = 1
(infinity), 0.83, and 0.33, respectively (Table 2). Notably, the reac-
tion rates and TOFs remained the same for the catalysts containing
0.05% and 0.01% gold, showing that the isolated Au3+ ions in these
two samples were equally active for the hydrogenation reaction.
Fig. 13 also compares the performance of 0.76%Au/ZrO2-
CP-673-473 with its cyanide-leached counterpart (0.08%Au/ZrO2-
CP-673-473). The cyanide-leached catalyst with only 0.08% gold
in the Au3+ state surprisingly demonstrated much higher activity
[31]. Specifically, the reaction rate and TOF for the cyanide-leached
catalyst were enhanced to 2.15 mmol sꢁ1 (g Au)ꢁ1 and 0.42 sꢁ1
from 0.20 mmol sꢁ1 (g Au)ꢁ1 and 0.17 sꢁ1 for the unleached
0
1
2
3
4
5
6
TOS (h)
Fig. 10. Effect of support calcination temperature (t) on the reaction rates for 1,3-
butadiene consumption by Au/ZrO2-CP-t-673 (A) and Au/ZrO2-AN-t-673 (B) of
t = 673 (e), 873 (N), 973 (s), and 1073 K (j).
0.25
A
0.20
0.15
0.10
0.05
0.00
0
1
2
3
4
5
6
25
20
15
10
5
TOS (h)
Fig. 11. Effect of catalyst calcination temperature (T) on the reaction rate for Au/
ZrO2-CP-673-T of T = 393 (e), 473 (j), 573 (s), 673 (N), and 773 K (h).
would cause Au3+ reduction and lower the Au3+/Au0 ratio from 0.42
to 0 (Table 2). The much lower activity of the T = 773 catalyst than
of T = 673 was probably due to the loss of hydroxyl groups at the
support surface as aforementioned.
Note that the butadiene conversion over the most active Au/
ZrO2-CP-673-393 was as high as 98.4% when the reaction was car-
ried out with the ‘‘standard’’ reaction space velocity,
GHSV = 8100 ml hꢁ1 (g cat.)ꢁ1 (Fig. S3). Such high butadiene conver-
sion could heavily mask or undervalue the catalytic activity because
the reaction deviated severely from the situation in a differential
reactor. We therefore increased the feed space velocity to measure
the reaction rate of this Au/ZrO2-CP-673-393 catalyst. We observed
0
0
1
2
3
4
5
6
TOS /h
Fig. 12. Effect of catalyst calcination temperature (T) on butadiene conversion for
Au/ZrO2-CP-1073-T at T = 473 (h), 573 (d), and 673 K ( ).
D

84
X. Zhang et al. / Journal of Catalysis 279 (2011) 75–87
100
80
60
40
20
0
2.5
2.0
1.5
1.0
0.5
0.0
0
1
2
3
4
5
6
0
1
2
3
4
5
6
TOS (h)
TOS (h)
Fig. 14. Effect of reduction pretreatment on the reaction rate over 0.05%Au/ZrO2-
CP-673-473 catalyst: (e) 473 K in Ar, (j) 473 K in H2, (s) 523 K in H2.
Fig. 13. Effect of gold loading on 1,3-butadiene conversion at 393 K over Au/ZrO2-
CP-673-473 containing 0.76 (ꢀ), 0.23 (j), 0.08 (e), 0.05 (s), and 0.01% Au (h).
for the hydrogenation of 1,3-butadiene. The former two factors
were not always independent and were sensitively related to the
gold loading, the calcination temperature of ZrO2 before gold load-
ing (t), and the catalyst calcination temperature (T).
catalyst. For the hydrogenation of 1-pentene over Au/SiO2, Sermon
et al. [11] also found that the activity of gold increased remarkably
when the gold loading was lowered to less than 0.1% (e.g., 0.01% or
0.05%). The active gold species in those Au/SiO2 catalysts could
most possibly be Au3+ rather than Au0.
All of the present Au/ZrO2 catalysts except the cyanide-treated
one were prepared by deposition–precipitation at pH 9.0. The
dominant gold species in solution would be AuðOHÞꢁ4 and/or
AuClðOHÞꢁ3 [53] and the surface of the ZrO2 support would carry
negative charges, as its isoelectronic point is pH 6.7. Scheme 1 is
proposed to illustrate possible interactions of AuðOHÞꢁ4 at the sur-
face. The deposition of AuðOHÞꢁ4 would lead to surface gold oxyhy-
droxide or hydroxyl gold species (e.g., AuO(OH) and/or Au(OH)3).
These species could easily dehydrate to give Au2O3 at 413–423 K.
This latter product would then decompose to give metallic Au
when calcined at slightly higher temperatures; decomposition of
unsupported gold oxides to metallic Au occurred even at 353 K
[2]. Thus, the presence of unreduced Au3+ ions in our samples
would be associated with oxy-gold species, such as AuO(OH),
Au(OH)3, or Au2O3, stabilized on the support surface [26,27,31,43].
When the gold loading on ZrO2-CP-673 (120 m2/g) was lowered
from 0.76% to 0.23% and then 0.08%, the density of gold atoms at
the support surface would decrease from 19.4 to 5.9 and then 2.1
gold atoms per 100 nm2. The lower density of gold atoms on the
support surface would probably limit the sintering and Ostwald
ripening of gold [53], which accounts for the observed increase
in the average Au particle size with increasing gold loading. The
addition of gold titrated, in a sense, the stronger binding sites
and then weaker ones on the support surface. Thus, the oxy-gold
species at very low gold loadings (e.g., <0.1%) would be strongly
bound and their decomposition or reduction to Au0 would require
some higher temperatures. This issue is supported by our present
data (Fig. 5) that the Au3+/Au0 ratio and TPR peak temperature
for Au3+ reduction both declined with increasing gold loading for
the m%Au/ZrO2-CP-673-473 samples (Table 2). A recent DFT calcu-
lation study showed that isolated Au3+ ions on a stepped ZrO2 sur-
face are more stable than those on a flat surface [54].
Interestingly, the reaction over the Au/ZrO2-CP-673-473 cata-
lysts with various gold loadings was found to be 100% selective
for butenes (Fig. S4); no butane was detected even over the most
active 0.08% Au/ZrO2-CP-673-473 catalyst (butadiene conversion
85.4%). The thermodynamic equilibrium composition of butenes
at 393 K is calculated to be 8.3%, 63.8%, and 27.8% for 1-butene,
trans-2-butene, and cis-2-butene. The quantity of the major com-
ponent of butenes obtained in the present study (ca. 60% 1-butene)
is much higher than the value predicted from the thermodynamic
equilibrium composition of butenes. The high concentration of
1-butene could indicate that the hydrogenation of 1,3-butadiene
on Au/ZrO2 catalysts mainly occurs according to the 1,2-addition
pathway, which is a feature markedly different from thermal
hydrogenation via 1,4-addition to trans-2-butene as the main
product [19]. The trans/cis ratio of 2-butenes was in the range
0.3–1.0 over all Au/ZrO2 catalysts, which is lower than the ratio
of 1.5–2 (close to the equilibrium value) obtained over Pt and other
Group VIII noble metals [19]. According to a mechanism proposed
by Meyer and Burwell [51], the semihydrogenated 1,3-butadiene
would be adsorbed either as syn- or anti-p-allyl species. The ratio
of these two surface intermediates should be close to that in the
gas phase, which is about 1:10 [19]. If this was the case, the inter-
conversion from an anti to a syn intermediate would occur easily
on Au/ZrO2, resulting in a low proportion of trans-2-butene. This
would mean kinetically favorable trans- to cis- isomerization dur-
ing the reaction over a gold catalyst [52].
The key role of Au3+ ions in determining the hydrogenation
activity was further confirmed by investigating the effect of
hydrogen reduction on the catalytic activity of 0.05%Au/ZrO2-
CP-673-473 (Fig. 14). H2-reduction pretreatment at 473 K led to a
sharp decrease in the reaction rate (from ca. 2.01 to 0.88 mmol sꢁ1
(g Au)ꢁ1), and complete reduction of Au3+ ions at 523 K further
decreased the rate to 0.24 mmol sꢁ1 (g Au)ꢁ1. These results clearly
suggest that the catalyst suffered severe activity losses after the
reduction of Au3+ ions.
Due to the much lower surface area (14 m2/g) and concentra-
tion of hydroxyl groups for ZrO2-CP-1073, compared with ZrO2-
CP-673, the density of gold atoms at the support surface increased
to as high as 158 per 100 nm2 in the Au/ZrO2-CP-1073-T samples.
This would significantly reduce the stability of cationic gold spe-
cies while at the same time increasing the probability of gold sin-
tering during the catalyst preparation (Scheme 1), leading to lower
Au3+/Au0 ratios and larger Au particle sizes for the Au/ZrO2-
CP-1073-T catalysts.
4. Discussion
The present data demonstrate that the Au3+/Au0 ratio, ZrO2-
based hydroxyl groups, and moisture in the reaction feed are
important factors to the catalytic activity of a Au/ZrO2 catalyst
Differences in the size and size distribution of Au and ZrO2
nanoparticles could also influence their activity, as documented

X. Zhang et al. / Journal of Catalysis 279 (2011) 75–87
85
HO
HO
OH
HO
OH
-
Au
-
Au
-
[Au(OH)4]
[AuCl4]
+
NH4OH
HO
OH
-
O
OH
Zr
O
OH
Zr
-
H2O
Au(OH)3 + OH
Zr
Zr
Zr
O
O
O
2H2O
O
O
O
2H2O
OH
-
HO
HO
OH
Au
Au
Au0
ZrO2
H2O
O
calcination
Au2O3
ZrO2
-
O
O
O
OH
Zr
Zr
Zr
Zr
Zr
O
O
O
O
O
Scheme 1. Possible interaction of AuðOHÞꢁ4 ions with the negatively charged ZrO2 surface and their subsequent transformations during catalyst preparation.
for CO oxidation [1–6]. Okumura et al. [12] found that the activity
of Au/Al2O3 catalysts for butadiene hydrogenation was almost
insensitive to the size of Au nanoparticles (2.5–37.0 nm). In the
work of Hugon et al. [55], the TOF rates on exposed Au0 atoms in
both Au/TiO2 and Au/Al2O3 catalysts for butadiene hydrogenation
also varied slightly (within a factor of 1.3) when the average Au
particle size changed from 5 to 2 nm. By changing the calcination
temperature (373–923 K) of an Au/Al2O3 catalyst to effect changes
in the Au particle size, Jia et al. [13] observed that the optimum Au
particle size for the hydrogenation of acetylene was ca. 3 nm. How-
ever, these authors made no consideration of the possible involve-
ment of oxy-gold species and surface hydroxyl groups. By carefully
controlling the size of metallic Au particles (average 4–5 nm) on
ZrO2-AN-t support, the catalytic activity of Au/ZrO2-AN-t-673 also
decreased with increasing t (see Fig. 10B and Table 2). So the sizes
of Au nanoparticles in Au/ZrO2-CP-t catalysts would not be a main
cause for the sharp decrease in the TOF from 0.12 to 0.01 sꢁ1 when
t was increased from 673 to 1073 K (Table 2).
25–40 nm) or in Au/ZrO2-CP-1073-673 and -AN-1073-673 cata-
lysts appeared inactive for butadiene hydrogenation but became
remarkably active after being subjected to a water treatment
(e.g., Fig. 9). This kind of catalyst ‘‘activation’’ by water treatment
did not involve any Au3+, as the TPR measurement detected no for-
mation of Au3+ ions in the water-treated sample (Fig. 5B). These re-
sults would instead indicate that the size of the zirconia support
was not among the keys to the hydrogenation activity. Therefore,
the remarkably enhanced activity after the water treatment should
be associated with the partial regeneration of hydroxyl groups on
the ZrO2 surface, as indicated by the IR and TG measurements,
which convincingly demonstrates the vital role of the surface hy-
droxyl groups in the hydrogenation of 1,3-butadiene.
OH groups on ZrO2 were proposed to supply protons, either di-
rectly or indirectly [57,58]. However, H/D exchange reactions be-
tween D2 and the OH groups occurred only at above 323 K [59],
giving a maximum H/D exchange rate at 418 K [60]. Indeed, both
butadiene hydrogenation and H2/D2 exchange reactions were ob-
served using a closed recirculation reactor at 323–348 K on ‘‘pure’’
ZrO2 catalysts pretreated at varying temperatures (673–1173 K)
[61]. The specific surface reaction rates for both reactions increased
with increasing catalyst pretreatment temperature up to 873 K.
However, the catalytic rates declined with further increase of the
pretreatment temperature and became zero when the pretreat-
ment temperature was elevated to 1073 K or higher. These earlier
data [61] could imply that the surface hydroxyl groups could be a
key to H2 activation on zirconia. By using a H2/D2 mixture for buta-
diene hydrogenation, Buchanan and Webb [10] found earlier that
the activation of H2/D2 on Au/Al2O3 catalysts would involve a
At the present stage, it remains difficult to decouple the two ef-
fects on gold activity when they coexist: hydroxyl groups and
nanosize of ZrO2 particles. Obviously, the Au–ZrO2 contact bound-
ary or interface would increase with the decline in ZrO2 particle
size if Au particles could be made in similar sizes. In this sense,
the Au–ZrO2 interface would be maximized in nanocomposite
Au/ZrO2-AN-673-673 composed of comparably sized Au (4–5 nm)
and ZrO2-AN-673 (5–10 nm) particles [5]. Those metallic gold
atoms in close contact with the ZrO2 support or located at the
Au–ZrO2 interface have proven to be the catalytic active sites for
CO oxidation [1,2,5], and H2–D2 exchange reactions [22]. At first
glance, the obtaining of the highest 1,3-butadiene conversion
(49.0%) and specific rate (0.13 mmol sꢁ1 (g Au)ꢁ1) on Au/ZrO2-
AN-673-673 among all of the Au/ZrO2-AN-t-673 and -CP-t-673
catalysts (673 6 t 6 1073 K) could indicate a possible nanosize
effect of the support particles on the activity of Au particles [56].
In this case, the slightly higher activity of Au/ZrO2-AN-973-673
than of Au/ZrO2-CP-973-673 was explicable by the smaller average
size of ZrO2 particles in the former. Nevertheless, Au/ZrO2-AN-
873-673 with smaller ZrO2 particles compared to Au/ZrO2-CP-
873-673 exhibited lower activity, which indicates that the higher
concentration of hydroxyl groups in the latter catalyst could
outweigh the inferiority of larger ZrO2 particles.
hydrogen source (surface OH groups) on the
c-Al2O3 support,
and metallic Au particles on OH-free -Al2O3 support were inactive
a
for the hydrogenation reaction. Recently, metallic Au clusters on
partially dehydroxylated MgO (pretreated in vacuum at 673 K)
were also found inactive in ethylene hydrogenation [30].
In this study, nanosized ZrO2 powders without gold showed no
activity for the hydrogenation reaction. The TOF data for butadiene
consumption based on exposed Au0 atoms over our Au/ZrO2 cata-
lysts were in the range of 0.01–0.12 sꢁ1 (Table 2). These data are
close to those (ca. 0.01–0.09 sꢁ1) at 423 K for the Au/TiO2 and
Au/Al2O3 catalysts under similar reaction conditions [12]. Gluhoi
et al. [62] showed that the D/H exchange rate between D2 and
The Au3+-free metallic Au nanoparticles supported on OH-free
ZrO2 (ZrO2-CP-1073 of 40–200 nm and ZrO2-AN-1073 of
the surface hydroxyl groups of
c-Al2O3 was significantly enhanced
when the oxide ( -Al2O3) was loaded with nanosized Au particles.
c

86
X. Zhang et al. / Journal of Catalysis 279 (2011) 75–87
Thus, it can be argued that the presence of metallic Au particles can
promote the activation/dissociation of H2 on OH-carrying ZrO2 sur-
faces. Indeed, our H2-TPD experiments evidenced that the presence
of Au particles improved the chemisorption of H2, since a weak H2-
desorption peak was detected for Au/ZrO2-CP-673-673 but not for
ZrO2-CP-673 (Fig. 7d and f). Au/ZrO2-CP-1073-673 with no surface
OH-groups was also inactive for H2 chemisorption. In a catalyst
containing no Au3+ ions, the metallic Au particles and the OH-car-
rying ZrO2 have to work together for H2 activation.
The Au/ZrO2 catalysts summarized in Table 2 can be grouped
into four categories. The Au/ZrO2-CP-1073-673 and -AN-1073-
673 catalysts, which feature metallic gold particles supported on
an OH-free ZrO2 surface as depicted in Scheme 2a, are ascribed
to the first category. The surface Au0 atoms in these catalysts were
essentially inactive for the selective hydrogenation of 1,3-butadi-
show that even so, interaction between metallic Au particles and
zirconia can be important to the H2 activation on the Au/ZrO2 cat-
alysts; participation of surface hydroxyl groups is, however, quite
indispensable. A solid piece of evidence for this argument is our
observation that the inactive Au/ZrO2-CP-1073-673 catalyst can
also become highly active after it is subjected to simple water
treatment (Fig. 9). In a very recent report of Fujitani et al. [23],
the activity of Au/TiO2 for H2 dissociation could be related to the
number of Au0 atoms located at the circumference of the sup-
ported Au particles. This does not violate our point that H2 activa-
tion in the absence of Au3+ ions would need cooperation between
the Au particles and the oxide-based OH groups, since it does not
exclude the possibility that some OH groups residing in the sup-
port oxide (TiO2) could be also involved in the H2 dissociation
chemistry.
ene, as in the cases of OH-free Au/
a
-Al2O3 for the H/D exchange
Au/ZrO2 catalysts containing no more than 0.08% Au, which fea-
ture isolated Au3+ ions on the support surface (Scheme 2c), are as-
cribed to the third category. The isolated Au3+ in the catalysts of
this category could be surface immobilized AuO(OH) or Au(OH)3
[50]; they showed the same highest catalytic activity by the reac-
tion rate (2.1 mmol sꢁ1 (g Au)ꢁ1) and TOF (0.41 sꢁ1), despite their
varied gold loadings (0.01–0.08%). As we have discussed earlier
[27,31], these results suggest that the isolated Au3+ ions are equally
active for the hydrogenation of 1,3-butadiene. It should be pointed
out that the reaction rate and TOF of these isolated Au3+ ions were
1–2 orders of magnitude higher than those of the Au particles in
the samples of the second category (Table 2). The TOF of these iso-
lated Au3+ ions is even comparable to those of Group VIII noble me-
tal catalysts (e.g., 0.2–1.0 sꢁ1 for supported Pd catalysts) [19]. The
remarkable difference in activity by TOF between these surface-
type Au3+ ions and Au0 atoms could mean different mechanisms
for the required dissociative activation of H2. DFT calculations
[21,54] indicate that H2 dissociation on cationic gold species would
be easy, although metallic Au0 atoms are noble to H2 chemisorp-
tion. The capability of surface Au3+ ions for H2 activation was
clearly evidenced in our H2-TPD experiments, where the amount
of chemisorbed hydrogen was sensitively related to the sample
Au3+/Au0 ratio (Fig. 7).
reaction [10] and dehydroxylated Au/MgO for the hydrogenation
of ethene [30]. Therefore, metallic Au nanoparticles on OH free
ZrO2 surfaces are unlikely to possess active sites for H2 activa-
tion/dissociation. The surface of these Au particles resembles those
of bulk Au in that they show no activity for H2 dissociation [21].
This explanation is further supported by our observation that the
Au/ZrO2-CP-1073-673 catalyst showed no activity for H2 chemi-
sorption, as revealed by the H2-TPD data (Fig. 7e).
The Au/ZrO2-CP-t-T and -AN-t-T catalysts (t < 1073 K) with
T = 673 or 773 K, which feature metallic Au nanoparticles sup-
ported on ZrO2 surface with more or less residual OH-groups (see
Scheme 2b), are assigned to the second category. It should be
pointed out that the catalysts in this category have been the sub-
ject of many investigations. A number of models have been pro-
posed toward understanding of the catalytically active gold sites
for the hydrogenation catalysis [9,10,12,23,63]. Exposed metallic
gold atoms were proposed to be the active sites in most cases,
including those on very small Au clusters not sensed by TEM [9]
or less than 2 nm [23], low-coordinated edge or corner atoms at
the surfaces of Au nanoparticles [12,63]. However, a model pro-
posed by Buchanan and Webb [10] claimed that the active sites
responsible for H2 activation were originated from the oxide
(c
-Al2O3) support surface but not the Au particles. Our data clearly
The catalysts in the fourth category include 0.23% Au/ZrO2-CP-
673-473 and Au/ZrO2-CP-673-T with T = 393, 473, and 573 K,
which feature coexisting Au0 atoms and Au3+ ions on OH-carrying
ZrO2 surface (Scheme 2d). These catalysts may be regarded as mix-
tures of the catalysts in categories 2 and 3 in varying proportions.
There are thus three types of surface sites for H2 activation/disso-
ciation on the catalysts of this category, one being at the interfaces
between the Au nanoparticles and the OH-carrying ZrO2, the sec-
ond being the isolated surface Au3+ ions, and the third involving
ensembles of Au0 atoms and Au3+ ions on the OH-carrying ZrO2
surface. This third type of catalytic sites was also proposed previ-
ously to operate in gold-catalyzed CO oxidation reactions [3,64]
but had not yet been considered in gold-catalyzed hydrogenation
reactions. For the catalysts in this category, the TOF data based
on exposed Au0 atoms (numbers outside the parentheses in
Table 2) appeared not very different from those based on the
Au3+ ions (numbers inside the parentheses in Table 2). However,
the activity of these Au3+ ions by TOF (0.16–0.25 sꢁ1) is much lower
than that (0.41 sꢁ1) of the isolated Au3+ ions in the catalysts of the
third category (Table 2 and Fig. 13). This difference could hint that
the Au0 atoms could lower the catalytic activity of their coexisting
Au3+ ions, either by burying some of the Au3+ ions or by electronic
interaction [27].
a
b
c
d
O2-,
Zr4+,
Au0,
Au3+,
H
Scheme 3 is proposed to highlight the richness of catalytic sites
for hydrogen activation and their subsequent reaction with 1,3-
butadiene over Au/ZrO2 catalysts. Though little information is
available at this stage for the adsorption of 1,3-butadiene, this sub-
strate would become adsorbed by nucleophilic interaction via
Scheme 2. Proposed structure models for Au/ZrO2 catalysts: (a) metallic Au0
particles deposited on OH-free ZrO2-CP-1073 or ZrO2-AN-1073; (b) metallic Au0
particles deposited on OH-carrying ZrO2; (c) immobilized isolated Au3+ species on
ZrO2; (d) dispersed Au0 particles and Au3+ species on ZrO2 with varying OH
concentrations.

X. Zhang et al. / Journal of Catalysis 279 (2011) 75–87
[3] G.C. Bond, D.T. Thompson, Gold Bull. 33 (2000) 41.
87
[4] M.C. Kung, R.J. Davis, H.H. Kung, J. Phys. Chem. C 111 (2007) 11767.
[5] X. Zhang, H. Wang, B.-Q. Xu, J. Phys. Chem. B 109 (2005) 9678.
[6] M.S. Chen, D.W. Goodman, Catal. Today 111 (2006) 22.
[7] P. Claus, Appl. Catal. A 291 (2005) 222.
[8] A. Comas-Vives, C. Gonzalez-Arellano, A. Corma, M. Iglesias, F. Sánchez, G.
Ujaque, J. Am. Chem. Soc. 128 (2006) 4756.
[9] G.C. Bond, P.A. Sermon, G. Webb, D.A. Buchanan, P.B. Wells, J. Chem. Soc. Chem.
Commun. (1973) 444.
[10] D.A. Buchanan, G. Webb, J. Chem. Soc. Faraday Trans. 71 (1975) 134.
[11] P.A. Sermon, G.C. Bond, P.B. Wells, J. Chem. Soc. Faraday Trans. 75 (1979) 385.
[12] M. Okumura, T. Akita, M. Haruta, Catal. Today 74 (2002) 265.
[13] J. Jia, K. Haraki, J.N. Kondo, K. Domen, K. Tamaru, J. Phys. Chem. B 104 (2000)
11153.
[14] J.E. Bailie, G.J. Hutchings, Chem. Commun. (1999) 2151.
[15] J.E. Bailie, H.A. Abdullah, J.A. Anderson, C.H. Rochester, N.V. Richardson, N.
Hodge, J.-G. Zhang, A. Burrows, C.J. Kiely, G.J. Hutchings, Phys. Chem. Chem.
Phys. 3 (2001) 4113.
[16] J.E. Bailie, G.J. Hutchings, Catal. Commun. 2 (2001) 291.
[17] C. Mohr, H. Hofmeister, P. Claus, J. Catal. 213 (2003) 86.
[18] H. Shi, N. Xu, D. Zhao, B.-Q. Xu, Catal. Commun. 9 (2008) 1949.
[19] H. Arnold, F. Döbert, J. Gaube, in: G. Ertl, H. Knözinger, J. Weitkamp (Eds.),
Handbook of Heterogeneous Catalysis, vol. 5, VCH Weinheim, 1997, p. 2165.
[20] S. Lin, M.A. Vannice, Catal. Lett. 10 (1991) 47.
[21] B. Hammer, J.K. Norskov, Nature 376 (1995) 238.
[22] S.A. Varganov, R.M. Olson, M.S. Gordon, G. Mills, H. Metiu, J. Chem. Phys. 120
(2004) 5169.
[23] T. Fujitani, I. Nakamura, T. Akita, M. Okumura, M. Haruta, Angew. Chem. Int.
Ed. 48 (2009) 9515.
Scheme 3. Proposed reaction routes over zirconia-supported gold catalyst for the
selective hydrogenation of 1,3-butadiene. Note: the activated hydrogen atoms are
shown as the HI species.
either C@C bonds with the surface Au0 and/or Au3+ sites [10,52], or
Zr4+ or Zr3+ sites at the surface of zirconia [61]. Further reaction be-
tween the adsorbed butadiene and the activated hydrogen pro-
duces the product butenes.
The stability of surface Au3+ ions in the reducing atmosphere of
the hydrogenation reaction can be critical for applications. The
slightly changed catalytic rates, as observed in Figs. 11 and 13,
and Fig. S3, would demonstrate that the active Au3+ ions in our cat-
alysts were stable enough to survive the reaction. Our TPR data
(Fig. 5) evidenced that the onset reduction temperature for the
Au3+ ions was always higher than the reaction temperature
(393 K). Stabilization of Au3+ ions by interaction with the support
oxide was also documented in the literature, especially when
CeO2 [28,29] and ZrO2 [45,65] were used as the support oxides.
Our quantitative TPR measurement revealed that the quantity of
Au3+ ions in the Au/ZrO2-CP-673-473 catalyst did not change after
it was used to catalyze the hydrogenation reaction for 6 h. This is a
piece of strong evidence for the catalytic stability of Au3+ ions in
the hydrogenation reaction.
[24] M. Boronat, F. Illas, A. Corma, J. Phys. Chem. A 113 (2009) 3750.
[25] G.J. Hutchings, M.S. Hall, A.F. Carley, P. Landon, B.E. Solsona, C.J. Kiely, A.
Herzing, M. Makkee, J.A. Moulijn, A. Overweg, J.C. Fierro-Gonzalez, J. Guzman,
B.C. Gates, J. Catal. 242 (2006) 71.
[26] E.D. Park, J.S. Lee, J. Catal. 186 (1999) 1.
[27] X. Zhang, H. Shi, B.-Q. Xu, Catal. Today 122 (2007) 330.
[28] S. Carrettin, A. Corma, M. Iglesias, F. Sánchez, Appl. Catal. A 291 (2005) 247.
[29] Q. Fu, H. Saltsburg, M. Flytzani-Stephanopoulos, Science 301 (2003) 935.
[30] J. Guzman, B.C. Gates, J. Catal. 226 (2004) 111.
[31] X. Zhang, H. Shi, B.-Q. Xu, Angew. Chem. Int. Ed. 44 (2005) 7132.
[32] X. Zhang, H. Shi, B.-Q. Xu, Science and Technology in Catalysis 2006, in: K.
Eguchi, M. Machida, I. Yamanaka, Eds., Kondansha-Elsevier, Stud. Surf. Sci.
Catal., vol. 172, 2007, p. 481.
[33] M. Daté, M. Haruta, J. Catal. 201 (2001) 221.
[34] C.K. Costello, J.H. Yang, H.Y. Law, Y. Wang, J.N. Lin, L.D. Marks, M.C. Kung, H.H.
Kung, Appl. Catal. A 243 (2003) 15.
[35] B.-Q. Xu, J.-M. Wei, J.-L. Li, Q.-M. Zhu, Top. Catal. 22 (2003) 77.
[36] J. Kondo, Y. Sakata, K. Domen, K. Maruya, T. Onishi, J. Chem. Soc. Faraday Trans.
86 (1990) 397.
[37] G.K. Chuah, S. Jaenicke, B.K. Pong, J. Catal. 175 (1998) 80.
[38] S.L. Jones, C.J. Norman, J. Am. Ceram. Soc. 71 (1988) C190.
[39] Y. Nakano, T. Iizuka, H. Hattori, K. Tanabe, J. Catal. 57 (1978) 1.
[40] L.L. Hench, J.K. West, Chem. Rev. 90 (1991) 33.
[41] K.T. Jung, A.T. Bell, J. Mol. Catal. A 163 (2000) 27.
[42] M.G. Mason, Phys. Rev. B 27 (1983) 748.
[43] N.A. Hodge, C.J. Kiely, R. Whyman, M.R.H. Siddiqui, G.J. Hutchings, Q.A.
Pankhurst, F.E. Wagner, R.R. Rajaram, S.E. Golunski, Catal. Today 72 (2002)
133.
[44] J. Guzman, B.C. Gates, J. Phys. Chem. B 107 (2003) 2242.
[45] L. Ilieva, J.W. Sobczak, M. Manzoli, B.L. Su, D. Andreeva, Appl. Catal. A 291
(2005) 85.
5. Conclusions
This work shows that the gold oxidation state and hydroxyl
groups on the support oxide surface played vital roles in hydroge-
nation catalysis by Au/ZrO2. A continued activity decrease is seen
when the density of surface hydroxyl groups is lowered by elevat-
ing the pre-calcination temperature of ZrO2. Au nanoparticles sup-
ported on OH-free ZrO2 are inactive, but become highly active after
they are subjected to a simple water treatment. Catalytic ensem-
bles involving metallic Au0 atoms and Au3+ ions on OH-carrying
ZrO2 have been proposed to address the structural features of ac-
tive sites in these Au/ZrO2 catalysts for H2 activation.
[46] J.T. Calla, R.J. Davis, Catal. Lett. 99 (2005) 21.
[47] Y. Guan, E.J.M. Hensen, Phys. Chem. Chem. Phys. 11 (2009) 9578.
[48] P. Claus, A. Brücker, C. Mohr, H. Hofmeister, J. Am. Chem. Soc. 122 (2000)
11430.
[49] H. Liu, L. Feng, X. Zhang, Q. Xue, J. Phys. Chem. 99 (1995) 332.
[50] L. Zhao, H. Liu, W. Bu, J. Gao, C. Xu, X. Zhang, Y. Chen, Function of Water for the
Adsorption of H2 and 1,3-butadiene on Au (1 1 1): A DFT Study, unpublished
results.
Acknowledgment
[51] E.F. Meyer, R.L. Burwell, J. Am. Chem. Soc. 85 (1963) 2881.
[52] X.F. Yang, A.Q. Wang, Y.L. Wang, T. Zhang, J. Li, J. Phys. Chem. C 114 (2010)
3131.
[53] F. Moreau, G.C. Bond, A.O. Taylor, J. Catal. 231 (2005) 105.
[54] Z.P. Liu, C.M. Wang, K.N. Fan, Angew. Chem. Int. Ed. 45 (2006) 6865.
[55] A. Hugon, L. Delannoy, C. Louis, Gold Bull. 41 (2008) 127.
[56] X. Zhang, B.-Q. Xu, Chem. J. Chin. Univ. 26 (2005) 106.
[57] T. Onishi, H. Abe, K. Maruya, K. Domen, J. Chem. Soc. Chem. Commun. 9 (1985)
617.
We thank the NSF of China (20921001 and 20903119) and the
Program for New Century Excellent Talents in University (NCET)
for financial support.
Appendix A. Supplementary material
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.jcat.2011.01.002.
[58] J. Kondo, K. Domen, K. Maruya, T. Onishi, Chem. Phys. Lett. 188 (1992) 443.
[59] K. Tanabe, Mater. Chem. Phys. 13 (1985) 347.
[60] D. Martin, D. Duprez, J. Phys. Chem. B 101 (1997) 4428.
[61] Y. Nakano, T. Yamaguchi, K. Tanabe, J. Catal. 80 (1983) 307.
[62] A.C. Gluhoi, H.S. Vreeburg, J.W. Bakker, B.E. Nieuwenhuys, Appl. Catal. A 291
(2005) 145.
[63] E. Bus, J.T. Miller, J.A. van Bokhoven, J. Phys. Chem. B 109 (2005) 14581.
[64] H.H. Kung, M.C. Kung, C.K. Costello, J. Catal. 216 (2003) 425.
[65] X. Zhang, A. Corma, Angew. Chem. Int. Ed. 47 (2008) 4358.
References
[1] G.C. Bond, C. Louis, D.T. Thompson (Eds.), Catalysis Science Series 6, Catalysis
by Gold, Imperial College Press, United Kingdom, 2006. p. 1.
[2] M. Haruta, S. Tsubota, T. Kobayashi, H. Kageyama, M.J. Genet, B. Delmon, J.
Catal. 144 (1993) 175.