Selective hydrogenation of acetylene in the presence of ethylene on zeolite-supported bimetallic catalysts

Article abstract of DOI:10.1016/j.jcat.2006.11.013

Novel catalysts were synthesized by supporting Pd-based bimetallic catalysts on Na⁺-exchanged β-zeolites to achieve higher selectivity for the hydrogenation of acetylene in the presence of ethylene at room temperature. Batch reactor studies carried out using infrared spectroscopy show that Pd-Ag bimetallic catalysts have higher selectivity for acetylene hydrogenation in the presence of ethylene than either Pd or Ag monometallic catalysts. Kinetic modeling of the reaction data reveals significant differences in the hydrogenation rate constants and adsorption equilibrium constants. The influence of the Na⁺-β-zeolite support is compared with traditional γ-Al₂O₃-supported catalysts. The Na⁺-β-zeolite-supported catalysts exhibit higher selectivity than their γ-Al₂O₃ counterparts. Overall, zeolite-supported Pd-Ag bimetallic catalyst was found to have the highest selectivity of the catalysts studied.

Full text of DOI:10.1016/j.jcat.2006.11.013

Journal of Catalysis 246 (2007) 40–51
www.elsevier.com/locate/jcat
Selective hydrogenation of acetylene in the presence of ethylene
on zeolite-supported bimetallic catalysts
Wei Huang a, John R. McCormick b, Raul F. Lobo a, Jingguang G. Chen a,∗
a
Department of Chemical Engineering, University of Delaware, Newark, DE 19711, USA
b
Department of Materials Science and Engineering Center for Catalytic Science and Technology (CCST), University of Delaware, Newark, DE 19711, USA
Received 20 June 2006; revised 12 October 2006; accepted 11 November 2006
Available online 21 December 2006
Abstract
+
Novel catalysts were synthesized by supporting Pd-based bimetallic catalysts on Na -exchanged β-zeolites to achieve higher selectivity for the
hydrogenation of acetylene in the presence of ethylene at room temperature. Batch reactor studies carried out using infrared spectroscopy show
that Pd–Ag bimetallic catalysts have higher selectivity for acetylene hydrogenation in the presence of ethylene than either Pd or Ag monometallic
catalysts. Kinetic modeling of the reaction data reveals significant differences in the hydrogenation rate constants and adsorption equilibrium
+
+
constants. The influence of the Na -β-zeolite support is compared with traditional γ -Al O -supported catalysts. The Na -β-zeolite-supported
2
3
catalysts exhibit higher selectivity than their γ -Al O counterparts. Overall, zeolite-supported Pd–Ag bimetallic catalyst was found to have the
2
3
highest selectivity of the catalysts studied.
© 2006 Elsevier Inc. All rights reserved.
Keywords: Hydrogenation; Acetylene; Ethylene; Pd–Ag; Pd–Ni; Bimetallic catalysts; Zeolite
1. Introduction
Previous studies [2,8–12] showed that acetylene hydrogena-
tion produces ethylene, which is then hydrogenated to ethane.
The initial steps of the reaction mechanism include dissociative
adsorption of hydrogen on the catalyst surface. Acetylene and
ethylene are also co-adsorbed on the catalyst before reaction.
Addition of the first hydrogen atom to acetylene or ethylene
is the rate-determining step [9–14]. Ethylene–acetylene hydro-
genation is a classical case of a series-parallel reaction. The
reaction is in series in terms of acetylene and ethylene but par-
allel with respect to hydrogen, which means there is no direct
formation of ethane from acetylene [2,12]. The hydrogenation
mechanisms can be written as
Ethylene is a crucial intermediate in many industrial reac-
tions. One of the most important of these is ethylene polymer-
ization. Ethylene is produced by thermal or catalytic cracking of
higher hydrocarbons, which also generates impurities, such as
acetylene, that can poison traditional polymerization catalysts.
To prevent catalyst poisoning, the concentration of acetylene
must be <5 ppm [1–4]. The typical method used in industry
for reducing acetylene concentration is selective hydrogenation
of acetylene using Pd catalysts [5–7]. Due to the poor selectiv-
ity at high acetylene conversion and oligomer formation during
acetylene hydrogenation, the current commercial Pd-based cat-
alysts have considerable room for improvement. The objective
of the current work is to develop new catalysts by combining
Pd-based bimetallic catalysts and zeolite supports to achieve
higher selectivity for the selective hydrogenation of acetylene
in a stream containing ethylene.
Zeolite-supported metal catalysts often show superior se-
lectivity and activity in many reactions [15,16]. In zeolites,
extra-framework cations can bind to the π face of a double
*
Corresponding author.
E-mail address: jgchen@udel.edu (J.G. Chen).
0021-9517/$ – see front matter © 2006 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2006.11.013

W. Huang et al. / Journal of Catalysis 246 (2007) 40–51
41
or triple carbon–carbon bond structure through a noncovalent
2. Experimental
force known as cation–π interaction [17,18]. The interaction
can be considered an electrostatic attraction between a posi-
tive charge and the quadrupole moment of the π-bond. The
interaction is usually strong and changes with cation identity.
Alkenes generally are more polarizable than alkynes, yet show
lower binding energies to cations [19] in the gas phase be-
cause of this π–quadruple interaction. One possible advantage
of zeolite-supported catalyst is to use cation–π interactions
to preferentially adsorb acetylene in an ethylene-rich environ-
ment, thus enhancing the local concentration of acetylene and
achieving higher hydrogenation selectivity. To avoid blocking
of the zeolite pores by metal particles and reaction byprod-
ucts, the preferred zeolite is expected to be a structure contain-
ing a multidimensional pore system with large pore apertures
(12 rings). Two promising candidates are the zeolite Y or X
(FAU) and the zeolite β (*BEA). We studied zeolite β in the
present work.
A major challenge is that using preferential adsorption of
acetylene to increase selectivity requires carrying out the exper-
iments at relatively low temperatures. There is no distinguish-
able preferential adsorption at high temperatures, because the
difference in adsorption equilibrium constants decreases expo-
nentially with temperature. The bimetallic system is a plausible
solution to achieve acceptable levels of low-temperature hydro-
genation activity. Bimetallic catalysts also provide a method
of tuning catalytic materials to obtain highly selective and
highly active catalysts. Comparisons of recombinative desorp-
tion of H2 and D2 on the bimetallic and the corresponding
monometallic surfaces have been studied previously [20–25].
The dissociative adsorption of H2 on transition metal surfaces
is a nonactivated process [26]; thus, the observation of a lower
desorption temperature on bimetallic surfaces indicates that hy-
drogen atoms bind more weakly than on either of the parent
metals. In principle, the presence of weakly bonded hydrogen
atoms should favor low-temperature hydrogenation, as demon-
strated for the hydrogenation of cyclohexene [21] and other
alkenes [27]. Thus, in the current study, bimetallic catalysts are
supported on the zeolites to potentially reduce the activation
energy of the rate-determining hydrogenation step and simulta-
neously maintain the preferential adsorption of acetylene at low
temperatures. Pd-based catalysts have shown superior selectiv-
ity and activity for the selective hydrogenation reaction [5–7];
thus a second metal, such as Ni or Ag, has been added to the
Pd catalyst in an attempt to achieve higher selectivity and activ-
ity.
2.1. Catalyst preparation
Supported catalysts were synthesized using Pd, Ni, and
Ag supported on either γ -Al2O3 (Alfa-Aesar, 99.7%) or
Na+-exchanged β-zeolite support using the incipient wetness
method [1–4]. The β-zeolite (SiO2:Al2O3 = 25, Zeolyst Com-
pany) was originally obtained in its ammonia form. Before
the addition of metal precursor, the zeolite was ion-exchanged
twice with 500 mL of 0.01 M NaNO3 solutions to obtain a
Na+-formed zeolite. After the ion-exchange process, the zeo-
lite was washed with DI water and dried at 383 K. The γ -Al2O3
support was used as received.
The Pd–Ni bimetallic catalysts (Table 1) were prepared by
co-impregnation using Pd(NO3)2 (Alfa-Aesar) and Ni(NO3)2 ·
xH2O (Alfa-Aesar), whereas the Pd–Ag bimetallic catalysts
were prepared by sequential impregnation using Pd(NO3)2 and
then AgNO3 (Alfa-Aesar), because of the low solubility of the
Ag salts. The impregnated supports were dried in air overnight.
The catalysts were then dried at 383 K for 5 h and calcined in
air at 623 K for 6 h.
The dispersion of supported Pd and PdAg catalysts was
determined by carbon monoxide (CO) chemisorption experi-
ments, whereas the dispersion of supported Ni and PdNi cat-
alysts was studied by hydrogen (H2) chemisorption experi-
ments to avoid the stoichiometric problem of CO–Ni. Pulse CO
chemisorption/H2 chemisorption experiments were carried out
using an Altamira Instruments AMI-200ip. The catalysts were
reduced in dilute hydrogen at 723 K for 1 h and cooled to room
temperature in a helium/argon stream. Pulses of CO/H2 in a he-
lium/argon carrier gas were injected into the quartz reactor, and
the signal was monitored by a thermal conductivity detector
(TCD). The mean stoichiometry CO–XPd was determined ac-
cording to the iterative method derived by Lambert et al. [28]:
In brief, the starting value of X = 0.5 was chosen for CO–XPd,
and a first value of Pd dispersion was calculated from the CO
uptake. A polynomial function was fitted based on the table es-
tablished by Joyal and Butt [29], who determined CO–XPd as
a function of dispersion in Pd/SiO2 catalysts. Using this fitted
function, a new value of CO–XPd was determined correspond-
ing to the precedent obtained dispersion value; this new value of
CO–XPd was used to calculate a new value of dispersion. The
iterative cycle was repeated until convergence. Using this itera-
tive method, the values of CO–XPd were determined to be 0.62
for Pd/γ -Al2O3 and 0.69 for Pd/Na+-β-zeolite; these values
were used to determine the dispersion of the two Pd catalysts.
According to the literature, pure Ag does not chemisorb CO
[30–32]. Because in monometallic catalysts CO chemisorption
The primary objective of the current study is to synthesize
and evaluate zeolite-supported Pd–Ni and Pd–Ag bimetallic
catalysts for the selective hydrogenation of acetylene in the
presence of ethylene. We performed batch reactor studies to
confirm the novel and promising catalytic properties of zeolite-
supported bimetallic catalysts compared with those supported
on γ -Al2O3 and to those of monometallic catalysts. Further-
more, we performed kinetic modeling to provide a more quan-
titative comparison of the hydrogenation activity and selectivity
of the various monometallic and bimetallic catalysts.
Table 1
+
Catalysts compositions in wt% on γ -Al O and Na -β-zeolite
2
3
Support
γ -Al O
Pd
Pd–Ni
Pd–Ag
1.36
1.36, 0.75
Atomic ratio: 1:1
1.36, 0.75
1.36, 1.38
Atomic ratio: 1:1
1.36, 1.38
2
3
β-Zeolite
1.36
Atomic ratio: 1:1
Atomic ratio: 1:1

42
W. Huang et al. / Journal of Catalysis 246 (2007) 40–51
Table 2
couple to monitor the temperature. To remove water and other
impurities the cell was evacuated to a pressure <10−6 Torr at
room temperature for 30 min and then reduced at 723 K in
30 Torr hydrogen for 30 min, which was followed by evacua-
tion and a high-temperature flash (723 K) to remove any surface
species generated during the reduction. The reduction cycle was
repeated three times before performing IR experiments.
The gas-phase reaction products were monitored by record-
ing gas-phase spectra every 60 s during the reaction. A mixture
of 10% acetylene in 90% helium mixture (Matheson) was puri-
fied by passing it through a dry ice–acetone trap. The effects of
purification were tested by comparing 60-Torr purified mixture
and 60-Torr unpurified mixture; the latter showed trace amounts
of acetone. Research-grade ethylene (Matheson) and hydrogen
(Keen) were used without further purification. The IR cell was
filled with 10-Torr 10% acetylene and 1-Torr ethylene. After
equilibration and the formation of a stable baseline, 6-Torr H2
was introduced to initiate reaction at room temperature, and the
IR spectra were recorded. The concentrations of the three main
species—acetylene, ethylene, and ethane—were determined by
the classical least squares method using the C–H stretching vi-
brational features from 2600 to 3500 cm−1 [41].
Catalyst dispersion from CO or H chemisorption measurements
2
Sample
CO chemisorption
H chemisorption
2
Dispersion
(%)
Uptake
per gram
catalyst
Dispersion
(%)
Uptake
per gram
catalyst
Pd/γ -Al O
Pd/β-zeolite
PdAg/γ -Al O
PdAg/β-zeolite
PdNi/γ -Al O
PdNi/β-zeolite
Ni/γ -Al O
Ni/β-zeolite
Ag/γ -Al O
Ag/β-zeolite
16.4
23.1
14.2
20.9
13.0
20.3
13.6
20.2
2
3
2
3
11.3
12.4
10.6
11.6
–
14.4
15.9
6.8
7.4
–
2
3
2
3
–
–
–
–
2
3
–
–
occurs on Pd but not on Ag, it was assumed that on the surface
of alloy particles in Pd–Ag bimetallic catalysts, CO chemisorp-
tion occurred only on Pd atoms [33,34], which was supported
by Somanoto et. al. [31]. Results from our current study also
confirmed no strong interaction between CO and Ag/γ -Al2O3
and Ag/β-zeolite. Therefore, the dispersion of Pd–Ag bimetal-
lic catalysts was determined based on the assumption that CO
adsorbed only on surface Pd atoms. The dispersions of the Pd–
Ag catalysts were also determined using CO chemisorption.
Balerna et al. found that Pd–Ag bulk alloys (0.5 wt% Pd and
0.6 wt% Ag) consist of a phase of almost pure Ag and another
phase with similar Pd and Ag composition [35]. Because the
stoichiometry of CO–XPd is 0.5 [29] for low-dispersed Pd and
<1 for a Ag-rich Pd–Ag alloy [31], a mean stoichiometry of
CO/Pd of 0.75 was assumed based on the average of 0.5–1 for
the Pd–Ag catalysts in the current study.
3. Results
3.1. Calibration of gas-phase IR intensity
The concentrations of gas-phase C2H2, C2H4, and C2H6
in the IR cell were calibrated using the classical least squares
method [41] with a 16-member training set. The Beer–Lambert
law states that the absorbance is proportional to the concentra-
tion,
The mean stoichiometry of H2 chemisorption was assumed
to be 0.5 on supported Ni catalysts [36,37] due to the dissociate
adsorption of H2 on transition metals [38]. In H2 chemisorp-
tion analysis done to determine the dispersion of Pd catalysts at
28 ◦C, a stoichiometric number of 0.5 was also assumed [39].
Therefore, in the current study the dispersion of Ni and Pd–Ni
catalysts was determined by H2 chemisorption analysis assum-
ing a stoichiometry of 0.5. Table 2 shows the metal dispersions
on the corresponding catalysts.
A = KC,
(1)
where A is the training set of absorbance matrix, K is the pro-
portionality constant, and C is the matrix of component concen-
trations. The subscript “unknown” refers to the reaction data.
The coefficient K can be calculated as
ꢀ
ꢁ−1
K = ACT CCT
.
(2)
Therefore, an unknown mixture concentration can be predicted
by
ꢀ
ꢁ−1
2.2. Reaction rate measurements
Cunknown = KTK KTAunknown
.
(3)
Fourier transform infrared (FTIR) spectroscopy was used to
examine the kinetics of the hydrogenation reactions. The IR ex-
periments were performed using a stainless steel IR cell with of
BaF2 windows, which allowed in situ reduction of samples and
spectroscopic measurements of surface species and gas-phase
products in a batch reactor based on the design reported pre-
viously [40]. Spectra were recorded at room temperature with
4 cm−1 spectral resolution by collecting 128 scans for gas-
phase and 512 scans for surface species using a Nicolet-510
FTIR spectrometer equipped with a MCT-A (mercury cadmium
telluride) detector.
In our calibration, a set of 16 gas mixtures of varying but known
ratios of acetylene, ethylene, and ethane were each recorded at
room temperature using 128 scans and a 4-cm−1 spectral reso-
lution. These 16 spectra formed the absorbance matrix A. The
concentration values were included in the matrix of component
concentrations C in the corresponding order as the spectra were
stored in matrix A. The coefficient K was then calculated using
Eq. (2).
The spectrum of an unknown sample was truncated from
3500 to 2600 cm−1, and an automatic baseline correction was
applied. The resulting spectrum was labeled Aunknown, and the
concentration of the unknown spectrum was then calculated us-
ing Eq. (3).
Powder catalysts samples of ∼22 mg were pressed onto
a square tungsten mesh with a spot-welded K-type thermo-

W. Huang et al. / Journal of Catalysis 246 (2007) 40–51
43
3.2. Parameter fitting
and
dCH
2
rH = −
The equilibrium constants for the adsorption of acetylene
and ethylene and the dissociative adsorption of hydrogen are
defined as K1, K2, and K3, respectively; k1 and k2 are the rate
constants for the hydrogenation of acetylene and ethylene, re-
spectively. The rate-determining step is the addition of the first
hydrogen atom to an adsorbed acetylene or ethylene molecule
[9–14]. The equations are as follows:
2
dt
ꢂ
√
k1K1 K3P
P
C2H2
H2
= −
−
ꢂ
2
(1 + K1PC H + K2PC H
+
K3PH
)
2
2
2
4
2
ꢂ
√
k2K2 K3P
P
C2H4
H2
ꢂ
.
(16)
2
(1 + K1PC H + K2PC H
+
K3PH
)
2
2
2
4
2
For convenience, the rate constants for acetylene hydrogenation
and ethylene hydrogenation can be lumped as
C2H2 + ∗ ꢀ C2H∗2 K1,
C2H4 + ∗ ꢀ C2H∗4 K2,
H2 + 2∗ ꢀ 2H∗ K3,
(4)
(5)
(6)
(7)
ꢂ
k1∗ = k1K1 K3
(17)
and
ꢂ
k2∗ = k2K2 K3.
(18)
C2H2∗ + H∗ → C2H∗3 + ∗ k1,
The equations used by the fitting program are (11), (12), (15),
and
and (16).
Because the FTIR spectrometer cannot detect gas-phase H2,
the initial concentration of hydrogen was estimated as six times
the initial acetylene concentration based on the experimental
parameters. The concentration of hydrogen during the reaction
was then calculated based on the reaction stoichiometry. The
hydrogen equilibrium constant was set at 0.56 cm3/mol based
on [42]. The kinetic parameters were then calculated using
Eqs. (11), (12), (15), and (16). The fitting algorithm also de-
pends on the initial guess values. However, despite the changes
C2H4∗ + H∗ → C2H∗5 + ∗ k2.
The rates of acetylene hydrogenation and ethylene hydrogena-
tion can be written as
(8)
rC H = −k1θC H θH
(9)
2
2
2 2
and
rC H = −k2θC H θH + k1θC H θH,
(10)
2
4
2
4
2 2
k K
1
of those fitted parameters, the ratio of
1 , defined as the rel-
where θC H and θC H represent the surface coverage of acety-
k K
2
2
2 4
2
2
lene and ethylene, respectively. Using Langmuir–Hinshelwood
classical approximations, the rate can be defined as
ative selectivity of acetylene hydrogenation (S), remains nearly
constant with respect to the initial guess values.
dC
C2H2
rC H = −
3.3. Hydrogenation over γ -Al2O3-supported PdNi and PdAg
catalyst
2
2
dt
ꢂ
√
k1K1 K3P
P
C2H2
H2
= −
ꢂ
(11)
2
(1 + K1PC H + K2PC H
+
K3PH
)
2
2
2
4
2
Fig. 1 compares the concentrations of gas-phase C2H2,
C2H4, and C2H6 as a function of reaction time at room tem-
perature (298 K) over three γ -Al2O3-supported catalysts, Pd,
PdNi, and PdAg. The sharp decrease of the total mass bal-
ance at t = 20 min is due to the introduction of 6.0-Torr hy-
drogen and an unavoidable consequence of back-mixing into
the dosing lines. Due to the presence of excess H2, the fi-
nal product of all reactions was ethane. Based on the reaction
time, the PdAg/γ -Al2O3 exhibited the highest hydrogenation
activity among the three γ -Al2O3-supported catalysts, followed
by PdNi/γ -Al2O3 and then Pd/γ -Al2O3. The acetylene concen-
tration decreased during the reaction while ethane concentra-
tion increased until all acetylene and ethylene were consumed.
In contrast, the ethylene concentration remained nearly con-
stant for PdAg/γ -Al2O3 until the acetylene was almost com-
pletely consumed. The ethylene concentration on the Pd and
PdNi catalysts decreased from the beginning of the reaction, al-
though the rate of decrease was accelerated after the complete
consumption of acetylene. The fitted curves, using the equa-
tions described earlier, are compared with experimental data in
Fig. 2. The equilibrium and rate constants, derived from the fit-
ting and normalized per surface active site, are listed in Table 3.
and
dC
C2H4
rC H = −
2
4
dt
ꢂ
√
k1K1 K3P
P
C2H2
H2
=
ꢂ
2
(1 + K1PC H + K2PC H
+
K3PH )
2
2
2
2
4
ꢂ
√
k2K2 K3P
P
C2H4
H2
−
ꢂ
.
(12)
2
(1 + K1PC H + K2PC H
+
K3PH
)
2
2
2
4
2
In addition, the rates of ethane generation and hydrogen con-
sumption can be expressed as
rC H = k2θC H θH
(13)
2
6
2 4
and
rH = −k1θC H θH − k2θC H θH.
(14)
2
2
2
2 4
Therefore,
ꢂ
√
k2K2 K3P
P
dC
C2H4
H2
C2H6
rC H
=
=
ꢂ
2
6
2
dt
(1 + K1PC H + K2PC H
+
K3PH )
2
2
2
2
4
(15)

44
W. Huang et al. / Journal of Catalysis 246 (2007) 40–51
Fig. 1. Concentrations of gas-phase C H , C H , and C H as a function of reaction time at room temperature on (a) Pd/γ -Al O , (b) PdNi/γ -Al O , and
2
2
2
4
2
6
2
3
2 3
(c) PdAg/γ -Al O .
2
3
Fig. 2. Fitted concentrations of gas-phase C H , C H , and C H on (a) Pd/γ -Al O , (b) PdNi/γ -Al O , and (c) PdAg/γ -Al O .
2
2
2
4
2
6
2
3
2
3
2 3
Table 3
Values of reaction rate constants and adsorption equilibrium constants fitted from experimental data
−1
−1
3
3
∗
2
Catalysts
Pd/γ -Al O
k
(min
)
k
2
(min
)
K
(cm /mol)
K
(cm /mol)
S = (k /k ) (K /K )
1
1
2
1
1
2
43.4 1.1
33 0.5
35.2 1.0
21.4 7.1
20.5 5.5
5.1 0.6
5.5
17.3 1.8
6.8 0.1
16.4 2.0
36.5 0.9
28.6 1.3
26.5 5.9
0.8
1.9 0.5
0.6 0.2
0.8 0.2
2.7 0.2
3.3 0.3
5.2 0.3
0.4
3.3 0.3
2.2 0.5
0.8 0.1
0.6 0.3
0.6 0.2
0.15 0.05
6.1
1.4
1.3
2.1
2.6
3.9
6.8
2
3
PdNi/γ -Al O
2
3
PdAg/γ -Al O
Pd/Na-β-zeolite
PdNi/Na-β-zeolite
PdAg/Na-β-zeolite
2
3
a
Ag/γ -Al O
−
2
3
Ni/Na-β-zeolite
Ag/Na-β-zeolite
4.8
1.7
2.0
0.9
3.1
5.9
1.8
3.3
−
−
a
a
Is calculated per gram of catalysts.

W. Huang et al. / Journal of Catalysis 246 (2007) 40–51
45
3.4. Hydrogenation over β-zeolite-supported catalysts
acetylene and ethylene was consumed. However, on all of the
β-zeolite-supported catalysts, the ethylene concentration ini-
tially increased and then began to decrease as the acetylene was
consumed. The fitted curves are compared with experimental
data in Fig. 4, and the kinetic parameters based on surface ac-
tive sites are listed in Table 3.
The reactions on monometallic Ag/γ -Al2O3, Ni/γ -Al2O3,
Ag/β-zeolite, and Ni/β-zeolite were also tested; these materials
showed significantly lower hydrogenation activity (IR spectra
not shown). The rate and equilibrium constants are compared in
Table 3. Due to the low hydrogenation activity on Ni/γ -Al2O3,
we were unable to find an acceptable solution to Eqs. (9)–(18)
for this catalyst.
Fig. 3 compares the reaction kinetics of acetylene and eth-
ylene hydrogenation over β-zeolite-supported Pd, PdNi, and
PdAg catalysts at room temperature (298 K). The experiment
conditions were the same as the γ -Al2O3-supported catalysts.
Based on the reaction time, the β-zeolite-supported Pd and
PdNi catalysts showed higher hydrogenation activities than
those on the γ -Al2O3-supported Pd and PdNi catalysts, al-
though the hydrogenation activity of PdAg/β-zeolite is lower
than that of PdAg/γ -Al2O3. Similar to the γ -Al2O3-supported
catalysts, the acetylene concentration decreased during the re-
action while the ethane concentration increased until all of the
+
+
Fig. 3. Concentrations of gas-phase C H , C H , and C H as a function of reaction time at room temperature on (a) Pd/Na -β-zeolite, (b) PdNi/Na -β-zeolite,
and (c) PdAg/Na -β-zeolite.
2
2
2
4
2 6
+
+
+
+
Fig. 4. Fitted concentrations of gas-phase C H , C H , and C H on (a) Pd/Na -β-zeolite, (b) PdNi/Na -β-zeolite, and (c) PdAg/Na -β-zeolite.
2
2
2
4
2 6

46
W. Huang et al. / Journal of Catalysis 246 (2007) 40–51
3.5. Adsorbed species on different surfaces
features for the sp-, sp2-, and sp3-hybridized C2Hx species.
The expected frequencies for these species are also labeled in
Fig. 5 for the main vibrational features. In the current study
we are interested mainly in the relative concentrations of ad-
sorbed species on the different catalyst surfaces. As shown in
Fig. 5, the vibrational features for the CH, CH2, and CH3 func-
tional groups are detected on Pd/γ -Al2O3 and PdNi/γ -Al2O3,
indicating the formation of various surface C2Hx species on
the two catalysts. In contrast, the vibrational features on the
PdAg/γ -Al2O3 surface are relatively weak and can be as-
signed to weakly π-bonded species. As reported in previ-
ous studies, π-bonded species are more easily hydrogenated
[9,10], which is likely an important factor contributing to
Fig. 5 shows the IR spectra of the adsorbed mixture on
γ -Al2O3-supported Pd, PdNi, and PdAg, the intensities of
which are renormalized based on the number of active sites of
metal. The IR spectra were recorded after exposing the cata-
lysts to C2H2 and C2H4, followed by evacuation for ∼10 min.
As found in previous vibrational studies [51,52], the adsorption
and reaction of C2H2 and C2H4 can lead to a wide range of hy-
drocarbon fragments on surfaces, including π- and σ-bonded
C2H2 and C2H4, and C2 fragments, such as vinyl and ethyli-
dyne. As a result, the vibrational assignment is rather com-
plex. Table 4 provides a general assignment of vibrational
Fig. 5. IR spectra of mixture adsorption on (a) Pd/γ -Al O , (b) PdNi/γ -Al O , and (c) PdAg/γ -Al O .
2
3
2
3
2 3
Table 4
Assignment of IR spectra features of adsorbed species on catalysts supported on γ -Al O
2
3
Assignment
Pd/γ -Al O
PdAg/γ -Al O
PdNi/γ -Al O
Gas phase [64,65]
C H
2
3
2
3
2
3
C H
2 4
2
2
ν
(≡C–H)
3298
3295
3274
3293
3374
3287
C–H
ν
ν
(CH )
3192
2969
3266
2965
2930
2872
2002
1945
1689
1631
1554
1472
1411
as
2
(CH )
as
3
ν (CH )
ν (CH )
2960
1650
s
2
2863 [66]
2009
s
3
ν(C≡C)
1999
1945
1688
1649
1534
1974
ν(C≡C) [67]
ν(C=C) [67]
1688
1616
1534
1474
ν(C=C) + δ(CH ) [68]
2
δ
as
(CH )
3
δ(CH)
1355
1298
δ(CH ) (ethylidyne)
δ(CH)
1320
1245
1347
1248
3

W. Huang et al. / Journal of Catalysis 246 (2007) 40–51
47
Fig. 6. IR spectra of mixture (C H and C H ) adsorption on (a) Pd/β-zeolite, (b) PdNi/β-zeolite, and (c) PdAg/β-zeolite.
2
2
2 4
Table 5
only, and of C2H4 only. The comparison reveals that the adsorp-
tion of C2H4 is rather weak, as indicated by the relatively weak
vibrational features after exposure of the catalyst to C2H4. In
contrast, the adsorption of C2H2 is stronger, as demonstrated by
the more intense vibrational features after exposure of the cat-
alyst to C2H2. Although the IR spectrum of mixture adsorption
is generally similar to that of C2H2, several additional features,
particularly in the ν(CHx) region, differ from those of either
C2H2 or C2H4. More detailed IR studies are needed to clarify
the origin of these intermediates, which are beyond the scope
of the current paper.
Assignment of IR spectra features of adsorbed species on catalysts supported
+
on Na -β-zeolite
Assignment
Pd/β-zeolite
PdAg/β-zeolite
PdNi/β-zeolite
ν
(≡C–H)
3204
3258
3212
3183
2964
2933
2875
2034
1952
1712
1674
1608
1542
1519
1476
3200
C–H
ν
ν
(CH )
(CH )
3180
2964
2930
2879
1948
2999
2960
2926
2872
1954
1936
1712
1677
1619
1561
as
2
as
3
ν (CH )
s
2
ν (CH )
s
3
ν(C≡C)
ν(C=C)
1712
1677
1604
1573
1515
1476
4. Discussion
ν(C=C) + δ(CH )
2
4.1. Effect of bimetallic formation
δ
as
(CH )
1476
1418
1372
1332
3
As shown in Table 3, PdAg/γ -Al2O3 has a higher selectiv-
ity for acetylene hydrogenation than Pd/γ -Al2O3 and PdNi/-
γ -Al2O3. The equilibrium constants for ethylene adsorption
are larger on Pd/γ -Al2O3 and PdNi/γ -Al2O3 than on acety-
lene (Table 3). This indicates that the Pd/γ -Al2O3 and PdNi/-
γ -Al2O3 surfaces can adsorb ethylene more readily than acety-
lene. In comparison, on the PdAg/γ -Al2O3 catalyst, the equi-
librium constants of ethylene and acetylene are comparable.
Therefore, the selectivity to acetylene hydrogenation increases
on PdAg/γ -Al2O3 mainly because the relative ratio of K1/K2
increases, which means that alloying Pd with Ag can help facil-
itate acetylene adsorption.
δ(CH)
1376
1372
the higher hydrogenation activity on the PdAg/γ -Al2O3 cata-
lyst.
Fig. 6 shows the adsorbed mixture (C2H2 and C2H4) on
β-zeolite-supported Pd, PdNi, and PdAg, the intensities of
which are renormalized based on the active sites of surface. The
assignment of the frequencies is listed in Tables 5. Similar to
Fig. 5, π-bonded species are observed on the zeolite-supported
catalysts (Fig. 6). Based on the normalized IR intensities, the
concentration of adsorbed species is higher on PdAg/β-zeolite
than on Pd/β-zeolite and PdNi/β-zeolite. To further understand
the competitive adsorption of C2H2 and C2H4, Fig. 7 shows
the IR spectra after the adsorption of C2H2 and C2H4, of C2H2
The calculated rate constant for acetylene hydrogenation
is larger than that for ethylene hydrogenation on the three
γ -Al2O3-supported catalysts, consistent with the hydrogena-
tion barrier calculated in previous studies by Gislason et

48
W. Huang et al. / Journal of Catalysis 246 (2007) 40–51
Fig. 7. IR spectra of adsorption on PdAg/β-zeolite, (a) mixture of C H and C H , (b) C H only, and (c) C H only.
2
2
2
4
2
2
2 4
al. [43]. Our fitted data in Table 3 indicate that the rate con-
stant of acetylene hydrogenation is smaller on PdAg/γ -Al2O3
than on Pd/γ -Al2O3, which is also consistent with literature re-
ports indicating that alloying the Pd catalyst with Ag increases
selectivity but reduces catalyst activity [43,44]. However, the
higher ratio of equilibrium constants of acetylene over ethyl-
ene adsorption on PdAg/γ -Al2O3 results in an increase in the
overall reaction rate in the batch reactor. From Eq. (11) and the
constants given in Table 3, the rate of acetylene hydrogenation
on Pd/γ -Al2O3 and PdAg/γ -Al2O3 can be written as
supports this effect (Fig. 5). Our models show that alloying Ni
reduces the rate constant, k, for ethylene hydrogenation more
than for acetylene but has the opposite effect for the equilib-
rium constant, K (Table 3). These opposing changes might be
responsible for the observation that the selectivity to acetylene
hydrogenation is similar for PdNi/γ -Al2O3 and Pd/γ -Al2O3.
4.2. Effect of catalyst supports
The β-zeolite-supported Pd and PdNi catalysts show signif-
icantly higher activity than their γ -Al2O3-supported counter-
parts. We attribute this observation to the greater dispersion
on β-zeolite support than on γ -Al2O3 support (Table 2), as
well as to the reduced ethylene adsorption equilibrium constant
(Table 3). It is interesting to point out that on the β-zeolite-
supported catalysts, the equilibrium constant of ethylene is
smaller than that of acetylene, opposite to what is observed
on the γ -Al2O3-supported catalysts. We attribute this to the
higher cation–π interactions between alkali metal Na+ and the
π-bond of acetylene than with ethylene. The strong interaction
between the acetylene quadruple moment and Na+ leads to an
increase in the local concentration of acetylene and a decrease
in the local ethylene concentration. The preferential adsorption
of acetylene over ethylene leads to a higher selectivity for acety-
lene hydrogenation.
The rate constant for acetylene hydrogenation is lower on
Pd/β-zeolite than on Pd/γ -Al2O3, whereas the rate constant
for ethylene is higher on Pd/β-zeolite than on Pd/γ -Al2O3
(Table 3). A previous study [45] also found that the activity
of acetylene hydrogenation decreases while the dispersion in-
creases on the Pd/α-Al2O3. The β-zeolite-supported Pd shows
higher dispersion than the γ -Al2O3-supported Pd (Table 2),
ꢂ
√
53.5 K3PC H PH
2
2
2
rC H = −
ꢂ
2
2
2
(1 + 1.26PC H + 2.06PC H + 0.79 K3PH
)
2
2
2
2
4
for Pd/γ -Al2O3
and
ꢂ
√
45.0 K3PC H PH
2
2
2
rC H = −
ꢂ
2
2
2
(1 + 0.84PC H + 0.96PC H + 1.2 K3PH
)
2
2
2
2
4
for PdAg/γ -Al2O3.
Although the numerator is larger for the rate on Pd/γ -Al2O3
than on PdAg/γ -Al2O3, the denominator can also become
larger on Pd/γ -Al2O3 than on PdAg/γ -Al2O3, depending on
the values of the partial pressures of reactants (e.g., the ini-
tial partial pressure of 1:1:6 for C2H2:C2H4:H2). This in turn
causes the overall rate on Pd/γ -Al2O3 to be smaller than that
on PdAg/γ -Al2O3, as observed experimentally and shown in
Fig. 1.
Alloying Pd with Ni weakens the acetylene and ethylene
binding energies, as indicated by the decreased K1 and K2 val-
ues shown in Table 3. The reduced IR intensities for adsorbed
species on PdNi/γ -Al2O3 compared with those on Pd/γ -Al2O3

W. Huang et al. / Journal of Catalysis 246 (2007) 40–51
49
suggesting that the particle size of Pd clusters on β-zeolite is
smaller than that on γ -Al2O3. Larger particle size favors the
formation of palladium β-hydride [45,46], which plays an im-
portant role in the hydrogenation activity of acetylene.
Previous studies [47] have found that ethylene molecules
preferentially adsorb in the di-σ configuration on the larger par-
ticles, whereas small particles favor the weakly bounded π eth-
ylene molecules. Because the π-bonded ethylene is generally
considered the intermediate hydrogenation species [48–50],
smaller particles should promote the hydrogenation of ethylene
over larger particles. These observations are a possible expla-
nation of why the β-zeolite-supported Pd exhibits a higher rate
constant for ethylene hydrogenation.
PdAg/β-zeolite exhibits the highest selectivity of the three
β-zeolite-supported catalysts, followed by PdNi/β-zeolite and
then Pd/β-zeolite. The higher selectivity of PdAg/β-zeolite is
achieved by the selective adsorption of acetylene resulting from
alloying with Ag on zeolite support. As shown in Table 3,
the equilibrium constant of acetylene is much larger than the
equilibrium constant of ethylene. This is consistent with the
IR results in Fig. 7, where the IR spectrum following the ad-
sorption of C2H2 + C2H4 mixture on PdAg/β-zeolite is very
similar to the adsorption spectrum of pure acetylene. In con-
trast to the γ -Al2O3-supported catalysts, the reaction rate on the
PdAg/β-zeolite is lower than Pd/β-zeolite and PdNi/β-zeolite.
This may be due to the strong chemisorption of acetylene on
the PdAg/β-zeolite. This explanation is also supported by the
IR spectra of adsorption on PdAg/γ -Al2O3 and PdAg/β-zeolite
(Figs. 5 and 6), which show that the IR intensities of adsorbed
species are much greater on PdAg/β-zeolite than on PdAg/γ -
Al2O3.
It is interesting to note that the selectivity for acetylene
hydrogenation is about 50% higher on PdNi/β-zeolite than
on Pd/β-zeolite, whereas the selectivity for PdNi/γ -Al2O3
and Pd/γ -Al2O3 are similar. Because the ratio of equilibrium
constants, which is dominated by zeolite support, on PdNi/
β-zeolite is similar to that on Pd/β-zeolite, the selectivity dif-
ference is most likely resulting from a difference in the rate
constants for acetylene and ethylene hydrogenation, which is
dominated by active metal particles. This supports our point of
view that the bimetallic effects resulting from alloying Pd with
Ni reduces the rate constant of ethylene more than that of acety-
lene, as is shown in Table 3. In this case, the bimetallic effects
and support effects work together to enhance the catalyst selec-
tivity toward acetylene hydrogenation without any significant
loss in activity.
nation of 1,2-dichloroethane into ethylene over Pd–Ag/SiO2
catalysts. Zhu et. al. [44] found that on 3 wt% Pd/Al2O3, the
conversion of acetylene reaches 100% at 325.5 K with a nega-
tive selectivity of −4%, with the definition of
increase of ethylene
S =
× 100%,
decrease of acetylene
in a flow reactor system [44]. The conversion of acetylene over
3 wt% Pd–6 wt% Ag/Al2O3 reaches 100% at 378 K with a se-
lectivity of 38%, much higher than that over Pd/Al2O3 catalyst.
Another study found that the selectivity of acetylene hydro-
genation is 45% on a 0.45 wt% Pd/Stöber silica sphere at 333 K
and 80% on a 0.45 wt% Pd–0.25 wt% Ag/Stöber silica sphere
when both catalysts are reduced at 773 K [56]. To the best of
our knowledge, the effect of adding Ni as a promoter has not
yet been reported.
In this paper, we studied the bimetallic effects of alloying
Pd with Ag or Ni. Consistent with reports in the literature, we
found that alloying Pd with Ag enhances the hydrogenation
selectivity in a batch reactor. Our study also shows that alloy-
ing Ni reduces the rate constant (k) for ethylene hydrogenation
more than acetylene, whereas the equilibrium constants (K)
exhibit the opposite trend (Table 3), which leaves room for
using support effects to control the relative concentrations of
adsorbed acetylene and ethylene.
Most previous studies have been performed on Al2O3 or
SiO2 supports [5,6,44,46,52]; there are few studies on the ef-
fects of supports, such as pumice [57,58] and zeolite [15].
Duca et al. [57] found that pumice-supported Pd catalysts ex-
hibited good activity and excellent selectivity and stability in
the selective hydrogenation of acetylene. Weiss et al. [15] con-
cluded that the reaction rates and selectivity for Pd on silicalite
and Pd incorporated into ZSM-5 during synthesis compared
favorably with those of a commercial reference catalyst. In
the present study, we investigated the support effects of Na+-
exchanged β-zeolite compared with traditional Al2O3 supports.
We found that the cation–π interaction between Na+ and acety-
lene or ethylene helps the catalysts selectively adsorb acetylene
over ethylene, so that the selectivity on the Na+-exchanged
β-zeolite-supported catalysts is enhanced compared with the
γ -Al2O3 counterparts. In particular, PdAg/Na+-β-zeolite has
the highest selectivity of the catalysts studied in the present
work. It has also been found that aided by competitive adsorp-
tion of the Na+-exchanged β-zeolite supports, PdNi/Na+-β-
zeolite shows 30% higher selectivity than Pd/Na+-β-zeolite but
maintains similar activity. Moreover, the β-zeolite-supported
catalysts have higher dispersions than the γ -Al2O3-supported
catalysts, leading to an increase in the overall reaction rate in
the batch reactor system.
4.3. Comparison with previous results
Previous studies of the selective hydrogenation of acetylene
have focused on low loading (0.038–3 wt%) Pd-based catalysts
[2,6,8,43,45]. Recently, emphasis has been on tuning catalyst
selectivity by the formation of alloys, especially alloying with
Ag [44,51–53]. It has been found that alloying with Ag en-
hances the selectivity of acetylene hydrogenation reaction but
decreases the activity. Heinrichs et al. [54] and Lambert et al.
[55] observed the same trend for the selective hydrodechlori-
It is widely reported in the literature that the adsorption of
acetylene should be stronger than that of ethylene on Pd sur-
faces. This is consistent with the equilibrium constant (K1 and
K2) values for the adsorption of acetylene and ethylene on
zeolite-supported Pd (Table 3). However, such a trend is not ob-
served in the K values on alumina-supported Pd. This is most
likely due to the involvement of the alumina substrate in the
competitive adsorption of acetylene and ethylene. As reported

50
W. Huang et al. / Journal of Catalysis 246 (2007) 40–51
in the literature, there are three types of active sites for the
hydrogenation of the acetylene–ethylene mixture on alumina-
supported Pd [59,60]. A1 and A2 sites are highly selective in
the hydrogenation of acetylene to ethylene via a mechanism
involving the competitive (A1) and noncompetitive (A2) ad-
sorption of acetylene and hydrogen. The E sites are active only
for the hydrogenation of ethylene, with both acetylene and eth-
ylene competitively adsorbing on the E sites. Furthermore, on
γ -Al2O3-supported Pd, ethylene can be hydrogenated on the
γ -Al2O3 support by means of hydrogen spilled over from Pd
[59,61–63]. In this case, the E sites are as important as A sites
and play an important role in the selective hydrogenation of
acetylene in ethylene. Compared with the active Pd metal sites,
the support sites are more abundant with the catalyst loadings
used in the present study. As a result, the adsorption of ethyl-
ene could be stronger than that of acetylene by involving the
support, because the equilibrium constants in Table 3 include
contributions from both Pd metals and the γ -Al2O3 support.
[12] H. Nakatsuji, M. Hada, T. Yonezawa, Surf. Sci. 185 (1987) 319–342.
[13] J. Horiuchi, M. Polanyi, Trans. Faraday Soc. 30 (1934) 1164–1172.
[14] R.D. Cortright, S.A. Goddard, J.E. Rekoske, J.A. Dumesic, J. Catal. 127
(1991) 342–353.
[15] W.L. Kranich, A.H. Weiss, Z. Schay, L. Guczi, Appl. Catal. 13 (1985)
257–267.
[16] D.R. Corbin, L. Abrams, C. Bonifaz, J. Catal. 115 (1989) 420–429.
[17] D.A. Dougherty, Science 271 (1996) 163–168.
[18] R.I. Keir, D.W. Lamb, G.L.D. Ritchie, J.N. Watson, Chem. Phys. Lett. 279
(1997) 22–28.
[19] T. Kar, R. Ponec, A.B. Sannigrahi, J. Phys. Chem. A 105 (2001) 7737–
7744.
[20] N.A. Khan, H.H. Hwu, J.G. Chen, J. Catal. 205 (2002) 259–265.
[21] H.H. Hwu, J. Eng, J.G. Chen, J. Am. Chem. Soc. 124 (2002) 702–709.
[22] N.A. Khan, L.E. Murillo, Y.Y. Shu, J.G. Chen, Catal. Lett. 105 (2005)
233–238.
[23] J.R. Kitchin, J.K. Norskov, M.A. Barteau, J.G. Chen, Phys. Rev. Lett. 93
(2004) 156801.
[24] N.A. Khan, J.G. Chen, in: Nanotechnology in Catalysis, vol. 1, 2004,
pp. 17–32.
[25] M.B. Zellner, A.M. Goda, O. Skoplyak, M.A. Barteau, J.G. Chen, Surf.
Sci. 583 (2005) 281–296.
[26] B. Hammer, J.K. Norskov, in: Advances in Catalysis, vol. 45, 2000,
pp. 71–129.
5. Conclusion
[27] L.E. Murillo, N.A. Khan, J.G. Chen, Surf. Sci. 594 (2005) 27–42.
[28] S. Lambert, C. Cellier, P. Grange, J.P. Pirard, B. Heinrichs, J. Catal. 221
(2004) 335–346.
[29] C.L.M. Joyal, J.B. Butt, J. Chem. Soc. Faraday Trans. I 83 (1987) 2757–
2764.
[30] V. Ponec, G.C. Bond, Catalysis by Metals and Alloy, Elsevier, Amsterdam,
1995.
[31] Y. Somanoto, W.M. Sachtler, J. Catal. 32 (1974) 315–324.
[32] D. Cormack, J. Pritchard, R.L. Moss, J. Catal. 37 (1975) 548–552.
[33] B. Heinrichs, F. Noville, J.P. Schoebrechts, J.P. Pirard, J. Catal. 192 (2000)
108–118.
[34] W.M.H. Sachtler, Catal. Rev. Sci. Eng. 14 (1976) 193–210.
[35] A. Balerna, G. Deganello, L. Liotta, A. Longo, A. Martorana, C. Menegh-
ini, S. Mobilio, A.M. Venezia, J. Non-Cryst. Solids 293 (2001) 682–687.
[36] J.M. Wei, E. Iglesia, J. Catal. 224 (2004) 370–383.
[37] M.C.J. Bradford, M.A. Vannice, Appl. Catal. A 142 (1996) 97–122.
[38] H. Yang, J.L. Whitten, J. Chem. Phys. 98 (1993) 5039–5049.
[39] D.J. Moon, M.J. Chung, K.Y. Park, S.I. Hong, Appl. Catal. A 168 (1998)
159–170.
Novel selective hydrogenation catalysts have been synthe-
sized and evaluated using an FTIR batch reactor. The results
show that among the γ -Al2O3-supported catalysts, the PdAg
catalyst has the highest selectivity, whereas PdNi behaves sim-
ilar to the Pd monometallic catalyst. Preferential adsorption of
acetylene over ethylene on Na+-β-zeolites increases the sur-
face concentration of acetylene, resulting in a further increase
in the hydrogenation selectivity of acetylene. Finally, the PdAg/
Na+-β-zeolite catalyst shows significantly higher selectivity
than the other catalysts.
Acknowledgments
Financial support was provided by from the US Department
of Energy, Office of Science, Division of Chemical Sciences
(Grant FG02-03ER15468).
[40] P. Basu, T.H. Ballinger, J.T. Yates, Rev. Sci. Instrum. 59 (1988) 1321–
1327.
References
[41] R. Kramer, Chemometric Techniques for Quantitative Analysis, Dekker,
New York, 1998.
[42] A.B. Mhadeshwar, J.R. Kitchin, M.A. Barteau, D.G. Vlachos, Catal.
Lett. 96 (2004) 13–22.
[1] J.H. Kang, E.W. Shin, W.J. Kim, J.D. Park, S.H. Moon, Catal. Today 63
(2000) 183–188.
[43] J. Gislason, W.S. Xia, H. Sellers, J. Phys. Chem. A 106 (2002) 767–774.
[44] Q.W. Zhang, J. Li, X.X. Liu, Q.M. Zhu, Appl. Catal. A 197 (2000) 221–
228.
[45] C.E. Gigola, H.R. Aduriz, P. Bodnariuk, Appl. Catal. 27 (1986) 133–144.
[46] R.K. Nandi, P. Georgopoulos, J.B. Cohen, J.B. Butt, R.L. Burwell, D.H.
Bilderback, J. Catal. 77 (1982) 421–431.
[2] N.S. Schbib, M.A. Garcia, C.E. Gigola, A.F. Errazu, Ind. Eng. Chem.
Res. 35 (1996) 1496–1505.
[3] P. Praserthdam, S. Phatanasri, J. Meksikarin, Catal. Today 63 (2000) 209–
213.
[4] W.J. Kim, J.H. Kang, I.Y. Ahn, S.H. Moon, Appl. Catal. A 268 (2004)
77–82.
[47] S. Shaikhutdinov, M. Heemeier, M. Baumer, T. Lear, D. Lennon, R.J. Old-
man, S.D. Jackson, H.J. Freund, J. Catal. 200 (2001) 330–339.
[48] P.S. Cremer, G.A. Somorjai, J. Chem. Soc. Faraday Trans. 91 (1995)
3671–3677.
[49] P.S. Cremer, X.C. Su, Y.R. Shen, G.A. Somorjai, J. Am. Chem. Soc. 118
(1996) 2942–2949.
[50] T. Miura, H. Kobayashi, K. Domen, J. Phys. Chem. B 104 (2000) 6809–
6814.
[51] D.C. Huang, K.H. Chang, W.F. Pong, P.K. Tseng, K.J. Hung, W.F. Huang,
Catal. Lett. 53 (1998) 155–159.
[5] H. Molero, B.F. Bartlett, W.T. Tysoe, J. Catal. 181 (1999) 49–56.
[6] A. Sarkany, A. Beck, A. Horvath, Z. Revay, L. Guczi, Appl. Catal. A 253
(2003) 283–292.
[7] B.M. Choudary, M.L. Kantam, N.M. Reddy, K.K. Rao, Y. Haritha,
V. Bhaskar, F. Figueras, A. Tuel, Appl. Catal. A 181 (1999) 139–144.
[8] H.R. Aduriz, P. Bodnariuk, M. Dennehy, C.E. Gigola, Appl. Catal. 58
(1990) 227–239.
[9] S.G. Podkolzin, R. Alcala, J.A. Dumesic, J. Mol. Catal. A Chem. 218
(2004) 217–227.
[10] P.A. Sheth, M. Neurock, C.M. Smith, J. Phys. Chem. B 107 (2003) 2009–
2017.
[52] R.N. Lamb, B. Ngamsom, D.L. Trimm, B. Gong, P.L. Silveston,
P. Praserthdam, Appl. Catal. A 268 (2004) 43–50.
[11] J.W. Medlin, M.D. Allendorf, J. Phys. Chem. B 107 (2003) 217–223.

W. Huang et al. / Journal of Catalysis 246 (2007) 40–51
51
[53] G. Meitzner, J.H. Sinfelt, Catal. Lett. 30 (1995) 1–10.
[60] A. Borodzinski, A. Cybulski, Appl. Catal. A 198 (2000) 51–66.
[61] A. Sarkany, L. Guczi, A.H. Weiss, Appl. Catal. 10 (1984) 369–388.
[62] S. Leviness, V. Nair, A.H. Weiss, Z. Schay, L. Guczi J. Mol. Catal. 25
(1984) 131–140.
[54] B. Heinrichs, J.P. Schoebrechts, J.P. Pirard, J. Catal. 200 (2001) 309–320.
[55] S. Lambert, F. Ferauche, A. Brasseur, J.P. Pirard, B. Heinrichs, Catal. To-
day 100 (2005) 283–289.
[56] Y.M. Jin, A.K. Datye, E. Rightor, R. Gulotty, W. Waterman, M. Smith,
M. Holbrook, J. Maj, J. Blackson, J. Catal. 203 (2001) 292–306.
[57] D. Duca, F. Frusteri, A. Parmaliana, G. Deganello, Appl. Catal. A 146
(1996) 269–284.
[63] A. Borodzinski, Pol. J. Chem. 69 (1995) 111–117.
[64] A.V. Kiselev, V.I. Lygin, Infrared Spectra of Surface Compounds, Wiley,
New York/Toronto, 1975.
[65] B.E. Bent, Chem. Rev. 96 (1996) 1361–1390.
[66] T.P. Beebe, J.T. Yates, J. Phys. Chem. 91 (1987) 254–257.
[67] S.S. Randhava, A. Rehmat, Trans. Faraday Soc. 66 (1) (1970) 235–241.
[68] M.P. Lapinski, J.G. Ekerdt, J. Phys. Chem. 94 (1990) 4599–4610.
[58] D. Duca, F. Arena, A. Parmaliana, G. Deganello, Appl. Catal. A 172
(1998) 207–216.
[59] A. Borodzinski, A. Golebiowski, Langmuir 13 (1997) 883–887.