Journal of Colloid and Interface Science 354 (2011) 100–108

Contents lists available at ScienceDirect

Journal of Colloid and Interface Science

www.elsevier.com/locate/jcis

Fabrication and characterization of bimetallic Pt–Au nanowires supported

on FSM-16 and their catalytic activities toward water–gas shift reaction
Mohamed Mokhtar Mohamed ⇑, K.S. Khairou
Umm Al-Qura University, Faculty of Applied Science, Chemistry Department, Makkah, Saudi Arabia

a r t i c l e i n f o a b s t r a c t

Article history: A facile, previously unexplored, method to synthesize bimetallic Pt–Au nanowires (20 nm diame-
Received 30 July 2010 ter  120–170 nm long) on mesoporous FSM-16 (2.7 nm) was fabricated by co-impregnation of H2PtCl6
Accepted 7 October 2010 with HAuCl4 followed by evacuation at 300 K and finally exposure to the CO/H2O gas mixture (60:5 Torr)
Available online 13 October 2010
at 323 K for 1.0 h. On the other hand, spherical monometallic nanoparticles of pure Pt (7.0 nm diameter)
and Au (7–26 nm diameter) were synthesized as well, by impregnation, at the same reaction conditions.
Keywords: The catalysts were characterized by in situ FTIR spectroscopy, UV–vis absorption spectroscopy, TEM, TPR
Bimetallic Pt–Au clusters
and TPCOR. The catalytic activities toward the water–gas shift reaction (WGSR) were also examined
Nanowires
FSM-16
under atmospheric pressure and at the margin of 323–373 K. The optical absorption spectra showed a
FTIR remarkable shift and broadening of Pt–Au surface Plasmon resonance band at 515 nm apart from those
UV–vis of individual analogue emphasizing bimetallic formation. Results from in situ FTIR spectroscopy indi-
TPR cated that incorporation of Au assisted and stabilized the formation of carbonyl clusters of Pt–Au–CO
TPRSR (2084 cm 1) and Pt–CO (1888 cm 1) inside the host FSM-16. The Pt–Au carbonyl clusters built up at
TEM the moment of vanishing the linear carbonyl band of the charged Au (Au+–CO, 2186 cm 1) along with
WGSR a concomitant increase in the reduced gold (Au0–CO, 2124 cm 1) species. TPR profiles showed that the
H2 consumed was higher for Pt/FSM-16 than for Pt–Au/FSM-16 verifying the facile reduction of Pt moi-
eties after addition of Au. The CO adsorption peak maximum, in TPCOR, for Pt/FSM-16 occurred at higher
temperature than that of Pt–Au/FSM-16, which exhibited higher amounts of CO2 produced. The relative
decrease in CO bindings on bimetallic surface was responsible for increasing the CO oxidation activity
mainly through an association mechanism. Accordingly, the activity of Pt–Au/FSM-16 towards WGS
showed a marked increase (8–23 times) compared with those of monometallics emphasizing the depen-
dence of this reaction on the electronic defects of the nanowires. A straightforward reduction mechanism
was deduced for Pt–Au alloy formation in view of the results obtained.
Ó 2010 Elsevier Inc. All rights reserved.

1. Introduction itself is not a catalytically active metal towards some reactions,

Bimetallic clusters and alloys have drawn considerable atten-
tion lately due to their optical, magnetic and catalytic properties
[1,2]. The interactions between the two components in bimetallic
clusters introduce a mutual influence on the neighboring atoms
and lead to unique properties for these clusters, which show differ-
ent catalytic behaviors from those of the monometallic clusters.
Due to the multivalent bonding character and low-lying excited
state, the geometric and electronic structures of most bimetallic
clusters are complicated and remain to be elucidated [3]. Among
the transition bimetallic clusters, platinum–gold binary ones at-
tract strong attention for their potential applications in catalysis
[4,5]. It is well known that the addition of gold can have important
effects on the catalytic properties of platinum [6,7]. Since gold by
⇑ Corresponding author. Address: Benha University, Fac. Sci., Benha, Egypt.
E-mail address: mohmok2000@yahoo.com (M.M. Mohamed).
the study of Pt–Au alloys offers a good opportunity to test the
importance of the geometric effects as well as electronic ones on
their catalytic activities.
Much of the previous effort in preparing supported mixed-me-
tal Pt–Au catalysts was mainly directed to CVD, deposition–precip-
itation, sol–gel and co-impregnation of Pt and Au salts [7–10].
Phase separation is a common problem with these methods since
the Pt–Au alloys that contain between 18% and 97% Pt are thermo-
dynamically unstable. In addition, the presence of Au is known to
affect Pt particle size. Furthermore, the low melting point of Au
and therefore the increased mobility of the Au atoms at high tem-
peratures result in broad Au particle size distributions and difficul-
ties in stabilizing Au in a small cluster form [7]. Such segregation
problems can be resolved – at least in principle – by the use of pre-
formed bimetallic cluster precursors, frequently stabilized by or-
ganic ligands [11–19]. Past studies in this field involved the use
of carbonyl clusters as precursors [14,17–21]. In these studies,

0021-9797/$ - see front matter Ó 2010 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcis.2010.10.008
 M.M. Mohamed, K.S. Khairou / Journal of Colloid and Interface Science 354 (2011) 100–108
metal cluster compounds were supported on various oxide sup-
ports (silica, alumina, or zeolites) and their carbonyl ligands were
removed by heating under vacuum or inert atmosphere. Further
examinations showed that in some cases the structure of the clus-
ter’s metal core remained substantially intact [22]. On the other
hand, there are a few examples of phosphine stabilized bimetallic
Au-containing clusters (i.e., [Pt(Au(PPh3)8](NO3)2, [Pt(PPh3)(AuPPh
3)6](NO3)2, [Pt(H)(PPh3)(AuPPh3)7](NO3)2) that have been immobi-
lized and thermally activated on solid supports [23,24]. The pres-
ence of phosphorus in these potential catalyst precursors may
have substantial effects on the resulting catalysts. These clusters
have high Au content and the presence of phosphorous in the li-
gands represents a potential poison for several reactions. Recently,
Chandler et al. [25,26] reported the preparation of supported Pt–Au
catalysts from a cluster precursor that entirely contains acetylide
ligands and they concluded that true bimetallic Pt–Au particles
can be formed via this synthetic route.
The synthesis of mesoporous solids such as MCM-41 and FSM
(folded-sheet mesporous material)-16 [27] attract a great deal of
attention because of the large pore diameter of these materials
(20–100 Å). Using FSM-16, which has appreciable textural meso-
porosity as a support, the robust carbonyl cluster [Pt15(CO)30]2
has been synthesized [27,28] that exhibited high catalytic activities
for the water–gas shift (WGS) reactions compared with NaY en-
trapped Pt9–Pt12 carbonyl clusters [29]. Nowadays, it is known that
catalysts containing gold with a partially reducible oxide (e.g.,
ceria, titania) are active for water–gas shift reaction [30]. The
water–gas shift reaction is an industrially important reaction for
H2 production as well as CO; one of the major air pollutants, deple-
tion. A relatively unexplored method to fabricate Pt–Au nanowires
encapsulated inside FSM-16, throughout the reaction of platinum
carbonyl clusters with gold(1) under the influence of interacting
CO and water mixture, was demonstrated. Temperature-pro-
gramed reduction (TPR), temperature-programed carbonylation
reaction (TPCOR), transmission electron microscopy (TEM) and
in situ FTIR and UV–vis spectroscopy techniques were used to
characterize Pt–Au/FSM-16 catalysts in an attempt to elucidate
the bimetallic alloy formation and the reasons for their over
bifunctional activities toward the WGS reaction compared with
Au/FSM-16 and Pt/FSM-16 catalysts.
2. Experimental
2.1. Sample preparation
The host FSM-16 of 2.7 nm size was synthesized using a polysi-
licate (Kanemite; NaHSi2O5
3H2O) with C16H33NMe3Cl as a surfac-
tant template according to published procedures [31]. After
calcination in air at 823 K, the resulting material (SiO2/
Al2O3 = 320) presented well-defined hexagonal walls with silanol
groups (3745 cm 1) and a characteristic X-ray powder pattern in
the low angle region (Cu Ka, 2h = 2.30°, 4.00°, 4.60°, 6.10°). The
BET surface area of the catalysts as determined by N2 adsorption
measurements was 1015 m2/g for FSM-16, 939 m2/g for Pt/FSM-
16, 870 m2/g for Au/FSM-16 and 820 m2/g for Pt–Au/FSM-16.
Au/FSM-16, Pt/FSM-16 and Pt–Au/FSM-16 catalysts (2.5 wt.%
loading for each) were prepared at room temperature by either
impregnation (for the monometallic catalysts) or co-impregnation
(for the bimetallic ones) of HAuCl4
4H2O and H2PtCl6
6H2O, by dis-
solving the metal salt in distilled water. During stirring of the sub-
strate slurry, the aqueous solution(s) of the salt was added in drops
followed with stirring at 300 K for 24 h. The samples were slowly
evacuated using a rotary evaporator at 300 K to remove water for
almost 12 h, and then vacuum dried (10 4 Torr) at the same tem-
perature. For comparison purposes the 2.5 pint-2.5 Au/NaY catalyst
101
(Si/Al = 5.6) was prepared by the co-impregnation method as pre-
viously described.
2.2. Catalyst characterization
2.2.1. Temperature-programed carbonylation reaction (TPCOR)
Temperature-programed carbonylation reaction (TPCOR) exper-
iments were performed using an automated gas desorption ana-
lyzer (Quadrasorb SI-MP). The sample (50 mg) was loaded in a u-
shaped quartz microreactor and heated from 298 to 323 K (heating
rate 2 K/min) in flowing helium (20 ml/min). The sample was then
subjected to the flowing CO/He mixture (15 ml/min, 2.48% CO in
He); CO consumption and CO2 evolution were monitored using a
thermal conductivity detector. The TPCOR experiments were
accomplished following CO adsorption at low temperatures (LTR)
as well as at high temperatures (HTR). The adsorption of CO was
performed at room temperature (using a liquid nitrogen trap) fol-
lowed by increasing the temperature linearly to 450 K at a ramping
rate of 2 K/min. To monitor the amount of CO2 evolved and the
temperature at which CO2 occurred apart from those of CO, a dry
ice/acetone trap was used.
2.2.2. Temperature-programed reduction (TPR)
TPR experiments were performed using the same apparatus.
After heating the sample at 323 K for 1 h, switching on H2/Ar mix-
ture (15 ml/min, 9.98% H2 in Ar) on the sample was admitted and
followed by increasing the temperature linearly into 873 K at a
constant rate of 2 K/min. These experiments were performed using
a dry ice/acetone trap to prevent any interference peaks.
2.2.3. In situ FTIR
In situ FTIR spectra of the samples were produced by a closed
circulation system with a dead volume of 168 cm3. An IR cell
equipped with NaCl windows was used for pretreatment and mea-
surements. A self-supported wafer (20 mg) was prepared under
nitrogen atmosphere and mounted to the sample holder in the IR
cell. The wafer was evacuated at the desired temperature for
30 min before adsorption at room temperature. The IR spectra
were recorded with a resolution of 2 cm 1 using a Jasco double
beam FTIR-40.
2.2.4. UV–vis spectroscopy
The UV–vis NIR absorption spectra were recorded with a dou-
ble-beam spectrophotometer (Perkin Elmer, Lambda 20) at the
wavelength error of ±0.1 nm.
2.2.5. Transmission Electron Microscope (TEM)
In preparation for TEM imaging approximately 15 mg for each
sample was finally dispersed in 3 ml deionized water by ultrason-
ication for 30 min. A drop of this fine dispersion was deposited on a
carbon foil with microgrid and then dried at 313 K for 10 min until
excess water was removed. TEM micrographs were measured
using a Philips Model Tecani Feil2 instrument at an accelerating
voltage of 200 kV. TEM images were observed with minimum elec-
tron irradiation to prevent damage to the sample structure.
2.2.6. Adsoprtion measurements
The surface properties including BET surface area, total pore
volume and mean pore radius were determined from N2 adsorp-
tion isotherms measured at 77 K; using the BJH model, using con-
ventional volumetric apparatus.
2.2.7. Catalytic reaction
Catalytic performance for the water–gas shift reaction (WGSR)
was studied using a closed circulating system with a Pyrex glass
u-shaped reactor. The amount of catalyst used for catalytic testing
102 M.M. Mohamed, K.S. Khairou / Journal of Colloid and Interface Science 354 (2011) 100–108

was 100 mg. The WGSR preceded at the temperature of 323 K and
at pressures of 60 Torr for CO and 5 Torr for H2O [32]. The catalysts
were treated by heating the samples under vacuum (10 4 Torr) at
323 K for 1.0 h prior to introducing the reacting gas mixture. The
gas phase reactants and products were analyzed by online gas-
chromatography (Shimadzu-14C) using Porapak q and molecular
sieve 13 columns, with a TCD and a FID detector operated at
333 K for the separation of CO, CO2 and H2.

3. Results and discussion

3.1. Adsorption measurements and TEM micrographs

Fig. 1 shows the electron micrograph of Pt/FSM-16 after expo-

sure to the CO/H2O gas mixture (60/5 Torr) at 323 K for 1.0 h.
This sample showed that nano spherical particles of ca. 4.0 nm
diameter were uniformly distributed along the ordered mesopor-
ous surfaces of FSM-16 crystals. The particles formed were discrete
nanoparticles with a narrow size distribution. At higher magnifica-
tion (not shown) Pt nanoparticles on FSM-16 were also seen as cir-
cular in shape. The TEM view given for Au loaded FSM-16, prepared
under the same reaction conditions, showed larger particles than
those of Pt, of nearly spherical shape and diameter of 7–26 nm.
Fig. 2 shows the micrograph of Pt–Au/FSM-16 exposed to the same
reacting gas mixture for 1.0 h. The specific BET surface areas, pore
volume and average pore diameters calculated from the N2 adsorp-
tion isotherms using the BJH model summarized in Table 1 for Pt
and Au supported samples did not change considerably indicating
Fig. 2. TEM image of Pt–Au nanowires obtained following exposure of CO/water
(60/5 Torr) at 323 K.

filling of mesopores by either Au or pt nanoparticles keeping the

Fig. 1. TEM images of Pt and Au nanoparticles encapsulated inside FSM-16
obtained at 323 K after exposure to CO/water (60/5 Torr, 1.0 h).
mesoporous structure intact.
The Pt–Au sample showed major morphological changes from
those of the monometallic ones. It shows nano-structured wires
of 20 nm diameter  120–170 nm long formed on the FSM-16 sub-
strate as well as irregularly arranged structures of little aggrega-
tion of 23–40 nm diameter. These two structures correspond,
respectively, to the bimetallic clusters of Pt–Au and to the non-
complete transformation of either one of monometallics or both.
Accommodating either nanowires or nanoparticles with higher
dimensions than the pore aperture of the host FSM-16 (2.7 nm) im-
plied that they occupy major parts of the FSM-16 surface. The large
diameter of the nano-wire structure could lead to its coincidence
with the cylindrical mesoporous channels of FSM-16, and conse-
quently a significant loss in surface area was expected. However,
Table 1 showed that the decrease in surface area for the Pt–Au wire
sample is not large (24%). This indicates that the wires were not ex-
tended throughout the pore diameter of the host FSM-16 (Table 1)
but mostly deep inside the pore volume that suffers reduction
comprises of only 20%. The presence of very small amounts of
nanoparticles in this image, specifically in the upper part and lower
part, evidences that the formation of wires was originally resulted
from nanoparticles. Investigating the structure images belong to
Au/FSM-16 and Pt/FSM-16 and compare them with that of Pt–
Au/FSM-16, let us presume that monometallic particles showed a
change from nanospheres; in former samples, to a nanowire in
the latter, which showed much wider particles size ranges and thus
showed wide spacing between each other when compared with Pt/
FSM-16. This change in gold and platinum particles morphology
can be attributed to the strong contact between Au and Pt particles
suggesting the alloy formation between them. The existence of the
individual Pt particles as a spherical uniformed aggregates after
exposure to the same reacting gas mixture at the same
temperature for 1.0 h and its morphological change following Au
 M.M. Mohamed, K.S. Khairou / Journal of Colloid and Interface Science 354 (2011) 100–108
Table 1
Surface characteristics of parent FSM-16, Pt/FSM-16, Au/FSM-16 and Pt–Au/FSM-16
catalysts.
Sample Surface area Pore diameter Pore volume
(m2/g) (nm) (cm3/g)
FSM-16 1020 2.95 1.14
2.5 wt.% Pt/FSM-16 952 2.89 1.05
2.5 wt.% Au/FSM-16 938 2.80 1.0
2.5 wt.% Pt-2.5 wt.% Au/ 759 2.65 0.92
FSM-16
incorporation is rarely observed unless exposing to repeated oxida-
tion and reduction steps (e.g. at 923 K) to form e.g. onion like struc-
ture as in case of Pt3Si [33]. However, evidencing the formation of
metal nanowire alloy without any pretreatment undergone by the
catalyst except heating it with the reacting gas mixture at 323 K for
1.0 h will open up a facile novel fabrication of nano-wires for var-
ious potential applications.
The electron micrograph of the Pt–Au/FSM-16 sample exposed
to the reacting gas mixture for 5.0 h (not shown) exhibited the per-
sistence of the nano-wire shape with recognizing a little decompo-
sition into particle ones. One cannot deny the probability of
presence of some aggregated nanoparticles from Au and Pt beside
the nanowires of both.
3.2. TPR profiles of Au/FSM-16, Pt/FSM-16 and Pt–Au/FSM-16 samples
Fig. 3 shows the TPR patterns of FSM-16 encapsulated Au, Pt
and Pt–Au catalysts.
Based on the previously reported TPR study of Au/mordenites
[34] typical two peaks are existed in Au/FSM-16, consisting of
two partially overlapping peaks of maxima at 393 and 423 K char-
acteristic of reduction of gold precursor, chloroauric acid, into zero
valent gold. The degree of reduction of AuCl3, as estimated by inte-
grating the area under the curve, is approximately 32.2% for the
393 K peak, corresponding to AuCl species. This was in agreement
with the reported data for the reduction of AuCl3 at typical temper-
atures [34]. Since there is no peak appeared further during ramping
the temperature till 873 K, one can conclude that the peak at 423 K
represents the metallic state of Au species. This peak is not a result
of the sublimation of chloroauric as AuCl3, which has it self a peak
around 528 K as estimated by Schwank et al. for the Au/SiO2 sys-
tem [7].
For Pt/FSM-16, a strong peak at 423 K and other two minor
peaks at 606 and 806 K were clearly observed, characteristic of
reduction of different Pt species. A quantitative estimation of the
low temperature signal indicates that the amount of H2 consumed
exceeds that required for the reduction of metal chlorides of type
PtCl2 to the corresponding metal. The H2/Pt ratio, which relates
Fig. 3. Temperature programed reduction profiles of mono and bimetallic Au and Pt
supported FSM-16.
103
the moles of H2 consumed per mole of metal atoms, gives a value
of 4.6 to indicate that the main reduction peak is present as PtCl4.6.
This pattern clearly shows that most of the chloroplatinic acid is
greatly reduced, in a very low temperature margin, except very
few amounts constituting a ratio of 0.5 for each one of those
peaked at higher temperatures; 606 and 806 K. These latter peaks
could be caused by the formation kinetics of platinum particles
wide size distributions, in line with those seen in Pt/SiO2 or those
could be hidden in interior surfaces of FSM-16 host [35].
The TPR pattern of the Pt–Au/FSM-16 catalyst shows peaks at
304, 393, 423 and 623 K with a very small minor one at 806 K sim-
ilar to that seen in Pt/FSM-16. The low temperature peak at 304 K
was not depicted for monometallic catalysts. Thus, it might repre-
sent a bimetallic contribution or it can be associated with the
reduction of some chloride bonds. Comparison suggests that the
peaks at 393 K and 423 K in the bimetallic catalyst are exactly
where those of Au/FSM-16 and Pt/FSM-16 appeared, respectively,
but of lower hydrogen consumption. The broad strong peak at
623 K suggests largely the contact of the two metals and their dis-
persions that can be confirmed from the peak broadness extended
from 573 to 673 K. Additionally, this peak is shifted to higher tem-
peratures compared with the one appeared at 606 K for Pt/FSM-16
to further support the bimetallic combination. The H2/M ratio indi-
cates a value of 2.65 for the peak centered at 623 K. The decrease in
the total H2 consumption for Pt–Au/FSM-16 compared with the
summation of monometallic ones could be attributed to the bime-
tallic alloy formation. In a DFT study, it has been revealed that the
binding energy of H2 on Pt–Au surface was weaker than on Pt [36].
From the foregoing results, one can interpret that the reduction
mechanism proceeds initially with the charged Au species convert-
ing them into metallic moieties through dissociating hydrogen;
atomic hydrogen migrates to Pt chloride in contact with metallic
Au and reduces the chloride instantaneously. This is validated
through comparing the patterns of monometallics showing that
Au chloride reduces earlier into Au0 at 423 K. At that temperature,
chloroplatinic reduction was started. Indeed, the reduction of Au
commenced first followed by Pt; as will also be affirmed later from
IR observations, comprehending that; in this specific case, Au is a
core enclosed by a Pt as a shell. Emphasizing this phenomenon is
beyond the scope of this work since it necessitates further work.
It is interesting to observe that the rate of reduction was ini-
tially fast for the individual catalyst and slowed down as time
elapsed, whereas the bimetallic catalyst showed a different behav-
ior. These different types of reduction models implied that Pt and
Au entities are in intimate contact with each other possibly in
the form of a bimetallic alloy.
3.3. Temperature-programed carbonylation reaction (tpcor)
TPCOR of Au/FSM-16, Pt/FSM-16 and Pt–Au/FSM-16 catalysts
were carried out to evaluate the consumption of CO during ramp-
ing the temperature as well as monitoring the amount of CO2 for-
mation (Fig. 4).
The TPCOR pattern of Au/FSM-16 shows a broad peak centered
at 393 K that compromises a value of 9.8 lmol CO. CO consump-
tion was appreciably increased on Pt/FSM-16 giving a value of
20.2 lmol at a maximum of 406 K. This can be due to increasing
the adsorption energy of CO on Pt than on Au. The TPCOR pattern
of Pt–Au/FSM-16 shows a peak at 393 K, and a broad peak centered
at 415 K. The amount of CO consumed calculated for latter peaks
measured a value of 27.0 lmol. Decreasing the amount of CO con-
sumed on Pt–Au compared with those obtained from combination
of Au/FSM-16 and Pt/FSM-16 evidences the formation of surface al-
loy formation. It has been acknowledged that bimetallic alloy com-
position produces a synergistic effect that involves the suppression
of adsorption [37,38]. The carbonylation of Pt/FSM-16 occurred at a
104 M.M. Mohamed, K.S. Khairou / Journal of Colloid and Interface Science 354 (2011) 100–108

Fig. 4. Temperature programed carbonylation reaction (TPCOR) of CO on Pt nanoparticles, Au nanoparticles and Pt–Au nano-wires encapsulated in FSM-16 while using liq. N2
trap & dry-ice acetone trap (CO2).

significantly lower temperature than that of Pt–Au/FSM-16, sug-

gesting that electronic effects are perceived on adding Au to Pt
rather than coverage effects [36].
The desorbed amount of CO2 from Pt–Au/FSM-16 was the high-
est comparatively where it detected in small amounts on Au/FSM-
16, considering that the adsorbed CO only reacts via an associative
mechanism (most probably via involving surface formates;
through surface hydroxyls, and carbonates) [29,32]. The lower
activity of Au for CO oxidation could be due to lower stability Au
nanoparticles forming large catalytically inactive clusters which
in line with pervious studies [39]. Extending the CO consumption
during ramping the temperature from 393 to 423 k in Pt–Au/
FSM-16 reflects the broad distribution of adsorption sites of differ-
ent adsorption strengths. This is emphasized by dividing the
amount of CO2 produced, in lmol/g, for each catalyst by the initial
amount of the metal, in lmol/g, of each catalyst, on the assumption
that CO2/Ms = 1. That was simple to calculate for the monometallic
Au/FSM-16 and Pt/FSM-16 catalysts those disclosed values equal
0.9 and 1.65, respectively. However, the dispersion for bimetallic
catalyst based on CO2/Pts showed a value of 2.1 that somewhat lar-
ger than the value of Pt/FSM-16 giving a decisive modification of Pt
upon addition of Au.

3.4. In situ FTIR spectra

3.4.1. Pt–Au nanowire/FSM-16 sample

Fig. 5 shows the IR spectra of Pt–Au/FSM-16 after exposing to
the CO/H2O gas mixture (60/5 Torr), at 323 K, as a function of time.
The obtained spectra were subtracted from background for each
spectrum. Trace a shows the bands recorded after admission of
the reacting gas mixture for 3 min.
The bands at 2184 and 2124 cm 1 are observed and ascribed to
the chemisorbed CO onto Au+ and Au0, respectively coincided with
those reported for CO coordinated Au/mordenites [34]. Besides, a
band at 1440 cm 1 is also revealed (not shown) and agreed well with
the features of free carbonate species depicted by Davydov et al. [40].
After 20 min, the 2184 cm 1 band showed a marked decrease in
intensity with a harmonized developing of the Au0–CO
(2128 cm 1) band. After 40 min, an evolution of small band at
2080 cm 1 together with a shoulder at 2165 was clearly distin-
guished with simultaneous declining in intensity for the band at
2184 cm 1, which appeared as a shoulder. That band at 2080 cm 1
is intensified and blue shifted to give M–CO stretching band at
2084 cm 1 attributable to the preliminary formation of the terminal
vibrational carbonyl coordinated Pt–Au nanowires. Comparison of
latter spectrum with those of CO coordinated Pt and Au [41,42]
was accomplished to undoubtedly confirm the former assignment.
Fig. 5. In situ FTIR spectral changes of Pt–Au nanowires encapsulated in FSM-16
after adsorption of 60 Torr CO and 5 Torr H2O mixture at 323 K as a function of time.
From bottom to top after: (a) 3 min, (b) 10 min, (c) 20 min, (d) 40 min, (e) 60 min,
(f) 80 min, (g) evacuation for 5 min, (h) evacuation for 20 min.
Appearing the linear CO stretching vibration of Pt–Au nanowires
at 2084 cm 1 while that of Pt nanowires at 2092 cm 1 [29] pre-
sumes the formation of Pt–Au alloy i.e. addition of Au by co-impreg-
nation forming the bimetallic Pt–Au sample resulted in a red shift of
the CO–Pt peak from 2092 to 2084 cm 1. This result agrees with the
finding of Woo et al. [43] while investigating the formation of Pt–Ru
nanowire. The shift in the CO stretching band of the Pt–Au nanowire
is attributed to electron transfer from Au to Pt in the Pt–Au alloy,
which can be explained in terms of back donation of electrons from
Pt d-bands into 2P* molecular orbital of adsorbed CO [37]. This spec-
 M.M. Mohamed, K.S. Khairou / Journal of Colloid and Interface Science 354 (2011) 100–108
trum showed an increase in the Au0–CO (2124 cm 1) band intensity
at the expense of that at 2186 cm 1 (Au+–CO); which suffered a great
diminishing, emphasizing an increase in Au metal ratio. The
2165 cm 1 band can be attributed to the CO gas phase condensed in-
side FSM-16 channels resemble that of CO adsorbed gas inside zeo-
lite cages [44].
Spectrum d signified the transition state between the predom-
inant influence of the metallic Au carbonyls and the consequent
developing of the Pt–Au carbonyl species. After 60 min, the
2084 cm 1 band showed an increase in intensity with the perma-
nent existence of the band at 2124 cm 1, along with a concomitant
development of a broad band at 1888 cm 1, characterizing bridged
carbonyl coordinated Pt species [27].
To date, no Pt–Au carbonyl clusters are formed without any
treatment in advance as illustrated in this work, however, cluster
formation usually necessitates a high reductive carbonylation
(e.g. CO = 200–760 Torr) [45] as well as high heating temperatures
(473 K), which extend sometimes for 24 h. The cluster formation
under the reduced CO/H2O pressure suggests that Au played a un-
ique role on Pt clusters formation inside the FSM-16 framework.
The bands at 2084 and 1888 cm 1 were successively intensified
with time with fading that of Au0–CO at 2124 cm 1. Upon evacua-
tion at 323 K for 20 min, the Pt–Au carbonyl clusters were still
prominent (spectrum h).
One can notice that, there is no band observed for Pt carbonyls
during the first 35 min. however, the Pt–Au carbonyl cluster formed
Fig. 6. In situ FTIR spectral changes of Pt–Au encapsulated in NaY zeolite after adsorption of 60 Torr CO and 5 Torr H2O mixture at 323 K as a function of time. From bottom to
top after (a) 3 min, (b) 10 min, (c) 30 min, (d) 50 min, (e) 60 min.
105
at the time when Au+-CO was almost reduced to Au0–CO. This ex-
plains that the Pt–Au carbonyl cluster formation is very much asso-
ciated with the reduction of Au+-CO species. More precisely, water–
gas shift reaction took place initially on Au+ sites; to form CO2 and H2.
Then H2 spilled over into Pt to further reduce it into metal particles
that by its turn form clusters. The facile reaction of gold at very
beginning to form carbonyls followed by Pt could indicate that Pt
forms like a shell enclosing Au. More research work is required to
accomplish the latter suggestion. More interestingly, no decomposi-
tion for Pt–Au carbonyl clusters is noted on evacuation (10 5 Torr) at
323 K. However, only a small shift to lower wavenumbers is de-
picted (2084–2070 and 1888–1850 cm 1) without a substantial
intensity change [27,46]. Thus, suggesting that the structure of the
Pt carbonyl clusters bounded with Au as Au–Pt–CO is substantially
expected; together with the coexistence of Pt–CO clusters (1888–
1850 cm 1) [27,29] as emerged from DFT studies on Pt–Au microcl-
usters calculation using CO molecule as a probe [35,37]. It has thus
been envisaged through FTIR that CO bound in a different way to
the bimetallic catalyst than it did to either parent metal, suggesting
a bimetallic combination.
3.4.2. In situ FTIR of Pt–Au/NaY sample
Fig. 6 shows the spectra recorded after dosing CO/H2O (60/
5 Torr) gas mixture onto Pt–Au/NaY, at 323 K, as a function of time.
After 3 min contact time, bands at 2184, 2128, 1635 and
1440 cm 1 were observed.
106 M.M. Mohamed, K.S. Khairou / Journal of Colloid and Interface Science 354 (2011) 100–108
With increasing time, the intensities of the bands at 2128 (Au0–
CO), 1638 and 1440 cm 1 were significantly enhanced. Besides, a
band at 2584 cm 1 was also formed to display an intensive in-
crease in intensity with time. The band at 2184 cm 1 (Au+–CO)
exhibited a marked decrease in intensity during evolving of a small
band at 2080 cm 1. In addition, a small band at 1730 cm 1 was de-
tected with time. The bands at 2584 and 1730 cm 1 are developed
synchronously with increasing time, indicating that they might
arise from the same species. These bands can be assigned to the
bicarbonate species [40]. Upon evacuation, the 2584 and
1730 cm 1 bands disappeared while a marked decrease in inten-
sity is noted for the 1440 cm 1 band.
On comparing the various surface species formed on Pt–Au/
FSM-16 with those of the corresponding analogue on NaY, one
can figure out that the complete formation of Pt–Au carbonyl clus-
ters was not existed in NaY cavities emphasizing the stabilizing ef-
fect offered by the host FSM-16. In addition, the band at 2080 cm 1
(Pt–Au) showed a shift to lower frequencies as well as a decrease in
intensity compared with that in FSM-16. The deformation band of
water (1638 cm 1) as well as the carbonate ones [38] (1440 cm 1)
on NaY sample was enormously bigger than those seen on FSM-16.
The competitive adsorption between H2O and carbonates/bicar-
bonates on the active sites could obstruct the generation of Pt–
Au carbonyl cluster inside NaY cavities of higher hygroscopsity
than FSM-16. Further, the presence of Au near by Pt could prohibit
the formation of Pt–Au carbonyl clusters due to the micropore con-
straint. The two metals could also remain aggregated due to their
miscibility gap and thus prohibiting the bimetallic alloy formation.
3.5. UV–vis absorption spectra
Fig. 7 shows the UV–vis absorption spectra of pure Au, pure Pt,
and Pt–Au supported by FSM-16 substrate. The peaks at 450 nm
Fig. 7. UV–vis spectra of Pt/FSM-16, Au/FSM-16 and Pt–Au/FSM-16 catalysts.
and 567 nm belong to surface plasmon absorption of pure Pt/
FSM-16 and Au/FSM-16, respectively, those previously submitted
to the CO/H2O gas mixture at 323 K.
According to the Mie theory the position of the SPR band max-
ima (kmax) of spherical noble metal (Au0 and Pt0) NPs is sensitive to
changes in refractive index (RI, n) of the embedding medium [47].
It is known that increasing refractive index of matrix shifts the SPR
band towards longer wavelength. However, our FSM-16 owns ultra
low RI (n = 1.07), and thus does not cause any shift for Au0 and Pt0-
SPR than what already known for them as individual nanoparticles
[48].
The alloy formation is indicated from the fact that the optical
absorption spectrum shows two broad composition-sensitive plas-
mon bands one of them is in between the absorption peaks of the
two pure metal NPs (at 515 nm) [49] where the other is at 400 nm
close to that of pure Pt. The plasmon band at 515 nm is slowly red-
shifted from Pt peak position (450 nm) towards Au peak position
(567 nm) depending on increasing Au/pt ratio nanoparticles
(Fig. 7). It has been observed that the bimetallic particles show a
completely different spectral profile from the monometallic ones
in the sense that in Pt–Au the Plasmon absorption bands are de-
creased and became wider. The blue-shift of the surface Plasmon
band of the peak appearing at 400 nm might arise from the plas-
mon band of Au/Pt composite clusters and/or due to the chemical
interaction of the two metals. These facts from Fig. 7 clearly indi-
cate the formation of Pt–Au bimetallic nanoparticles rather than
physical mixtures of pure Au and Pt nanoparticles.
3.6. Catalytic activity
The initial formation rate of CO2 produced from performing
WGS reaction under different pressures of CO and H2O and at dif-
ferent temperatures for FSM-16 supported Pt–Au is compared in
 M.M. Mohamed, K.S. Khairou / Journal of Colloid and Interface Science 354 (2011) 100–108 107

Table 2
Catalytic activity of Pt–Au/ FSM-16 catalysts toward water–gas shift reaction.

Catalyst Reaction temp. (K) CO:H2O (Torr) H2 and CO2 formation rate (lmol/g s) Activation energy (kJ/mol)
Au wt.% Pt wt.% Support
3
2.5 2.5 FSM-16 323 60:5 2.8  10 0.086 56.4
3
2.5 2.5 FSM-16 350 60:5 5.0  10 0.14
2.5 2.5 FSM-16 373 60:5 0.037 0.67
2.5 2.5 FSM-16 323 20:5 0.045
2.5 2.5 FSM-16 323 40:5 0.073
2.5 2.5 FSM-16 323 60:10 0.053
1.0 2.5 FSM-16 323 60:5 0.033
2.5 1.0 FSM-16 323 60:5 0.027
2.5 – FSM-16 323 60:5 0.013
1.25 2.5 FSM-16 323 60:5 0.02
1.25 FSM-16 323 60:5 0.039
2.5 2.5 NaY (CVD) 323 60:5 0.0025
2.5 2.5 NaY (Co-imp) 323 60:5 0.0023

Fig. 8. The influence of bimetallic Pt–Au supported on FSM-16 towards enhancing
the activity of WGSR at 323 K. Reaction conditions; 0.1 g catalyst, PCO = 60 Torr,
PH2O = 5 Torr.
Table 2. The increased pressure of CO at 323 K resulted in a linear
increase of CO2 formation rate (not shown) with a slope near to 1.
The dependence of CO2 formation on CO pressure indicates that
WGS reaction is first order with respect to CO. On testing the
dependence of the activity of the catalyst on temperature, we
found that the activity is enhanced at 373 K to almost 8 times that
at 323 K. Observations of the function of Pt–Au bulk alloy catalysts
include the decrease of activation energy to 56.4 kJ/mol and thus
facilitating oxidative desorption as well as suppressing the CO
adsorption i.e. sufficient high adsorptivity to support catalytic oxi-
dation reactions. This value is low if compared with those of Pt/
Al2O3 (122 kJ/mol), Cu–Zn–Al2O3 (110 kJ/mol) and [Pt12(CO)18]2 /
NaY (70 kJ/mol) while performing the same reaction [50,51]. The
increased pressure of water from 5 to 10 Torr at constant CO pres-
sure (60 Torr) decreased the activity towards CO2 formation by a
factor of 1.6 than that at 323 K for 5 Torr. This could be due to
masking of some active sites when vapor pressure of water in-
creased. Furthermore, the competitive adsorption between CO
and H2O on the active site might be another plausible explanation.
The preliminary formation rate of CO2 over FSM-16 containing Pt–
Au is remarkably increased (ca. 35 times) compared with that in
NaY zeolite. More interestingly, the former catalyst exhibited high-
er activity (20 times) in the WGS reaction than that on
[pt15(CO)30]2 entrapped in FSM-16, [27] based on measuring the
initial formation rate of CO2 produced (Fig. 8).
Besides, the role played by the bimetallic structure of Pt–Au on
FSM-16 surfaces in the sense that lowering the bond strength of
Pt–Au (2.30 eV) than those of Pt2 (3.29 eV); and its closer to those
of Au2 (2.04 eV), [52] indicates an expected weakened interaction
between CO and Pt–Au; than with Pt2, but strong enough to partic-
ipate with H2O forming CO2 and H2 rather than Au2. The latter pos-
sesses stronger relativistic effect than Pt2 that owned more 5d
orbitals participating in the bonding of Pt2 [53] of higher strength
than Au2. Thus, an anticipated strong interaction of CO with Pt–Pt
will indeed affect on the liability of CO to react further with water
molecules to perform effectively WGSR since each CO molecule
effectively blocks two adsorption sites [29]. Contrarily, it has been
acknowledged that the adsorption of CO on Pt atom adjacent to Au
ones is the preferred adsorption center [54] i.e. the strength of CO
cluster bond depends on the environment of atom that CO interact
with. A small production of H2 is depicted in Table 2 probably due
to the interaction between CO and OH groups present on the sur-
face of the support [55]. This suggests that at low temperature,
CO oxidation on Pt–Au based catalysts involves the removal of sur-
face OH groups.
Testing the activity on varying the loadings of Au and Pt sup-
ported by FSM-16 indicated that exceeding the Au loading over
Pt decreases the activity. This could be due to Au covers the Pt;
in this specific case, and thus a decrease in CO adsorption at such
low temperature (323 K) is expected [56]. This is also validated
from decreasing the frequency shift of CO bonded Pt–Au nanowires
compared with Pt nanowires [29] giving a hint on decreasing the
electron deficiency of the former comparatively. This exhibited de-
crease in CO bounded Pt–Au surface than Pt one could also help
reducing CO poisoning. The positive synergistic effect offered by
Pt–Au was mainly based on the structure and particle morphology
of nanowires of considerable electron deficiency rather than sur-
face area values. This infers that the surface atoms of the wires
are highly active for WGSR compared to those of particles. Results
presented in Table 2 showed that monometallic Pt exposes a little
increase in WGSR than the corresponding Au reflecting that this
reaction does depend on neither dispersion nor crystallites size.
Catalysts with equal total mutual contents such as 1.25Au-
1.25Pt/FSM-16 presented higher activities than that of 2.5Pt/
FSM-16 comprehending the synergistic effect of bimetallic cata-
lysts. The presence of small nanoparticles in Fig. 2 besides the
nanowires proved without doubt that the latter are the active com-
ponents since the former; that owned spherical particles as de-
picted in Fig. 1, offered very low activity comparatively. This was
in agreement with the study carried out the same reaction on Pt
nanowires, which showed higher activity as compared with Pt
nanoparticles [29].
4. Conclusions
The present study was undertaken to synthesize nanowire Pt–
Au clusters using non-conventional technique via introducing
Au+ into in situ designed Pt carbonyl clusters. That gold showed
108 M.M. Mohamed, K.S. Khairou / Journal of Colloid and Interface Science 354 (2011) 100–108
positive effects on decreasing the temperatures required for reduc-
ing Pt as well as stabilizing its carbonyl clusters formed. Character-
ization results provided evidence of Au–Pt bimetallic formation.
The presence of Au together with Pt showed a dramatic effect on
the activity towards WGSR compared with the monometallics. This
was mainly due to the effectiveness of the surface alloy at such low
temperatures (323 K) and to the morphology of the wire structure.
Gold shows more than a simple site blockage on Pt surface; as con-
ceived from lowering the amounts of H2 consumed, as depicted
from TPR results, and CO consumption; illustrated from TPCOR,
but however, stimulates an electronic effect in which Au affects
the adsorption properties of several adjacent Pt sites i.e. such mod-
ifications cause changes in surface interatomic distances and thus
the electronic structures. An interesting feature that emerged was
that the catalytic activity towards WGSR is merely dependent on
surface area, particles size and dispersion. A reduction mechanism
of forming Pt–Au alloy was evidenced based on reduction of Au+
moieties through performing WGSR; to form gold metal, and thus
evolve H2 that can diffuse across the shell to reduce Pt species.
References
[1] B.C. Gates, J.R. Katzer, G.C.A. Schuit, Chemistry of Catalytic Processes, McGraw-
Hill, New York, 1979.
[2] V. Ponec, G.C. Bond, Catalysis by metals and alloys, in: B. Delmon, J.T. Yates
(Eds.), Studies in Surface Science and Catalysis, vol. 95, Elsevier, Amsterdam,
1995.
[3] X. Li, A.E. Kuznetsov, H.-F. Zhang, A.I. Boldyrev, L.-S. Wang, Science 291 (2001)
859.
[4] H. Tada, F. Suzuki, S. Ito, T. Akita, K. Tanaka, T. Kawahara, H. Kobayashi, J. Phys.
Chem. B 106 (2002) 8714.
[5] D. Bazine, C. Mottet, G. Treglia, Appl. Catal. A 200 (2000) 47.
[6] J.W.A. Sachtler, J.A. Somorjai, J. Catal. 81 (1983) 77.
[7] (a) A. Sachdev, J. Schwank, J. Catal. 120 (1989) 353;
(b) C. Mihut, B.D. Chandler, M.D. Amiridis, Catal. Commun. 3 (2002) 91;
(c) L.D. Marks, Rep. Prog. Phys. 57 (1994) 603.
[8] J.M. Dumas, C. Geron, H. Hadrane, P. Marecot, J. Barbier, J. Mol. Catal. 77 (1992)
87.
[9] J.C. Menezo, M.F. Denanot, S. Peyrovi, J. Barbier, Appl. Catal. 15 (1985) 353.
[10] J. Barbier, P. Marecot, G. Mabilon, D. Durant, M. Prigent, French Patent 90/
15,750, 1990.
[11] B.C. Gates, Chem. Rev. 95 (1995) 511.
[12] R. Psaro, S. Recchia, Catal. Today 41 (1998) 139.
[13] B.C. Gates, L. Guczi, H. Knozinger, in: B. Delmon, J.T. Yates (Eds.), Metal
Clusters in Catalysis, Studies in Surface Science and Catalysis, vol. 29, Elsevier,
Amsterdam, 1986.
[14] S. Kawi, O. Alexeev, M. Shelef, B.C. Gates, J. Phys. Chem. 99 (1995) 6926.
[15] R. Ugo, C. Dossi, R. Psaro, J. Mol. Catal. A 107 (1996) 13.
[16] Y. Iwasawa, M. Yamada, J. Chem. Soc. Chem. Commun. (1985) 675.
[17] J.R. Anderson, D.E. Mainwaring, J. Catal. 35 (1974) 162.
[18] J.R. Anderson, P.S. Elmes, R.F. Howe, D.E. Mainwaring, J. Catal. 50 (1977) 508.
[19] F.S. Xiao, M. Ichikawa, Cuihua Xuebao 14 (1993) 87.
[20] O. Alexeev, S. Kawi, M. Shelef, B.C. Gates, J. Phys. Chem. 100 (1996) 253.
[21] P. Braunstein, J. Rose, in: I. Bernal (Ed.), Stereochemistry of Organometallic and
Inorganic Compounds, vol. 3, Elsevier, Amsterdam, 1988, p. 3.
[22] I.V.G. Graf, J.W. Bacon, M.E. Curley, L.N. Ito, L.H. Pignolet, Inorg. Chem. 35
(1996) 689.
[23] J.S. Beck, J.C. Vartuli, W.J. Roth, M.E. Leonowicz, C.T. Kresge, K.D. Schmitt, C.T.-
W. Chu, J.L. Schlenker, J. Am. Chem. Soc. 114 (27) (1992) 5834.
[24] D.K. Captain, K.L. Roberts, M.D. Amiridis, Catal. Today 42 (1998) 93.
[25] B.D. Chandler, L.I. Rubinstein, L.H. Pignolet, J. Mol. Catal. 133 (1998) 267.
[26] B.D. Chandler, A.B. Schabel, F. Blanford, L.H. Pignolet, J. Catal. 187 (1999) 367.
[27] T. Yamamoto, T. Shido, S. Inagaki, Y. Fukushima, M. Ichikawa, J. Am. Chem. Soc.
118 (24) (1996) 5810.
[28] S. Inagaki, Y. Fukushima, K. Kuroda, J. Chem. Soc. Chem. Commun. (1993) 680.
[29] M.M. Mohamed, M.S. Thabet, J. Phys. Chem. C 112 (24) (2008) 8890.
[30] (a) M.M. Mohamed, K.S. Khairou, Energy Fuels 23 (9) (2009) 4413;
(b) G. Jacobs, S. Ricote, P.M. Patterson, U.M. Graham, A. Dozier, S. Khalid, E.
Rhodus, B.H. Davis, J. Appl. Catal. A 292 (2005) 229.
[31] L.-F. Rao, A. Fukuoka, N. Kosugi, H. Kuroda, M. Ichikawa, J. Phys. Chem. 94 (13)
(1990) 5317.
[32] M.M. Mohamed, T.M. Salama, M. Ichikawa, J. Colloid Inter. Sci. 224 (2000) 366.
[33] A. Fukuoka, N. Higashimoto, Y. Sakamoto, M. Sasaki, N. Sugimoto, S. Inagaki, Y.
Fukushi, M. Ichikawa, Catal. Today 66 (2001) 23.
[34] M.M. Mohamed, T.M. Salama, R. Ohnishi, M. Ichikawa, Langmuir 17 (2001)
5678.
[35] B.D. Chandler, L.I. Rubinstein, L.H. Pignolet, J. Mol. Catal. A 133 (3) (1998) 267.
[36] H. Ren, M.P. Humbert, C.A. Menning, J.G. Chen, Y. Shu, U.G. Singh, W.S. Chang,
Appl. Catal., A 375 (1) (2010) 303.
[37] H. Gronbeck, W. Andreoni, Chem. Phys. 262 (2000) 1.
[38] W.Q. Tian, M. Ge, F. Gu, T. Yamada, Y. Aoki, J. Phys. Chem. A 110 (2006) 6285.
[39] A. Abd El-Moemen, G. Kucěrova, R.J. Behm, Appl. Catal. B 95 (2010) 57.
[40] A.A. Davydov, Infrared Spectroscopy of Adsorbed Species on the Surface of
Transition Metal Oxides, John Wiley & Sons, 1990.
[41] C. Mihut, C. Descorme, D. Duprez, M. Amiridis, J. Catal. 125 (2002) 212.
[42] R.W.J. Scott, C. Sivadinarayana, O.M. Wilson, Z. Yan, D.W. Goodman, R.M.
Crooks, J. Am. Chem. Soc. 127 (2005) 1380.
[43] W.C. Choi, S.I. Woo, J. Power Sources 124 (2003) 420.
[44] M. Mihaylov, B.C. Gates, J.C. Fierro-Gonzalez, K. Hadjivanoy, H. Knozinger, J.
Phys. Chem. C 111 (6) (2007) 2548.
[45] M. Chatterjee, I. Iwasaki, Y. Onodera, T. Nagase, Catal. Lett. 61 (1999) 199.
[46] J.C. Calabrese, L.F. Dahl, P. Chini, G. Longoni, S. Martinengo, J. Am. Chem. Soc.
96 (8) (1974) 2614.
[47] M. Yamaguchi, H. Nakayama, K. Yamada, H. Imai, Opt. Lett. 34 (13) (2009)
2060.
[48] Y. Luo, H.O. Seo, K.-D. Kim, M.J. Kim, W.S. Tai, M. Burkhart, Y.D. Kim, Catal. Lett.
134 (2010) 45.
[49] L. Chen, W. Zhao, Y. Jiao, X. He, J. Wang, Y. Zhang, Spectrochim. Acta A 68
(2007) 484.
[50] R.-J. Wang, T. Fujimoto, T. Shido, M. Ichikawa, J. Chem. Soc. Chem. Commun. 13
(1992) 962.
[51] M. Sasaki, M. Osada, N. Higashimoto, T. Yamamoto, A. Fukuoka, M. Ichikawa, J.
Mol. Catal. A 141 (1999) 223.
[52] L. H. Pignolet; in R.D. Adams, F.A. Cotton (Eds.), Catalysis by Di- and
Polynuclear Metal Cluster Complexes, VCH Publishers, New York, 1997.
[53] P. Pyykko, Chem. Rev. 88 (1988) 563.
[54] Q. Ge, C. Song, L. Wang, Comput. Mater. Sci. 35 (3) (2006) 247.
[55] P.O. Graf, D.J.M. De Vlieger, B.L. Mojet, L. Lefferts, J. Catal. 262 (2009) 181.
[56] C.-H. Kim, L.T. Thompson, J. Catal. 244 (2) (2006) 248.