Monodisperse and polydisperse platinum nanoclusters supported over TiO2 anatase as catalysts for catalytic oxidation of styrene

Article abstract of DOI:10.1016/j.molcata.2015.12.011

Very small platinum nanoclusters have been synthesized by reduction of H₂PtCl₆·6H₂O salt with NaBH₄ in the presence and absence of water-soluble l-cysteine ligand. The synthesized platinum clusters were supporte

Full text of DOI:10.1016/j.molcata.2015.12.011

Journal of Molecular Catalysis A: Chemical 413 (2016) 67–76
Contents lists available at ScienceDirect
Journal of Molecular Catalysis A: Chemical
journal homepage: www.elsevier.com/locate/molcata
Monodisperse and polydisperse platinum nanoclusters supported
over TiO anatase as catalysts for catalytic oxidation of styrene
2
Mostafa Farrag
Chemistry Department, Faculty of Science, Assiut University, 71515 Assiut, Egypt
a r t i c l e i n f o
a b s t r a c t
Article history:
Very small platinum nanoclusters have been synthesized by reduction of H PtCl ·6H O salt with NaBH
2
6
2
4
Received 8 November 2015
Received in revised form 2 December 2015
Accepted 16 December 2015
Available online 22 December 2015
in the presence and absence of water-soluble l-cysteine ligand. The synthesized platinum clusters were
supported over anatase titanium dioxide (TiO2) by 1% wt/wt. The spectroscopic properties of Ptn(l-Cys)m
clusters were studied by UV–vis spectroscopy and Fourier transform infrared spectroscopy (FTIR) and
its chemical structure were determined by thermogravimetric analysis (TGA) and elemental analysis.
The particle sizes of supported and non-supported platinum clusters were characterized by transmis-
sion electron microscopy (TEM) and powder X-ray diffraction analysis. As well as their specific surface
areas were determined using nitrogen adsorption/desorption measurements, pore volume and average
pore diameter were also calculated. Finally, oxidation of styrene is used as a simple model reaction to
demonstrate that these systems exhibit the potential to be used as heterogeneous catalysts. The doped
monodisperse platinum clusters (1% Ptn(l-Cys)m/TiO2) catalyst exhibited the highest catalytic activity
and selectivity (100% benzaldehyde) in oxidation of styrene. The non-doped platinum clusters (Ptn(l-
Cys)m) showed more catalytic activity than doped polydisperse platinum clusters (1% Pt/TiO2). Hydrogen
peroxide was used as oxidizing agent, which exhibited stronger oxidation potential than oxygen gas.
The unique atom-packing structure and electronic properties of Ptn(l-Cys)m nanoclusters (∼1 nm) are
rationalized to be responsible for their extraordinary catalytic activity observed in oxidation of styrene.
Keywords:
Monodisperse platinum nanoclusters
Non-protected platinum clusters
Oxidation of styrene
Doped TiO2 anatase
Hydrogen peroxide
©
2015 Elsevier B.V. All rights reserved.
1
. Introduction
chemical reaction such as selective oxidation and hydrogenation
processes [4,14]. A colloidal method was used to prepare unpro-
tected platinum clusters of well-defined size that are subsequently
functionalized in a separate step with some hydrophilic chiral
ligands [11]. These platinum clusters were supported over alu-
mina to use as potential asymmetric catalysts in hydrogenation of
2-butanone [11]. Schrader et al. prepared Pt nanoparticles func-
tionalized with l-proline ligand, these protected nanoparticles
showed high catalytic activity in selective hydrogenation of ace-
tophenone [12]. In my previous work, I synthesized some platinum
nanoclusters protected by l-glutathione, N-acetyl-l-cysteine and
l-penicillamine in regime ∼1 nm [13]. These clusters showed high
photocatalytic activity in degradation of methylene blue dye after
deposition over Degussa TiO2 P25 [13].
Bulk platinum, as a noble metal, was initially thought to be cat-
alytically inert, but the nanoparticles platinum (<10 nm) has been
demonstrated to exhibit catalytic activity in a variety of chem-
ical reactions [1–3]. However, the origin of the catalytic power
of nanoscale platinum is still unclear and under intense debate
[
4]. This is primarily due to the polydispersity issue of platinum
nanoparticles. To prepare monodisperse platinum clusters, a great
deal of work on model catalysts should do, such as deposition of
platinum atoms over single-crystal surfaces under ultrahigh vac-
uum (UHV) conditions 10 mbar [5]. This model of catalysis is
very complicated and largely deviated from the real-world catalytic
conditions.
−
9
I reported my effort in creating a well-defined catalytic
system by preparing truly monodisperse, thiolate-capped gold
nanoclusters (Aun) [6–10] and protected platinum nanoclus-
ters (Ptn) [11–13] and utilizing them as homogeneous catalysts
Different sizes of gold clusters (Au25(SCH CH Ph)18,
2
2
Au38(SCH CH Ph)
and Au144(SCH CH Ph)60) were used as
2
2
24
2 2
catalysts in oxidation of styrene in presence of oxygen as oxidizing
agent [4]. The catalytic activity of these clusters decreases in this
order Au25 ᮀ Au38 ᮀ Au144 [15]. Coated titania with porous carbon
(
unsupported) and heterogeneous catalysts (supported) for many
(
PC@TiO ) showed the highest catalytic activity in oxidation of
2
styrene compared to bare TiO2 and carbon coated titania (C@TiO2)
in presence of hydrogen peroxide (H O ) as oxidizing agent [16].
E-mail address: mostafafarrag@aun.edu.eg
2
2
http://dx.doi.org/10.1016/j.molcata.2015.12.011
381-1169/© 2015 Elsevier B.V. All rights reserved.
1

6
8
M. Farrag / Journal of Molecular Catalysis A: Chemical 413 (2016) 67–76
As we confirmed in our previous work [17] hydrogen peroxide is a
stronger oxidizing agent than molecular oxygen. Size selected gold
clusters protected by l-glutathione (Au25(SG)18) was supported
distilled water) was added dropwise over the resulting solution
while stirring vigorously (∼1100 rpm), During the reduction step
the color of solution changed from canaries yellow to brown then
to deep brown gradually, indicating the reduction of the platinum
salt and the formation of nanoparticles. The reaction was allowed
to proceed under constant stirring for 1 h. The mixture was evapo-
rated under a vacuum to near dryness, and then the particles were
precipitated by adding ethanol. The resulting precipitate was then
collected through centrifugal precipitation and repeatedly washed
with ethanol to remove the unreacted material. The deep brown
solid consisting of Ptn(l-Cys)m nanoclusters was finally dried under
reduced pressure in a vacuum desiccator.
◦
over hydroxylapatite and calcined at 300 C in vacuum 2 h to
use as a catalyst in oxidization of styrene by using anhydrous
tert-butyl hydroperoxide (TBHP) as an oxidant [18]. This catalyst
oxidized styrene by 100% conversion and 92% selectivity to the
epoxide after 12 h stirring [18]. Lambert and co-workers have
demonstrated that Au55 clusters supported on silica (Au55/SiO2)
can activate the catalytic oxidation reaction of styrene by using
oxygen gas as oxidizing agent. This catalyst shows high catalytic
conversion of styrene because of the electronic structures of the
clusters, although the yield of the epoxide is not high (5%) [19].
This work hints at the possibility that there is an optimal cluster
size for the efficient and selective oxidation of alkenes [19].
2.3. Preparation of supported Ptn(l-Cys)m nanoclusters over TiO2
anatase
Herein, I report a new method to prepare monodisperse plat-
inum clusters protected by l-cysteine. The method used in this
work produces very narrow size distribution platinum clusters
around 1 nm. The yield of prepared clusters is high, since this
method based on a wet chemistry method. The optical properties of
protected platinum clusters were studied by UV–vis spectroscopy
and Fourier transform infrared spectroscopy (FTIR). The size and
composition of Ptn(l-Cys)m clusters were assessed by transmis-
sion electron microscopy (TEM), powder X-ray diffraction analysis
The 1% Ptn(l-Cys)m/TiO2 catalyst was prepared by well known
impregnation method. 1 g TiO anatase powder was dispersed into
2
a 40 ml aqueous solution contains 10 mg Ptn(l-Cys)m clusters. The
slurry was stirred one day at room temperature, the brown color
of clusters solution becomes colorless, confirming the transfer of
all clusters amount to the TiO2 surface. The suspension was evap-
◦
orated at 100 C until drying, getting size selected nanoclusters
supported over TiO , denoted as 1% Ptn(l-Cys)m/TiO .
2
2
(
XRD), thermogravimetric analysis (TGA) and elemental analysis.
The complete isotherm of the prepared clusters was measured
using nitrogen adsorption–desorption at −196 C, specific surface
2.4. Preparation of polydisperse platinum clusters over TiO2
◦
area SBET, pore volume and average pore diameter were calculated.
Powder X-ray diffraction analysis of protected clusters (unsup-
ported) and supported platinum clusters over TiO2 anatase were
studied. To study the catalytic activity of synthesized cluster, oxi-
dation of styrene is used as a model reaction to demonstrate that.
Polydisperse platinum metal clusters over TiO2 was prepared to
comparison between the catalytic activity of monodisperse clusters
and polydisperse clusters.
(I) A 1 g of TiO anatase was dispersed in 30 ml of water, and then
2
10 mg of H PtCl ·6H O dissolved in 10 ml water was added to the
2
6
2
suspension. Such mixture was stirred one day at room temperature,
◦
filtered, washed, and then dried at 100 C.
(II) 20 ml of 5 mg/ml NaBH4 aqueous solution was added to
100 mg of as-prepared powder (from first step). The mixture was
stirred at room temperature for 60 min, filtered, washed, and then
◦
dried at 100 C overnight, getting nanoparticles, denoted as 1%
Pt/TiO2.
2
. Experimental
2.5. Catalytic oxidation of styrene
2.1. Chemicals
1
00 mg powder from the supported platinum catalysts (1%
Ptn(l-Cys)m/TiO2 and 1% Pt/TiO ) and the bare support material
2
Chloroplatinic acid (H PtCl ·6H O, ≥37.50% Pt basis,
2
6
2
(TiO2 anatase) or 1 mg of unsupported platinum clusters (Ptn(l-
Sigma–Aldrich), Sodium borohydride (NaBH , ≥96%, Aldrich),
4
Cys)m) were mixed with 20 ml styrene solution in toluene (40 mM),
then the oxidizing agent was added in a 50 ml round bottom flask
with vigorous stirring. I used three types of oxidizing agents (1)
l-cysteine (99%, Sigma–Aldrich) were used for synthesizing
the protected platinum nanoclusters. Ethanol (HPLC grade,
Sigma–Aldrich) was used to separate the nanoclusters from the
aqueous mother solution. Styrene (99%, Sigma–Aldrich), hydrogen
0
.5 ml of hydrogen peroxide (2) 7.5 ␮l hydrogen peroxide (as initia-
tor) and pure O2 gas (as oxidizing agent) (3) pure O2 gas (99.99%).
peroxide (H O ) 30 wt.% in H O (Sigma–Aldrich) and oxygen
2
2
2
◦
Then the mixture was heated to 80 C for 12 h. To follow up the
gas were used in the catalytic reaction. Benzaldehyde (99%,
Sigma–Aldrich), acetophenone (98%, Sigma–Aldrich) and styrene
epoxide (97%, Sigma–Aldrich) were used as authentic samples.
conversion of styrene during the stirring time, several cuts were
extracted every one hour, and then the decreasing in the styrene
concentration and the yield of products were detected by gas chro-
matography after removing the catalysts by filtration. I identified
the products by authentic samples and gas chromatography/mass
spectrophotometer (GC/MS).
Titanium dioxide (TiO ) anatase, 99.7% trace metals basis. All
2
chemicals were used as received. All glassware was thoroughly
cleaned with aqua regia (HCl:HNO = 3:1 v/v), rinsed with 2nd
3
distilled water and ethanol, and then dried in an oven prior to use.
2.6. Instrumentation and characterization
2.2. Preparation of Ptn(l-Cys)m nanoclusters
To obtain the UV–vis absorption spectra of the platinum
The 50 mg chloroplatinic acid (H PtCl ·6H O, 96 ␮mol) was
clusters (Ptn(l-Cys)m) and bare ligand (l-cysteine), aqueous solu-
tion of approximately 1–2 mg/ml was prepared. The spectra were
recorded at ambient temperature from 200 to 800 nm with a
double-beam spectrophotometer (Evolution 300). Thermal gravi-
metric analysis (TGA, ∼ 2 mg sample tested) was conducted in a N2
atmosphere (flow rate ∼50 ml /min) with a ThermoStartTM TG/DAT
(Pfeiffer Vacuum). The measurement was performed with a heat-
2
6
2
dissolved in 10 ml 2nd distilled water, 7.8 mg l-cysteine (l-Cys,
6
5 ␮mol) was dissolved in 3 ml 2nd distilled water and added to the
platinum solution, then vigorously stirring (∼1100 rpm) at room
temperature, the yellow solution of platinum salt showed some tur-
bidity after adding the ligand, then the color changed to orange and
finally to canaries yellow color. After 30 min stirring, freshly pre-
pared aqueous solution of NaBH4 (36.3 mg, dissolved in 2 ml 2nd
◦
ing rate of 10 C/min, starting from room temperature and ramping

M. Farrag / Journal of Molecular Catalysis A: Chemical 413 (2016) 67–76
69
◦
Table 1
up to 1000 C. Analysis of C, S, N, and H content was performed by
an elemental analyzer (Vario EL) that allowed controlled combus-
tion of the samples with subsequent chromatographic separation
and the detection of the as-separated species with a TCD detec-
tor. The amount of Pt was analyzed by a fast sequential atomic
absorption spectrometer (ICAP 6200 ICP-OES analyzer—Thermo
Scientific) using a Pt lamp as light source. PtNCs were dissolved
in aqua regia, and the solution was then evaporated completely.
The sample was re-dissolved in 10% HCl. For calibration, Pt solu-
tions of different concentrations were prepared using the standard
matrix. For TEM measurements, solutions with a concentration of
Elementals ratios of Ptn(l-Cys)m clusters were determined by elemental analysis
(
EA) and atomic absorption spectroscopy (AAS).
Elements
C
H
N
S
O
Pt
Percentage
7.8
1.85
3.06
6.89
11.39
69.01
3.1. Chemical composition of protected platinum nanoclusters
(Ptn(l-Cys)m)
The effective ways to obtain the ratio between organic part
and the metal content is elemental analysis and thermogravimet-
ric analysis. With this technique the metal-to-ligand ratio (M/L)
and, hence, the average chemical formula of the protected nan-
oclusters (PNCs) can be determined. TGA curve of protecting ligand
(l-cysteine) shows several decomposition steps (Fig. 1b) [9]. Fig. 1a
shows the thermogram of Ptn(l-Cys)m nanoclusters, the clusters are
1
–2 mg/ml were prepared by dissolving or suspension the catalysts
materials in 2nd distilled water. A droplet of these catalysts solu-
tions was casted onto carbon-coated copper grids. The solvent was
then allowed to evaporate slowly. TEM images were obtained at
a magnification of 300,000, 200,000, 150,000 and 30,000 for 1%
Ptn(l-Cys)m/TiO , Ptn(l-Cys)m clusters, 1% Pt/TiO and TiO anatase
2
2
2
samples, respectively, with a high resolution transmission electron
found to be hygroscopic, which is indicated by the early detected
microscope (HRTEM) TECNAI G2 spirit TWIN at acceleration volt-
mass loss at 29 ◦C. A mass loss at this low temperature is attributed
age of 120 kV, conducted by VELETA Camera. The images were then
analyzed by using Image J software (version 1.44). Powder X-ray
diffraction (XRD) was performed on a philips X-ray powder diffrac-
tometer, model pw 2013/00. Ni-filtered Cu K␣ with a wavelength
of ꢀ = 1.541838 Å was used as a constant source of radiation. The
generator was operated at 35 kV and 20 mA, and diffractometer at
to the removal of adsorbed water molecules rather than to a decom-
position of the ligand, since the pure ligand does not show any mass
loss up to 200 C [9,20]. The other steps of the TGA curve correspond
◦
to the decomposition of the protecting ligand. The decomposition
◦
of the ligand is completed before 1000 C and the residue consists
of platinum metal atoms only (Table S1 and Fig. 1a).
5
0 diverting and receiving slits and a scan rate of 20 mm/min. Fine
From the relative weight of the residue and the total weight loss
of organic ligand, the average metal-to-ligand ratio (M/L) can be cal-
culated and thus the average molecular formula of the synthesized
clusters is derived as Pt3nL2n (Table S1) [9,20,21]. The size selected
gold and silver clusters Au25(SCH CH Ph) [6] and Ag25(L-GSH)18
powder samples were loaded on a quartz plate holder by spreading
the powders as a smooth thin layer on the plate. For all diffrac-
tograms, the following settings were used: scan range 4–80 (2ꢁ),
scan step 0.06 . Infrared spectra of all protected clusters and the
ligands were obtained using an FTIR spectrometer (Nicolet iS10,
resolution: 4 cm , Thermo Scientific corporation, transmission
mode) using KBr disks. Each spectrum was obtained by accumulat-
ing 200 scans. The system which investigates the textural studies of
the samples is of the type NOVA 3000, version 6.10 high-speed gas
sorption analyzer (Quantachrome corporation). Prior to analysis,
◦
◦
2
2
18
[9] showed the same metal to ligand ratio.
−1
TGA analysis shows that the organic weight loss of the plat-
inum nanoclusters is 30.97% (wt.). Therefore, the platinum content
of Ptn(l-Cys)m clusters is calculated as 69.03% (Tables S1). This
composition is further supported by the elemental analysis of the
Ptn(l-Cys)m nanoclusters, which show that the organic content
of these clusters consists of 7.8% C, 1.85% H, 3.06% N, 6.89% S
and 11.39% O (sum 30.99% wt.) (Table 1). In addition, by using
atomic absorption spectroscopy, the percentage of Pt was directly
obtained. It was found to be 69.01% (Table 1), thus confirming com-
position obtained from thermogravimetric analysis [20].
◦
the samples were outgases at 150 C for 2 h. Five points BET surface
areas, total pore volumes, and pore size distribution (BJH method)
were calculated from 24 points nitrogen adsorption–desorption
isotherms. GC analysis was performed employing Thermo Scien-
tific, Trace GC ultra using capillary column TG-5MS (5% Phenyl
Methylpolysiloxane), length 30 m, internal diameter 0.25 mm, and
film thickness 0.25 ␮m. The initial temperature of column is 50 C
and the final temperature is 300 C, the rate of heating is 10 C/min.
The injector temperature is 350 C and the detector temperature
is 350 C, flame ionization detector is the used detector. GC/MS
analysis were carried out by GC model: GC 7890A, using capil-
lary column DB-5 ms (5% phenyl—95% methyl polysiloxane), with
length 30 m, internal diameter 0.25 mm, film thickness 0.5 ␮m, ini-
tial temperature of column 50 C, final temperature 300 C, rate of
heating 10 C/min. and injector temperature 250 C, attached with
mass spectrometer: model 5975B.
◦
3.2. Optical properties of protected platinum manoclusters
(Ptn(l-Cys)m)
◦
◦
◦
◦
The origin of optical properties for protected metal clusters
comes from the excitation of electrons plasmon resonance or
interband transitions, the region and number of absorbance are
characteristic for each metal, particles size, chemical valence and
protected ligand.
The platinum nanoclusters protected by different ligands in
regime 3–20 nm show absorption band at 215 nm [22,23]. How-
ever, the intensity of this band decreases rapidly for particles
smaller than 3 nm, and is essentially undetectable for smaller clus-
ters less than 2 nm [24]. Fig. 2a shows the UV–vis absorption
spectrum of the synthesized platinum clusters (Ptn(l-Cys)m). This
spectrum shows featureless absorption curve in visible regions, in
agreement with its brown color. The capping agent (l-cysteine)
shows one absorption peak at 2 1 0 nm (Fig. 2b). The disappear-
ance of ligand peak in the UV–vis spectrum of protected clusters
◦
◦
◦
◦
3
. Results and discussion
In the following I present an approach for obtaining size
selected platinum clusters functionalized with l-cysteine ligand.
After having successfully synthesized the monolayer-protected
platinum nanoclusters, the sample was characterized by UV–vis
spectroscopy, Fourier transform infrared (FTIR), transmission elec-
tron microscopy (TEM), X-ray diffraction (XRD), thermogravimetric
analysis (TGA) and surface area measurements. The catalytic activ-
ity of the synthesized clusters and supported platinum clusters over
TiO2 anatase was evaluated for oxidation of styrene.
(Fig. 2a), confirm the prepared clusters are completely pure from
the unreacted ligand.
The ligand anchors to the cluster surface through the sulfur atom
[
13,20], this confirmed by Infra-red (IR) analysis (Fig. S1). By com-
paring the IR absorption spectrum of pure ligand with those of the
platinum clusters, the vibrational band of the S H function group at

7
0
M. Farrag / Journal of Molecular Catalysis A: Chemical 413 (2016) 67–76
Fig. 1. The thermogravimetric analysis of synthesized platinum clusters Ptn(l-Cys)m (curve a) and protecting ligand l-cysteine (curve b).
Fig. 2. UV–vis absorption spectrum of synthesized platinum clusters Ptn(l-Cys)m (curve a) and protecting ligand l-cysteine (curve b). Ptn(l-Cys)m clusters shows featureless
absorption curves in UV and visible region 200–800 nm and l-cysteine ligand shows one absorption peak at 2 1 0 nm.
−
1
2
535–2564 cm in the protected clusters disappears (Fig. 1S), con-
(TEM) (Fig. 3) and powder X-ray diffraction (XRD) analysis (Fig. 4).
TEM is particularly suited for obtaining the sizes of particles larger
than 1 nm. It is found that, the particles of Ptn(l-Cys)m clusters are
uniformly distributed and have a diameter of about 1 nm (Fig. 3-
I). This is a clear indication that particles with a very narrow size
distribution can be obtained with this synthesis method. According
to the TEM image (Fig. 3-I) I expect, the synthesized nanoparticles
are monodisperse clusters, especially this image is similar to Au25
clusters protected by 2-phenylethanethiol [6] and l-glutathione
firming the anchoring of ligand to the cluster surface through the
sulfur atom [13,20,25]. However, the rest of ligand bands appeared
in the same position with the protected clusters, but with low inten-
sity. This confirms the chemical structure of the capping ligands did
not change in protected clusters.
3
.3. Particles size of synthesized platinum clusters.
[
9]. However, the particle size distribution of Ptn(l-Cys)m clusters
To determine the particle size of the synthesized catalysts Ptn(l-
was measured using image J program. Around 70 particles have
size 1 nm and the rest 33 particles have size 1 ± 0.1 nm (Fig. 3-I).
Cys)m, 1% Ptn(l-Cys)m/TiO , 1% Pt/TiO and bare TiO2 anatase, two
2
2
analytical methods were applied transmission electron microscopy

M. Farrag / Journal of Molecular Catalysis A: Chemical 413 (2016) 67–76
71
Fig. 3. Transmission electron microscopy (TEM) images (I–IV) of synthesized platinum clusters Ptn(l-Cys)m (∼1 nm), 1% Ptn(l-Cys)m/TiO2, non-protected platinum clusters
over TiO2 (1% Pt/TiO2, 2 ± 0.5 nm) and TiO2 anatase (av. 25 nm), respectively.
◦
Fig. 4. (I) X-ray diffractograms of Ptn(l-Cys)m clusters (curve a), which exhibits a broad and intense (1 1 1) peak of platinum at 2ꢁ = ∼40 and the stick pattern (curve b) is an
indication of the diffraction peaks of the bulk Pt metal, which shows face-centered cubic (fcc) lattice (1 1 1), (2 0 0), (2 2 0), and (3 1 1) reflections. (II) X-ray diffractograms of
1
% Ptn(l-Cys)m/TiO2, 1% Pt/TiO2 and TiO2 catalysts (curves a–c), respectively.
◦
TEM analysis was used to confirm the Ptn(l-Cys)m clusters after
intense feature at 2ꢁ = 40 ± 0.5 , which is centered at almost the
deposition over TiO anatase still monodisperse and do not agglom-
same position as the peak of a bulk platinum (1 1 1) lattice (Fig. 4-
Ib) [26,27]. Fig. 4-Ib shows the diffraction pattern of bulk platinum
metal (stick spectrum), which shows face-centered cubic (fcc) lat-
tice 40 (1 1 1), 46.5 (2 0 0), 67.5 (2 2 0) and 81 (3 1 1) reflections
[26,27].
The particles size of the synthesized platinum nanoclusters
Ptn(l-Cys)m can be estimated according to the broadening of the
full width at half-maximum (fwhm) of the (1 1 1) diffraction peak
by the Debye–Scherrer equation [9]. The calculated particles size of
Ptn(l-Cys)m clusters from XRD analysis is 0.9 ± 0.05 nm, which in
good agreement with TEM analysis. According to the Bragg equa-
2
erate (Fig. 3-II). However, reduction of deposited platinum salt over
TiO2 by sodium borohydride in absence of l-cysteine ligand pro-
duces polydisperse platinum clusters in regime 2 ± 0.5 nm as seen
in image Fig. 3-III., this refer to the role of protecting ligands in pro-
duce monodisperse clusters. Fig. 3-IV shows the particles size of
the bare TiO2 anatase in average 25 nm.
Powder X-ray diffraction (XRD) was used to compare the
structures of the synthesized platinum nanoclusters protected by
l-cysteine with that of bulk platinum (Fig. 4-I). The X-ray diffrac-
togram of Ptn(l-Cys)m powder (Fig. 4-Ia) exhibits a broad and
◦
◦
◦
◦

7
2
M. Farrag / Journal of Molecular Catalysis A: Chemical 413 (2016) 67–76
tion, this broad and intense peak indicates an average interplane
spacing (d) of 2.255 ± 0.025 Å, which corresponds very well to the
interplane spacing of bulk Pt (1 1 1) 2.25 Å.
Table 2 gives the textural data obtained through the analysis of
N2 sorption data of all prepared catalysts. The values of SBET and St
for all investigated clusters are close to each other which justify the
correct choice of standard t-curves for pore analysis and indicate
the absence of ultra-micropores in this adsorbents [31,32].
Fig. 4-II illustrates XRD diffractograms of 1% Ptn(l-Cys)m/TiO2,
1
% Pt/TiO2 and TiO2 anatase catalysts. The characteristic diffrac-
tions peaks of TiO2 anatase are (1 0 1), (0 0 4), (2 0 0), (1 0 5), (2 0 4),
◦
◦
◦
(
5
1 1 6), (2 2 0) and (2 1 5) appear at 2ꢁ equal to 25.35 , 37.78 , 47.5 ,
3.5. Catalytic activity of synthesized platinum clusters
◦ ◦ ◦ ◦ ◦
3.92 , 62.72 , 68.99 , 74.49 , 82.21 , respectively (Fig. 4-IIc) [28].
Doping TiO2 by Ptn(l-Cys)m clusters and polydisperse platinum
clusters do not modify its pattern and no other peaks related to
the presence of Pt are detected (Fig. 4-IIa and b, respectively). The
only different between the diffractogram of bare TiO2 and plat-
inum doped TiO2 is relatively wider and lower intense peaks with
doped catalysts. This indicates a little distortion in the crystalline
structure of the bare TiO2 was happened, because the location of
platinum clusters within the TiO2 lattice is not possible owing to
the mismatch of ionic radius between TiO2 and platinum [29]. The
crystalline size (dTiO2) of bare TiO2 anatase is calculated by using
Debye–Scherrer equation as 28 nm for the broadening of (1 0 1)
peak reflection. The dTiO2 values in 1% Ptn(l-Cys)m/TiO2 and 1%
Pt/TiO2 catalysts are less than bare TiO2 16 and 17.2 nm, respec-
tively.
Herein, I chose selective oxidation of styrene by using hydro-
gen peroxide and/or oxygen gas as oxidizing agents to test the
utility of platinum nanoclusters as oxidation catalysts. The process
of selective oxidation of styrene is of industrial importance, since
the products of styrene oxidation (benzaldehyde, styrene epoxide
and acetophenone) are high-value chemicals widely used in fine
chemical industry (Scheme 1).
To elucidate the effects of particle sizes of platinum nanoclusters
and the effects of supports, monodisperse and polydisperse plat-
inum nanoclusters were prepared and supported over TiO anatase.
2
I chose three oxidant systems to oxidize styrene: (1) hydrogen per-
oxide (H O ) as the oxidant, (2) H O as an initiator and O2 as the
2
2
2
2
main oxidant, and (3) O2 as the oxidant. The temperature of this
◦
catalytic reaction is 80 C, below the thiolate-desorption onset tem-
perature of Ptn(l-Cys)m clusters (Fig. 1), I believe that the thiolate
ligands should remain on the protected clusters during the course
of the catalytic reaction. The catalytic reaction occurs at the surface
of metal clusters, whereas the molecules of styrene is small and can
penetrate the thiolate ligand shell of Ptn(l-Cys)m and reach to the
cluster surface [33].
3
.4. Texture analysis of synthesized platinum clusters
The surface texture properties of Ptn(l-Cys)m, 1% Ptn(l-
Cys)m/TiO2 and 1% Pt/TiO2 catalysts, as well as the bare
TiO2 anatase for comparison were studied by measuring the
adsorption–desorption isotherms (Fig. 5).
Firstly, I studied the effect of catalyst size and the effect of
support in the catalytic oxidation of styrene by using one type of
oxidizing agent (H O ). The catalytic activity of platinum clusters
The adsorption–desorption isotherms of bare TiO2 anatase and
% Pt/TiO2 show large hysteresis loops characterizing mesoporous
1
2
2
nature of H1 type (Fig. 5-IIa and b, respectively). However, unsup-
ported clusters (Ptn(l-Cys)m) and the doped TiO2 with this clusters
catalysts exhibits a strong dependence on size. The smaller plat-
inum clusters (Ptn(l-Cys)m, ∼1 nm) shows a much higher catalytic
activity than the polydisperse clusters 2 ± 0.5 nm (Fig. 6-I). More-
over, supporting of monodisperse clusters (Ptn(l-Cys)m) over TiO2
(
1% Ptn(l-Cys)m/TiO ) show sorption isotherm from type H4 and
2
H3 (Fig. 5-Ia and b, respectively) according to IUPAC classification
of hysteresis loops [30].
(1% Ptn(l-Cys)m/TiO ) leads to large increase in catalytic activity
2
The specific surface area and pore volume distribution of all cat-
alysts are summarized in Table 2. As I mentioned before in XRD
analysis, the doped metal creates defect sites and surface disor-
of styrene compared to 1% Pt/TiO2 and unsupported Ptn(l-Cys)m
catalysts (Fig. 6-I). This increasing in activity attributed to the elec-
tronic interaction between TiO and Ptn(l-Cys)m and increase of the
2
der in the support (TiO ), therefore the total pore volume of doped
exposed area from clusters to the reaction after deposition over the
support (TiO ) [14,15]. However, the bare TiO does not catalyze the
2
alumina (1% Pt/TiO ) is more than the bare TiO [29]. In concor-
2
2
2
2
dance, the doped titania with Ptn(l-Cys)m clusters shows decrease
in pore volume, since the particles size of these clusters are very
small (∼1 nm), therefore localized deep inside the pores affecting
the pore volumes that by its turn decrease the surface area values
styrene oxidation reaction under comparable conditions (Fig. 6-I).
The good linear correlation of (Ln Co/C) versus stirring time plot is
the proof for the pseudo-first order rate kinetics (Fig. 6-II). To deter-
mine the role of support (TiO ) in catalytic oxidation of styrene, the
2
(
Table 2). Therefore, This explain why the hysteresis loop of non-
activation energy (Ea) was calculated. The values of rate constants
measured at 70–100 C over 1% Ptn(l-Cys)m/TiO2 and Ptn(l-Cys)m
◦
doped TiO changed from H1 into H3 type by doping with protected
2
platinum clusters and keeps H1 type when doped by non-protected
catalysts have been used to calculate the activation energy values
by applying Arrhenius equation (Fig. 6-III). The activation energy
platinum clusters (1% Pt/TiO ).
2
The pore size distributions of all catalysts were determined by
the BJH method (Fig. 5-III). The results showed that the pore size
of TiO2 anatase was uniform in two domains maximizing at 185.4
and 22.7 Å; as a broad and sharp bands, respectively (Fig. 5-IIIc),
of styrene oxidation using H O2 as oxidizing agent over 1% Ptn(l-
2
Cys)m/TiO2 and Ptn(l-Cys)m catalysts are 57.16 and 70.15 KJ/mol,
respectively. From the activation energy values the support (TiO2)
reduces the required energy to activate this reaction. Thus I can
say the role of support is not only better disperse of platinum clus-
ters but also reduce activation energy of styrene oxidation reaction
by the electronic interaction between TiO2 and Ptn(l-Cys)m clus-
ters. This result is in good agreement with the previous published
work [4,14,15,34–36], in selective hydrogenation of benzalacetone
and oxidation of styrene over Au25/TiO , Au /Fe O , Au25/HAP and
also the doped catalyst (1% Pt/TiO ) showed two maximum peaks
2
at 235.5 and 20.1 Å (Fig. 5-IIId), whereas the Ptn(l-Cys)m and 1%
Ptn(l-Cys)m/TiO2 catalysts showed a one sharp maximum peak at
∼
27 Å (Fig. 5-IIIa and b, respectively). Fig. 5-III shows the 1% Pt/TiO2
and bare TiO catalysts have mesoporous and small number from
2
micropores, however, Ptn(l-Cys)m and 1% Ptn(l-Cys)m/TiO2 cata-
lysts have only micropores. The Va–t plots of Ptn(l-Cys)m and 1%
2
25
2
3
Au25/SiO2 catalysts [4,14,15]. These metal oxide supports (TiO2,
HAP and Fe O ) have a strong electron-donating tendency [34,35].
Ptn(l-Cys)m/TiO catalysts exhibited a downward deviation (Fig. 5-
2
2
3
IVa and b), this clearly suggests the presence microporous nature
in these two catalysts. However, 1% Pt/TiO and bare TiO catalysts
exhibited upward deviation (Fig. 5-IVc and d) confirming the pres-
ence of mesoporous. Therefore, I can say the Va–t calculation in good
agreement with the pore size calculation by the BJH method.
The electron transfer from the metal oxide support to gold pro-
motes H2 activation [36]. In contrast, the inert SiO2 support has no
electron-donating capability, hence, it shows no enhancement in
catalytic activity. On the other hand, the activation energy of water-
gas shift (WGS) reaction is stable for different supported platinum
2
2

M. Farrag / Journal of Molecular Catalysis A: Chemical 413 (2016) 67–76
73
Fig. 5. (I) Nitrogen adsorption–desorption isotherms of synthesized platinum clusters Ptn(l-Cys)m (curve a) and the supported protected clusters 1% Ptn(l-Cys)m/TiO2, the
two isotherms show sorption isotherm from type H2 and H3, respectively. (II) Nitrogen adsorption–desorption isotherms of bare TiO2 (curve a) and 1% Pt/TiO2 catalysts
(
curve b), the two isotherms show sorption isotherm from type H1. (III) Pore volume distribution curves of all prepared catalysts. (IV) The Va–t plots curves of all prepared
catalysts.
Table 2
Texture data obtained from the analysis of nitrogen sorption isotherms of as-prepared clusters catalysts.
SBET(m2
g
−1
)
St(m2
g
−1
)
Total pore volume(cc g × 10−3
−1
)
Average pore diameter (Å)
No.
Catalysts
1
2
3
4
1% Ptn(l-Cys)m/TiO2anatase
Ptn(l-Cys)m
1% Pt/TiO2 anatase
TiO2 anatase
41.07
5.23
42.41
45.82
41.07
5.24
42.41
45.83
59.36
6.67
218.24
182.76
58.26
95.59
205.83
159.54
Scheme 1. Catalytic oxidation of styrene over platinum clusters.

7
4
M. Farrag / Journal of Molecular Catalysis A: Chemical 413 (2016) 67–76
Fig. 6. (I) The change in concentration of styrene with stirring time during the oxidation reaction by H2O2 over all prepared catalysts. (II) Ln Co/C versus stirring time plots,
◦
which proof the styrene oxidation reaction, is pseudo-first order. (III) Arrhenius plots, obtained for oxidation of styrene carried out at 70–100 C over 1% Ptn(l-Cys)m/TiO2
and Ptn(l-Cys)m catalysts.
activity in comparison to another two oxidizing systems (Fig. 7). 1%
Ptn(l-Cys)m/TiO catalyst shows 92.5% conversion by using H O as
2
2
2
the oxidant, however by system 2 (H O2 initiator/O2 oxidant) and
2
system 3 (O2 as oxidant) shows 29% and 23% conversion, respec-
tively. A similar order of conversion was found for Ptn(l-Cys)m, 1%
Pt/TiO2 and TiO2 catalysts in the three different oxidant systems
(Fig. 7). Therefore, the constituent of the oxidant plays a critical
role in the catalytic activity of styrene oxidation.
Benzaldehyde is the major product yield in oxidation of styrene
by using the synthesized catalysts whatever the oxidizing agents
(
1
Fig. 8). Moreover, the selectivity of benzaldehyde can reach to
00% by using the supported monodisperse clusters (1% Ptn(l-
Cys)m/TiO ) catalyst and H O as oxidant or mixture from H O2
2
2
2
2
(initiator) and O2 (oxidant) (Fig. 8-I). When oxygen is used as
the main oxidant in system 2 and 3 the activity and selectivity
of styrene oxidation decrease, which indicates that the oxidizing
agent play a big role in oxidation of styrene (Figs. 7 and 8) [15].
To check the advantage of 1% Ptn(l-Cys)m/TiO2 catalyst and its
applicability to reuse [13], the oxidation reaction of styrene was
achieved with 100 mg 1% Ptn(l-Cys)m/TiO2 catalyst and 0.5 ml of
hydrogen peroxide as oxidizing agent. Then the catalyst was col-
lected at the end of the reaction and reused for a second cycle and
the process repeated so on till four cycles keeping all other param-
eters constant (Fig. 9). The results revealed that catalyst shows a
very good activity for four catalytic runs with a very small loss in the
Fig. 7. The catalytic conversion of styrene over all prepared catalysts (1–4) by
using different oxidizing agents after 12 h stirring at 80 C, (1–4) refer to 1% Ptn(l-
Cys)m/TiO2, Ptn(l-Cys)m, 1% Pt/TiO2 and TiO2 catalysts, respectively.
◦
catalysts (TiO , L-zeolites and mesoporous silica MCM-41) [37]. The
2
apparent activation energy of WGS reaction over Au/TiO2 catalyst
styrene conversion. It can be conclude that the 1% Pt (l-Cys)m/TiO2
n
[
38] is similar to those reported for gold on ceria, iron oxide, lan-
catalyst possesses high stability and it may be reusable for at least
thana and other supports [39–43], this confirm the support oxide
itself does not participate directly in the WGS reaction.
Fig. 7 shows the relationship between the conversion of styrene
and the used oxidizing systems after 12 h stirring. It is clearly
4
runs, showing a good potential for practical applications (Fig. 9).
The distinct size effect observed for 1% Ptn(l-Cys)m/TiO2 cat-
alyst in selective oxidation of styrene is interesting. It should be
noted that the observed size dependency is not a simple conse-
observed that system 1 (i.e., H O2 as the oxidant) gives the highest
2
◦
Fig. 8. (I–III) The catalytic selectivity of styrene oxidation products by using different oxidizing agents after 12 h stirring at 80 C over the prepared catalysts 1% Ptn(l-
Cys)m/TiO2, Ptn(l-Cys)m and 1% Pt/TiO2 catalysts, respectively.

M. Farrag / Journal of Molecular Catalysis A: Chemical 413 (2016) 67–76
75
showed extremely catalytic activity in oxidation of styrene with
00% benzaldehyde selectivity. Polydisperse platinum metal clus-
1
ters over TiO2 was prepared to comparison between the catalytic
activity of monodisperse clusters and polydisperse clusters. The
core/shell nature of the protected platinum clusters, i.e., the pres-
ence of a surface layer of platinum atoms bearing partial positive
charges and the electron-rich core, is believed to be primarily
responsible for their catalytic activities. H O2 as oxidizing agent
2
can achieve the oxidation of styrene even with monodisperse and
polydisperse platinum clusters by high conversion and selectivity
more than oxygen gas. The role of support is not only better dis-
perse of platinum clusters but also reduce the activation energy of
styrene oxidation reaction
Acknowledgments
This work is supported by Assiut University, Egypt. I kindly
acknowledge Prof. Dr. H. Gasteiger for carrying out TGA measure-
ment in technical university of Munich (TUM), Germany.
Fig. 9. Repeated cycles of oxidation reaction up to 4 times over 1% Ptn(l-Cys)m/TiO2
catalyst by using H2O2 as oxidizing agent.
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.molcata.2015.12.
quence of the higher surface area of smaller nanoclusters. I can
attribute the extraordinary activity of monodisperse clusters to
their electronic properties, which induce HOMU-LUMO gap in their
electronic structures [4,44].
0
11.
References
Many researchers confirmed that the supported gold catalysts
show high activity in epoxidation reactions, this activity attributed
to the cationic gold species located at the interface between gold
nanoparticles and the reducible supports [45–47]. Density func-
tional theory (DFT) calculations have shown distinct differences in
charge distribution on the Au13 core (electron-rich) and the Au12
shell (electron-deficient) for Au25(SR)18 clusters [48–51]. I expect
the electronic structures of the synthesized nanoclusters (Ptn(l-
Cys)m) like Au25, moreover, the electronegativity of sulfur atom
in ligand is higher than the platinum atom in the surface of clus-
ters (S-Pt). Therefore, I can say there are partial positive charge
in the surface platinum atoms and partial negative charge in the
sulfur atom. The presence of partial positive charges on the sur-
face Pt atoms in Ptn(l-Cys)m clusters should facilitate activation of
[
[
[
1] M. Haruta, Nature 437 (2005) 1098.
2] A.S.K. Hashmi, G.J. Hutchings, Angew. Chem. Int. Ed. 45 (2006) 7896.
3] J. Luo, V.W. Jones, M.M. Maye, L. Han, N.N. Kariuki, C.-J. Zhong, J. Am. Chem.
Soc. 124 (2002) 13988.
4] B.Y. Zhu, H. Qian, M. Zhu, R. Jin, Adv. Mater. 22 (2010) 1915–1920.
5] S. Kunz, K. Hartl, M. Nesselberger, F. Schweinberger, G. Kwon, M. Hanzlik, K.
Mayrhofer, U. Heiz, M. Arenz, Phys. Chem. Chem. Phys. 12 (2010)
10288–10291.
[
[
[
[
6] Z. Wu, J. Suhan, R. Jin, J. Mater. Chem. 19 (2009) 622–626.
7] H. Qian, Y. Zhu, R. Jin, Nano 3 (2009) 3795–3803.
[8] H. Qian, R. Jin, Nano Lett. 9 (2009) 4083–4087.
[9] M. Farrag, M. Tschur, U. Heiz, Chem. Mater. 25 (2013) 862.
10] M. Farrag, M. Tschurl, A. Dass, U. Heiz, Phys. Chem. Chem. Phys. 15 (2013)
[
[
12539.
11] S. Kunz, P. Schreiber, M. Ludwig, M.M. Maturi, O. Ackermann, M. Tschurl, U.
Heiz, Phys. Chem. Chem. Phys. 15 (2013) 19253.
[12] I. Schrader, J. Warneke, J. Backenköhler, S. Kunz, J. Am. Chem. Soc. 137 (2015)
05–912.
13] M. Farrag, J. Photochem. Photobiol. A: Chem. 318 (2016) 42–50.
14] Y. Zhu, H. Qian, B.A. Drake, R. Jin, Angew. Chem. Int. Ed. 49 (2010) 1295–1298.
[15] Y. Zhu, H. Qian, R. Jin, Chem. Eur. J. 16 (2010) 11455–11462.
16] S. Lubis, L. Yuliati, S. Ling Lee, I. Sumpono, H. Nur, Chem. Eng. J. 209 (2012)
86–493.
17] O.S. Mohamed, S.A. Ahmed, M.F. Mostafa, A.A. Abdel-Wahab, Int. J.
Photoenergy 2008 (2007) 1–11, Article ID 205358.
18] Y. Liu, H. Tsunoyama, T. Akita, T. Tsukuda, Chem. Commun. 46 (2010)
550–552.
9
the nucleophilic group ( CH CH ) in styrene since the positive Pt
2
[
[
atoms are electrophilic. In contrast, the electron-rich platinum core
should facilitate O2 activation.
[
[
[
4
4
. Conclusion
In this work I present a new synthesis method to prepare
monodisperse platinum clusters protected by l-cysteine. This
method can produce a very narrow size distribution platinum
clusters ∼1 nm. The optical properties of the prepared clusters
were studied by UV–vis spectroscopy. The platinum nanoclusters
anchored to the ligand through the thiol group, which confirmed
[19] M. Turner, V.B. Golovko, O.P.H. Vaughan, P. Abdulkin, A. Berenguer-Marcia,
M.S. Tikhov, B.F.G. Johnson, R.M. Lambert, Nature 454 (2008) 981.
[
20] M. Farrag, M. Thämer, M. Tschurl, T. Bürgi, U. Heiz, J. Phys. Chem. C 116 (2012)
034.
[21] B.Z. Wu, J. Chen, R. Jin, Adv. Funct. Mater. 21 (2011) 177.
8
[
[
[
22] J.A. Creighton, D.G. Eadon, J. Chem. Soc. Faraday Trans. 87 (1991) 3881–3891.
23] S. Chen, K. Kimura, J. Phys. Chem. B 105 (2001) 5397–5403.
24] M.J. Hostetler, J.E. Wingate, C.-J. Zhong, J.E. Harris, R.W. Vachet, M.R. Clark, J.D.
Londono, S.J. Green, J.J. Stokes, G.D. Wignall, G.L. Glish, M.D. Porter, N.D.
Evans, R.W. Murray, Langmuir 14 (1998) 17–30.
−
1
by disappearance of S H vibrational band at 2535–2570 cm in
IR spectra of clusters. The metal-to-ligand ratio (M/L) and the aver-
age chemical formula of the prepared clusters were calculated from
thermogravimetric analysis (TGA) and elemental analysis (EA). The
particle sizes of all synthesized catalysts were assessed by trans-
mission electron microscopy (TEM) and powder X-ray diffraction
analysis (XRD. The complete isotherm of the prepared clusters was
[
25] E.S. Shibu, M.A. Habeeb Muhammed, T. Tsukuda, T. Pradeep, J. Phys. Chem. C
112 (2008) 12168.
[26] R. Sharma, L.-S. Bouchard, Sci. Rep. 2 (2012) 277.
[
[
[
27] T. Hyde, Platinum Met. Rev. 52 (2008) 129–130.
28] A. Zielinska-Jurek, A. Zaleska, Catal. Today 230 (2014) 104–111.
29] A. Neren Ökte, Appl. Catal. A: Gen. 475 (2014) 27–39.
◦
measured using nitrogen adsorption–desorption at –196 C, spe-
[30] I.U.P.A.C. Recommendations, Pure Appl. Chem. 66 (1994) 1739.
[
31] M.T.h. Makhlouf, B.M. Abu-Zied, T.H. Mansoure, Appl. Surf. Sci. 274 (2013)
5–52.
cific surface area SBET, pore volume and average pore diameter were
calculated. To study the catalytic activity of synthesized clusters,
oxidation of styrene is used as a model reaction to demonstrate that.
Supported monodisperse platinum clusters (1% Ptn(l-Cys)m/TiO2)
4
[
32] W.M. Shaheen, A.A. Zahran, G.A. El-Shobaky, Colloids Surf. A: Physicochem.
Eng. Asp. 231 (2003) 51–65.
[33] Y. Zhu, H. Qian, B.A. Drake, R. Jin, Angew. Chem. 122 (2010) 1317–1320.

7
6
M. Farrag / Journal of Molecular Catalysis A: Chemical 413 (2016) 67–76
[
[
[
34] M. Boronat, F. Illas, A. Corma, J. Phys. Chem. A. 113 (2009) 3750–3757.
35] P. Winiarek, J. Kijenski, J. Chem. Soc. Faraday Trans. 94 (1998) 167–172.
36] P. Clause, A. Bruckner, C. Mohr, H. Hofmeister, J. Am. Chem. Soc. 122 (2000)
[44] M. Zhu, C.M. Aikens, F.J. Hollander, G.C. Schatz, R. Jin, J. Am. Chem. Soc. 130
(2008) 5883.
[45] N.S. Patil, B.S. Uphade, P. Jana, R.S. Sonawane, S.K. Bhargava, V.K. Choudhary,
Catal. Lett. 94 (2004) 89.
1
1430–11439.
[
[
37] M. Yang, J. Liu, S. Lee, B. Zugic, J. Huang, L.F. Allard, M.
Flytzani-Stephanopoulos, J. Am. Chem. Soc. 137 (2015) 3470–3473.
38] M. Yang, L.F. Allard, M. Flytzani-Stephanopoulos, J. Am. Chem. Soc. 135 (2013)
[46] E.E. Stangland, K.B. Stavens, R.P. Andres, W.N. Delgass, J. Catal. 191 (2000) 332.
[47] A. Abad, P. Concepcion, A. Corma, H. Garcia, Angew. Chem. Int. Ed. 44 (2005)
4066.
3
768–3771.
[48] J. Akola, M. Walter, R.L. Whetten, H. Hakkinen, H. Gronbeck, J. Am. Chem. Soc.
130 (2008) 3756.
[49] M. Walter, J. Akola, O. Lopez-Acevedo, P.D. Jadzinsky, G. Calero, C.J. Ackerson,
R.L. Whetten, H. Grönbeck, H. Hákkinen, Proc. Natl. Acad. Sci. U. S. A. 105
(2008) 9157.
[50] M. Zhu, C.M. Aikens, M.P. Hendrich, R. Gupta, H. Qian, G.C. Schatz, R. Jin, J. Am.
Chem. Soc. 131 (2009) 2490.
[51] C.M. Aikens, J. Phys. Chem. C 112 (2008) 19797.
[
[
39] Q. Fu, H. Saltsburg, M. Flytzani-Stephanopoulos, Science 301 (2003) 935.
40] Q. Fu, W. Deng, H. Saltsburg, M. Flytzani-Stephanopoulos, Appl. Catal. B 56
(
2005) 57.
[
[
[
41] M.B. Boucher, S. Goergen, N. Yi, M. Flytzani-Stephanopoulos, Phys. Chem.
Chem. Phys. 13 (2011) 2517.
42] W.D. Williams, M. Shekhar, W.S. Lee, V. Kispersky, W.N. Delgass, F.H. Ribeiro,
S.M. Kim, E.A. Stach, J.T. Miller, L.F. Allard, J. Am. Chem. Soc. 132 (2010) 14018.
43] J.D. Lessard, I. Valsamakis, M. Flytzani-Stephanopoulos, Chem. Commun. 48
(
2012) 4857.