Influence of support on the aerobic oxidation of HMF into FDCA over preformed Pd nanoparticle based materials

Article abstract of DOI:10.1016/j.apcata.2014.03.020

Here, the preparation and evaluation of supported nanoparticle based catalytic material is reported. Polyvinylpyrrolidone (PVP) stabilized palladium nanoparticles with a mean particle size of 1.8 nm were synthesized in ethylene glycol and subsequently deposited onto different metal oxide supports (TiO ₂, γ-Al₂O₃, KF/Al₂O ₃, and ZrO₂/La₂O₃). The prepared catalysts were applied to the aerobic oxidation of 5-hydroxymethylfurfural (HMF) to 2,5-furandicarboxylic acid (FDCA) in aqueous solution at atmospheric pressure (T = 90 °C, pO₂ = 1 bar) and compared regarding their catalytic performance and stability. The highest FDCA yield (>90%) was obtained for the Pd/ZrO₂/La₂O₃ catalysts which additionally showed a relatively stable catalytic performance when the material was reused. Various characterization methods including XRD, TEM, XPS, and AAS were applied to obtain information about the Pd NP before and after utilization in HMF oxidation. For the Pd/TiO₂ the least changes in Pd NP structure were observed after using the material in HMF oxidation. This was attributed to a stronger interaction between the Pd NP and the TiO₂ support compared to other supports used in the studies.

Full text of DOI:10.1016/j.apcata.2014.03.020

Applied Catalysis A: General 478 (2014) 107–116
Contents lists available at ScienceDirect
Applied Catalysis A: General
journal homepage: www.elsevier.com/locate/apcata
Influence of support on the aerobic oxidation of HMF into FDCA over
preformed Pd nanoparticle based materials
Baraa Siyoa,b, Matthias Schneidera, Jörg Radnika, Marga-Martina Pohla, Peter Langera,b
,
Norbert Steinfeldta,∗
a Leibniz-Institut für Katalyse e.V., Albert-Einstein-Str. 29a, 18059 Rostock, Germany
b Institut für Chemie, Universität Rostock, Albert-Einstein-Str. 3a, 18059 Rostock, Germany
a r t i c l e i n f o
a b s t r a c t
Article history:
Here, the preparation and evaluation of supported nanoparticle based catalytic material is reported.
Polyvinylpyrrolidone (PVP) stabilized palladium nanoparticles with a mean particle size of 1.8 nm were
synthesized in ethylene glycol and subsequently deposited onto different metal oxide supports (TiO2,
␥-Al2O3, KF/Al2O3, and ZrO2/La2O3). The prepared catalysts were applied to the aerobic oxidation of 5-
hydroxymethylfurfural (HMF) to 2,5-furandicarboxylic acid (FDCA) in aqueous solution at atmospheric
pressure (T = 90 ◦C, pO2 = 1 bar) and compared regarding their catalytic performance and stability. The
highest FDCA yield (>90%) was obtained for the Pd/ZrO2/La2O3 catalysts which additionally showed a
relatively stable catalytic performance when the material was reused. Various characterization methods
including XRD, TEM, XPS, and AAS were applied to obtain information about the Pd NP before and after
utilization in HMF oxidation. For the Pd/TiO2 the least changes in Pd NP structure were observed after
using the material in HMF oxidation. This was attributed to a stronger interaction between the Pd NP and
the TiO2 support compared to other supports used in the studies.
Received 20 December 2013
Received in revised form 21 February 2014
Accepted 18 March 2014
Available online 25 March 2014
Keywords:
Pd
Nanoparticles
Supported catalysts
5-Hydroxymethylfurfural
2 ;5-Furandicarboxylic acid
Oxidation
© 2014 Elsevier B.V. All rights reserved.
1. Introduction
can e.g. increase the stability of the NP against coarsening, alter the
structure and shape of NP, encapsulate the active NP at high tem-
Metallic nanoparticles dispersed onto oxide supports have
been used as catalysts for the synthesis of chemicals, pharma-
ceuticals and for energy applications [1]. Mostly, conventional
salt-impregnation methods were used for catalyst preparation,
where the metal particles are formed in the presence of the sup-
port. In this case metal particle size, shape and dispersion of the
catalytically active compound are influenced by a large number of
parameters, e.g. the support and can therefore vary strongly [2].
In the last two decades colloid chemistry based methods
have been developed to synthesize differently sized metallic NP
with narrow size distribution [3–6]. Deposition of the pre-formed
metal nanoparticles onto carrier materials influences the adjusted
particle size distribution in many cases only moderately [7–9].
Therefore, catalytic materials synthesized by colloidal deposition
methods might be considered as an interesting alternative for cat-
alyst preparation [10]. Furthermore, this method can be also used
to study the effect of the support on catalytic activity and selectivity
isolated from other factors of catalyst preparation [11]. The support
perature, transfer charge from/to the NP, supply additional reaction
sites (e.g. oxygen vacancies), and stabilize intermediate reaction
species [12].
5-Hydroxymethylfurfural (HMF), is considered as a platform
chemical and, after its reduction or oxidation, as a potential start-
ing material for several industrial applications [13]. It can be
synthesized via acid catalyzed dehydration of carbohydrates (e.g.
fructose or glucose) [14–17]. 2,5-Furandicarboxylic acid (FDCA)
which can be formed by oxidation of HMF, is considered as one
of the twelve most potentially useful biomass derived chemicals
[18], and might be used as a renewable intermediate for polymers,
fine chemicals, pharmaceuticals and agrochemicals [19]. In this
reaction, 5-hydroxymethyl-furan-2-carboxylic acid (HMFCA) and
5-formyl-furan-2-carboxylic acid (FFCA) are formed as intermedi-
ates (Scheme 1). The main difficulties in this reaction originate from
the fact that a base has to be present in the reaction mixture, while
the substrate (HMF) is unstable in alkaline aqueous solution [20].
The Influence of the support on the oxidation of HMF to FDCA
was investigated with Au [21,22], Pt [23,24] and ruthenium hydrox-
ide based catalysts [25]. Pd based materials have already been
successfully applied to the aerobic oxidation of alcohols to carbonyl
groups, both in an organic and aqueous reaction media [26–31].
∗
Corresponding author. Tel.: +49 0 381 1281 319; fax: +49 0 381 1281 51319.
E-mail address: norbert.steinfeldt@catalysis.de (N. Steinfeldt).
http://dx.doi.org/10.1016/j.apcata.2014.03.020
0926-860X/© 2014 Elsevier B.V. All rights reserved.

108
B. Siyo et al. / Applied Catalysis A: General 478 (2014) 107–116
Scheme 1. Formal reaction scheme of the aerobic oxidation of HMF with molecular oxygen in water.
However, studies on supported Pd based materials for the HMF
oxidation were explored merely using Pd/C as catalyst [32].
Previously, it was shown that the catalytic performance of PVP
protected Pd NP in HMF oxidation depends on the particle size [33].
Small Pd NP with a mean diameter of 1.8 nm were more active and
selective than larger ones. However, stabilization of these unsup-
ported Pd NP by PVP in alkaline solution decreased during the
reaction. Deposition of the Pd NP on solid surfaces can be consid-
ered as an option to obtain materials with higher catalytic stability.
Here, the most active unsupported Pd NP species with a mean diam-
eter of 1.8 nm delivering the highest yield of FDCA was chosen
for the deposition on reducible (TiO2) and passive (ZrO2/La2O3, ␥-
Al2O3, KF/Al2O3) support materials. The aim of the presented work
is to elucidate the role of metal oxide support on the catalytic per-
formance and stability of Pd NP based materials in HMF oxidation.
Therefore, particle size and electronic structure of the NP both after
NP deposition on the respective support and after using the mate-
rials as catalyst in HMF oxidation were investigated. Additionally,
the reusability of the catalysts were studied.
25 ◦C over 1 h. Depending on the respective support, different times
passed until the solution became clear. Finally, the solid material
was separated from the solvent by centrifugation (3500 rpm, for
20 min) and dried in a dry box at 40 ◦C.
2.4. Oxidation of HMF
HMF oxidation was carried out in a 25 mL three-necked flask
equipped with a reflux condenser at 90 ◦C. The reactor was charged
with 0.4 mmol of HMF in 20 mL of water. The molar ratio of HMF
to Pd was 100:1. The reactor was then placed in a preheated oil
bath (100 ◦C). After few seconds oxygen and aqueous NaOH (0.5 M)
were introduced through a septum. The oxygen flow rate was con-
trolled by a mass flow controller (Bronkhorst) and aqueous solution
of NaOH (0.5 M) was continuously added to the reaction mixture
using a Hamilton syringe pump. To follow the course of the reac-
tion, samples of 50 ␮L were taken from the reaction mixture in
regular time intervals using a 100 ␮L Hamilton syringe. The sam-
ples were diluted with a 0.2% aqueous solution of H3PO4 (total
volume of 10 mL). Then, the samples were centrifuged to remove
solids and subjected to HPLC analysis. Concentrations of observed
products were obtained from calibration curves formed by injecting
solutions with known concentration.
2. Experimental
2.1. Materials
HPLC analysis: Merck-Hitachi with L-4500 diode array detec-
tor, column: Rezex ROA-Organic Acid H+ (8%) 300 × 7.8 mm
(phenomenex), injection: 5 ␮L, mobile phase: 0.2% H3PO4/water,
temperature: 25 ◦C, flow rate: 0.8 mL/min.
5-Hydroxymethylfurfural (HMF) (>99%), 2,5-furandicarboxylic
(FDCA) (>97%), polyvinylpyrrolidone K30 (PVP), sodium tetra-
chloropalladate Na2PdCl4 (>98%) were purchased from Sigma-
Aldrich, while 5-hydroxymethyl-furan-2-carboxylic acid (HMFCA)
was obtained from Santa-Cruz Biotechnology. 5-Formyl-furan-
2-carboxylic acid (FFCA) was purchased from endothermic life
science molecules, and oxygen was obtained from air–liquide.
As supports P25 (TiO2, Evonik,), Al2O3 (Alfa-Aesar,), KF/Al2O3
(Sigma-Aldrich) and ZrO2/La2O3 (MEL, 7 wt% La2O3) were used. All
chemicals were used as received.
2.5. Characterization methods
The amount of Pd on the different supports was determined by
atom absorption spectroscopy (AAS) using a Perkin Elmer AAS-A
Analyst 300 after pulping the solid using sulphuric acid and potas-
sium bisulfate. In order to determine the Pd amount in the solution,
10 mL of the post reaction mixture were mixed with 10 mL aqua
regia before being analyzed by AAS.
2.2. Synthesis of Pd NP
Nitrogen adsorption–desorption isotherms were collected on a
BELSORP-mini II (BEL Japan, Inc.). Specific surface areas and pore
size distributions were calculated from the adsorption and desorp-
tion branches of the isotherm, respectively, applying the Brunauer,
Emmet and Teller (BET) equation for N2.
The phase composition of the catalyst was determined by XRD
using a theta/theta diffractometer (X’Pert Pro from Panalytical,
Almelo, Netherlands) with Cu K␣ radiation (ꢀ = 0.15418 nm, 40 kV,
40 mA) and a X’Celerator RTMS detector.
Transmission electron microscopy (TEM) and scanning trans-
mission electron microscopy (STEM) images of Pd based materials
were obtained at 200 kV with a JEM-ARM200F (JEOL Ltd) which is
aberration-corrected by a CESCOR (CEOS). High-angle annular dark
field (HAADF) imaging was operated with a spot size of 6c (approx-
imately 0.13 nm) and a 40 ␮m condenser aperture. The samples
were deposited on holey carbon supported Cu-grids (mesh 300)
and transferred to the microscope.
The oxidation states and the surface compositions were deter-
mined by X-ray photoelectron spectroscopy (XPS). The measure-
ments were performed with an ESCALAB 220iXL (ThermoFisher
Scientific) with monochromatic Al K␣ radiation (E = 1486.6 eV). The
Pd NP were synthesized in a round bottom flask equipped
with a reflux-condenser with argon bubbling through the reac-
tion mixture. Sodium tetrachloropalladate (Na2PdCl4) (0.236 g,
0.80 mmol) and polyvinylpyrrolidone (0.266 g, 2.3 mmol, PVP
monomer unit = 111.4 g/mol) were dissolved in 33.6 mL of ethylene
glycol at room temperature. The solution was heated up to 90 ◦C.
At this point, 6.4 mL of NaOH dissolved in ethylene glycol (0.5 M)
were added dropwise to the Pd salt containing solution. The final
concentration of NaOH in the synthesis solution was 80 mM. The
stirred reaction mixture was kept constant at 90 ◦C under argon
atmospheres for 3 h. Finally, the obtained black solution was cooled
and stored at ambient temperature.
2.3. Deposition of Pd NP
2.3 mL of the Pd NP containing synthesis solution was washed
with 10 mL acetone. The precipitate, separated from the liquid
phase by centrifugation (3500 rpm for 10 min), was re-dissolved in
10 mL ethanol or water and added to a support/solvent suspension
(0.995 g of support in 25 mL solvent) by means of a syringe pump at

B. Siyo et al. / Applied Catalysis A: General 478 (2014) 107–116
109
Table 1
Characteristics of NP deposition.
Catalyst
Solvent
Time (h)
24
Pd (wt%)
–
BET (m2/g)
Pore volume (mL/g)
d(NP) (nm)
Pd NP
Al2O3
Pd/Al2O3
KF/Al2O3
Pd/KF/Al2O3
–
220
176
25
1.8 0.5
2.7 0.6
0.62
EtOH
0.45
0.6
0.06 [47]
EtOH
H2O
–
48
101
50
35
35
279
n.d.*
TiO2
Pd/TiO2
0.25 [48]
0.08
EtOH
H2O
1
1
0.53
0.5
2.1 0.7
ZrO2/La2O3
Pd/ZrO2/La2O3
EtOH
H2O
–
48
0.39
252
2.6 0.5
*
Not determined.
samples were fixed on a stainless steel sample holder with double
adhesive carbon tape. For charge compensation a flood gun was
used, the spectra were referenced to the C1s peak at 284.8 eV. The
error range for the determination of the electron binding energy
is estimated to be 0.2 eV. After background subtraction the peaks
were fitted with Gaussian–Lorentzian curves to determine the pos-
itions and the areas of the peaks. The surface composition was
calculated from the peak areas divided by the element- specific
Scofield factor and the transmission function of the spectrometer.
The particle size distribution of Pd NP in ethylene glycol solu-
tion and on the different supports was determined using HRTEM
and STEM-HAADF for unsupported NP and supported materials,
respectively (see Fig. 5 and Figs. S1 and S2 in Supplementary mate-
rials). Differences in mean size between Pd NP in ethylene glycol
solution and on support surface were slightly larger for Al2O3 and
ZrO2/La2O3 compared to TiO2 (see Table 1). The small increase in
particle size after deposition is consistent with previous reports,
which suggest that the size of pre-formed NP increases only mod-
erately during deposition [35].
Since deposition was carried out with PVP stabilized NP, Pd NP
can adsorb on the respective support with and without direct metal
support interaction as shown in Fig. 1. To get an indication which of
these Pd species might dominate on the single support, the supports
were treated with pure PVP (15.33 mg PVP in 10 mL water corre-
sponding to 2.3 mL PVP in NP synthesis solution) in water for 1 h
(the same procedure used as for NP deposition but without metal).
The amount of PVP adsorbed on catalyst surface was determined
from the difference of carbon content in the water phase before and
after support addition. On alumina, about 50% of the available PVP
amount was adsorbed after 1 h, on titania only 7%. Using ZrO2/La2O3
as support, no PVP adsorption was detected after 1 h. This finding
indicates that on titania most of the deposited Pd NP should be
in direct interaction with the support (species i) since in this case
the Pd NP were completely adsorbed on the support after 1 h. In
contrast, on alumina more Pd NP should be embedded in adsorbed
PVP (species ii) compared to TiO2 because of the better adsorption
of PVP on the alumina surface.
A better adsorption of Au NP on TiO2 compared to that on
ZrO2 was previously reported for 2 nm negatively polarized Au
NP at pH < IEP (IEP—isoelectronic point). This was explained by a
higher number of positively charged hydroxyl surface groups on
titania. Here, it was assumed that the adsorbed layer of ligands
necessary for stabilization of the Au NP in solution was destroyed
during the adsorption process. Furthermore, it was suggested that
besides electrostatic interactions additional factors (e.g. van der
3. Results and discussion
3.1. Deposition of pre-formed PVP stabilized Pd NP on metal
oxide surfaces
The deposition of the pre-formed PVP stabilized Pd NP on the
different supports was carried out both in water and ethanol as sol-
vent with a nominal Pd loading of 0.5 wt%. The change in the color
of the PVP-Pd NP containing solvent from dark brown to approx-
imately colorless was used as indication that most of the Pd NP
were adsorbed onto the support. Depending on the respective sup-
port, different times passed until the solution became clear (see
Table 1). On TiO2, Pd NP were deposited both in ethanol and water
within of 1 h with the nominal loading of 0.5 wt%. The deposition
of the Pd NP on KF/Al2O3 and ZrO2/La2O3 could only be achieved in
water but after clearly longer time intervals (48 h) in comparison
to titania. A faster NP deposition on TiO2 compared to other sup-
ports (e.g. ␥-Al2O3, ZrO2) was already observed previously for Au
NP and it was shown that the NP deposition time did not depend on
the point of zero charge (p.z.c.) of the particular support [11]. The
same result was found here, when the deposition time was com-
pared with the p.z.c. of the used support (p.z.c: TiO2 = 6–7 [34];
ZrO2/La2O3 ∼ 7 (derived from measurement of the zeta potential);
␥-Al2O3 = 8–9 [34]). As reason for the independence of the Au NP
adsorption time from the p.z.c., the negative charge of the protec-
ting agent (polyvinylalcohol) was suggested [11].
Table 1 summarizes BET surface and pore volume of the differ-
ent supports and BET surfaces of the support after Pd loading. As
expected, BET surface of the Pd/support material was moderately
lower compared to the unloaded support since Pd loading was rel-
atively low [25]. The increase in BET surface for the Pd/KF/Al2O3
support is attributed to the loss of KF from the alumina surface
which increases the accessibility of the pores for nitrogen (see Sec-
tion 3.3). The mean pore size of TiO2, ZrO2/La2O3, and ␥-Al2O3
was 17.5 nm, 18 nm, and 7 nm, respectively. The obtained results
indicate no correlation of BET surface, pore volume, and pore
diameter with the deposition time. Because of the small pore vol-
ume (ZrO2/La2O3, TiO2) and the low pore diameter (␥-Al2O3) it
is believed that the deposited Pd NP were located mainly on the
external surface of the support.
Fig. 1. Formal scheme of Pd NP species which could be formed during NP deposition
on the support surface.

110
B. Siyo et al. / Applied Catalysis A: General 478 (2014) 107–116
Table 2
Catalytic results of HMF oxidation (X—HMF conversion, Y—product yield) on different supports and Pd/support materials with continuous addition of a homogeneous base
(aq. NaOH).
Entry
Catalyst
X (HMF) (%)
Y (HMFCA) (%)
Y (FFCA) (%)
Y (FDCA) (%)
C-Balance
1
2
3
4
5
6
7
8
Without Pd
Na2PdCl4
ZrO2/La2O3
Al2O3
>99 (5 h)
>99 (5 h)
>99 (5 h)
>99 (6 h)
>99 (5 h)
>99 (5 h)
>99
>99
>99
>99
>99
3
10
5
5
4
8
7
7
9
0
0
<1
<1
1
0
0
0
3
2
0
3
<1
0
<1
90
90
91
78
62
52
0.03
0.13
0.07
0.06
0.06
0.98
0.97
0.98
0.90
0.83
0.86
TiO2
Pd-NP*
Pd/ZrO2/La2O3
Pd/KF/Al2O3
Pd/Al2O3
Pd/TiO2*(EtOH)
Pd/TiO2 (H2O)
9
10
11
19
32
2
Reaction conditions: HMF (0.4 mmol), H2O (20 mL), molar ratio HMF/Pd = 100, O2 flow rate = 35 mL/min, 363 K, aq. NaOH addition (0.5 mmol/h, 4 h), reaction time 8 h (*7 h).
Waals interactions) might also play a role in the NP adsorption
processes [9].
during the oxidation as recently observed for HMF oxidation with
Au/CeO2 and Au/TiO2 catalysts [21].
Interestingly, in contrast to HMFCA transformation, oxidation
of FFCA proceeded until it was completely converted. In order to
exclude that FFCA might be converted to FDCA without partici-
pation of Pd NP, an experiment was performed under identical
conditions (35 mL/min O2, 1 mL/h NaOH (0.5 M, 4 h) and 90 ◦C)
using FFCA (0.15 mmol) as substrate in the absence of the solid cat-
alyst. Here, only a very slow amount of FFCA was converted (12%
conversion of FFCA in 6 h) and FDCA was not formed. These findings
strongly suggest that the supported Pd NP catalyze the transforma-
tion of FFCA to FDCA even when oxidation of HMFCA was already
diminished. The results suggest that the oxidation of HMFCA to
FFCA and that of FFCA to FDCA could occur on different Pd surface
sites. It is worth noting, that this behavior was already observed
using unsupported Pd NP [33].
The stability of FDCA was studied under identical reaction con-
ditions. In this case FDCA (0.4 mmol) was used instead of HMF as
reactant in the presence of Pd/ZrO2/La2O3, Pd/TiO2, or Pd/Al2O3.
The FDCA concentration remained unchanged after a time interval
of 5 h. Thus, FDCA degradation did not seem to occur to a mea-
surable extent under these conditions as recently observed with
Au/TiO2 catalyst [21].
To study the influence of leached Pd on HMF oxidation, the pro-
cess was stopped after 1 h, the catalyst (Pd/TiO2, Pd/ZrO2/La2O3)
was separated from reaction mixture and the remaining solution
was further reacted up to 8 h. The results showed that after cat-
alyst removal the amount of HMFCA, FFCA and FDCA remained
unchanged (Fig. S3a and b in SM). These results indicate that either
no Pd NP leached from the catalyst or that those leached Pd NP
were not catalytically active anymore. Analysis of the final reac-
tion mixture revealed that after 8 h about 3% Pd had leached from
the Pd/TiO2 catalyst into the solution. In case of Pd/ZrO2/La2O3
about 6% leached Pd were found. This solution, which contained
higher amount of leached Pd than in former experiments, was fur-
ther reacted for 8 h after catalyst separation. Additionally, HMFCA
(0.08 mmol) and FFCA (0.156 mmol) were added to this solution
to investigate whether the leached Pd can oxidize the interme-
diates. Also in this case the concentration of both intermediates
FFCA and HMFCA remained unchanged (Fig. S4 in SM). The results
obtained indicate that the HMF transformation was catalyzed by
the Pd/support materials and not by leached Pd.
3.2. HMF oxidation
3.2.1. Influence of the support
Previous studies have shown that the presence of high amounts
of base in the reaction mixture foster the formation of byproducts
through side reactions of HMF [32]. In order to ascertain the extent
of these side reactions, a HMF oxidation experiment with addition
of base and oxygen but without catalyst was monitored by HPLC.
Under these conditions, the HMF was completely converted after
5 h, but only a very small amount of HMFCA could be detected
(Table 2, entry 1). It was reported that at absence of catalyst in
basic solution finally humins type products were formed by reac-
tions of HMF degradation compounds such as formic acid, levulinic
acid, and furan [36]. These products (levulinic acid and formic acid)
were also observed by Pasini et al. [37]. FDCA was also not formed
when the Pd precursor was used as catalyst (entry 2). This suggests
that Pd2+ ions are not able to catalyze the reaction of HMF with
oxygen to the desired products. Similar results were obtained for
the unloaded supports (entries 3–5). In all cases a strongly colored
reaction mixture was obtained. In contrast to that, the solution was
only slightly yellow colored in experiments with high FDCA yield.
The strong coloring might indicate degradation of HMF in the reac-
tion solution [24]. In the presence of Pd NP and Pd NP/support
materials, the formation of byproducts decreases (entries 6–11)
leading to higher molar carbon mass balances between 83 and 98%
(combined yield of HMFCA, FFCA, and FDCA). Unsupported Pd NP
afforded 90% yield of FDCA after 7 h at full conversion of HMF (entry
6). Pd/KF/Al2O3 and Pd/ZrO2/La2O3 also gave FDCA yields higher
than 90% (entries 7 and 8). Utilization of Pd NP deposited on Al2O3
or TiO2 (entries 9 and 10) led to lower FDCA yields of 78% and 62%,
respectively, and lower mass balances. The influence of the support
on FDCA yield is further demonstrated when comparing results for
TiO2 in Table 2 (entries 10 and 11). Also the yields of HMFCA varied
between 1 and 32% depending on the catalyst material.
The temporal evolution of HMF conversion and product yield
using the different materials are compared in Fig. 2. Generally,
differences in HMF conversion are very small and HMF was fully
consumed after 1.5 h (NaOH/HMF ratio = 1.8). The highest concen-
tration of the intermediate HMFCA was always found at slightly
earlier times than in the case of FFCA (Fig. 2b and c). TiO2 sup-
ported Pd NP showed the highest HMFCA yield. However, while
FFCA was completely oxidized to FDCA, HMFCA oxidation stopped
after 2 h. During this time interval also most of FDCA was formed. As
no additional HMFCA was formed at this time (X (HMF) = 1), it can be
assumed that active Pd surface sites lost their capability to oxidize
HMFCA after 2 h. This might be due to the partial blocking of active
Pd surface sites by adsorbed organic molecules which are formed
HMF oxidation on these materials was also conducted without
addition of homogeneous base (Table 3). Without base, HMF con-
version on the Pd based materials was slow and FDCA was formed
only in trace amounts.
3.2.2. Influence of rate of base addition
Since the Pd/ZrO2/La2O3 catalyst gave the highest FDCA yields,
this catalyst was chosen to study the effect of the rate of base

B. Siyo et al. / Applied Catalysis A: General 478 (2014) 107–116
111
Fig. 2. Plots of HMF conversion (a) and product yields (b–d) versus time on Pd/support catalysts (reaction conditions: HMF (0.4 mmol), H2O (20 mL), HMF/Pd = 100 (mol/mol),
O2 flow rate (35 mL/min), 363 K, aq. NaOH addition [0.5 mmol/h, 4 h]).
addition on the HMF conversion and product formation over
time. In these experiments, the aqueous NaOH solution (0.5 M)
was continuously added to the reaction mixture with different
flow rates (0.5, 1 and 2 mL/h). Base is believed to deprotonate
the alcohol side chain to form an alkoxy intermediate [38] and to
increase the solubility of FDCA which is low in base free solution.
Hence, precipitation of the formed FDCA onto the catalyst surface
may hamper the reaction [39].
The conversion and yield versus time plots (Fig. 3) show that
the rate of all steps in HMF oxidation depends strongly on base
concentration in the reaction mixture and the rates of formation
and conversion of all investigated intermediates increased with
faster base addition. At the highest flow rate (2 mL/h), the inter-
mediates HMFCA and FFCA were almost completely transformed
after 1 h (NaOH/HMF ratio = 2.5) and a FDCA yield of 80% was
obtained. Constant values of carbon mass balance (0.82) after
0.5 h reaction time (X(HMF) = 1) suggest that the formation of
byproducts occurred at the early stage of the reaction (<0.5 h).
The highest FDCA yield (90%) was obtained when the base was
added with at a rate of 1 mL/h. In contrast to Au based materials
FDCA can be obtained in high yields on Pd already at relatively low
NaOH/HMF ratios (<2.5) [32]. At the slowest NaOH addition rate
(0.5 mL/h), HMF conversion was completely after 2 h and FDCA
was formed in 70% yield after 8 h. The influence of the nature
of base on HMF oxidation was studied recently in presence of
Pt/C catalyst by Rass et al. They observed heavy HMF degradation
from the beginning of the reaction using a NaOH/HMF ratio of 4
[24]. When the reaction was carried out under moderately basic
conditions (using NaHCO3 instead of NaOH) higher yields of FDCA
were obtained and byproduct formation was suppressed.
The rate of base addition also has an influence on the amount of
Pd leached from the support. When 4 mL of 0.5 M aq. NaOH were
immediately added to the Pd/Al2O3 or Pd/ZrO2/La2O3 catalyst sus-
pended in pure water (20 mL, 90 ◦C), the Pd/Al2O3 lost about 17%
and the Pd/ZrO2/La2O3 catalyst about 12% of its available Pd amount
after 8 h. In contrast, if the base was added continuously with differ-
ent rate base addition of aq.NaOH (0.5 M) 2, 1, and 0.5 mL/h to the
reaction mixture over 2, 4 and 8 h, respectively. Under these con-
ditions, the loss of Pd from the Pd/ZrO2/La2O3 catalyst amounted
to 12, 6 and 2%, respectively.
3.2.3. Catalyst stability and reusability
In order to obtain information about the catalytic stability of the
Pd based materials, the Pd/ZrO2/La2O3, Pd/KF/Al2O3, and Pd/TiO2
catalysts were reused. For recycling, the solid material was sep-
arated from the reaction mixture by centrifugation, washed once
with water and dried at 40 ◦C. Material loss due to sampling for a
single run on Pd/ZrO2/La2O3 and Pd/TiO2 was at about 7% found
by weighing of the material before and after reaction. Plots of HMF
conversion and product yields versus time are shown in Figs. S5–S7
Table 3
Catalytic results of HMF oxidation (X—HMF conversion, Y—product yield) on different Pd NP/support catalysts without addition of homogeneous base.
Entry
Catalyst
X (HMF) (%)
Y (HMFCA) (%)
Y (FFCA) (%)
Y (FDCA)(%)
C-Balance
1
2
3
Pd/Al2O3
Pd/KF/Al2O3
Pd/ZrO2/La2O3
10
60
26
2
30
4
3
15
6
3
3
1
0.8
0.8
0.4
Reaction conditions: HMF (0.4 mmol), H2O (20 mL), molar ratio HMF/Pd = 100, O2 flow rate = 35 mL/min, 363 K, aq. NaOH addition (0.5 mmol/h, 4 h), reaction time (7 h).

112
B. Siyo et al. / Applied Catalysis A: General 478 (2014) 107–116
Fig. 3. Plots of HMF conversion (a) and product yields (b–d) versus reaction time at different rates of base addition. (conditions: HMF (0.4 mmol), H2O (20 mL), HMF/Pd—100
(mol/mol), O2 flow rate—35 mL/min, 363 K, aq. NaOH (0.5 M), final NaOH/HMF ratio = 5).
of SM. In all runs, the HMF was completely converted after approx-
imately 1.5 h and the formed FFCA was complete oxidized to FDCA,
while oxidation of the first intermediate HMFCA did not continue
after reaction times longer than 2 h. In the following experiment,
the HMFCA intermediate was converted again as indicated by the
concomitant yield of FDCA. The results suggest that the loss of capa-
bility to oxidize the HMFCA is a reversible effect. However, a slightly
lower FDCA yield was observed in each consecutive run. Results of
different runs observed after 8 h reaction time are summarized in
Table 4. For the Pd/ZrO2/La2O3 material the yield of FDCA decreased
slightly from 90% to 86% after three runs. A lower FDCA yield was
also observed for Pd/TiO2 (62% → 54%). In this context it should be
mentioned that Pd leaching from ZrO2/La2O3 support decreased
with increasing number of runs (6% in first run, 3% in third run).
For the Pd/KF/Al2O3 catalyst the difference in FDCA yield between
the first and the second run was larger (91 → 55%). In all cases, an
increase in HMFCA yield was detected with an increasing number of
runs. A similar behavior, a slight decrease of the FDCA yield during
recycling experiments, has previously been observed for Au/CeO2
catalysts [21]. A bimetallic Au/Cu/TiO2 catalyst showed more sta-
ble catalytic performances, however, the catalyst was tested under
conditions leading to high yields of HMFCA compared to FDCA [37].
A stable catalytic performance was also observed for a Pt/Bi/C cat-
alyst using NaHCO3 as base [24].
3.3. Characterization of fresh and used catalysts
The fresh and used catalysts were subjected to XRD to determine
possible structural changes of the materials (Fig. 4). Reflections
from metallic Pd NP were not observed because of the low Pd load-
ing (0.5 wt%). Diffraction patterns of TiO2, ZrO2/La2O3, and Al2O3
were neither influenced by the Pd deposition nor by HMF oxida-
tion at 90 ◦C in basic solution. In case of Al2O3 and ZrO2/La2O3
only very broad XRD signals appeared which might be caused by
a small crystallite size and/or by a high portion of amorphous
phases. SEM images of the ZrO2/La2O3 material, before and after
application in HMF oxidation (Fig. S8, SM) imply that the morphol-
ogy of the ZrO2/La2O3 particles did not change during the HMF
oxidation.
For Pd/KF/Al2O3 (see Fig. 4d) the diffraction pattern changed
both during Pd loading and during HMF oxidation but the alumina
structure remained intact. The K3AlF6 phase (PDF-Nr. 000-0615)
observed after Pd NP deposition disappeared and a PdF4 phase
(PDF-Nr. 070-0717) was formed during the reaction. Additional
peaks can be attributed to a K2O phase (PDF-Nr. 013-0373). The
observed differences in phase compositions might be responsible
for the relatively high loss of material (33%) in a single experiment
as well as for the differences in the catalytic performance between
first and second use (FDCA yield: 1st use—91%; 2nd use—55%).
Table 4
Results of reusability tests on Pd/support catalysts with continuous addition of homogeneous base; X (HMF) = 100%.
First use
Second use
Third use
Catalyst
Y% HMFCA
Y% FFCA
Y% FDCA
Y% HMFCA
Y% FFCA
Y% FDCA
Y% HMFCA
Y% FFCA
Y% FDCA
Pd/ZrO2/La2O3
Pd/KF/Al2O3
Pd/TiO2
7
7
19
0
0
2
90
91
62
12
43
19
0
0
2
87
55
59
12
–
26
2
–
0
86
–
54
*
Reaction conditions: HMF (0.4 mmol), H2O (20 mL), HMF/Pd = 100 (mol/mol), O2 flow rate: 35 mL/min, 363 K, aq. NaOH addition (0.5 mmol/h, 4 h), reaction time: 8 h (*7 h).

B. Siyo et al. / Applied Catalysis A: General 478 (2014) 107–116
113
Fig. 4. Influence of Pd NP deposition and HMF oxidation on the XRD pattern of the Pd/support materials (a) TiO2, (b) ZrO2/La2O3, (c) Al2O3, and (d) KF/Al2O3 (*—K3AlF6,
ꢀ—K2O, ꢀ—PdF4).
Fig. 5. STEM-HAADF images (a, b, d and e) and results of Pd mapping (c and f) on Pd NP supported on TiO2 (a and d), Al2O3 (b and e), and ZrO2/La2O3 (c and f) after NP
deposition (a–c) and after third use (d–f) in HMF oxidation (Pd/Al2O3 after first use, insert in (b) shows structure of a large Pd aggregate; scale bar: 5 nm).

114
B. Siyo et al. / Applied Catalysis A: General 478 (2014) 107–116
Pd 3d
1/2
Zr 3p
3/2
Zr 3p
1/2
a)
b)
c)
used
Pd 3d
3/2
used
used
fresh
fresh
fresh
350
345
340
335
330
360
350
340
330
320
570
565
560
555
550
545
Electron Binding Energy / eV
Electron Binding Energy / eV
Electron Binding Energy / eV
Fig. 6. XPS spectra of (a) the Pd3d electrons from the fresh and used (3×) Pd/TiO2 catalysts, (b) the Zr3p and Pd 3d electrons and (c) the Pd3p electrons from the fresh and
used (3×) Pd/ZrO2/La2O3 catalysts.
It is generally accepted that the catalytic performance of NP
based catalysts depends on the geometric and electronic structure
of the active sites [40] and the nature of support [41,42]. The influ-
ence of HMF oxidation on Pd NP structure for the different supports
was studied by TEM. Fig. 5 shows STEM-HAADF images of freshly
prepared Pd/TiO2, Pd/Al2O3, and Pd/ZrO2/La2O3 catalysts and after
utilization in the HMF oxidation. For Pd/ZrO2/La2O3 more reliable
results were obtained by element mapping with EDX because the
electron density of the ZrO2/La2O3 support (La: atomic number 57,
Zr: atomic number 40) is higher and similar compared to Pd (atomic
number 46), respectively.
The utilization of the materials in HMF oxidation led to an
increase in particle size, broadening of the size distribution and
changing of the shape (see also Table 1 and Fig. S2, SM). The extent
of the observed changes depended on the support and seems to
be highest for Al2O3. Here large Pd aggregates with mean sizes
larger than 20 nm were observed already after the first use (see
inset Fig. 5e). On the ZrO2/La2O3 support the formed Pd aggregates
were smaller although the material was used three times. Further-
more, the Pd aggregates seem to be more elongated. Such elongated
Pd structures were also obtained when the PVP was substituted by
a molecular stabilizer by means of a phase transfer to toluene at
100 ◦C [43]. The lowest change in particle size and shape of Pd NP
after reaction was observed on titania.
of metallic Pd in the XPS spectra of Pd3d electrons. Bearing in
mind that identical metallic Pd NP were used for the deposition
on the ZrO2/La2O3 and on the TiO2 the presence of bivalent Pd
after adsorption of Pd NP on the TiO2 support (Fig. 6) is remark-
able, because on passive ZrO2/La2O3 only Pd NP in metallic state
were detected. This observation indicates an influence of the sup-
port on the oxidation state of the deposited NP. Strong interactions
between the Pd clusters and TiO2 were already observed previously
[44]. An additional evidence for this strong interaction is given in
Fig. 7 showing the XP spectra of Ti2p electrons of the fresh and used
sample. While for the fresh sample only a Ti4+ peak was detected,
the XPS signal of the used sample is composed of peaks of Ti4+ and
Ti3+. The formation of Ti3+ during HMF reaction shows clearly the
interaction between the Pd NP and the TiO2 support. Moreover,
such interactions might be also responsible for the faster adsorp-
tion of the Pd NP on TiO2 compared to the ZrO2/La2O3 and Al2O3
supports.
During HMF reaction the population of bivalent Pd decreased
probably because the Pd NP accept electrons released during the
HMF oxidation. For Pt and Au based materials it was suggested
that molecular oxygen played an indirect role in HMF oxidation
by accepting electrons deposited to the supported metal particles
[38]. The formation of Ti3+ during the reaction might be explained
as follows: as shown previously (Table 2) HMF oxidation on TiO2
In order to obtain information about the electronic structure of
the supported Pd NP, the Pd/ZrO2/La2O3 and Pd/TiO2 catalysts were
characterized by XPS both after NP deposition and after utilization
in HMF oxidation. In general, the positions of the Pd XPS signals
are influenced by the valence state of Pd, e.g. the binding energy
for bivalent Pd is usually higher than for metallic Pd. Moreover, the
electronic interaction of Pd with the support and the size of the
Pd species could lead to a shift of the binding energy [44–46]. For
the ZrO2/La2O3 support the interpretation of the Pd3d spectra was
complicated, due to the overlap of peaks of the Pd3d and Zr3p elec-
trons. Nevertheless, it was possible to determine the maximum of
the Pd3d3/2 peak (340.4 eV for the fresh sample, 341.4 eV after use).
Additionally, the Pd3p1/2 electrons were also recorded to obtain the
Pd peak position without overlapping by other peaks. The maxi-
mum at 558.6 eV for both samples, fresh and utilizes, confirm that
the most part of Pd is metallic and that the electronic structure of
the Pd NP is not changed by the oxidation reaction (Fig. 6b and c).
In contrast to ZrO2/La2O3, on the TiO2 support, metallic Pd
(Fig. 6a, grey area, 3d5/2 = 335.4 eV, 5d3/2 = 340.6 eV) and bivalent
Pd (dark grey area, 3d5/2 = 337.0 eV, 3d3/2 = 342.4 eV) were detected
both in fresh and used Pd/TiO2 samples. After the utilization in HMF
oxidation, a reduction of a part of the bivalent Pd was observed
in the Pd/TiO2 sample, indicated by an increase of the peak area
Ti4+
Ti 2p
Ti3+
used
fresh
485 480 475 470 465 460 455 450 445
Electron Binding Energy / eV
Fig. 7. XPS spectra of the Ti2p electrons from the fresh and used (3×) Pd/TiO2
catalyst.

B. Siyo et al. / Applied Catalysis A: General 478 (2014) 107–116
115
Table 5
strong interaction makes the TiO2 support very interesting for fur-
ther studies, e.g. by adsorption of surface modified Pd nanoparticles
which could lead to a decrease of byproduct formation.
Composition of the near-surface region of the fresh and used catalysts determined
by XPS.
Pd
N
Ti
Two different processes of deactivation were observed using the
Pd NP based materials: (i)— a reversible short-term deactivation by
blocking of active Pd surface sites and, (ii)— an irreversible long-
term deactivation by changes in Pd structure and by Pd leaching.
The extent of leaching is heavily influenced by the base concentra-
tion in the reaction mixture. Particle size and shape of the NP on the
support surface changed during HMF oxidation, mainly because of
formerly isolated small NP aggregates. The influence of the changes
in Pd structure on the catalytic performance seems to be relatively
low probably because the surface of the small Pd NP in the formed
aggregates remains largely intact.
Fresh
Used (3×)
0.009
0.008
0.039
0.014
1
1
Pd/TiO2
Pd
0.194
0.112
N
0.295
0.129
Zr
1
1
Pd/ZrO2/La2O3
Fresh
Used (3×)
proceeds with byproduct formation. These byproducts might block
active surface sites and make them inaccessible for molecular oxy-
gen. Electrons which get released during the oxidation process
might get transferred from the NP to the support and reduce Ti4+
sites in TiO2.
Acknowledgments
During the oxidation reaction not only a low amount of Pd
leached from the support but also the amount of PVP on the sup-
port surface decreased, probably by degradation, as indicated by the
lower amount of nitrogen on the support surface after the reaction
(see Table 5). Those NP which were formerly firmly embedded in
the polymer could now migrate to either the support surface or into
the liquid phase. On the support surface they could undergo, e.g.
aggregation reactions, and thus form aggregated Pd structures as
observed from TEM images. Because of the good adsorption of PVP
on alumina (see Section 3.1), the amount of Pd NP which are embed-
ded in adsorbed polymer (species ii in Fig. 1) should be the highest.
This might be one reason that the largest aggregated Pd structures
were observed on the alumina support after HMF oxidation. Dif-
ferences in surface charge between the supports ZrO2/La2O3 and
Al2O3 in basic solution are indicated by the more negative zeta
potential of the respective support (pH 10: −43 mV compared to
pH 10: −24 mV, measurement conditions are given in SM). This
might be another reason for the lower degree of aggregation of Pd
NP on ZrO2/La2O3 compared to Al2O3. Degradation of PVP might
also serve as explanation for comparable activities in the reusabil-
ity experiments, despite the observed Pd leaching. Surface sites
formerly blocked by PVP might become accessible during these pro-
cesses. The low differences in FDCA yield between the successive
runs in the reusability experiments might be attributed also to the
formation of Pd aggregates composed of formerly smaller particles.
In such structures the Pd NP would only lose a part of their surface.
Hence, large parts of the former structure should still remain intact.
Financial support by Tishreen University Latakia (Syria, grant
number 1122/L) and the German Academic Exchange Service
(DAAD) (scholarship for B.S., grant number A/09/95233) is grate-
fully acknowledged.
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.apcata.
2014.03.020.
References
[1] R.A. Van Santen, P.W.N.M. van Leeuwen, J.A. Moulijn, B.A. Averill, Catalysis An
Integrated Approach, Elsevier, 2000.
[2] F. Pinna, Catal. Today 41 (1998) 129–137.
[3] A. Roucoux, J. Schulz, H. Patin, Chem. Rev. 102 (2002) 3757–3778.
[4] J. Park, J. Joo, S.G. Kwon, Y. Jang, T. Hyeon, Angew. Chem. Int. Ed. 46 (2007)
4630–4660.
[5] T. Teranishi, M. Miyake, Chem. Mater. 10 (1998) 594–600.
[6] E. Ramirez, S. Jansat, K. Philippot, P. Lecante, M. Gomez, A.M. Masdeu-Bultó, B.
Chaudret, J. Organomet. Chem. 689 (2004) 4601–4610.
[7] J. Schäffer, V.A. Kondratenko, N. Steinfeldt, M. Sebek, E.V. Kondratenko, J. Catal.
301 (2013) 210–216.
[8] J. Croy, S. Mostafa, J. Liu, Y. Sohn, H. Heinrich, B. Cuenya, Catal. Lett. 119 (2007)
209–216.
[9] J.-D. Grunwaldt, C. Kiener, C. Wögerbauer, A. Baiker, J. Catal. 181 (1999)
223–232.
[10] A.R. West, Solid State Chemistry and its Applications, Wiley, Chichester, 2014.
[11] M. Comotti, W.-C. Li, B. Spliethoff, F. Schüth, J. Am. Chem. Soc. 128 (2005)
917–924.
4. Conclusions
[12] B.R. Cuenya, Thin Solid Films 518 (2010) 3127–3150.
[13] A.A. Rosatella, S.P. Simeonov, R.F.M. Frade, C.A.M. Afonso, Green Chem. 13
(2011) 754–793.
[14] C. Moreau, M.N. Belgacem, A. Gandini, Top. Catal. 27 (2004) 11–30.
[15] S.G. Wettstein, D.M. Alonso, E.I. Gürbüz, J.A. Dumesic, Curr. Opin. Chem. Eng. 1
(2012) 218–224.
[16] C. Rasrendra, J. Soetedjo, I. Makertihartha, S. Adisasmito, H. Heeres, Top. Catal.
55 (2012) 543–549.
[17] M. Kröger, U. Prüße, K.-D. Vorlop, Top. Catal. 13 (2000) 237–242.
[18] T. Werpy, G. Petersen, US Department of Energy Report, August 2004, 2004,
<http://www.osti.gov/bridge/>.
[19] A. Gandini, M.N. Belgacem, Prog. Polym. Sci. 22 (1997) 1203–1379.
[20] S.E. Davis, M.S. Ide, R.J. Davis, Green Chem. 15 (2013) 17–45.
[21] O. Casanova, S. Iborra, A. Corma, ChemSusChem 2 (2009) 1138–1144.
[22] N.K. Gupta, S. Nishimura, A. Takagaki, K. Ebitani, Green Chem. 13 (2011)
824–827.
Pd NP based materials prepared via colloidal deposition of
pre-formed PVP stabilized Pd NP were used as catalysts in HMF
oxidation at presence of homogeneous base. The yield of FDCA is
influenced both, by the reaction conditions and by the particular
support. The strength of interaction between Pd NP and the support
seems to be higher for TiO2 than for ZrO2/La2O3 or Al2O3 because
of interactions between the TiO2 support and the NP as evidenced
from XPS. This interaction might be also responsible for the lower
alteration of particle size and shape of the NP during HMF oxidation
on TiO2, and for the lower Pd amount in the post reaction mixture
in compared to the ZrO2/La2O3 and Al2O3 supports.
The highest yield of FDCA and a relatively stable catalytic per-
formance was observed with the Pd/ZrO2/La2O3 catalyst. Here,
byproduct formation was generally low. In case of titania, with
a stronger interaction between the Pd NP and the support, cat-
alytic performance was also stable, but lower yields of FDCA were
obtained. This might be explained by a stronger adsorption of HMF
and reaction intermediates on the catalyst surface, leading to a
blocking of active surface sites during the first stage of the reaction
as indicated by the lower carbon mass balance. Nevertheless, this
[23] M. Lilga, R. Hallen, M. Gray, Top. Catal. 53 (2010) 1264–1269.
[24] H.A. Rass, N. Essayem, M. Besson, Green Chem. 15 (2013) 2240–2251.
[25] Y. Gorbanev, S. Kegnæs, A. Riisager, Top. Catal. 54 (2011) 1318–1324.
[26] K. Mori, T. Hara, T. Mizugaki, K. Ebitani, K. Kaneda, J. Am. Chem. Soc. 126 (2004)
10657–10666.
[27] M. Mifsud, K.V. Parkhomenko, I.W.C.E. Arends, R.A. Sheldon, Tetrahedron 66
(2010) 1040–1044.
[28] W. Hou, N.A. Dehm, R.W. Scott, J. Catal. 253 (2008) 22–27.
[29] B. Feng, Z. Hou, H. Yang, X. Wang, Y. Hu, H. Li, Y. Qiao, X. Zhao, Q. Huang,
Langmuir 26 (2009) 2505–2513.
[30] Z. Ma, H. Yang, Y. Qin, Y. Hao, G. Li, J. Mol. Catal. A: Chem. 331 (2010) 78–85.

116
B. Siyo et al. / Applied Catalysis A: General 478 (2014) 107–116
[31] H. Alex, N. Steinfeldt, K. Jähnisch, S. Hübner, M. Bauer, Nanotechnol. Rev. 3
(2014) 99–110.
[39] Y.Y. Gorbanev, S.K. Klitgaard, J.M. Woodley, C.H. Christensen, A. Riisager, Chem-
SusChem 2 (2009) 672–675.
[32] S.E. Davis, L.R. Houk, E.C. Tamargo, A.K. Datye, R.J. Davis, Catal. Today 160 (2011)
55–60.
[33] B. Siyo, M. Schneider, M.-M. Pohl, P. Langer, N. Steinfeldt, Catal. Lett. 144 (2013)
498–506.
[34] M. Kosmulski, J. Colloid Interface Sci. 275 (2004) 214–224.
[35] C. Berger-Karin, M. Sebek, M.M. Pohl, U. Bentrup, V.A. Kondratenko, N. Stein-
feldt, E.V. Kondratenko, ChemCatChem 4 (2012) 1368–1375.
[36] K.R. Vuyyuru, P. Strasser, Catal. Today 195 (2012) 144–154.
[37] T. Pasini, M. Piccinini, M. Blosi, R. Bonelli, S. Albonetti, N. Dimitratos, J.A. Lopez-
Sanchez, M. Sankar, Q. He, C.J. Kiely, G.J. Hutchings, F. Cavani, Green Chem. 13
(2011) 2091–2099.
[40] H. Li, L. Li, Y. Li, Nanotechnol. Rev. 2 (2013) 515–528.
[41] H.-J.F.T. Risse, Top. Organomet. Chem. 16 (2005) 117–149.
[42] A.K. Santra, D.W. Goodman, J. Phys.: Condens. Matter 14 (2002) 31–62.
[43] J. Baumgard, M.-M. Pohl, U. Kragl, N. Steinfeldt, Nanotechnol. Rev. 3 (2014)
87–98.
[44] W.E. Kaden, T. Wu, W.A. Kunkel, S.L. Anderson, Science 326 (2009) 826–829.
[45] J.H. Kang, E.W. Shin, W.J. Kim, J.D. Park, S.H. Moon, J. Catal. 208 (2002)
310–320.
[46] K. Okumura, M. Niwa, Catal. Surv. Asia 5 (2002) 121–126.
[47] V. Raju, R. Radhakrishnan, S. Jaenicke, G.K. Chuab, Catal. Today 164 (2011)
139–142.
[38] S.E. Davis, B.N. Zope, R.J. Davis, Green Chem. 14 (2012) 143–147.
[48] K.J.A. Raj, B. Viswanathan, Indian J. Chem. 48A (2009) 1378–1382.