Platinum supported on nanosilica and fibrous nanosilica for hydrogenation reactions

Article abstract of DOI:10.1016/j.poly.2020.114769

Platinum nanoparticles supported on nanosilica (NP) and fibrous nanosilica (dendritic fibrous nano-spheres, DFNS) were prepared by direct grafting of the Pt precursor onto the silanol groups or via a polyethylenimine (PEI) linker. From the SEM and TEM images the average diameter of the nanosilica and fibrous nanosilica (DFNS), was determined to be 21.4 and 503 nm, respectively. While surface areas as measured by ASAP is 463.4 m² g^?1 for DFNS and 142.5 m² g^?1 for the nanosilica. For the four Pt containing catalysts (Pt/NP, Pt/DFNS, Pt/PEI/NP and Pt/PEI/DFNS), a Pt loading between 1.35 × 10¹⁷ and 8.46 × 10¹⁷ Pt atoms per gram support were determined. The PEI-containing catalyst gave higher Pt-loading than the direct anchoring of the Pt onto the silanol groups of the support. The catalysts were further characterised ATR FTIR and XPS. After oxidation of the pre-catalysts 85% of the Pt was in the oxide form. While after reduction, ca. 82% the Pt supported on DFNS was in the metallic form. Reduction of the Pt supported on NP, resulted in 100% of the Pt in the Pt⁰ oxidation state. These catalysts were tested for the hydrogenation of C[dbnd]C and/or C[dbnd]O bonds in cyclohexene, benzaldehyde and cinnamaldehyde. The % conversion and product distribution will be discussed in term of diameter, surface area and Pt-loading.

Full text of DOI:10.1016/j.poly.2020.114769

Polyhedron 193 (2021) 114769
Contents lists available at ScienceDirect
Polyhedron
journal homepage: www.elsevier.com/locate/poly
Platinum supported on nanosilica and fibrous nanosilica for
hydrogenation reactions
⇑
Z. Xantini, E. Erasmus
Department of Chemistry, University of the Free State, PO Box 339, Bloemfontein 9300, South Africa
a r t i c l e i n f o
a b s t r a c t
Article history:
Platinum nanoparticles supported on nanosilica (NP) and fibrous nanosilica (dendritic fibrous nano-
spheres, DFNS) were prepared by direct grafting of the Pt precursor onto the silanol groups or via a
polyethylenimine (PEI) linker. From the SEM and TEM images the average diameter of the nanosilica
and fibrous nanosilica (DFNS), was determined to be 21.4 and 503 nm, respectively. While surface areas
Received 3 August 2020
Accepted 21 August 2020
Available online 11 September 2020
2
ꢀ1
for DFNS and 142.5 m2
ꢀ1
as measured by ASAP is 463.4 m
g
g
for the nanosilica. For the four Pt con-
Keywords:
17
taining catalysts (Pt/NP, Pt/DFNS, Pt/PEI/NP and Pt/PEI/DFNS), a Pt loading between 1.35 ꢁ 10 and
Fibrous silica (DFNS)
Polyethylenimine (PEI)
XPS
Selective hydrogenation
Platinum
17
8
.46 ꢁ 10 Pt atoms per gram support were determined. The PEI-containing catalyst gave higher
Pt-loading than the direct anchoring of the Pt onto the silanol groups of the support. The catalysts were
further characterised ATR FTIR and XPS. After oxidation of the pre-catalysts 85% of the Pt was in the
oxide form. While after reduction, ca. 82% the Pt supported on DFNS was in the metallic form.
Reduction of the Pt supported on NP, resulted in 100% of the Pt in the Pt0 oxidation state. These cat-
alysts were tested for the hydrogenation of C@C and/or C@O bonds in cyclohexene, benzaldehyde
and cinnamaldehyde. The % conversion and product distribution will be discussed in term of diameter,
surface area and Pt-loading.
Ó 2020 Published by Elsevier Ltd.
1
. Introduction
Metal-containing heterogeneous catalysts usually consist of
the amount of metal ions deposited onto the support, the particle
size and metal distribution [13,14].
The fundamental properties that describe the efficiency of a
heterogeneous catalyst are: activity, selectivity, reproducibility,
affordability, thermal and mechanical stability. This also include
morphological characteristics like high surface area, high number
of active sites per unit area and crystallinity [15]. Consequently,
it is important to maximize the surface area and the number of
the active sites to improve the catalytic performance.
Porous material (known to have a large surface area) have been
used for various applications, such as biological and pharmaceuti-
cal systems [16–18], gas storage and chemical sensing [19]. One of
the major uses of porous materials is in heterogeneous catalysis
[20]. The pores increase the surface area dramatically, which
allows for anchoring of more catalytically active sites (metals)
[21], than a smooth surface with the same volume. Pores are not
the only way to increase the surface area, fibrous extensions pro-
truding from the core of a particle can also increase surface area.
In 2010, Polshettiwar et al. [22,23] discovered fibrous silica or den-
dritic fibrous nano-spheres (DFNS). Fibrous silica has high surface
transition metals deposited onto the surface of the solid support
as the active phase of the catalyst [1]. When the catalytically active
species are supported by metal oxides, the support may also offer
additional catalytic functionality that may also improve the overall
catalytic performance [2]. Upon choosing a suitable catalytic sup-
port, one has to consider the surface area, thermal, mechanical
and chemical stability [1,3,4]. Silica has been used as a starting pre-
cursor due to its benign nature, low toxicity, ease of surface mod-
ifications, chemical and mechanical stability [5].
The oxide groups on silica are unreactive towards binding with
metal precursors. The first important surface modification on silica
supports is, therefore, the hydroxylation of the silicon oxide, which
affords reactive silanol groups (Si-OH) [6]. The silanol groups can
then be used for direct binding of the metal precursor or to bind
other entities with different functional groups, which can act as
anchoring sites. These include alkoxysilanes [4,7–10], and den-
drimers [1,11,12],which have been used as a way of controlling
area owing to its fibrous morphology, ranging from 500 to
2
6
00 m per g [24,25]. This material also exhibits physical proper-
⇑
Corresponding author.
E-mail address: Erasmus@ufs.ac.za (E. Erasmus).
ties such as good thermal, hydrothermal and mechanical stability
[4].
https://doi.org/10.1016/j.poly.2020.114769
277-5387/Ó 2020 Published by Elsevier Ltd.
0

Z. Xantini and E. Erasmus
Polyhedron 193 (2021) 114769
Fibrous silica (DFNS) has also been used as a solid support in
heterogeneous catalysis. The fibrous morphology of DFNS not only
increases the surface area of the silica particle but also the accessi-
bility of active metals during catalysis, making it a very useful sup-
port [1,22]. Fibrous silica have a wide variety of applications such
2.2. Catalyst preparation
2.2.1. Synthesis of fibrous nanosilica (DFNS) [50]
A mixture of urea (1.213 g, 0.020 mol), water (60 ml) and
cetyltrimethylammonium bromide, CTAB, (2.006 g, 0.00550 mol)
was stirred for 15 min. Tetraethyl orthosilicate, TEOS) (5.00 g,
0.024 mol) was dissolved in cyclohexane (60 ml) and this was
added dropwise in the urea mixture solution. The mixture was stir-
red for a further 15 min and then 1-pentanol (3 ml) was added
dropwise. The resultant solution was stirred for 30 min at ambient
temperature, transferred to a Teflon-lined autoclave reactor and
heated for 5 h at 120 °C. The suspension was cooled to room tem-
perature and the product was recovered by centrifugation, washed
with water and acetone and dried in air for 24 h followed by calci-
nation at 550 °C for 6 h [50].
as CO
2
adsorbent, catalytic support as well as solar energy harvest
[
26,27]. Because fibrous silica is nontoxic and biocompatible, it has
2
6
found applications as a component for efficient drug delivery.
To improve the stability of supported catalysts, research has
also focussed on the use of polyethylenimine (PEI) as an alternative
to organofunctional silane to modify the support surface
[
18,25,28,29]. PEI is a polycationic molecule with repeating units
of primary, secondary and tertiary amino groups separated by a
CH -CH spacer [30,31]. The amino groups on PEI are active and
2
2
form strong metal-N ligand complexes and hydrogen bonds with
surface silanols [32,33]. PEI is often used as a pseudoligand, which
helps in homogeneously dispersing metal nanoparticles and main-
tains their stability against leaching [18,34].
IR
m
[cmꢀ1]: 807 (Si-O-Si) symmetric stretch, 1040 (Si-O-Si)
asymmetric stretch, 3500 (Si-OH),
Platinum is one of the most commonly used metals in chemical
catalytic reactions such as hydrogenation and dehydrogenation
2.2.2. Functionalisation of silica supports with branched
polyethylenimine
[
35]. It can easily donate or accept electrons. This result in changes
The silica supports, either nanosilica or DFNS (5.0 g) was
degassed in vacuo at 120 °C for 12 h and then cooled to 60 °C under
Argon. (3-Glycidyloxypropyl)trimethoxysilane, GTMS, (5.2 ml) dis-
solved in hot methanol (30 ml) was added to the degassed silica
supports with continuous stirring for 1.5 h. Branched polyethylen-
imine (PEI) (5.00 g) was dissolved in hot methanol (30 ml) and
then added to the reaction mixture. The resultant solution was stir-
red for 5 h at 60 °C. The functionalised supports was recovered by
centrifugation, washed with hot methanol three times and then
dried in vacuo for 12 h.
0
II
IV
in the oxidation states (Pt , Pt and Pt ) during the catalytic cycle
36–38]. The catalytic performance of platinum is enhanced when
[
it is present as nanosized particles that are homogeneously dis-
tributed on the solid support [39,40].
The hydrogenation of unsaturated aldehydes to saturated alco-
hols was first introduced a century ago and has been an important
industrial process ever since [41]. This type of a reaction can pro-
duce three products: unsaturated alcohols due to the reduction
of the C@O bond, saturated aldehydes due to the reduction of the
C@C bond and saturated alcohols due to the complete reduction
of both functionalities [42]. Unsaturated alcohols are important
intermediates for the production of fine chemicals, perfumes and
pharmaceuticals and as such, have attracted a lot of attention
around them [41,43,44].
The purpose behind this study is to compare nanosilica powder
and DFNS as catalytic supports for platinum. The effect that the
hydroxylated- and polyethylenimine (PEI)-modification have on
surface area, Pt-loading and catalytic activity towards hydrogena-
tion will be investigated. Each modification will be characterised
by different surface, elemental and structural characterisation
methods. A brief analysis on the catalytic activity, product distribu-
tion of the catalyst will also be conducted.
ꢀ1
DFNS: IR
m [cm ]: 807 (Si-O-Si) symmetric stretch, 1040
◦
(Si-O-Si) asymmetric stretch, 1500 (CAN), 1650 (1 NAH), 2800–
2950 (CAH), 3600–3000 (Si-OH, 1 and 2 NAH).
◦
◦
ꢀ1
Nanosilica: IR
m [cm ]: 807 (Si-O-Si) symmetric stretch, 1040
◦
(Si-O-Si) asymmetric stretch, 1500–1650 (CAN), 1650 (1 NAH),
2800–2950 (CAH), 3600–3000 (Si-OH, 1 and 2 NAH).
◦
◦
2.2.3. Grafting of PtCl onto the silica supports
2
The as prepared DFNS or nanosilica (100 mg) was added to a
solution of PtCl2 (10 mM, 5 ml) in chloroform. The suspension
was stirred at room temperature for 24 h. The platinated surfaces
were recovered by centrifugation followed by continuous washing
with CHCl3 to remove the unbound PtCl2 and dried in a nitrogen
stream.
2
.2.4. Calcination and reduction of the Pt-grafted silica supports
The silica supports grafted with Pt were calcined in a reduction
2
. Experimental
oven in air by increasing the temperature with 5 °C per minute to
250 °C and the same temperature was maintained for 3 h. This was
followed by subsequent cooling to room temperature. For the
reduction step, the calcined sample were heated in a reduction
oven under flowing H2 by increased the temperature with 5 °C
per minute to 350 °C and the same temperature was maintained
for 3 h. The samples were stored in a glove box.
2
.1. Materials and spectroscopic characterisation
Liquid and solid reagents were purchased from Sigma Aldrich
and Merck Chemicals and were used without further purification.
X-ray photoelectron Spectroscopic (XPS) measurements was con-
ducted on a PHI 5000 Versaprobe system, equipped with a
monochromatic Al Ka X-ray source (Al Ka = 1486.6 eV). Operating
conditions and settings are similar as reported in previous work
from this lab [45–49]. SEM images was captured using a Shimadzu
Superscan ZU SXX-550 electron microscope. TEM was performed
with a Phillips (FEI) CM100 equipped with a Megaview III digital
2
2
.3. Catalytic test reactions
.3.1. Hydrogenation of 1-cyclohexene
The silica supported Pt catalysts were activated before the
2
1
camera and coupled to an Oxford X-Max (80 nm ), energy-disper-
catalysis was commenced. 1-Cyclohexene (1 ml) dissolved in
sive X-ray spectroscope (EDS). ATR-FTIR spectra were recorded
from neat samples on a Digilab FTS 2000 Fourier transform spec-
trometer utilizing a He-Ne laser at 632.6 nm. TGA experiments
were performed on a TGA/SDTA851e instrument. All porosity and
surface area properties were determined on a Micromeritics ASAP
n-hexane (3 ml) were added to the activated catalyst in the reactor
followed by flushing of H three times. The reactor was sealed under
2
1
The silica supported Pt catalysts (10 mg) was loaded in a Parr Reactor. The
three
catalyst was activated by flushing the reactor containing the catalyst with H
times. The Pt was reduced at 150 ºC and 15 bar H pressure overnight.
2
2
020.
2
2

Z. Xantini and E. Erasmus
Polyhedron 193 (2021) 114769
H
2
at 15 bar. The reaction was carried at 60 °C for 6 h which stirring
77 K, see Table 1 for the data. The N
nanosilica surfaces are shown in the Supplementary information.
DFNS gave the highest surface area, 463.4 ± 3.24 m
2
isotherms for the DFNS and
1
at 250 rpm. The crude reaction mixture was analysed by H NMR.
2
ꢀ1
g , while
2
ꢀ1
the nanosilica had a surface area of 142.50 ± 1.28 m ꢂg
.
2.3.2. Hydrogenation benzaldehyde
From the substantial difference in size and surface area, we will
be able to investigated the influence of the smaller size of the
nanosilica and the increased surface area of the DFNS on Pt-loading
and eventually catalysis.
Benzaldehyde (1 ml) dissolved in n-hexane (3 ml) were added
to the activated Pt catalyst* in the reactor followed by flushing of
three times. The reactor was sealed under H at 15 bar. The
H
2
2
reaction was carried at 60 °C for 6 h which stirring at 250 rpm.
The crude reaction mixture was analysed by H NMR.
1
The silica supports are functionalised with Pt, using different
two methods (see Scheme 1). Direct grafting of PtCl on the silanol
2
groups of the supports (to produce Pt-O bonds), this resulted in the
Pt-supported nanosilica, Pt/NP, and the Pt-supported fibrous silica,
2
.3.3. Hydrogenation cinnamaldehyde
Cinnamaldehyde (1 ml) dissolved in n-hexane (3 ml) were
Pt/DFNS. The second route involves the immobilisation of PtCl
2
added to the activated Pt catalyst* in the reactor followed by flush-
ing of H three times. The reactor was sealed under H at 15 bar.
The reaction was carried at 60 °C for 6 h which stirring at
onto the surfaces functionalised with polyethylenimnine (PEI)
linkers (this produces Pt-N bonds). This resulted in the Pt-sup-
ported via PEI nanosilica, Pt/PEI/NP, and the Pt-supported via PEI
fibrous silica, Pt/PEI/DFNS.
2
2
1
2
50 rpm. The crude reaction mixture was analysed by H NMR.
The PEI functionalised supports was prepared by first attaching
the (3-glycidyloxypropyl) trimethoxysilane (GTMS) onto the sila-
nol groups. The branched polyethylenimine (PEI) was anchored
via an epoxy-amine cross-linking reaction of the epoxide group
on the GTMS and an amine group on the PEI.
3
. Results and discussion
Fibrous silica was prepared according to the adapted method of
Dong et al. [54] The preparation of fibrous nanosilica (DFNS)
involves the formation of microemulsions. Cetyltrimethylammo-
nium (CTAB) is used as a surfactant, which directs the structure
of the silicate molecules (TEOS is used as a silica precursor) and
is eventually removed during calcination [51], while the urea is
used as the hydrolysing agent [52]. The transmission electron
microscope (TEM) and scanning electron microscope (SEM) images
of the as-prepared DFNS, Fig. 1, shows well-defined and ordered
fibres, which are evenly distributed around the particle. The aver-
age particle size is 503 nm and from the TEM image it appears that
the fibres cover the outer ca. 120 nm of the particles. The particle
size distribution curves and statistical data obtained from the TEM
images are also shown in Fig. 1.
The functionalisation of the silica supports with PEI, decreased
the surface area. PEI-functionalised DFNS has a surface area of
2
ꢀ1
7
0.34 ± 0.57 m
g , (an 85% decrease in surface area compared
to unmodified DFNS. While the PEI-functionalised nanosilica has
2
ꢀ1
a surface area of 126.91 ± 0.00 m g , an 11% decrease in surface
area as compared to the unmodified nanosilica. After PEI-modifica-
tion the nanosilica has a larger surface area than the DFNS. The BET
surface area and pore volume of all four of the Pt-functionalised sil-
ica supports and their precursors are summarised in Table 1. The
2
N isotherms for PEI- and Pt-functionalised supports are shown
in the Supplementary information. From the comparison of the
SEM images of the DFNS before (Fig. 1) and after PEI-functionalisa-
tion, it looks like the PEI pseudo-ligand closes the pores (which
dramatically deceased the surface area), however, the structure
remains intact.
Since the chemisorption technique measures only the exposed
metal not the bulk, the amount of exposed Pt atoms per g of cata-
lysts could be determined and as a consequence per 10 mg, which
is the amount used during the catalytic reactions (this will be used
to determine the product formation frequency, PFF). The atomic
For comparison with the purchased nanosilica, the TEM images
along with the particle size distribution curves and statistical data
are presented in Fig. 2. The mean particle size of the nanosilica is
ca. 21.4 nm. The nanosilica is one order of magnitude smaller than
the DFNS.
The BET surface area and pore volume was determined using
2
accelerated surface area and porosity analyses (ASAP) with N at
0
ꢀ10
radius of Pt is 1.77 ꢁ 10
m, thus the surface area of a Pt atom
ꢀ
20
2
2
is 9.84 ꢁ 10
m (as calculated with a = pr ). The metallic area of
the sample (of the catalyst, Pt on the support) was determined
from ASAP, dividing this metallic area with surface area of one Pt
atom, the amount of Pt atoms per gram of sample (the catalyst)
could be determined. As an example, the calculations for Pt/DFNS
will be shown below.
amountofPtatomspergramofcatalyst
2
ꢀ1
Metallicareaofsampleðm g
Þ
¼
2
surfaceareaofonePtatomðm Þ
2
ꢀ1
amountofPtatomspergramofcatalyst ¼ 0
:0133m g
1
7
9
:84x10 m2
1
7
ꢀ1
¼
1:35x10
g
1
7
ꢀ1
amountofPtatomsper10mgofcatalyst ¼ 1:35x10
g
x0:01
1
5
ꢀ1
¼
1:35x10 ð10mgÞ
2
The amount of Pt-atoms per m , was determined by dividing
the amount of Pt-atoms per g by the BET surface area (measured
g ). Even though the nanosilica has a smaller surface area
than DFNS, more Pt atoms was loaded onto it per m . After
2
ꢀ1
in m
Fig. 1. Top: SEM (Left) and TEM (Right) images of the as prepared DFNS. Bottom:
The particle diameter (ø), and the histogram of DFNS.
2
3

Z. Xantini and E. Erasmus
Polyhedron 193 (2021) 114769
Fig. 2. The TEM (Left) of the purchase nanosilica, the particle diameter (ø), and the histogram.
Table 1
BET surface area, micropore volume and dispersion values obtained from accelerated surface area and porosity analyses (ASAP) with N
2
at 77 K. As well as the surface area data
obtained from chemisorption of CO at 35 °C, the % dispersion, metallic area of the sample, metallic area of the metal, Pt atoms per g, Pt atom per 10 mg (the amount of catalyst
used), mol Pt per 10 mg of catalyst and amount of Pt atoms per m2.
Sample
Physisoption data
BET Surface
Chemisorption data
Metallic area of
Pt atoms /
m
2
t-plot Micropore
Dispersion
(%)
Metallic area of
metal (m /g)
Pt atoms /g Pt atoms /
10 mg
mol Pt /
10 mg
2
ꢀ1
3
ꢀ1
2
2
Area (m .g
)
Volume (cm .g
)
sample (m /g)
DFNS
463.4 ± 3.24
0.086
0.0002
0.006
0.023
0.013
0.018
0.025
0.065
17
17
17
17
15
15
15
15
2.24 ꢁ 10ꢀ9 2.91 ꢁ 1014
5.65 ꢁ 10ꢀ9 4.83 ꢁ 1015
1.06 ꢁ 10ꢀ9 4.49 ꢁ 1015
1.40 ꢁ 10ꢀ9 6.67 ꢁ 1015
Pt/DFNS
PEI/DFNS
Pt/PEI/DFNS
NP
Pt/NP
PEI/NP
134.15 ± 0.85
70.34 ± 0.57
92.04 ± 0.71
142.50 ± 1.28
146.63 ± 1.55
126.91 ± 0.00
99.27 ± 0.86
0.1535
0.2883
0.2250
0.3371
0.0133
0.0335
0.0630
0.0832
0.3790
0.7119
0.6297
0.8324
1.35 ꢁ 10
3.40 ꢁ 10
6.40 ꢁ 10
8.46 ꢁ 10
1.35 ꢁ 10
3.40 ꢁ 10
6.40 ꢁ 10
8.46 ꢁ 10
Pt/PEI/NP
PEI-functionalisation, the PEI/DFNS showed a ca. 16 times increase
Adventitious oxygen and carbon are always present on all sam-
ples. This adventitious carbon is used to correct for any charging
effects observed in all the other measured peaks of importance.
The lowest binding energy of the fitted adventitious C 1 s peak
was set at 284.8 eV. The Si 2p photoelectron envelope showed a
maximum counts per seconds at ca. 102.4 eV for all four catalysts.
The PEI-containing catalyst showed the additional N 1 s photoelec-
tron line at ca. 398.5 eV. The N 1 s photoelectron envelope could be
simulated with three peaks representing the 1°, 2° and 3° amines
at ca. 397.6, 398.9 and 400.7 eV, respectively. The XPS spectra of
the C 1 s, Si 2p and N 1 s photoelectron lines from PEI/DFNS is
shown in the supplementary information as an example.
2
2
in Pt loading per m , while PEI/NP Pt loading per m increased ca.
.5 times. This clearly indicates that the functionalisation with the
1
2
PEI increases Pt loading per m .
To further confirm the success of the covalent attachment of
PEI, ATR FTIR analysis was conducted. The ATR FTIR spectra of
the as-prepared DFNS, the nanosilica as well as the PEI-modified
supports are shown in Fig. 3. The absorption bands of the asym-
metric stretch and distortion forms of the siloxane (AOASiAOA)
ꢀ1
groups are in the range 1040–1050 cm , while, the band at
ꢀ
1
8
[
07 cm
is assigned to the symmetric stretch of AOASiAOA
ꢀ1
53]. The broad absorption band at the 3000–3600 cm region is
characteristic of the OH peak from the Si-OH atoms and the
HAOAH anti-symmetric and symmetric stretching frequencies
from the water molecules adsorbed on silica. From the ATR FTIR
of DFNS and nanosilica, it is evident that silanol groups are already
present on the surface, most probably formed during preparation.
The appearance of additional bands upon GTMS and PEI func-
tionalisation of the silica supports is indicated in the blocks in
the ATR FTIR spectra (Fig. 3). The CAH vibrations due to the pres-
ence of the GTMS and PEI were observed in the range 2800–
XPS was also used to analyse the Pt-modified supports. Fig. 4
shows high resolution XPS spectra of the Pt 4f photoelectron lines
and tabulated binding energies of the as prepared Pt/DFNS as well
as in the oxidized and reduced state.
The XPS spectra of the Pt 4f photoelectron envelopes shows well
defined peaks with a spin–orbit-splitting of ca. 3.3 eV between the
Pt 4f7/2 and Pt 4f5/2. Depending on the sample and treatment, these
photoelectron envelopes were fitted with either one, two or three
sets of Gaussian peaks. The Pt 4f 7/2 photoelectron lines of the PtCl
2
ꢀ
1
ꢀ1
0
2
950 cm . The absorption bands at 1500 and 1650 cm
are
is present at ca. 73.6 eV, the PtO at ca. 72.3 eV and the Pt at ca.
71.0 eV, all in correlation with published values [59–61]. For all
the samples, oxidation only leads to the formation of ca. 85% PtO,
while the reduction seemed to be support dependent. Only ca
assigned to CAN stretching vibrations and bent NAH asymmetric
stretching from the 1° amines [54,55].
The modified silica supports were characterised by X-ray photo-
electron spectroscopy (XPS). XPS is a surface sensitive technique
that is used to determine the elemental composition and chemical
states of the elements present on solid materials. The analysis
involves the measurement of the kinetic energy of the photoelec-
trons emitted from the sample surface and translates it to the bind-
ing energy of the photoelectrons [56–58]. These binding energies
are unique for each element and also depend on the oxidation state
and chemical environment of the element. One can then deduce
the elements present on the surface and their oxidation state based
on the corresponding binding energy value.
0
82% of the Pt on the DFNS and PEI/DFNS reduced to Pt , while
the nanosilica gave 100% reduction of the Pt on both umodified
and PEI-modified nanosilica. The 100% reduction of the Pt on the
nanosilica is probably due to better exposure of the Pt to the H
2
atmosphere, since it does have a large micropore volume. For the
DFNS, some of the Pt-moieties are ‘‘hidden” in the fibres, also
explaining the presence of some PtCl
reduction.
2
after oxidation and
The Pt-supported catalysts were tested for the hydrogenation of
C@C and/or C@O bonds in cyclohexene, benzyl alcohol and
4

Z. Xantini and E. Erasmus
Polyhedron 193 (2021) 114769
Scheme 1. The anchoring of PtCl
2
directly on the silanol groups of the supports (route A) or via polyethylenimnine (PEI) linkers on the silica supports (route B).
1
cinnamaldehyde. This was done to investigated the influence of the
different solid supports (the size and surface area) as well as the
influence of the PEI addition onto the support on the hydrogena-
tion of double bonds.
Cyclohexene (3 M in n-hexane) was hydrogenated over the
in situ reduced Pt-supported catalysts (it is assumed that all Pt is
The analysis of the H NMR of the crude reaction mixtures col-
lected after 6 h showed the characteristic methylene protons of the
benzyl alcohol at 4.7 ppm and methyl protons of toluene at
2.2 ppm (see Supplementary Information). The characteristic
methyl protons of methylcyclohexane at 0.9 ppm was absent, indi-
cating that over reduction did not occur.
1
now in the metallic state) at 60 °C and 15 bar H
2
pressure. After
However, a doublet at 8.12 ppm appears in the H NMR spectra
1
6
h, the crude reaction mixture was subject to H NMR spec-
troscopy. The disappearance of the signals at 5.70, 2.0 and
.63 ppm representing the protons cyclohexene as well as the
appearance of the singlet at 1.5 ppm representing the cyclohexane,
confirmed the 100% conversion of cyclohexene to cyclohexane (see
supporting information for the H NMR spectra).
Benzaldehyde (3 M in n-hexane) was hydrogenated over the
in situ reduced Pt-supported catalysts at 60 °C and 15 bar H pres-
2
of the crude reaction mixture catalysed by the Pt/PEI/NP, Pt/DFNS
and Pt/PEI/DFNS. This doublet is assigned to the two –CH- protons
ortho to the carboxylic acid group of benzoic acid (confirmed by
1
1
comparison of with an H NMR spectra of pure benzoic acid)
[62]. The presence of the benzoic acid is unexpected in a hydro-
genation reaction. This imply that during the hydrogenation some
trace amounts of water present (probably present in the reagent or
due to complete hydrogenation of the –OH groups on the alcohols)
oxidised the benzaldehyde via the classical hydration to form a
germinal diol preceding the dehydrogenation to benzoic acid as
illustrated in Scheme 2 [63]. The characteristic singlet at 6.1 ppm
1
sure. The possible hydrogenation products that could form is ben-
zyl alcohol (form by hydrogenation of the carbonyl group), toluene
formed by the follow-up hydrogenation of the –OH of the benzyl
(
alcohol) and the over reduced product methylcyclohexane.
5

Z. Xantini and E. Erasmus
Polyhedron 193 (2021) 114769
The % conversion of benzaldehyde and % selectivity towards
benzyl alcohol, toluene and benzoic acid for each catalyst is given
ꢀ1
in Table 2 as well as the product formation frequency, PFF, (in s ).
PFF is calculated using the following equation:
!
amountofmoleculesofproductformed
amountofPtatoms
PFF ¼
timeðsÞ
The calculation of the PFF for benzyl alcohol over Pt/PEI/NP will
be discussed as an example. The total amount of benzaldehyde
2
1
molecules is 8.42 ꢁ 10 (1 ml benzaldehyde = 1.04 g = 9.8 ꢁ
ꢀ
3
21
1
0
mol = 8.42 ꢁ 10 atoms). Multiplying this value with the %
converted to benzyl alcohol, which is 21.3% (38% total conver-
sion ꢁ 56% selectivity towards benzyl alcohol = 21.3%) gives
2
1
2
.72 ꢁ 10 molecules which was converted to form benzyl alco-
hol. From the chemisorption data in Table 1 (where the amount
of Pt atoms per 10 mg was determined), the amount of Pt atom
present in Pt/PEI/NP was 6.40 ꢁ 10 . Inserting these values in
the above equation, where time is 21 600 s (6 h) the PFF could
be calculated.
1
5
Pt/PEI/DFNS gave the highest total conversion, best selectivity
and PFF towards benzyl alcohol. The second catalytic cycle using
Pt/PEI/DFNS gave only a 6.5% conversion to benzyl alcohol. This
is one order of magnitude smaller than the fresh catalyst, indicat-
ing poor recyclability. The used catalyst (after the first catalytic
cycle) was characterised by SEM and XPS (see Supplementary
Information). Visually the structure of the DFNS remains intact,
but the fibres seem to be more separated from each other. While
the XPS of the used catalysts showed that all Pt present was in
the metallic form. This is expected to increase the catalytic activity,
which was not the case. However, from the ratios (obtained from
the XPS) of Si:N:Pt of Pt/PEI/DFNS before (1:1.4:0.88) and after
Fig. 3. ATR FTIR spectra from the top to the bottom, of the GTMS and PEI-modified
nanosilica (PEI/NP), nanosilica (NP), the PEI-modified fibrous silica (PEI/DFNS), and
fibrous silica (DFNS).
of the methylene proton of the diol intermediate was not detected
1
in H NMR.
(
1:0.2:0.2) catalysis, it is clear that there was a dramatic decrease
Fig. 4. XPS data of Pt 4f photoelectron lines for the as-prepared Pt/DFNS, as well as after calcination and reduction.
6

Z. Xantini and E. Erasmus
Polyhedron 193 (2021) 114769
Scheme 2. Schematic representation of the benzaldehyde dehydration to benzoic
acid.
in N- and Pt-content after catalysis. This could possible explain the
decrease in catalytic activity.
Pt/DFNS gave the second highest conversion, however consider-
ing that it was 88.6% selective towards the formation of benzoic
acid, which is an uncatalysed hydration/dehydrogenation side-
reaction, the total catalysed conversion is only 4.4%.
The general trend seems to be the higher that amount of Pt
atoms present on the support, the higher the catalysed conversion
Scheme 3. Reaction scheme for the hydrogenation of cinnamaldehyde.
similar to the formation of the benzyl alcohol (the scheme is given
in the supplementary information) [68,69]. The intermediated diol
was not detected in the NMR spectra.
(
not considering the formation of benzoic acid). The exception to
The bar graphs in Fig. 5 indicates the % selectivity and PFF (sꢀ1)
toward the different products of the hydrogenation of cinnamalde-
hyde over each of the catalyst. With the exception of Pt/DFNS,
which favour the formation of 3-phenyl propanol, the selectivity
towards cinnamyl alcohol (the most desired product) is favoured
by all the catalysts. Pt/DFNS showed the highest PFF for both
3-phenyl propanol and cinnamyl alcohol.
this rule is Pt/PEI/DFNS, which gave much higher conversion. The
bar graphs showing the % selectivity and PFF is given in the supple-
mentary information.
Cinnamaldehyde has two double bonds (C@C and C@O), which
are susceptible and in competition for hydrogenation. The hydro-
genation of cinnamaldehyde can proceed via three different path-
ways as shown in Scheme 3. The hydrogenation of the carbonyl
group (C@O) to form cinnamyl alcohol, an unsaturated alcohol
The success of a catalyst is measured by its % selectivity and/or
the PFF towards the desired product. For the hydrogenation of ben-
zaldehyde and cinnamaldehyde, the desired products are benzyl
alcohol and cinnamyl alcohol, respectively. The factors that can
influence the % selectivity and/or the PFF towards the desired pro-
(
pathway 1). The hydrogenation of the carbon–carbon double bond
to form phenyl propionaldehyde, a saturated aldehyde (pathway
), or the complete hydrogenation of both functional groups to
2
form 3-phenyl propanol (pathways 3).
2
ꢀ1
The selectivity for the hydrogenation of cinnamaldehyde has
been reported as challenging in literature [64–66]. because of the
different intermediates and by products that may form. The hydro-
genation of the C@C bond is thermodynamically favoured by about
duct is: the total BET surface area in m
g
of the support (A),
2
ꢀ1
diameter of the support (D) and/or the Pt surface area in m
g
(P). Each individual factor’s influence on % selectivity and/or the
PFF towards the desired product as well as different combinations
of the factor’s influence on % selectivity and/or the PFF towards the
desired product was compared (see supplementary information). A
significant influence is measured by how good a linear relationship
3
5 kJ/mol over the hydrogenation of C@O [67], which makes the
selectivity of this reaction even more difficult to achieve.
A 2.6 M cinnamaldehyde in isopropanol was hydrogenated for
2
6
h in a pressurised vessel at 15 bar H
2
and 150 °C, over pre-
could be fitted to the data, a R value > 0.8 is considered significant.
1
reduced Pt/NP, Pt/DFNS, Pt/PEI/NP and Pt/PEI/DFNS. H NMR
analysis was used to determine the % conversion and product
distribution of the crude reaction mixture. See the supplementary
From this the data compared it can be concluded that the indi-
vidual or combination of factor’s (A, D and P) has no significant
influence on the % selectivity towards the desired products.
The PFF for benzyl alcohol is influenced by the combination of
the D ꢁ P as well as the combination of A ꢁ D ꢁ P. The combination
of the selectivity and PFF towards benzyl alcohol is influenced by
the combinations of D ꢁ P as well as A ꢁ D ꢁ P.
1
information for the H NMR spectra and Table 3 for the data.
1
The characteristic H NMR signals used to determine the
product distribution is: the singlet aldehyde signal at 9.8 ppm of
3
-phenyl propionaldehyde, doublet signal of the aldehyde of cin-
namaldehyde at 9.7 ppm, two sets of the doublet signals centred
at 7.8 and 6.49 ppm representing the olefinic methine protons of
cinnamic acid, a set of doublets centred at 4.3 ppm assigned to
The PFF of the formation of cinnamyl alcohol is dependent on D,
as well as the combination of the A ꢁ D as well as A ꢁ P. The com-
bination of PFF and % selectivity towards the cinnamyl alcohol is
dependant D, as well as the combination of the A ꢁ D ꢁ P.
From this it can be concluded that obtain high selectivity
towards the desired product as well as a good PFF, all the factors
that can influence the activity is important to consider, no single
factor is more important than the others. Thus a high surface area
of the support, a small diameter of the support and high Pt-loading.
the methylene protons
and lastly a triplet centred around 3.7 ppm representing the
methylene protons to the –OH group of 3-phenyl propanol.
a to the –OH group of cinnamyl alcohol,
a
The presence of cinnamic acid in the crude reaction mixture
implies that during catalytic hydrogenation, cinnamaldehyde is
oxidised. This oxidation occurs probably via the classical hydration
to the germinal diol followed by dehydrogenation to cinnamic acid
Table 2
ꢀ
1.
The % conversion and % selectivity towards benzyl alcohol and toluene per catalyst of the hydrogenation of benzaldehyde as well as the product formation frequencies (PFF) in s
Catalyst
% Conversion
% Selectivity
PFF (s 1
ꢀ
)
toluene
PFF (s
ꢀ1
)
benzoic acid
PFF (s )
ꢀ1
benzyl alcohol
Pt/NP
Pt/PEI/NP
Pt/DFNS
Pt/PEI/DFNS
Used Pt/PEI/DFNS
12.4
38.0
38.5
62.9
6.5
76.7
56.0
6.5
73.4
100
5.78
9.82
7.22
52.97
149.05
23.3
>0.1
4.9
1.4
–
1.77
–
5.49
1.03
–
>0.1
44.0
88.6
20.2
>0.1
–
7.69
98.46
18.11
–
7

Z. Xantini and E. Erasmus
Polyhedron 193 (2021) 114769
Table 3
The % conversion and % selectivity towards (3-phenyl propionaldehyde, cinnamyl alcohol and 3-phenyl propanol) of each catalyst form cinnamaldehyde.
%
Conversion % Selectivity
-Phenyl propionaldehyde PFF (s 1
ꢀ
ꢀ1
ꢀ1
ꢀ1
3
)
Cinnamyl alcohol PFF (s
)
3-Phenyl propanol PFF (s
)
Cinnamic acid PFF (s )
Pt/NP
Pt/PEI/NP
Pt/DFNS
52.6
98.1
68.0
9.9
>0.1
0.3
1.80
–
0.33
2.61
67.2
70.7
44.4
70.9
12.26
18.19
49.61
42.79
22.9
25.4
50.0
24.8
4.19
6.53
55.85
14.94
>0.1
3.9
5.3
–
0.99
5.91
–
Pt/PEI/DFNS 92.5
4.3
>0.1
Fig. 5. Graphic representation of the % selectivity and PFF (sꢀ1) toward the different products of the hydrogenation od cinnamaldehyde over each of the catalyst.
4
. Conclusion
References
[
[
[
1] M. Dhiman, V. Polshettiwar, J. Mater. Chem. A 4 (2016) 12416–12426.
2] J. Niemantsverdriet, Appl. Surf. Sci. 144–145 (1999) 366–374.
3] V. Lordi, N. Yao, J. Wei, Chem. Mater. 13 (2001) 733–737.
We have prepared Pt supported on nanosilica and DFNS, which
either anchored directly on to the support or via a polyethylen-
imine (PEI) linker. The PEI linker increase the Pt-loading on the
supports up to 16 times than direct binding. These catalysts dis-
played 100% conversion for the hydrogenation of cyclohexene to
cyclohexane. For the hydrogenation of benzaldehyde, a conversion
of upto 62.9% was obtained with a selectivity of 73.4% towards the
[4] E. Finocchio, E. Macis, R. Raiteri, G. Busca, Langmuir 23 (2007) 2505–2509.
[
[
[
5] M.A. Bodaghifard, H. Moghanian, A. Mobinikhaledi, F. Esmaeilzadeh, Inorg.
Nano-Metal Chem. 47 (2017) 845–849.
6] E. Erasmus, P.C. Thüne, M.W.G.M. Verhoeven, J.W. Niemantsverdriet, J.C.
Swarts, Catal. Commun. 27 (2012) 193–199.
7] J. Xie, Y. Wang, Y. Li, Y. Wei, Springer. 102 (2011) 143–154.
ꢀ
1
[8] S. Dwivedi, S.S. Gujral, T. Taniike, M. Terano, Pure Appl. Chem. 85 (2012) 1–9.
[
formation of benzyl alcohol for Pt/PEI/DFNS, with a PFF of 52.9 s
.
9] X. Tu, Z. Li, J. Lu, Y. Zhang, G. Yin, W. Wang, D. He, RSC Advances. 8 (2018)
2887–2891.
The formation of benzoic acid was observed in same catalytic reac-
tion, which is attributed to the presence of small amount of water,
resulting in the classical hydration to form a germinal diol, fol-
lowed by dehydration to benzoic acid. The Pt/PEI/NP catalyst
resulted in a high conversion of 98.1% for the hydrogenation of cin-
namaldehyde. This catalyst was 70.7% selective towards cinnamyl
aldehyde, 25.4% selective towards the formation of 3-phenyl pro-
[10] G.L. Witucki, J. Coat. Technol. 65 (1993) 57–60.
11] W. Huang, J.N. Kuhn, C. Tsung, Y. Zhang, S.E. Habas, P. Yang, G.A. Somorjai, Am.
Chem. Soc. 8 (2008) 2027–2034.
12] P. Gautam, M. Dhiman, V. Polshettiwar, B.M. Bhanage, R. Soc. Chem. 18 (2016)
5890–5899.
13] E. Erasmus, J.W. Niemantsverdriet, J.C. Swarts, Langmuir 28 (2012) 16477–
[
[
[
[
16484.
14] S.B. Kalidindi, B.R. Jagirdar, ChemSuSChem 5 (2012) 65–75.
[15] J. Haber, Pure & Appl. Chem. 63 (1991) 1227–1246.
16] B.G. Trewyn, S. Giri, I.I. Slowing, V.S.Y. Lin, Chem. Commun. (2007) 3236–3245.
17] Y. Xia, B. Gates, Y. Yin, Y. Lu, Adv. Mater. 12 (2000) 693–713.
18] J.L. Gole, Z.L. Wang, Z.R. Dai, J. Stout, M. White, Colloid Polym. Sci. 281 (2003)
673–685.
ꢀ1
panol with a PFF 18.9 s after 6 h. Since no 3-phenyl propionalde-
hyde was formed, a shorter reaction time might lead to a higher
conversion of cinnamyl aldehyde. In a similar type of hydration,
followed by dehydration, cinnamic acid was produced in small
quantities.
[
[
[
[
[
[
19] S.M. Sadeghzadeh, R. Zhiani, M. Khoobi, S. Emrani, Micropor Mesopor Mater.
2
57 (2018) 174–1153.
20] M.J. Ortiz-Iniesta, H.J. Heeres, I. Melian-Cabrera, Micropor Mesopor Mater. 171
2013) 208–214.
21] Y. He, A. Jawad, X. Li, M. Atanga, F. Rezaei, A.A. Rownaghi, J. Catal. 341 (2016)
49–159.
[22] S.M. Sadeghzadeh, R. Zhiani, M. Khoobi, S. Emrani, Micropor Mesopor Mater.
57 (2018) 147–153.
23] A. Maity, V. Polshettiwar, ChemSusChem. 10 (2017) 3866–3913.
24] B. Singh, V. Polshettiwar, J. Mater. Chem. A (2016) 1–15.
25] M. Dhiman, B. Chalke, V. Polshettiwar, ACS Sustain. Chem. Eng. 3 (2015) 3224–
(
CRediT authorship contribution statement
1
Z. Xantini: Investigation. E. Erasmus: Conceptualization,
Validation, Formal analysis, Resources, Writing - review & editing,
Supervision.
2
[
[
[
3
230.
Declaration of Competing Interest
[26] C. D. Seaborn and F. H. Nielsen, Biological Trace Element Research., 3, 251-261.
27] D.S. Moon, J.K. Lee, Langmuir. 28 (2012) 12341–12347.
[
[
28] V.V. Pushkarev, Z. Zhu, K. An, A. Hervier, G.A. Somorjai, Top Catal. 55 (2012)
The authors declare that they have no known competing finan-
cial interests or personal relationships that could have appeared
to influence the work reported in this paper.
1257–1275.
[
29] K. Niknam, A. Deris, F. Panahi, Chin. J. Catal. 34 (2013) 718–722.
[30] X. Zhai, Y. Pan, J. Xu, A. Wang, S. Wu, J. Shen, RSC Adv. 5 (2015) 105855–
05861.
1
[
31] D. Xia, J. Quan, G. Wu, X. Liu, Z. Zhang, H. Ji, D. Chen, L. Zhang, Y. Wang, S. Yi, Y.
Zhou, Y. Gau, R. Jin, Nanomaterials 9 (2019) 1225.
Acknowledgement
[
[
[
32] M. Englisch, A. Jentys, J.A. Lercher, J. Catal. 166 (1997) 25–35.
33] J. Dai, M.L. Bruening, Nano Lett. 2 (2002) 497–501.
34] E.M.E. van Kimmenade, A.E.T. Kuiper, Y. Tamminga, P.C. Thune, J.W.
Niemantsverdriet, J. Catal. 223 (2004) 134–141.
The author would like to acknowledge generous financial sup-
port from Sasol, the NRF and UFS during the course of this study.
8

Z. Xantini and E. Erasmus
Polyhedron 193 (2021) 114769
[
35] F.A. Cotton, G. Wilkinson, P.L. Gaus, Basic Inorganic Chemistry, 3rd Ed., John
Wiley & Sons Inc, United States, 1976.
[53] A. Mobinikhaledi, H. Moghanian, S. Pakdel, Chin. Chem. Lett. 26 (2015) 557–
563.
[
[
[
36] V.I. Okafor, N.J. Coville, S. Afr, J. Sci. 95 (1999) 503–508.
37] W. Yu, M.D. Porosoff, J.G. Chen, Chem. Rev. 112 (2012) 5780–5817.
38] C. Lian, H. Liu, C. Xiao, W. Yang, K. Zhang, Y. Liu, Y. Wang, Chem. Commun. 48
[54] M. Yamaura, R.L. Camilo, L.C. Sampaio, M.A. Macedo, M. Nakamura, H.E. Toma,
J. Magn. Magn. Mater. 279 (2004) 210–217.
[55] R.A. Bini, R.F.C. Marques, F.J. Santos, J.A. Chaker, J. Magnet. Magnet. Mater. 324
(2012) 534–539.
(
2012) 3124–3126.
39] Z.X. Chen, G.C. Smith, C.A.J. Putman, E.J.M. ter Voert, Catal. Letters. 50 (1998)
9–57.
[
[
[56] H. Kanno, Mater. Sci. Eng. Carbon (2016) 153–171.
4
[57] C. Lee, Y. Hyun, J. Choi, Carbon Nano-Objects. (2019) 225–242.
[58] K. Zhu, C. Chen, Interf. Sci. Technol. (2019) 281–330.
[59] F.B. Li, X.Z. Li, Chemosphere 48 (2002) 1103–1111.
[60] R. Fu, H. Zeng, Y. Lu, S.Y. Lai, W.H. Chan, F. Ng, Carbon 33 (1995) 657–661.
[61] G. Zhang, D. Yang, E. Sacher, J. Phys. Chem. C 111 (2007) 565–570.
[62] S.M. Kim, Y.S. Kim, D.W. Kim, J.W. Yang, Green Chem. (2012) 2996–2998.
[63] P. Ferreira, A. Hern Andez-Ortega, B.B. Herguedas, J. Rencoret, A. Guti Errez, M.
Martinez, J. Jimez-barbero, M. Medina, A.T. Mertinez, Biochem. J. 425 (2010)
585–593.
[64] N. Magqi, Preparation, characterization and selective hydrocarbon
transformation applications of flat-model and zeolite –supported catalyst
PhD thesis, University of the Free State, Bloemfontein, 2014.
[65] R.M.B. Basale, Selective catalytic hydrogenation of,-unsaturated aldehyde to
unsaturated alcohol: investigation of the role of the promoter, Eindhoven
University of Technology, Netherlands, 2013.
40] N. Magqi, Preparation, characterization and selective hydrocarbon
transformation applications of flat-model and zeolite supported catalysts
PhD thesis, University of the Free State, Bloemfontein, 2014.
[
[
41] F. Zaera, ACS Catal. 7 (2017) 4947–4967.
42] K. Kaviyarasu, E. Manikandan, Z.Y. Nuru, M. Maaza, Mater. Lett. 160 (2015) 61–
6
3.
[
[
[
[
43] P. Gallezot, D. Richard, Catal. Rev. Sci. Eng. 40 (1998) 81.
44] Y. Yuan, S. Yao, M. Wang, S. Lou, N. Yan, Curr. Org. Chem. 17 (2013) 400.
45] L.K. Parrott, E. Erasmus, Catal. Lett. 48 (2018) 2008–2018.
46] E. Botha, M. Landman, P.H. van Rooyen, E. Erasmus, Inorg. Chim. Acta 482
(
2018) 514–521.
[
[
[
[
47] E. Erasmus, J. Elec. Spec. Rel. Phenom. 223 (2018) 84–88.
48] S.J. Gerber, E. Erasmus, Mater. Chem. Phys. 203 (2018) 73–81.
49] S.J. Gerber, E. Erasmus, Trans. Metal Chem. 43 (2018) 409–420.
50] Z. Dong, X. Le, X Li, W. Zhang, C. Dong and J. Mia, Appl. Catal. B, Enviro. 2014,
[66] H. Liu, Q. Mei, S. Li, Y. Yang, Y. Wang, H. Liu, L. Zheng, P. An, J. Zhang, B. Han,
Chem. Commun. 54 (2018) 908–911.
1
58-159, 135-159.
[
[
51] N. Bayal, B. Singh, R. Singh, V. Poahettiwar, Sci. Rep. 6 (2016) 1–11.
52] V. Polshettiwar, D. Cha, X. Zhang and J.M. Basset., Angew.Chem.Int.Ed., 2010,
[67] C. Mohr, P. Claus, Sci. Progr. 84 (2001) 311.
[68] T. Mallat, Z. Bodnar, P. Hug, A. Baiker, J. Catal. 153 (1995) 131–143.
[69] T. Mallat, A. Baiker, Chem. Rev. 104 (2004) 3037–3058.
4
9, 9652-9656.
9