Active site isolation in bismuth-poisoned Pd/SiO2 catalysts for selective hydrogenation of furfural

Article abstract of DOI:10.1016/j.apcata.2018.11.016

Active site isolation in furfural (FA) hydrogenation was studied by poisoning a Pd catalyst with bismuth. A solution of FA in water was hydrogenated over a 5 wt% Pd/SiO₂ catalyst in a batch reactor at various reaction temperatures and pressures. Furfuryl alcohol (FAL) was an intermediate product which was further hydrogenated into tetrahydrofurfuryl alcohol (TFAL) or cyclopentanone (CPA) and cyclopentanol (CPOL). While application of hydrogen pressure above 30 bar had little effect on the hydrogenation kinetics, a reaction temperature affected product distribution and the main product changed from TFAL (at 50 °C) to FAL (100 and 150 °C). Poisoning the catalyst with Bi decreased the number of available active sites but had little effect on the turn-over frequencies, most likely because of the absence of electronic effects of Bi on Pd nanoparticles. The main reaction product over the Bi-poisoned catalyst was FAL with no FA oligomerisation products. At a reaction temperature of 150 °C, CPA was formed with a 57% yield. Considering that Bi preferentially poisons step sites of Pd, the comparison of the product distribution between the Pd and Pd-Bi catalyst as well as the literature data for the alloy Pd-Cu catalysts indicates that the active site isolation observed in the Pd-Bi catalysts is responsible for the increasing FAL and CPA selectivities and elimination of oligomer by-products.

Full text of DOI:10.1016/j.apcata.2018.11.016

Accepted Manuscript
Title: Active site isolation in bismuth-poisoned Pd/SiO2
catalysts for selective hydrogenation of furfural
Authors: Nikolay Cherkasov, Antonio Jos e´ Exp o´ sito, Moom
Sinn Aw, Javier Fern a´ ndez-Garc ´ı a, Steve Huband, Jeremy
Sloan, Larysa Paniwnyk, Evgeny V. Rebrov
PII:
S0926-860X(18)30571-4
DOI:
Reference:
https://doi.org/10.1016/j.apcata.2018.11.016
APCATA 16885
To appear in:
Applied Catalysis A: General
Received date:
Revised date:
Accepted date:
24 August 2018
2 November 2018
20 November 2018
Please cite this article as: Cherkasov N, Exp o´ sito AJ, Aw MS, Fern a´ ndez-Garc ´ı a J,
Huband S, Sloan J, Paniwnyk L, Rebrov EV, Active site isolation in bismuth-poisoned
Pd/SiO catalysts for selective hydrogenation of furfural, Applied Catalysis A, General
2
(2018), https://doi.org/10.1016/j.apcata.2018.11.016
This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.

Active site isolation in bismuth-poisoned Pd/SiO2
catalysts for selective hydrogenation of furfural
a,b,
a
c
d
e
Nikolay Cherkasov *, Antonio José Expósito , Moom Sinn Aw , Javier Fernández-García , Steve Huband ,
e
c
a,b,f,
Jeremy Sloan , Larysa Paniwnyk , Evgeny V Rebrov
*
a
School of Engineering, University of Warwick, Coventry CV4 7AL, UK
Stoli Catalysts Ltd, Coventry, CV3 4DS, UK
School of Life Sciences, Department of Health and Life Sciences, Coventry University, Alison Gingell
b
c
Building, Priory Street, Coventry CV1 5FB, UK
d
School of Chemical and Process Engineering, University of Leeds, LS2 9JT, UK
Department of Physics, University of Warwick, Coventry CV4 7AL, UK
Department of Biotechnology and Chemistry, Tver State Technical University, 170026, Nab. A. Nikitina 22,
e
f
Russia, E-mails: n. cherkasov@warwick.ac.uk, e.rebrov@warwick.ac.uk
Graphical abstract
Highlights



The Pd and Bi-poisoned Pd catalysts were studied in liquid-phase furfural hydrogenation under a
range of temperatures and pressures
Bi-poisoned catalysts completely suppressed oligomerisation of furfural and increase the selectivity
to furfuryl alcohol hydrogenation
Detailed characterisation and comparison with the literature data suggests that active site isolation
plays the determinant role in the effect of Bi
Abstract
Active site isolation in furfural (FA) hydrogenation was studied by poisoning a Pd catalyst with bismuth. A
solution of FA in water was hydrogenated over a 5 wt% Pd/SiO catalyst in a batch reactor at various reaction
2
temperatures and pressures. Furfuryl alcohol (FAL) was an intermediate product which was further
hydrogenated into tetrahydrofurfuryl alcohol (TFAL) or cyclopentanone (CPA) and cyclopentanol (CPOL).
While application of hydrogen pressure above 30 bar had little effect on the hydrogenation kinetics, a
o
reaction temperature affected product distribution and the main product changed from TFAL (at 50 C) to
o
FAL (100 and 150 C). Poisoning the catalyst with Bi decreased the number of available active sites but had

little effect on the turn-over frequencies, most likely because of the absence of electronic effects of Bi on Pd
nanoparticles. The main reaction product over the Bi-poisoned catalyst was FAL with no FA oligomerisation
o
products. At a reaction temperature of 150 C, CPA was formed with a 57 % yield. Considering that Bi
preferentially poisons step sites of Pd, the comparison of the product distribution between the Pd and Pd-Bi
catalyst as well as the literature data for the alloy Pd-Cu catalysts indicates that the active site isolation
observed in the Pd-Bi catalysts is responsible for the increasing FAL and CPA selectivities and elimination of
oligomer by-products.
Keywords: hydrogenation; active site; palladium; bismuth; poison
1
. Introduction
Renewable fuels and chemicals attract considerable attention due to the depletion of traditional fossil
feedstocks [1]. Biomass, however, is highly oxygenated and de-deoxygenation is often have to be performed
in order to transform it into more useful or stable chemicals [2,3]. Therefore, biomass utilisation is closely
connected to the development of bio-refineries that are able to transform biomass into valuable products
[
4].
Furfural (FA), a platform chemical for biorefineries [5], can be obtained by the hydrolysis and dehydration of
hemicelluloses in agricultural waste [6,7]. Although utilisation of furfural as a biofuel is difficult, many
valuable products can be obtained by its hydrogenation [5,8], The FA hydrogenation products include
furfuryl alcohol (FAL)[9,10], tetrahydrofurfuryl alcohol (TFAL) [11,12], cyclopentanone (CPA) [8,13],
cyclopentanol (CPOL) [14,15], 2-methyltetrahydrofuran [16,17], 2-methylfuran [3,18], furan [19], and other
aliphatic molecules formed by the furan ring opening [20]. The number of possible products and complexity
of chemistry results in low selectivity to a particular compound [3,21]. A number of catalysts were studied to
increase the product selectivity such as Pt [13,22], Ru [3,4], Pd [23,24], Cu [25,26] and Ni [8,27]. Pd-based
catalysts showed one of the highest reaction rates and were studied in both vapour [2,28] and liquid phase
hydrogenation [13,17,29–31].
Poisoning of Pd catalysts with a second metal provides an opportunity to alter the product selectivity and
study the hydrogenation reaction pathway [7,28,31,32]. For example, alloying Pd with Ir increases
(compared to either Pd or Ir alone) the reaction rate and TFAL selectivity [31]. The change in selectivity over
the Pd-Ir alloy catalysts is caused by the adsorption via both the furan and aldehyde groups, while the
monometallic Pd and Ir catalysts adsorb furfural predominantly only via either the furan, or the aldehyde
groups. Plating Pd with Cu increases the selectivity to CPA in the aqueous-phase furfural hydrogenation [7].
The second metal can also play an active role in stabilisation of the Pd nanoparticle dimensions. A Pd-Sn/SiO
2
o
catalyst after calcination at the optimum temperature of 450 C shows an increased furfural hydrogenation
activity, which is attributed to the stabilisation of the 2.5 nm particles [33].
There is little data on the effect of Bi poisoning of Pd catalysts on FA hydrogenation in the literature. Bi,
however, is particularly interesting because the nature of the poisoning is well known. Bi atoms
predominantly block the step sites of the Pd nanoparticle leaving the flat terraces intact as confirmed by
both DFT calculations and infrared spectroscopy studies [34–36]. In our earlier reports [37,38], we
demonstrated that Bi-poisoned Pd catalysts show a significant decrease in hydrogenation rates for an
internal compared to a terminal alkene. Bi poisoning improves alkene selectivity in terminal alkynol semi-
hydrogenation with little electronic effects, thus allowing to decouple spatial and electronic effects of the
catalyst on the product selectivity. In this paper we study the effect of Bi poisoning on Pd catalysts to gain a
better understanding of spatial effects on FA hydrogenation selectivity.

2
2
. Experimental
.1.Catalyst preparation
A 5 wt% Pd/SiO
2
catalyst was prepared by wet impregnation. A solution of palladium (II) acetylacetonate
(
98%, Sigma Aldrich) in toluene (99%, VWR chemicals) filled the pores of the fumed silica (Alfa Aesar, BET
2
-1
specific surface area of 200 m g ). The impregnated silica was dried in a rotatory evaporator, calcined at
4
1
o
-1
-1
00 C for 2 hours in a tube furnace in the flow of air 100 mL min and reduced in H
2
flow of 20 mL min at
o
50 C for 1 hour. (All gas flow rates are presented at the normal temperature and pressure).
The Bi-poisoned Pd catalyst was prepared by adopting the methodology reported in references [35,37].
Briefly, the reduced 5 wt% Pd/SiO catalyst (700 mg) was dispersed in 20 mL of an aqueous 2 vol% acetic acid
solution. A freshly-prepared 20 mM aqueous Bi(NO solution in 2 vol% acetic acid was added on stirring at
500 rpm. The amount of the Bi solution was selected to obtain a Pd/Bi molar ratio of 7 assuming a
2
)
3 3
1
o
complete reduction of Bi. The slurry was then stirred for 16 hours at 80 C in a static H
cooling, H was replaced with N₂. Finally, the solvent was evaporated in a rotavapour and the sample was
passivated by slowly introducing air through a 24-gauge syringe needle into the vial with the catalyst.
2
atmosphere. On
2
2
.2.Catalyst characterisation
Transmission electron microscopy (TEM) study was carried out with a Jeol 2010 microscope. The catalysts
were suspended in ethanol under sonication and applied on carbon-coated copper grids. The electron
images were taken from 10-15 various regions of the grids and analysed using ImageJ [39]. High Resolution
TEM and Scanning Transmission Electron Microscopy (STEM) imaging studies were carried out with Jeol
ARM200F double aberration corrected instrument operated at 200 kV and equipped with a Gatan Orius
SC1000 CCD camera. Energy Dispersive X-ray (EDX) mapping studies were performed with probe currents of
approximately 200 pA and collected with an Oxford Instruments X-max Silicon Drift Detector with an area of
2
8
0 mm .
Elemental analysis was performed from the solutions of the catalysts dissolved using an inductively coupled
plasma spectrometer Perkin Elmer Optima 5300DV. The catalysts were dissolved in a mixture of HF, HCl, and
o
HNO
3
(1:1:3 volume ratio) under microwave irradiation at 200 C. The samples were diluted and compared
against standards in the concentration range of 0.1-10 ppm.
X-ray diffraction (XRD) studies were carried out with an Empyrean X-ray diffractometer equipped with
monochromatic Cu K
α
source and a linear PIXcel detector. The catalyst samples were studied in the 2θ range
o
o
of 20-85 , step length of 0.04 , and step time of 30 min. Small-angle X-ray scattering measurements were
made using a Xenocs Xeuss 2.0 equipped with a micro-focus Cu K source collimated with scatterless slits
α
placing the powder catalysts between two Kapton windows. A Pilatus 300k detector with a pixel size of 0.172
μm x 0.172 μm was used to measure scattering. The system was calibrated using silver behenate
(AgC22H
43
O ) and fitting of the data was performed in the Irena analysis package [40]. Further details are
2
provided in the Supplementary Material.
Chemisorption study of carbon monoxide was performed in a modified system described in reference [41].
-
1
-1
The samples were dried in a flow of He (20 mL min ), calcined in a flow of 5 vol% O
2
in He (20 mL min ) at
o
-1
o
3
00 C, reduced in a 1 vol% H
2
in He mixture (20 mL min ) at 350 C. On cooling to room temperature in He,
-
1
a flow of 0.1 vol% CO, 0.1 vol% Ar in He (6 mL min ) was introduced through the catalyst bed. A mass
spectrometer was used to monitor the concentrations at the outlet, with Ar appearing quickly (which
corresponded to the dead volume of the reactor), followed by CO later corresponding to chemisorption on
the sample. The same system was used for the temperature-programmed reduction experiments. The

-
1
o
-1
material was oxidised in a flow of 10 mL min 20 vol% O
2
/He at 400 C for 2 h and reduced in 10 mL min of
o
o
-1
1
.6 vol% H
2
/He heating to 550 C at a ramp rate of 5 C min .
The X-ray photoelectron spectra (XPS) were recorded with a Kratos Axis Ultra DLD spectrometer equipped
-
10
with a monochromated Al Kα X-ray source and operating at 2·10 mbar pressure. The powdered samples
were applied to the electrically-conductive tape and studied at a pass energy of 20 eV from the area of about
0
.3 x 0.7 mm. The binding energy and spectrometer work function were calibrated using a polycrystalline Ag
sample.
2
.3.Hydrogenation experiments
FA (98%, Sigma Aldrich) was distilled under vacuum prior to experiment to remove oligomerisation products.
The catalyst (10-50 mg) was introduced together with 90 mL of water into a 160 mL high-pressure stainless
Parr batch reactor. Prior to the experiments, the slurry was degassed by flushing N
The reactor was heated up to the desired temperature under a pressure of 2 bar H
2
, 5 times, and H₂, 3 times.
2
.
After the reactor had reached the desired temperature, H₂ was introduced till the desired hydrogen
pressure, and the reaction started when the concentrated degassed FA solution was added to form 100 mL
of a 50 mM FA solution in the reactor. Liquid samples (0.5 mL) were taken and analysed with a gas
chromatograph (Shimadzu GC-2010) equipped with a 30 m x 0.32 mm Stabilwax capillary column and am FID
detector.
Conversion of FA (XFA) and selectivity (S ) were calculated using equations 2 and 3
X
퐶ꢀꢁ
푋퐹퐴 = 1 − 퐶
,
ꢁ,푖푛
(2)
(3)
ꢀ
퐶푥
푆ꢂ = 퐶ꢀꢁ,푖푛ꢃ퐶ꢀꢁ,
where C is the concentration of a product X and CFA,in is the initial FA concentration in the reactor.
X
The absence of external mass transfer limitations was verified in a separate series of experiments where the
reaction rates did not change with a stirring speed above 600 rpm, while the reaction was performed at
1
000 rpm. The highest Weisz-Prater numbers were calculated to be 0.07 for FA and 0.03 for H – both are
2
below the threshold values of 1 and 0.3, respectively, thereby confirming the absence of internal mass
transfer limitations.
3
3
. RESULTS AND DISCUSSION
.1.Catalyst characterisation
Table 1 shows the elemental composition of the catalysts studied. The Pd content in the Pd/SiO
agrees with a nominal Pd loading of 5 wt%. The slight decrease in the Pd loading in the Pd-Bi/SiO
2
catalyst
catalyst
2
was likely caused by the increase in total catalyst mass from introducing Bi. The Bi content in the Pd-Bi/SiO
catalyst corresponded to the Pd/Bi molar ratio of 7.06, which is in excellent agreement with the nominal
ratio of 7. Therefore, both Pd and Bi were fully incorporated into the silica support on wet impregnation.
2
Complete introduction of Bi is not surprising, considering that the Bi solution was evaporated allowing no
routes for Bi removal from the catalyst. Therefore, the Bi-containing compounds might have been
precipitated in the original form of separate bismuth nitrate or oxide particles rather than poisoning the Pd
catalyst. Chemisorption of CO performed on the catalyst (Table 1) confirmed that the CO capacity for the Pd-
Bi/SiO
2
catalyst decreased substantially compared to the Pd/SiO
2
catalyst. Considering that CO adsorbs
strongly on Pd but not on Bi, the decrease in the CO chemisorption capacity indicates blockage of the Pd

surface with adsorbed Bi species. The average Pd nanoparticle diameter in the Pd/SiO
from the total CO capacity was 6.5 nm [42].
2
catalyst estimated
Table 1. Elemental composition and CO chemisorption capacity of the catalysts studied.
Pd/SiO
5.12
0
2
Pd-Bi/SiO
4.87
1.36
2
Pd loading (wt%)
Bi loading (wt%)
CO capacity (μmol g )
-
1
59.6
37.1
Fig. 1 shows the powder XRD patterns for the Pd/SiO
2
and Pd-Bi/SiO catalysts. The patterns demonstrate a
2
o
broad peak at 2Θ =22 corresponding to the silica support with its width attributed to small dimensions of
the fumed silica particles. Also, there are peaks corresponding to Pd nanoparticles with an average diameter
of about 5 nm based on Scherrer’s equation. There are no peaks corresponding to Bi in the patterns
indicating that Bi particles are either amorphous, or, likely, undetectably small.
Fig. 1 Powder XRD patterns for the Pd/SiO
2
and Pd-Bi/SiO
2
catalysts.
Fig. 2A shows a representative TEM image of the Pd nanoparticles observed in the Pd/SiO
2
catalyst. The Pd
nanoparticles are 7.2±2.0 nm in diameter, in good agreement with the XRD and CO chemisorption data. The
Pd-Bi/SiO catalyst has the same nanoparticle morphology as the Pd/SiO catalyst and statistically
2
2
indistinguishable particle size distribution (7.8±2.3 nm, Fig. 2B). The same particle dimensions can be
explained by the negligible thickness of the Bi layer. The cuboctahedron particles 7.2 nm in diameter have
only about 4.8% surface sites on steps and edges which corresponds to below 0.4% of all the Pd atoms in the
particles [43]. Hence, the Pd/Bi ratio of 7 provides an almost 40-fold excess of Bi to poison the step and edge
Pd sites.

Fig. 2 TEM image of the (A) Pd/SiO
2
and (B) Pd-Bi/SiO catalyst and the Pd particle size distribution.
2
The metal nanoparticle dimensions were compared using small-angle X-ray scattering (SAXS) – a method of
nanoparticle analysis that provides a more representative analysis compared to microscopy. The scattering
data obtained (shown in the Supplementary Material) were fitted using a model of spheres with a lognormal
distribution of the radius. It was impossible to distinguish contributions from the Pd and Bi nanoparticles,
likely because of the relatively low Bi content. Hence, a combined contribution from both metals was
presented. The particle size distributions in Fig. 3 show that the Pd/SiO
2
catalyst had an average nanoparticle
diameter of 11.5 nm, while it was 11.0 nm in the Pd-Bi/SiO catalyst. The results are in a fair agreement with
2
the TEM data with the deviation appearing likely from the model selection. Comparison of the SAXS data for
the two studied catalysts, however, should be less sensitive to the model selection because of comparable
contributions introduced. Therefore, the metal nanoparticle dimensions were very similar for both studied
catalysts.
Fig. 3 Metal nanoparticle distribution obtained from the small-angle X-ray scattering studies.
The surface oxidation state of the incorporated Bi and Pd was studied with ex-situ X-ray photoelectron
spectroscopy (Fig. 4). The spectra for the Pd lines studied in the Pd-Bi/SiO
2
and Pd/SiO catalysts were similar
2
with two peaks corresponding to a single Pd(II) oxidation stage, likely PdO. Formation of the surface oxide
was caused by air oxidation during sample transfer. The Pd 3d5/2 line in the Pd-Bi catalyst showed a positive
shift of 0.5 eV compared to the Pd catalyst. The shift may indicate some electronic effects of Bi onto Pd
catalyst. However, Bi is more electropositive compared to Pd, so their interaction was expected to results in
a negative shift in the Pd line. Therefore, we interpret the shift observed by a different ex-situ oxidation of
Pd metal which could result in the peak shift of up to 2 eV [44]. The Bi lines in the Pd-Bi/SiO
contained only one specie – Bi
2
catalyst also
2 3
O .

Fig. 4 X-ray photoelectron spectra of the Pd line (A, B) and Bi line (C) for the (A) Pd/SiO
catalysts.
2
and (B, C) Pd-Bi/SiO
2
The XPS data indicate that Pd and Bi in the catalysts were in the oxidised state, which might not be
representative of the catalyst under the reaction conditions of reducing atmosphere and high temperature.
Hence, we performed a temperature programmed reduction study of the catalysts as well as the reference
2
Bi O
3
/SiO material shown in Fig. 5.
2
The oxidised Pd catalyst shows high initial release of water, in agreement with literature data that confirm
o
complete surface PdO reduction at a room temperature [45]. Reduction was completed below 150 C likely
because of the limited diffusion rate of hydrogen into the bulk of PdO.
o
Bi
2
O
3
, on the contrary, showed the peak at the maximum temperature of 550 C with the noticeable
o
reduction only starting above 250 C. The literature data are conflicting and show the Bi
2
O reduction either
3
o
o
in the temperature range of 180 - 400 C [46], or above 450 C [47].
The Pd-Bi catalyst reduction profile combines features of both the Pd catalyst and Bi
2
O . The intensive peak
3
observed at the room temperature corresponds to reduction of PdO, while higher-temperature peaks likely
correspond to the reduction of Bi
to hydrogen activation in its vicinity, but it was still well above 200 C. Therefore, Bi
2
O
3
. The reduction temperature of Bi
2
O
3
in the catalyst decreased likely due
observed in the Pd-
o
2
O
3
Bi/SiO
2
catalyst likely stays in its oxidised form under the hydrogenation conditions. Bi adatoms on the Pd
catalyst might be reduced to metallic state, but the data are not conclusive due to its low quantity of Bi.

Fig. 5 Temperature programmed reduction profiles the Pd and Pd-Bi catalysts as well as a reference 1 wt%
Bi /SiO material.
2
O
3
2
Electron diffraction was used to probe the catalyst composition further (Fig. 6). Compared to surface-specific
XPS analysis, electron diffraction provides bulk-phase data showing the crystallite phase composition rather
than surface chemical environment. The electron wavelength is also significantly shorter than that of X-rays
(
Fig. 1), reducing broadening and allowing for more accurate phase identification. The diffraction data
presented in Fig. 6 confirm the presence of Pd, PdO, Bi and Bi phases in the Pd-Bi/SiO catalyst.
2
O
3
2
Fig. 6 TEM electron diffraction pattern observed from the metal particles in the Pd-Bi/SiO catalyst.
2
Further examination of the nanoparticle elemental composition was performed using high-resolution TEM.
Fig. 7 shows that the metal nanoparticles preferentially contain Pd. However, a small amount of Bi was also
detected in the particles with a Pb/Bi molar ratio of 50 ± 16. The amount of Bi observed over the Pd
nanoparticles is significantly lower compared to the bulk content indicating that around 14% of Bi introduced
poisoned the Pd nanoparticles.

Fig. 7 High-resolution elemental map observed in the Pd-Bi/SiO
2
catalyst.
catalysts contained the Pd nanoparticles of
A combination of the characterisation data indicates that Pd/SiO
2
about 7 nm in diameter. The addition of an acidic solution of Bi nitrate under a reducing atmosphere
resulted in complete deposition of Bi. However, only a relatively small amount (around 14%) covered the Pd
nanoparticles resulting in a negligible change in the metal nanoparticle dimensions. These Bi adatoms might
be either metal or surface oxide form, but the data are not conclusive. The majority of Bi was precipitated
separately as Bi
2
O particles and are expected to have little effect on the catalyst performance because of
3
their spatial separation from the Pd nanoparticles.
3
.2.Hydrogenation over the Pd/SiO catalyst
2
Fig. 8 shows the concentration profiles and the material balance for FA hydrogenation over the Pd/SiO
2
catalyst under various experimental conditions. The corresponding selectivities to the key intermediates and
reaction products are listed in Table 2.

Fig. 8 Concentration profiles for furfural (FA) hydrogenation and the formation of furfuryl alcohol (FAL),
tetrahydrofurfuryl alcohol (TFAL), cyclopentanol (CPOL) and cyclopentanone (CPA) as well as the carbon
o
balance (CB). Hydrogenation of a 100 mM FA solution over the Pd/SiO
2
catalyst at (A) 50 C and 30 bar; (B)
o
o
o
o
o
5
0 C and 50 bar; (C) 100 C and 10 bar; (D) 100 C and 50 bar; (E) 150 C and 30 bar; (F) 150 C and 50 bar.
o
At a reaction temperature of 50 C, the main reaction products were TFAL and CPOL with minor amounts of
FAL. The FAL concentration went through a maximum at the initial reaction stages indicating that FAL was an
intermediate reaction product. The comparison of the experiments performed at 30 and 50 bar (Fig. 8A, B)
shows that the increasing pressure had a minor effect on the reaction rates and product distribution. The
concentrations of TFAL and CPOL increased monotonously with time likely because these compounds were
terminal hydrogenation products obtained via parallel reaction routes.
o
At a reaction temperature of 100 C, FAL was the major product with a yield above 50%. The reaction rate of
FA hydrogenation increased significantly at 50 bar as compared to 10 bar. At a reaction temperature of 150
o
C, TFAL was the main product. The TFAL concentration at the later reaction stages was constant indicating
that TFAL could not be transformed into CPOL or CPA. For all the temperatures studied, the carbon balance
decreased to 85-90%, indicating a formation of undetected products, which are likely oligomers. No

significant C-C cleavage was observed as confirmed by negligible (<1%) selectivity to the linear alcohol and
diene products [48].
The hydrogenation reaction rates are known to be the first order reactions to hydrogen as observed in FA
hydrogenation over the Pt, Ru, and Pd catalysts.[33,49–53] The linear relation between pressure and the
reaction rate, however, is valid only up to a certain pressure after which the reaction rates become
independent on the hydrogen pressure because surface coverage with hydrogen reaches a maximum [54].
Indeed, the reaction pressure had a noticeable effect on the FA hydrogenation reaction rate only at 10 bar
(
Table 2).
The results observed agree with the literature. The reaction rates increased with the temperature, which is
common for Pd, Pt, Ni and Ru catalysts [3,31,33,53,55]. TFAL selectivity decreased with temperature. The
DFT calculations for furan hydrogenation and ring opening over the Pd(111) surface suggest that hydrofuran
is a reactive intermediate for both pathways [56]. The study suggest that furan ring opening is
thermodynamically favoured, while ring hydrogenation is kinetically driven, and hence TFAL formed as a
major product at a low temperature.
Similarly, the selectivity to CPA and CPOL is known to increase with temperature. Pt-based catalysts show a
o
rather high CPA selectivity in FA hydrogenation above a reaction temperature of 140 C [13], while Ni shows
o
the selectivity to CPOL of 50% at 160 C [27]. Some reports, however, demonstrated that CPA and CPOL
selectivity decreases over the Pd catalyst with temperature [31,53].
Table 2 Effect of the experimental conditions on the reaction rates and product selectivity in furfural
hydrogenation.
Selectivity at 90% conversion (%)b
FAL TFAL CPOL CPA Others
Entry T
H
2
Initial
TOF
o
a
-1
(
C) pressure reaction rate (s )
-
1
(bar)
(mmol gcat
-
1
s )
A
B
C
D
E
F
50 30
50 50
100 10
100 50
150 30
150 50
1.64
1.57
0.99
1.69
2.46
2.69
287 16.3 43.1 37.3
274 18.3 47.4 32.0
173 51.7 17.3 12.7
295 64.3 11.3 5.1
430 62.3 31.0 7.3
470 70.5 19.8 7.3
0
0
0
0
0.2
0.5
3.3
2.4
18.3
19.3
0
-
a
b
within the first 20 gcat·s of reaction. FAL: furfuryl alcohol; TFAL: tetrahydrofurfuryl alcohol; CPOL:
cyclopentanol; CPA: cyclopentanone.
Hydrogen pressure had a stronger effect on the product selectivity (Table 2). The TFAL selectivity slightly
o
increased with pressure at the temperature of 50 C. The selectivity to CPA and CPOL were the most affected
o
by the hydrogen pressure at a reaction temperature of 150 C. The experiment agrees with the literature
data of increasing TFAL selectivity with pressure over the Pd-based catalysts [13,29,31].
From the concentration profiles obtained, the furfural hydrogenation pathway can be summarised in Fig. 9.
The pathway agrees with that observed over Ni catalysts [57]. The catalytic behaviour between Pd and Ni
with furfural is similar, although Ni forms stronger bonds with the furan ring that can yield the opening of
the ring [2]. Side reactions such as furfural decarbonylation to form furan and CO were previously reported
by several authors [2,21,22]. The previous reports suggest that the ring opening presumably occurs via the
hydrodeoxygenation of FAL alcohol over the Pd surface [21,58]. Ring opening, however, was not observed

o
during the current study. Below 150 C, our results agree with the study of Bhogeswararao and Srinivas who
o
found that hydrogenation of furfural is favoured over Pd at temperatures below 180 C [59].
Fig. 9 Proposed pathway for furfural hydrogenation in aqueous phase over the 5% Pd/SiO catalyst.
2
The first step (Fig. 9) is FA hydrogenation to FAL. FAL can undergo further hydrogenation to TFAL. Hronec
and Fulajtarová found that FA hydrogenation leads to the production of CPA in water while FAL or TFAL are
generated in other solvents [13]. In an aqueous solution, the furan ring is attacked forming 2-
cyclopentenone [7]. 2-cyclopentenone was not detected in the current work, likely because it was rapidly
hydrogenated into CPA. A further hydrogenation of CPA to CPOL is also possible [13]. Furthermore, FAL and
FA can oligomerize forming non-volatile products undetectable by gas chromatography [7], which is the
likely reason for the decreasing carbon balance in the experiments (Fig. 8).
3
.3.Poisoning with bismuth
Isolation of the active sites and electronic effects can be introduced into the catalyst by adding a second
metal such as Pb, Au, Cu [60–65]. We have decided to investigate the effect of the addition of Bi because it
preferentially adsorbs on the step and edge sites of Pd [35,36]. Therefore, the Bi-poisoned Pd catalyst has
predominantly terrace active sites available for hydrogenation. The concentration profiles of FA
hydrogenation over the Bi-poisoned catalysts are presented in Fig. 10 and the corresponding selectivity to
the main reaction products are summarized in Table 3.
The Pd-Bi/SiO
2
catalyst, compared to Pd/SiO , showed considerably lower initial FA hydrogenation reaction
2
rates under all the temperatures studied (Table 3). It was expected considering that Bi blocked a significant
fraction of the Pd active sites. The comparison of the turn-over frequencies (TOF, which normalise the
reaction rate by the number of active sites), however, provides comparable values. The TOF for the Pd/SiO
catalyst was only 10-25% higher than that of the Pd-Bi/SiO catalyst. This indicates absence of significant
2
2
electronic effects, in agreement with the data for alkynol semi-hydrogenation [37]. The difference observed
might be caused by spatial limitations as blocking one Pd site may decrease the availability of nearby Pd sites
for adsorption of bulky molecules.

Fig. 10 Concentration profiles for furfural (FA) hydrogenation and the formation of furfuryl alcohol (FAL),
tetrahydrofurfuryl alcohol (TFAL), cyclopentanol (CPOL) and cyclopentanone (CPA) as well as the carbon
o
balance (CB). Hydrogenation of a 100 mM FA solution over the Pd-Bi/SiO
2
catalyst at (A) 50 C and 50 bar; (B)
o
o
1
00 C and 50 bar; (C) 150 C and 50 bar.
o
At a reaction temperature of 50 C (Fig. 10), the main reaction products were FAL and TFAL with a small
amount of CPOL. The Pd-Bi/SiO
2
catalyst, compared to Pd/SiO , showed a considerably higher selectivity to
2
FAL and a lower selectivity to CPOL. Moreover, the carbon balance for the reaction conducted with the
poisoned catalyst was stable over time indicating no significant product decomposition or polymerisation.
o
At a reaction temperature of 100 C, the main reaction product was FAL. Over the Bi-poisoned catalyst, the
FAL selectivity did not change significantly compared to the unpoisoned catalyst, but unidentified products
were completely eliminated producing TFAL instead.
o
At a reaction temperature of 150 C, FAL and TFAL were the main reaction products in the beginning of the
reaction over the Pd-Bi/SiO
reaction, however, CPA was the main product with a yield factor of 8.3 higher compared to the unpoisoned
catalyst. The carbon balance over the Pd-Bi/SiO catalyst was also considerably higher compared to Pd/SiO
2
catalyst – the behaviour similar to the Pd/SiO
2
catalyst. Later during the
2
2
decreasing somewhat only after 100 gcat·s. The likely reason for the reduction in the carbon balance is TFAL
evaporation during sampling. The maximum TFAL yield decreased from 78% to 37% on poisoning the Pd/SiO
catalyst.
2
o
Formation of CPA in FA hydrogenation is mainly favoured under a high reaction temperature of 150 C with a
maximum yield observed of almost 57%. Over the Pd/SiO
2
catalyst, the dominant reaction is hydrogenation
of FAL to TFAL, while the Pd-Bi/SiO catalyst showed a formation of both TFAL and CPA in comparable
2
quantities. Therefore, Bi-poisoning of the Pd catalyst affects the reaction of FAL with water to a lesser extent
compared to hydrogenation of FAL to TFAL. Additionally, the Bi-poisoned catalyst showed little CPOL
formation indicating that the CPA to CPOL hydrogenation was limited.

Table 3 Effect of the reaction conditions in the hydrogenation of furfural over the Bi-Pd/SiO
Entry T Initial
2
catalyst.
H
2
TOF Selectivity at 90 % conversion (%)
-
1
(
ºC) pressure reaction (s )
FAL TFAL CPOL CPA Others
a
(bar)
rate
(
g
mmol
cat s )
-
1 -1
A
B
C
50
100 50
150 50
50
0.79
0.96
2.11
222 60.1 29.9 14.9
270 76.8 23.2 5.7
0
0.5
7.2
0
0
0
592 72.9 20.8
0
a
b
within the first 20 gcat·s of reaction. FAL: furfuryl alcohol; TFAL: tetrahydrofurfuryl alcohol; CPOL:
cyclopentanol; CPA: cyclopentanone.
Table 4 shows the CPA selectivity obtained from FA hydrogenation in the current work compared to the
o
literature. All the studies were carried out at a reaction temperature above 130 C with the CPA yield above
5
0%. The highest yield of 96% was obtained over the Cu-Ni [66] and 92% over the Pd-Cu catalysts [7].
The comparison of the results obtained over the Pd-Cu and Pd-Bi catalysts is particularly interesting because
the effect of Cu onto Pd nanoparticles is well studied. Poisoning Pd with Cu, according to literature, increases
the FAL selectivity at low temperature and the CPA selectivity at a higher temperature [7,25,32,67].
Depending on the catalyst preparation method, however, the origins of the effect may be different. Lesiak et
al.[67] observed the formation of Pd-Cu alloys, which affected both the FAL selectivity and FA hydrogenation
rate via electronic effects. Sitthisa et al.[25] also formed an alloy which decreased the gas-phase reaction
rate by reducing the FA decarbonylation rate, but increased the hydrogenation rate. Hronek and co-workers
[
7,32] observed an increasing FAL selectivity in liquid-phase FA hydrogenation attributed to the proximity of
Cu(I) and Pd species. Therefore, the Pd-Bi and Pd-Cu catalysts showed very similar effects on FA
hydrogenation despite a different nature of Pd modification.
Table 4 A summary of the literature data on the yield (Y) of cyclopentanone (CPA) and reaction rate in the
furfural hydrogenation.
Ref.
Catalyst
T
(
P
t
Y (%) Reaction rate
o
-1 -1
C) (bar) (hours)
(mmol gcat s )
Hronec et al.[13]
Hronec et al.[13]
Hronec et al.[13]
Hronec et al.[7]
Yang et al.[68]
Zhou et al.[15]
Pt / C
Pd / C
Ru / C
160 80
160 30
175 80
160 30
160 40
160 40
0.5
1
1
1
4
76.5 0.32
67.1
57.3
-
-
Pd-Cu / C
Ni-Cu / SBA-15
Cu-Mg-Al
hydrotalcite
Ni / HY zeolite
Raney Ni
Cu-Ni / C
Cu-Co / C
92.1 0.66
60.0 0.14
-
3
10
7.4
1.9·10
-
-
-
3
3
3
Liu et al.[8]
Xu et al.[27]
Wang et al.[66]
Wang et al.[10]
Ma et al.[14]
Zhou et al.[69]
150 40
180 10
130 50
140 30
160 20
140 40
9
4
5
-
-
8
86.5 2.2·10
39.0 3.2·10
96.9 3.8·10
23.4
-
-
-
2
3
Co / ZrO
Cu / Carb.
nanotubes
2 2
-La O
3
56.0 1.1·10
65.5 1.1·10
This work
This work
Pd / SiO
Pd-Bi / SiO
2
150 50
150 50
1.7
2.3
7.4
54.6 2.1
2.7
2
Similarity in the behaviour despite a different Pd poisoning mechanism may be reconciled considering the
active site isolation. In fact, the Pd-Cu alloy catalyst (with the Cu atoms poisoning the Pd sites uniformly) [36]
and the Pd-Bi catalyst (with Bi blocking mainly the edge sites) [35,36] both isolate the Pd sites. The

decreasing FA oligomerisation on Pd poisoning becomes obvious considering that the probability of having
two nearby adsorbed FA molecules decreases along with the number of nearby active sites. Therefore, the
observed increase in the FAL selectivity over the Bi-poisoned catalyst is explained by the decreasing FA
oligomerisation caused by isolation of the Pd active sites.
o
The increasing CPA selectivity at 150 C can also be explained in the same terms by Pd active site isolation
over both the Pd-Cu and Pd-Bi catalysts. At the initial reaction stage, FA hydrogenation results in the
formation of the FAL molecules (Fig. 9). At a later stage, FAL may form either (i) TFAL via the furan ring
hydrogenation or (ii) CPA via the reaction with water followed by hydrogenation of the C=C bond in the
cyclopentene ring (Fig. 11). The furan ring hydrogenation requires adsorption over a significantly larger area
with the addition of several H₂ molecules in short succession [70]. The C=C bond hydrogenation, on the
contrary, requires only one H₂ molecule and a smaller adsorption space due to flexibility of the cyclopentene
ring.
Therefore, once the reaction temperature becomes sufficiently high to sustain FAL hydration, the
subsequent hydrogenation to CPA becomes dominant. In fact, Bi poisoning almost halved the TFAL
-
1
-1
-1 -1
formation rate – it was 85 μmol·gcat ·s over the Pd catalyst and only 36 μmol·gcat ·s over the Pd-Bi/SiO
catalyst. The poisoned Pd catalysts with isolated active sites show a decreased FAL hydrogenation rate
allowing for a higher CPA formation. Hydrogenation of one bond in the cyclopentene ring, on the other
hand, requires only one hydrogenation step that can be performed over an isolated Pd site.
2
Fig. 11 Scheme of the effect of Bi poisoning on furfural hydrogenation pathways. The oxidation state of Bi
adatoms was not conclusively confirmed – the sites may be either metallic or oxidised.
Conclusions
Bismuth was deposited onto the Pd catalysts from an aqueous solution. Characterisation of the catalysts
showed that the majority of Bi precipitated as Bi
2
O . But about 14% Bi deposited onto the Pd nanoparticles
3
blocking a fraction of the available Pd surface area. The Bi-poisoned Pd catalysts showed a different
behaviour in furfural hydrogenation. Compared to the initial Pd catalyst, the Bi-poisoned catalyst supressed
oligomerisation of furfural and increased the formation of cyclopentanol at a higher reaction temperature.
The change in the product selectivity on Bi poisoning can be explained in terms of active site isolation. The
addition of Bi decreases the size of the Pd active site ensembles reducing adsorption of furfural via the
aromatic ring. The lower site density decreases the probability of adsorption of furfural molecules nearby
and their inter-reaction leading to oligomerisation. Similarly, separation of active sites makes single-site
reactions such as cyclopentanone formation preferred over the aromatic ring hydrogenation.
Conflicts of interest
There are no conflicts to declare.

Acknowledgments
We acknowledge the support of the European Research Council Proof of Concept grant (project Micarf
6
93739). JS is indebted to support from EPSRC grants EP/M010643/1 and EP/R019428/1. The authors are
grateful to Jonathan Strong for the ICP and TEM analyses and to Dr Aruni Fonseka for assistance with the
elemental mapping on the ARM200F instrument.
References
[
[
[
[
1]
2]
3]
4]
S. Proskurina, J. Heinim, F. Schipfer, E. Vakkilainen, Renew. Energy 111 (2017) 265–274.
S. Sitthisa, D.E. Resasco, Catal. Letters 141 (2011) 784–791.
P. Panagiotopoulou, N. Martin, D.G. Vlachos, J. Mol. Catal. A Chem. 392 (2014) 223–228.
Q. Yuan, D. Zhang, L. van Haandel, F. Ye, T. Xue, E.J.. Hensen, Y. Guan, J. Mol. Catal. A Chem. 406
(2015) 58–64.
[
[
[
5]
6]
7]
K. Yan, G. Wu, T. La, C. Jarvis, Renew. Sustain. Energy Rev. 38 (2014) 663–676.
J. Lange, E. Van Der Heide, J. Van Buijtenen, R. Price, ChemSusChem 5 (2012) 150–166.
M. Hronec, K. Fulajtárová, I. Vávra, T. Soták, E. Dobročka, M. Mičušík, Appl. Catal. B Environ. 181
(2016) 210–219.
[
[
[
[
[
[
[
[
[
[
[
[
[
[
[
8]
9]
C. Liu, R. Wei, G. Geng, M. Zhou, L. Gao, G. Xiao, Fuel Process. Technol. 134 (2015) 168–174.
M.S. Kim, F.S.H. Simanjuntak, S. Lim, J. Jae, J. Ha, H. Lee, J. Ind. Eng. Chem. 52 (2017) 59–65.
10] Y. Wang, Y. Miao, S. Li, L. Gao, G. Xiao, Mol. Catal. 436 (2017) 128–137.
11] L. Liu, H. Lou, M. Chen, Int. J. Hydrogen Energy 41 (2016) 14721–14731.
12] X. Chen, W. Sun, N. Xiao, Y. Yan, S. Liu, Chem. Eng. J. 126 (2007) 5–11.
13] M. Hronec, K. Fulajtarová, Catal. Commun. 24 (2012) 100–4.
14] Y. Ma, H. Wang, G. Xu, X. Liu, Y. Zhang, Y. Fu, Chinese Chem. Lett. 28 (2017) 1153–1158.
15] M. Zhou, Z. Zeng, H. Zhu, G. Xiao, R. Xiao, J. Energy Chem. 23 (2014) 91–96.
16] W. Xu, H. Wang, X. Liu, J. Ren, Y. Wang, G. Lu, Chem. Commun. 47 (2011) 3924–6.
17] M. Hronec, K. Fulajtárova, M. Mičušik, Appl. Catal. A Gen. 468 (2013) 426–431.
18] B. Wang, C. Li, B. He, J. Qi, C. Liang, J. Energy Chem. (2017) 1–9.
19] M.F. Fellah, Appl. Surf. Sci. 405 (2017) 395–404.
20] H. Liu, Z. Huang, H. Kang, C. Xia, J. Chen, Chinese J. Catal. 37 (2016) 700–710.
21] V. Vorotnikov, G. Mpourmpakis, D.G. Vlachos, ACS Catal. 2 (2012) 2496–2504.
22] R. V. Maligal-Ganesh, C. Xiao, T.W. Goh, L.L. Wang, J. Gustafson, Y. Pei, Z. Qi, D.D. Johnson, S. Zhang,
F. Tao, W. Huang, ACS Catal. 6 (2016) 1754–1763.
[
[
[
23] M.D. Quian, J.L. Xue, S.J. Xia, Z.M. Ni, J.H. Jiang, Y.Y. Cao, J. Fuel Chem. Technol. 45 (2017) 34–42.
24] S.H. Pang, C.A. Schoenbaum, D.K. Schwartz, J. Will Medlin, ACS Catal. 4 (2014) 3123–3131.
25] S. Sitthisa, T. Sooknoi, Y. Ma, P.B. Balbuena, D.E. Resasco, J. Catal. 277 (2011) 1–13.

[
[
26] S.H. Pang, N.E. Love, J.W. Medlin, J. Phys. Chem. Lett. 5 (2014) 4110–4114.
27] Y. Xu, S. Qiu, J. Long, C. Wang, J. Chang, J. Tan, Q. Liu, L. Ma, T. Wang, Q. Zhang, RSC Adv. 5 (2015)
9
1190–91195.
[
[
[
[
[
28] N. Pino, S. Sitthisa, Q. Tan, T. Souza, D. López, D.E. Resasco, J. Catal. 350 (2017) 30–40.
29] W. Yu, Y. Tang, L. Mo, P. Chen, H. Lou, X. Zheng, Bioresour. Technol. 102 (2011) 8241–8246.
30] D. Scholz, C. Aellig, I. Hermans, ChemSusChem 7 (2014) 268–275.
31] Y. Nakagawa, K. Takada, M. Tamura, K. Tomishige, ACS Catal. 4 (2014) 2718–26.
32] K. Fulajtárova, T. Soták, M. Hronec, I. Vávra, E. Dobročka, M. Omastová, Appl. Catal. A Gen. 502
(2015) 78–85.
[
33] A. O’Driscoll, T. Curtin, W.Y. Hernández, P. Van Der Voort, J.J. Leahy, Org. Process Res. Dev. 20 (2016)
1
917–1929.
[
[
[
[
34] J. Sá, J. Montero, E. Duncan, J.A. Anderson, Appl. Catal. B Environ. 73 (2007) 98–105.
35] J.A. Anderson, J. Mellor, R.K.P.K. Wells, J. Catal. 261 (2009) 208–216.
36] N. López, C. Vargas-Fuentes, Chem. Commun. 48 (2012) 1379–1391.
37] N. Cherkasov, A.O. Ibhadon, A. McCue, J.A. Anderson, S.K. Johnston, Appl. Catal. A Gen. 497 (2015)
2
2–30.
[
[
[
[
[
[
[
[
[
[
38] N. Cherkasov, A.O. Ibhadon, E. V. Rebrov, Appl. Catal. A Gen. 515 (2016) 108–115.
39] C.A. Schneider, W.S. Rasband, K.W. Eliceiri, Nat. Methods 9 (2012) 671–675.
40] J. Ilavsky, P.R. Jemian, J. Appl. Crystallogr. 42 (2009) 347–353.
41] Y. Bai, N. Cherkasov, S. Huband, D. Walker, R. Walton, E. Rebrov, Catalysts 8 (2018) 1–18.
42] A. Borodziński, M. Bonarowska, Langmuir 13 (1997) 5613–5620.
43] R.V.A.N. Hardeveld, F. Hartog, Surf. Interface Anal. 15 (1969) 189–230.
44] A.M. Venezia, Catal. Today 77 (2003) 359–370.
45] K. Noack, H. Zbinden, R. Schlögl, Catal. Letters 4 (1990) 145–155.
46] H. Miura, Y. Morikawa, T. Shirasaki, J. Catal. 28 (1975) 22–28.
47] R. Irmawati, M.N. Noorfarizan Nasriah, Y.H. Taufiq-Yap, S.B. Abdul Hamid, Catal. Today 93–95 (2004)
7
01–709.
[
48] N. Cherkasov, V. Jadvani, J. Mann, Y.B. Losovyj, Z.B. Shifrina, L.M. Bronstein, E.V. Rebrov, Fuel
Process. Technol. 167 (2017) 738–746.
[
[
[
49] D. Duca, L.F. Liotta, G. Deganello, J. Catal. 154 (1995) 69–79.
50] N. Cherkasov, M. Al-Rawashdeh, A.O. Ibhadon, E. V Rebrov, Catal. Today 273 (2016) 205–212.
51] Z. Wu, N. Cherkasov, G. Cravotto, E. Borretto, A.O. Ibhadon, J. Medlock, W. Bonrath, ChemCatChem 7
(2015) 952–959.
[
[
52] S. Vernuccio, R. Goy, A. Meier, P. Rudolf von Rohr, J. Medlock, Chem. Eng. J. 316 (2017) 121–130.
53] R.M. Mironenko, O.B. Belskaya, T.I. Gulyaeva, A.I. Nizovskii, A. V Kalinkin, V.I. Bukhtiyarov, A. V

Lavrenov, V.A. Likholobov, Catal. Today 249 (2015) 145–152.
[
[
54] W. Rachmady, M. Vannice, J. Catal. 192 (2000) 322–334.
55] L. Baijun, L. Lianhai, W. Bingchun, C. Tianxi, Katsuyoshi Iwatani, Appl. Catal. A Gen. 171 (1998) 117–
1
22.
[
[
56] S. Wang, V. Vorotnikov, D.G. Vlachos, Green Chem. 16 (2014) 736–747.
57] B. Yang, J. Zuo, X. Tang, F. Liu, X. Yu, X. Tang, H. Jiang, L. Gan, Ultrason. Sonochem. 21 (2014) 1310–
1
317.
[
[
[
[
58] B. Chen, F. Li, Z. Huang, G. Yuan, Appl. Catal. A Gen. 500 (2015) 23–29.
59] S. Bhogeswararao, D. Srinivas, J. Catal. 327 (2015) 65–77.
60] C. Moreno-Marrodan, F. Liguori, P. Barbaro, Beilstein J. Org. Chem. 13 (2017) 734–754.
61] M. Armbrüster, M. Behrens, F. Cinquini, K. Föttinger, Y. Grin, A. Haghofer, B. Klötzer, A. Knop-Gericke,
H. Lorenz, A. Ota, S. Penner, J. Prinz, C. Rameshan, Z. Révay, D. Rosenthal, G. Rupprechter, P. Sautet,
R. Schlögl, L. Shao, L. Szentmiklósi, D. Teschner, D. Torres, R. Wagner, R. Widmer, G. Wowsnick,
ChemCatChem 4 (2012) 1048–1063.
[
[
[
62] A.J. McCue, A. Guerrero-Ruiz, C. Ramirez-Barria, I. Rodríguez-Ramos, J.A. Anderson, J. Catal. 355
(2017) 40–52.
63] S.K. Johnston, N. Cherkasov, E. Pérez-Barrado, A. Aho, D.Y. Murzin, A.O. Ibhadon, M.G. Francesconi,
Appl. Catal. A Gen. 544 (2017) 40–45.
64] S. Gonzalez, K.M. Neyman, S. Shaikhutdinov, H.-J. Freund, F. Illas, J. Phys. Chem. C 2 (2007) 6852–
6
856.
[
[
[
65] F. Zaera, ACS Catal. 7 (2017) 4947–4967.
66] Y. Wang, S. Sang, W. Zhu, L. Gao, G. Xiao, Chem. Eng. J. 299 (2016) 104–111.
67] M. Lesiak, M. Binczarski, S. Karski, W. Maniukiewica, J. Rogowski, E. Szubiakiewicz, Berlows, "Journal
Mol. Catal. A, Chem. 395 (2014) 337–348.
[
[
[
68] Y. Yang, Z. Du, Y. Huang, F. Lu, F. Wang, J. Gao, J. Xu, Green Chem. 15 (2013) 1932–40.
69] M. Zhou, J. Li, K. Wang, H. Xia, J. Xu, J. Jiang, Fuel 202 (2017) 1–11.
70] A. Stanislaus, H.C. Barry, Catal. Rev. 36 (1994) 75–123.