Furfural hydrodeoxygenation (HDO) over silica-supported metal phosphides – The influence of metal–phosphorus stoichiometry on catalytic properties

Article abstract of DOI:10.1016/j.jcat.2021.01.031

The gas-phase hydrodeoxygenation (HDO) of furfural, a model compound for bio-based conversion, was investigated over transition metal phosphide catalysts. The HDO activity decreases in the order Ni₂P ≈ MoP > Co₂P ≈ WP ? Cu₃P > Fe₂P. Nickel phosphide phases (e.g., Ni₂P, Ni₁₂P₅, Ni₃P) are the most promising catalysts in the furfural HDO. Their selectivity to the gasoline additives 2-methylfuran and tetrahydro-2-methylfuran can be adjusted by varying the P/Ni ratio. The effect of P on catalyst properties as well as on the reaction mechanism of furfural HDO were investigated in depth for the first time. An increase of the P stoichiometry weakens the furan-ring/catalyst interaction, which contributes to a lower ring-opening and ring-hydrogenation activity. On the other hand, an increasing P content does lead to a stronger carbonyl/catalyst interaction, i.e., to a stronger η²(C, O) adsorption configuration, which weakens the C1[sbnd]O1 bond (Scheme 1) in the carbonyl group and enhances the carbonyl conversion. Phosphorus species can also act as Br?nsted acid sites promoting C1[sbnd]O1 (Scheme 1) hydrogenolysis of furfuryl alcohol, hence contributing to higher production of 2-methylfuran.

Full text of DOI:10.1016/j.jcat.2021.01.031

Journal of Catalysis xxx (xxxx) xxx
Contents lists available at ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
Furfural hydrodeoxygenation (HDO) over silica-supported metal
phosphides – The influence of metal–phosphorus stoichiometry on
catalytic properties
⇑
Xuefang Lan, Robert Pestman, Emiel J.M. Hensen, Thomas Weber
Laboratory of Inorganic Materials and Catalysis, Department of Chemical Engineering and Chemistry, Eindhoven University of Technology, Groene Loper 3, 5612 AE Eindhoven,
the Netherlands
a r t i c l e i n f o
a b s t r a c t
Article history:
The gas-phase hydrodeoxygenation (HDO) of furfural, a model compound for bio-based conversion, was
investigated over transition metal phosphide catalysts. The HDO activity decreases in the order Ni2P ꢀ
MoP > Co2P ꢀ WP ꢁ Cu3P > Fe2P. Nickel phosphide phases (e.g., Ni2P, Ni12P5, Ni3P) are the most promising
catalysts in the furfural HDO. Their selectivity to the gasoline additives 2-methylfuran and tetrahydro-2-
methylfuran can be adjusted by varying the P/Ni ratio. The effect of P on catalyst properties as well as on
the reaction mechanism of furfural HDO were investigated in depth for the first time. An increase of the P
stoichiometry weakens the furan-ring/catalyst interaction, which contributes to a lower ring-opening
and ring-hydrogenation activity. On the other hand, an increasing P content does lead to a stronger car-
Received 14 September 2020
Revised 28 January 2021
Accepted 30 January 2021
Available online xxxx
Keywords:
Metal phosphide
Ni2P
bonyl/catalyst interaction, i.e., to a stronger
g
2(C, O) adsorption configuration, which weakens the C1AO1
Co2P
bond (Scheme 1) in the carbonyl group and enhances the carbonyl conversion. Phosphorus species can
also act as Brønsted acid sites promoting C1AO1 (Scheme 1) hydrogenolysis of furfuryl alcohol, hence
contributing to higher production of 2-methylfuran.
Ó 2021 The Author(s). Published by Elsevier Inc. This is an open access article under the CC BY license
(http://creativecommons.org/licenses/by/4.0/).
Hydrodeoxygenation
Furfural
Biomass
Mechanism
Infrared spectroscopy
1. Introduction
[5,22,23]. Furthermore, the production of biofuels from furfural
has received extensive attention during the last decades. Aldol-
Second generation biomass is a renewable feedstock with a
lower carbon footprint than fossil crudes and is thus a promising
alternative raw material to produce fuels and chemicals [1–3].
One of the interesting biomass-derived platform molecules is fur-
fural (Scheme 1), which can be produced on an industrial scale
by dehydration of hemicellulose from agricultural waste and forest
residue [4–7] and has been identified as one of the numerous oxy-
genated compounds in pyrolysis bio-oils [8–14].
Currently, the most relevant use of furfural as a chemical feed-
stock is in the production of furfuryl alcohol (FOL) and tetrahydro-
furfuryl alcohol (THFOL) from selective hydrogenation, which have
significant applications in furan resins industry as solvents or
chemical intermediates [5,15–21]. Besides, it is also feasible to
obtain furan and tetrahydrofuran (THF) via decarbonation and
hydrogenation as an environmental benign industrial solvent in
the synthesis of plastics, pharmaceuticals and agrochemicals
condensation of furfural with acetone, self-condensation of fur-
fural, or condensation of intermediate products (e.g. furan, 2-
methylfuran) with other substrates are promising approaches to
produce higher alkanes (C8-C15) for diesel fuel applications
[20,24–26]. Selective HDO of furfural, on the other hand, yields
2-methylfuran (MF) or tetrahydro-2-methylfuran (THMF) which
are useful blending components for gasoline [6,27,28].
In the present study, we focus on the production of MF from
furfural HDO, because the involved hydrogen consumption is
lower, while the carbon yield is higher compared with products
like THMF, THFOL, and furan [5,6]. Moreover, MF has favorable
properties in fuel applications. Cu-based catalysts have been found
being active in the MF production [29–32]. However, Cu-Cr
catalysts [29] used in the industrial MF production are highly toxic,
while Cu supported on SiO2, Al2O3 or ZnO [30] suffers from sinter-
ing during long reaction periods. Group 9–10 metal catalysts (Ni
[5,33], Co [33], Pt [13,34] and Pd [4,5,35]) are also active in the
MF production. However, the strong interaction between the furan
ring and the transition metal surface leads to undesired
⇑
Corresponding author at: Solid-State Inorganic Chemistry, Inorganic Materials
and Catalysis, Department of Chemical Engineering and Chemistry, Eindhoven
University of Technology, Groene Loper 3, 5612 AE Eindhoven, the Netherlands.
E-mail address: th.weber@tue.nl (T. Weber).
ring-hydrogenation or ring-opening reactions; the unstable
O) adsorbate configuration tends to turn into an
g
g
2(C,
1(C)
https://doi.org/10.1016/j.jcat.2021.01.031
0021-9517/Ó 2021 The Author(s). Published by Elsevier Inc.
This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
Please cite this article as: X. Lan, R. Pestman, Emiel J.M. Hensen et al., Furfural hydrodeoxygenation (HDO) over silica-supported metal phosphides – The
influence of metal–phosphorus stoichiometry on catalytic properties, Journal of Catalysis, https://doi.org/10.1016/j.jcat.2021.01.031

X. Lan, R. Pestman, Emiel J.M. Hensen et al.
Journal of Catalysis xxx (xxxx) xxx
99%), (NH4)W12O39ꢂH2O (Aldrich, 99%), Co(NO3)2ꢂ6H2O (Aldrich,
99%), Fe(NO3)2ꢂ9H2O (Aldrich, 99%), Cu(NO3)2ꢂ3H2O (Aldrich, 99%)
were used to prepare catalyst samples of Ni2P, MoP, WP, Co2P,
Fe2P or Cu3P.
Scheme 1. Structure of furfural.
2.2. Preparation of different metal phosphides
Metal and phosphorus species were loaded on silica by impreg-
nation. For NiO-P/SiO2, MoO3-P/SiO2, WO3-P/SiO2, Fe2O3-P/SiO2,
Co3O4-P/SiO2, and CuO-P/SiO2 samples, metal precursors were first
loaded on silica support by incipient wetness impregnation. After
drying overnight at 110 °C in air and calcination at 550 °C for
1 h, the silica supported metal oxides were impregnated with
H3PO3 by incipient wetness impregnation and then air-dried at
110 °C overnight. Before reaction or characterization, the prepared
samples were reduced for 2 h in hydrogen at the desired reduction
temperature as determined by temperature programmed reduc-
tion measurements (TPR) with a heating rate of 5 °C/min.
The metal loading is 1.5 mmol/g SiO2; the atomic ratios of P/M
(phosphorus to metal) are 2 for Ni2P; 1 for MoP; 1.5 for WP; 0.50
for Fe2P, Co2P and Cu3P, which were empirically found to be
needed for the formation of the respective phases. The unreduced
precursors are referred to as NiO-P/SiO2, MoO3-P/SiO2, WO3-P/SiO2,
Fe2O3-P/SiO2, Co3O4-P/SiO2, and CuO-P/SiO2. A H3PO3/SiO2 sample
with a phosphorus loading of 3 mmol/g was prepared as reference
sample. NixMoyP/SiO2 catalysts with atomic Ni/Mo ratios of 1:1
and 2:1 were prepared similarly by co-impregnation with a solu-
tion containing both Ni and Mo. The loading of Ni/Mo/P on the Nix-
MoyP/SiO2 samples was 1.5/1.5/3 or 1.5/0.75/3 mmol (for the
elements Ni/Mo/P, respectively) per gram of SiO2.
NiP(x)/SiO2 and CoP(x)/SiO2 were prepared by incipient wet-
ness impregnation as in our previous work [42]. Catalysts were
prepared by a one-step impregnation of metal nitrates and H3PO3
followed by direct reduction of as-prepared precursors. The molar
P/Ni ratio was varied from 0 to 2 (0, 0.33, 0.5, 1.0, 2.0) in the pre-
cursors. Catalyst with different P/Ni ratio were labeled as Ni/SiO2,
NiP(0.33)/SiO2, NiP(0.5)/SiO2, NiP(1)/SiO2, and NiP(2)/SiO2, respec-
tively. Molar ratios of P/Co were varied from 0 to 2 (0, 0.5, 1.0, 1.5,
2.0) in the catalyst precursors. Catalysts are denoted as Co/SiO2,
CoP(0.5)/SiO2, CoP(1.0)/SiO2, CoP(1.5)/SiO2, and CoP(2.0)/SiO2,
respectively.
configuration at high temperature enabling undesired decarbony-
lation reactions [5].
To optimize MF production over pure transition metals (e.g. Ni,
Pt), more oxophilic metals (e.g. Fe, Zn) have been added to design
alloy catalysts (e.g. PtZn [7] and NiFe [6,27]) for furfural HDO. The
addition of more oxophilic elements provides more stable
g
2(C, O)
configurations to the catalyst surface, in which the furan ring is
tilted away from the catalyst surface and the carbonyl group is
bonded stronger to the catalyst surface by the additional interac-
tion of the carbonyl O to the oxophilic metal [6,7]. Such
g
2(C, O)
configurations suppress furan-ring hydrogenation or ring-
opening, and enhance the conversion of the carbonyl group by
weakening the C1AO1 bond (Scheme 1) [6,7]. By adjusting the Fe
loading from 0 to 5 wt% in FeNi/SiO2 catalysts at a constant Ni load-
ing of 5wt%, the MF yield was improved from 20% for Ni/SiO2 to
80% for 5/5 wt% NiFe/SiO2[6]. Besides alloying metals, ceramic
materials like Mo2C [36–39] have also been explored for the fur-
fural to MF conversion. Mo2C shows a high MF selectivity of 70%
[36–38], but suffers from rapid deactivation with the furfural con-
version decreasing from 90% to 18% over a period of 3 h [38]. Metal
borides like NiB, CoB and NiPB [14,19,40,41] have also been consid-
ered as catalysts in the furfural conversion. The electron-deficient
B can attract the O atom of the carbonyl group, which weakens
the C@O bond and thus promotes its hydrogenation to the alcohol
[14,19]. Nevertheless, only FOL, instead of MF, has been obtained as
main products on these metal boride catalysts.
Metal phosphides have extensively been investigated in
hydrodesulfurization (HDS) [42–44] and hydrodenitrogenation
(HDN) [44–48] reactions in recent years. Due to their promising
catalytic performance, they are also increasingly applied in
hydrodeoxygenation (HDO) reactions [49–52]. So far, there was a
focus on investigating HDO model compounds[53], such as phenol
[50,54–56], anisole [2,56], or guaiacol [1,57–62], furfural [63–65]
and some aliphatic model compounds [51,66–69]. However, stud-
ies of the furfural HDO over metal phosphides were limited to
Ni2P-based catalysts. Hence, we decided to explore the catalytic
performance of different metal phosphides in the furfural HDO
reaction with a special interest in obtaining MF as the most desired
product. In addition to changing the metal component, we varied
the phosphorus/metal ratios in our samples and explored catalytic
properties in response to the changed P/M stoichiometries for both
Ni and Co. Previous literature reports the influence of the P/Ni ratio
on the performance of nickel phosphides in HDS, HDN and HDO
[70–73]. However, only the relationship between the active nickel
phosphide phase and the catalytic performance has been estab-
lished, while the presence of P is considered to only determine or
maintain the composition of the active phases. In the present work,
we provide deeper insight into how the P/Ni ratio influences in
each of the active phases the catalytic performance during furfural
HDO, by means of furfural-IR and CO-IR spectroscopy.
2.3. Characterization
The metal loading was determined by inductively coupled
plasma atomic emission spectrometry (ICP-AES) analysis per-
formed on a Spectro Blue apparatus. Prior to analysis, samples
were dissolved in an HF/HNO3/H2O solution with a volumetric
ratio of 1:1:1. HF (VWR, 40%) and HNO3 (VWR, 65%) were used
to prepare the acidic solvent for ICP measurements.
X-ray diffraction (XRD) patterns were acquired on a Bruker D2
Phaser powder diffraction system using Cu
Ka radiation
(1.5406 Å). Scans were taken at a rate of 1°/min in the range of
10° ꢃ 2h ꢃ 80°.
BET surface area measurements were performed on
a
Micromeritics ASAP3020 Tristar system and a nitrogen stream
was applied for N2 physisorption at ꢄ196 °C. The specific surface
area was determined by the Brunauer-Emmett-Teller (BET) equa-
tion using the adsorption data.
2. Experimental
X-ray photoelectron spectroscopy (XPS) measurements were
performed on a Thermo Scientific K-Alpha XPS apparatus equipped
2.1. Materials and preparation
with a monochromatic Al Ka X-ray source. Samples were reduced
and sealed in a tubular quartz reactor, transferred into a glovebox
without exposure to air, then loaded in an airtight transfer vessel
and introduced into the XPS analysis chamber for analysis. The
background pressure prior to analysis was 2 ꢅ 10ꢄ9 mbar. Survey
Commercial silica (silica gel Davisil, particle size 90–125 um,
SBET = 305 m2/g) was used as support material. H3PO3 (Aesar,
98%), Ni(NO3)2ꢂ6H2O (Aldrich, 97%), (NH4)6Mo7O24ꢂ4H2O (VWR,
2

X. Lan, R. Pestman, Emiel J.M. Hensen et al.
Journal of Catalysis xxx (xxxx) xxx
scans were collected at constant pass energy of 200 eV, and region
scans at 50 eV. The spectra were calibrated to the Si 2p line at
103.3 eV and fitted with the CasaXPS program.
verify reproducibility. The product gas stream was analyzed after
2 h on stream.
Conversion and selectivity were calculated as follows:
Temperature programmed reduction measurements in H2 (H2-
TPR) were carried out between 50 °C and 700 °C. Typically,
15 mg catalyst precursor was heated in a 20 ml/min H2 flow from
50 °C to 700 °C at a heating rate of 5 °C /min. The effluent gas com-
position was analyzed by an online mass spectrometer. The mass
signals of 2 (H2), 18 (H2O), 34 (PH3), and 30 (NO) were monitored.
Temperature programmed desorption of NH3 (NH3-TPD) was
conducted between 25 °C and 400 °C. Typically, 15 mg catalyst pre-
cursor was reduced at 550 °C for 1 h and cooled down to 25 °C in a
20 ml/min flow of H2. Afterwards, the reduced samples were first
flushed with He for 30 min to remove physisorbed species, then
with NH3/He (10 v/v%) for 30 min to adsorb NH3, followed by He
for 30 min to remove weakly absorbed NH3. Subsequently, the pre-
treated samples were heated from 25 °C to 400 °C in a 20 ml/min
He flow with a heating rate of 5 °C /min. The effluent gas flow was
monitored by a TCD detector.
mol of the products
mol of furfural fed
Conversionð%Þ ¼
ꢅ 100
mol of one product
vityð%Þ ¼
mol of all products
Selecti
ꢅ 100
3. Results and discussions
3.1. Basic properties
The metal and phosphorus content of unreduced precursors as
determined with ICP and the textural properties are listed in
Table S1. Each precursor shows a P/metal molar ratio close to the
target value. BET surface areas are between 150 m2/g and
250 m2/g, which are all lower than the surface area of bare SiO2
(305 m2/g) and decrease with increasing P content due to pore
blocking [72].
Infrared (IR) spectroscopy of CO absorbed samples were carried
out with a Bruker Vertex V70v Fourier Transform IR spectrometer
equipped with a DTGS detector. The catalysts were pressed into a
self-supporting wafer with a diameter of 13 mm and then mounted
in an in-situ cell equipped with CaF2 windows. Prior to CO adsorp-
tion, the samples were reduced at 500 °C for 2 h, applying a heating
rate of 10 °C/min, followed by evacuating the cell to a pressure
below 2 ꢅ 10ꢄ6 mbar for 0.5 h at 500 °C. Afterwards, the reduced
sample was cooled down to 25 °C under evacuation. CO was intro-
duced into the cell with a metering valve until the CO pressure
reached 5 mbar. The catalyst was exposed to CO for 5 min, subse-
quently evacuated for 5 min to 2 ꢅ 10ꢄ6 mbar, followed by heating
to 425 °C at a rate of 10 °C /min under evacuation. The IR spectrum
3.2. TPR
Temperature programmed reduction (TPR) was used (Figure S1)
to determine the right reduction temperature for each catalyst.
As revealed by Figure S1, the NiO-P/SiO2, WO3-P/SiO2 and
MoO3-P/SiO2 samples are reduced at 500 °C, 600 °C and 550 °C,
respectively. A higher reduction temperature (650 °C) is required
to reduce Co3O4-P/SiO2 and CuO-P/SiO2, while Fe2O3-P/SiO2
reduces at 680 °C.
was recorded by accumulating 64 scans at a resolution of 2 cmꢄ1
.
In case of the NiP(x)/SiO2 catalysts (Figure S1b), the reduction
temperature is determined as 450 °C for Ni/SiO2, and 550 °C for
NiP(x) (x = 0.33, 0.5, 1, 2). Although the TPR profiles of NiP(1)
and NiP(2) show the most significant signals at 650 °C as a result
of PO34ꢄ reduction, we have chosen 550 °C as reduction tempera-
ture since P species formed by reduction of excessive H3PO3 suffice
for nickel phosphide formation.
In the case of CoP(x)/SiO2 catalysts (Figure S1c), Co/SiO2 is
reduced at 400 °C, while a high temperature of 650 °C is required
for CoP(x) reduction.
Spectra of the freshly reduced as well as CO saturated catalysts
were recorded during the heating process.
In case of furfural-IR, catalyst pretreatment and experimental
parameters were the same as those of the CO-IR experiments. Liq-
uid furfural was expanded into the evacuated cell at a pressure of
2 ꢅ 10ꢄ3 mbar for 5 min at 25 °C and then evaporated until a pres-
sure of 2 ꢅ 10ꢄ6 mbar was reached. Afterwards, the catalyst was
heated to 300 °C at a rate of 10 °C/min at 2 ꢅ 10ꢄ6 mbar. Spectra
of the furfural-saturated catalysts were recorded during the heat-
ing process. Spectra of freshly reduced catalysts were taken to
use as background correction reference.
The particle size of active Ni-phosphide phase was investigated
with a FEI Tecnai 20 transmission electron microscopy (TEM).
3.3. XRD & TEM
X-ray diffraction (XRD) reveals that each catalyst formed crys-
talline phases after reduction (Fig. 1). All catalysts show a broad
diffraction peak in the 2h range of 15–35°, typical of amorphous sil-
ica [60,74,75]. In Fig. 1a, all reduced samples contain pure phases
of the corresponding metal phosphides (i.e. Ni2P, MoP, Co2P, Fe2P,
and Cu3P) except WP/SiO2, which displays reflections of both WP
and unphosphided W0 phases. The sharp diffraction peaks of WP/
SiO2, Fe2P/SiO2, Co2P/SiO2 and Cu3P/SiO2 indicate that the metal
phosphides consist of relatively large particles due to the rather
high reduction temperature. For MoP/SiO2 and Ni2P/SiO2, milder
phosphidation conditions were applied and consistently the XRD
patterns show broad diffraction peaks, suggesting smaller
particles.
2.4. Catalytic activity measurements
The catalytic performance of synthesized materials was evalu-
ated in a plug flow fixed-bed reactor with 4 mm internal diameter.
Furfural was fed by a syringe pump (Hewlett Packard 1050) and
gasified at 170 °C using H2 as carrier gas.
Prior to reaction, catalysts were reduced in-situ at the desired
temperature in a hydrogen flow for 2 h. After reduction, the reactor
was cooled to reaction temperature, then fed with a furfural/H2
flow (H2/furfural molar ratio = 74). The products were analyzed
by an in-line gas chromatograph using a flame ionization detector
(GC-FID) and a DB-1 (30 m, 0.32 mm, 1.00
l
m) column.
Fig. 1b shows XRD patterns of reduced NiP(x)/SiO2 (x = 0, 0.33,
0.5, 1.0, and 2.0). Pure metallic Ni0 is formed in Ni/SiO2. Ni3P is
obtained on NiP(0.33), Ni12P5 on NiP(0.5), and Ni2P on NiP(1) and
NiP(2). FWHM indicates comparable particle sizes.
Fig. 1c shows that reduced CoP(x)/SiO2 (x = 0, 0.5, 1.0, 1.5, and
2.0) contains a pure Co0 phase on reduced Co/SiO2. As the P/Co
ratio increases, the active phase shifts from Co2P to CoP with
The reaction was carried out at atmospheric pressure between
120 °C and 200 °C. In a typical run, the H2 flow, the feeding rate
of furfural, H2 and the loading amount of catalysts were 20 ml/
min, 0.001 ml/min in liquid phase, and 60 mg, respectively. This
corresponds to a weight hourly space velocity (WHSV) of ca.
3 hꢄ1. All activity measurements were carried out in duplicate to
3

X. Lan, R. Pestman, Emiel J.M. Hensen et al.
Journal of Catalysis xxx (xxxx) xxx
Fig. 1. XRD patterns of different metal phosphides.
Co2P present in CoP(0.5), CoP, CoP(1.5), and CoP(2.0), and both
phases in CoP(1).
Different from NixPy (Ni2P, Ni3P, Ni12P5) and CoxPy (Co2P, CoP),
only one single phosphide phase has been obtained for Mo, Fe,
Cu, and W. Hence we explored the effect of metal/P ratio on the
catalytic performance only for nickel and cobalt phosphides.
Reduced Ni, NiP(0.5), NiP(1) and NiP(2) samples were also
characterized by TEM to determine the particle size distributions
(Fig. 2). Besides some large particles observed in the Ni
catalyst, all catalyst show small nanoparticles of a comparable
size of ~ 5 nm, consistent with the XRD data for the NiP(x)
catalysts.
Fig. 2. TEM images of NiP(x) (x = 0, 0.5, 1, 2) catalysts.
4

X. Lan, R. Pestman, Emiel J.M. Hensen et al.
Journal of Catalysis xxx (xxxx) xxx
3.4. XPS
The XPS results reveal the type and amount of chemical species
on the surface of reduced catalysts, which confirm the formation of
metal phosphides.
Fig. 3 and Table S2 show the results of the reduced NiP(x) (x = 0,
0.33, 0.5, 1, and 2) catalysts. The contribution at 852.2 eV in Ni 2p
spectra is assigned to Ni0 in metallic nickel, the contribution at
852.7 eV is assigned to Nid+ (0 < d < 1) in nickel phosphide phases
(Ni2P, Ni5P2 and Ni3P), and the signal at 856.4 eV to Ni2+ species
[42]. Concerning the P 2p spectra, the three peaks at 134.4 eV,
133.5 eV and 129.1 eV are due to PO34ꢄ, PO33ꢄ and Pdꢄ (0 < d < 1) spe-
cies in nickel phosphides [42], respectively. PO33ꢄ originates from
unreduced H3PO3, while PO34ꢄ forms from the disproportionation
or reduction of H3PO3 [42]. The presence of Nid+ (0 < d < 1) and
Pdꢄ (0 < d < 1) in nickel phosphide phases points to an interphase
Ni ? P transfer of electron density as a consequence of the electron
withdrawing nature of P [76]. The electron density on Nid+ is
expected to decline at higher P content due to the electron with-
drawing nature of P atoms. Note that the presence of Ni0 species
in NiP(2), NiP(1), NiP(0.5) and NiP(0.33) catalysts cannot be
excluded, since the P/Ni ratios of these samples (Table S2) are all
lower than the theoretical P/Ni ratios of Ni2P, Ni5P2 and Ni3P
phases (0.5, 0.4, and 0.33, respectively).
Figure S2 displays the Co 2p and P 2p core-level spectra of
reduced CoP(x) catalysts. Peaks at 777.4–777.9 eV are assigned to
the Co0 or Cod+ (0 < d < 1) sites in catalysts, while peaks at
129.0 eV are the fingerprint of Pdꢄ (0 < d < 1) in cobalt phosphides.
Similar to NiP(x) catalysts, the presence of Cod+ (0 < d < 1) and Pdꢄ
(0 < d < 1) in Co2P and CoP phases points to an interphase Co ? P
transfer of electron density as a consequence of the electron with-
drawing nature of P [76]. The binding energy of Cod+ is slightly
higher than that of Co0 due to the slight positive charge it bears.
Different from NiP(x) catalysts, the P/Co molar ratios of the
reduced phases are higher than the elemental analysis in the
respective phases (i.e. 0.5 in Co2P and 1 in CoP), indicating a P rich
surface of Co2P and CoP phases.
Fig. 4. NH3-TPD profiles of reduced nickel phosphide samples.
one at low temperature in the range between 100 °C and 300 °C
and another at high temperature between 500 °C and 650 °C. The
low temperature peak in the profile of H3PO3/SiO2 corresponds to
the POAH groups with weak Brønsted acidity, while the high tem-
perature peak in the profile of Ni/SiO2 might be attributed to Lewis
acid sites stemming from Nid+ species bearing a small positive
charge or unreduced Ni2+ species [2]. Both Brønsted acid and Lewis
acid sites are present on NiP(x)/SiO2 samples due to the POAH
groups and Nid+/Ni2+ species, respectively. The amount of Brønsted
acid sites evidently increases as the P content increases.
3.6. CO-IR and furfural-IR
To get insight into how P influences the nature of active sites on
the catalyst surface, further characterization of the reduced NiP(x)
catalysts by CO-IR (Fig. 5) and furfural-IR (Fig. 6) was performed.
The strength of the CO adsorption is directly related to the transfer
3.5. NH3-TPD
NH3-TPD is used to characterize acid properties of the catalysts.
Fig. 4 shows the NH3-TPD profiles of reduced Ni/SiO2 and NiP(x)/
SiO2 (x = 0.33, 0.5, 1, 2). The TPD profiles consist of two peaks:
of electrons from the metal to the
More back-donation of electrons to the
ger metal-CO bond and a weaker CAO bond. The latter is reflected
p
* orbital of the CO molecule.
p
* orbital results in a stron-
Fig. 3. Ni 2p core-level (a) and P 2p core-level (b) spectra of reduced NiP(x)/SiO2 (x = 0, 0.33, 0.5, 1, and 2).
5

X. Lan, R. Pestman, Emiel J.M. Hensen et al.
Journal of Catalysis xxx (xxxx) xxx
Fig. 5. IR spectra of CO adsorbed on reduced nickel phosphide samples in the range of (a) 3000–4000 cmꢄ1 and (b) 1750–2250 cmꢄ1. Samples were exposed to 5.0 mbar CO at
25 °C.
Fig. 6. IR spectra of furfural adsorbed on (a) Ni/SiO2, (b) NiP(0.33)/SiO2, (c) NiP(0.5)/SiO2, (d) NiP(1)/SiO2,(e) NiP(2)/SiO2 of furfural during a TPD experiment under evacuation
(Black: 25 °C, red: 50 °C, blue: 75°C, pink: 100 °C, dark yellow: 125 °C, wine: 150 °C, olive: 150 °C, orange: 175 °C, dark blue: 200 °C), (f) intensity ratio
m(C@O)/m(C@C) of the
(C@O) adsorption band (1460–1480 cmꢄ1) and the(C@C) adsorption band (1665–1695 cmꢄ1).
in a lower frequency of the CO stretching vibration. CO-IR can thus
be used to probe the CAO bond strength after adsorption as well as
the electron density of the adsorption site.
1962 cmꢄ1 is attributed to CO molecules bonded to two Ni atoms
(bridge-CO species) [77,78]. The band at 2086 cmꢄ1 is assigned
to coordination of multiple CO molecules to a single coordinatively
unsaturated Ni0 site.
Fig. 5 shows the effect of varying P content on the chemical nat-
ure of catalysts by CO-IR. The bands at 3745 and 3670 cmꢄ1 are due
The infrared spectra of NiP(x)/SiO2 catalysts show the strong
to the
m
(OAH) stretching vibrations of SiAOH and PAOH surface
(POAH) vibration (3670 cmꢄ1
bands between 2096 and 2086 cmꢄ1 (linear-CO bonded on Ni sites
groups [58]. The intensity of the
m
)
[57,78–80]), NiP(0.33) and NiP(0.5) exhibit a weak band at
increases with the P content, indicating a higher amount of
Brønsted acid sites on catalyst surface. This is in alignment with
the NH3-TPD results shown above.
1905 cmꢄ1 (bridge-CO species [77,81]), and NiP(1) and NiP(2) a
very weak band at 2202 cmꢄ1 (P@C@O surface species) [77,80].
The peak broadening at 2096 cmꢄ1 is observed on NiP(0.33) and
NiP(0.5) samples likely due to the dipole-dipole interaction of
absorbed CO since the XPS results (Table S2) reveal that more
For the Ni/SiO2 catalyst, the band at 2045 cmꢄ1 is assigned to CO
molecules bonded to a single Ni atom [77,78]; the band at
6

X. Lan, R. Pestman, Emiel J.M. Hensen et al.
Journal of Catalysis xxx (xxxx) xxx
neighboring Ni sites (i.e. lower P/Ni ratios) are available on NiP
(0.33) and NiP(0.5) surfaces than on NiP(1) and NiP(2) surfaces
[82,83].
As the P/Ni molar ratio increases, the m(C@O) bands gradually
shift to lower frequency from 1675 cmꢄ1 of Ni/SiO2 to 1655 cmꢄ1
for NiP(2)/SiO2, which suggests that the interaction between the
carbonyl group and catalyst surface is notably enhanced by P addi-
tion. Furthermore, new bands at 1625 cmꢄ1 and 1622 cmꢄ1 are
observed on NiP(0.33) and NiP(0.5), indicating the formation of a
second adsorption mode of the carbonyl group. In case of NiP(1)
and NiP(2), the shoulder at 1635 cmꢄ1 may also result from this
new adsorption configuration of carbonyl groups.
The absorption band at 2086 cmꢄ1 can tentatively be assigned
to CO adsorbed on a single Ni in a phosphide matrix. As nickel in
NiP(x) is positively charged according to XPS results (see Fig. 3),
less back donation of electrons will occur creating a blue shift for
the vibrational frequency of linearly adsorbed CO from
2045 cmꢄ1 to 2086 cmꢄ1. If the P content is reduced, more adjacent
Ni will be available, increasing the dipole interaction and thus
shifting and broadening the peak towards 2096 cmꢄ1. Since both
adsorption on positively charged nickel as adsorption of multiple
CO on unsaturated nickel atoms lead to less back donation, it is
not unexpected that the adsorption band of CO adsorbed on Ni in
NiP(x) has a frequency (Fig. 5b colored lines) similar to that of mul-
tiple CO adsorbed on unsaturated sites of pure nickel (Fig. 5b, black
line). The frequency of bridging-CO bands in NiP(x)/SiO2
(1905 cmꢄ1) is lower than those in the Ni catalyst (1962 cmꢄ1),
indicating the presence of stronger bonded bridge-CO species in
NiP(x)/SiO2. This allows us to conclude that the addition of P alters
the properties of the adsorption site, apparently leading to a stron-
ger bridge-CO adsorption. The intensity of the bridge-CO bands
(1962 and 1905 cmꢄ1) decreases significantly with increasing P
content, indicating that P lowers the amount of bridge-CO species
(Fig. 5b). This is expected since P could block the availability of
adjacent Ni atoms, which would act as adsorption sites for
bridge-CO [77,78]. Since adjacent Ni sites are required for a planar
adsorption of furan, the planar adsorption of the furan ring is also
expected to decrease with increasing P content and so do the
furan-ring hydrogenation or the ring-opening reaction in the fur-
fural HDO.
Based on geometric and electronic considerations, the tendency
of planar furan ring adsorption is expected to decrease with
increasing P content. Nevertheless, no frequency shift of the
m
(C@C) band (at 1465 and 1475 cmꢄ1) is observed (Fig. 6). Likely
because the interaction between the furan-ring and the catalyst
surface is strong and the weakening effect of P on the furan-
catalyst interaction insufficient to cause a frequency shift of IR
bands. The intensity of the
decrease with increasing P content. We therefore calculated the
intensity ratios of (C@O)/ (C@C) during a temperature pro-
m(C@C) IR bands, however, clearly
m
m
grammed heating process to further investigate the adsorption sit-
uation of carbonyl groups and furan-rings on the catalyst surface.
As temperature increases above 150 °C, a significant decrease is
observed in the
m(C@O)/m(C@C) ratios on Ni/SiO2 and NiP(0.5)/
SiO2, while only a slight decrease is detected on NiP(1)/SiO2 and
NiP(2)/SiO2. This does again confirm that NiP(1)/SiO2 and NiP(2)/
SiO2 with higher P content exhibit a stronger carbonyl adsorption
than Ni/SiO2 and NiP(0.5)/SiO2 do.
The CO-IR investigation of CoP(x) was also conducted (Fig-
ure S4). As in the case of P in the NiP(x) samples, the P addition
does suppress the presence of adjacent Co sites, i.e. bridging-CO
adsorption sites. As argued above, P atoms do likely ‘‘dilute” Co
atoms at the surface hence suppressing the bridge-CO adsorption
[77]. Besides, the intensity of the 2066 cmꢄ1 band assigned to mul-
tiple CO adsorbed on unsaturated sites, declines with increasing P
content, suggesting saturation of coordinated unsaturated Co sites
by P. Furthermore, a decrease in electron density of Cod+ is
expected at higher P content due to the electron withdrawing nat-
ure of P, in line with the declining adsorption strength of linear-CO
at higher P content (Figure S5f). Since adjacent Co sites are neces-
sary for the planar adsorption of the furan ring and a lowered elec-
tron density of Cod+ contributes to weaker adsorption of furan ring,
the P addition exerts a suppressing effect on the planar adsorption
of furan rings on CoP(x) surface for reasons of geometric constrains
and electronic structure, which is very similar to the situation in
NiP(x).
The strength of CO adsorption on the catalyst surface is also
explored by IR spectra obtained during temperature programmed
desorption (Figure S3). As temperature increases, the intensity of
the linear-CO band declines due to CO desorption. Increasing P
content facilitates the desorption of linear-CO (Figure S3f), indicat-
ing a weaker CO adsorption on Ni sites at higher P content. This
confirms the conclusions drawn from Fig. 5b that the adsorption
band of linearly absorbed CO shifts to higher frequencies at higher
P content as the adsorption is weaker due to the lower electron
density on Nid+. Based on this understanding, the increasing P con-
tent is also expected to weaken the furan adsorption on the cata-
lyst surface from an electronic point of view, because it
contributes to a lower electron density on Nid+ sites providing a
lower back donation of electrons from to the d-band of Ni to the
p* bond of the furan ring.
The adsorption bands (Fig. 5b) as well as the CO desorption
behaviors (Figure S3f) are similar for NiP(0.33) and NiP(0.5) cata-
lysts, as well as for NiP(1) and NiP(2), which is consistent with
the XPS results (Table S2) showing that NiP(0.33) and NiP(0.5) have
a similar Ni reduction degree of 0.67 and 0.68, respectively, and
NiP(1) and NiP(2) of 0.74 and 0.75, respectively. This suggests that
NiP(0.33) and NiP(0.5), as well as NiP(1) and NiP(2), possess active
sites with similar geometrical and electronical properties.
In summary, P addition decreases the amount of adjacent Ni
sites and the electron density on Ni sites, which is expected to sup-
press the planar adsorption of furan ring from both geometrical
and electronical view.
4. Catalytic activity measurements
4.1. Catalyst screening
The activity of different metal phosphides was tested in furfural
HDO at a WHSV of 3 hꢄ1, a pressure of 1 bar and a temperature of
200 °C (Fig. 7a). MoP, Ni2P, Co2P and WP show promising furfural
conversions in the order Ni2P ꢀ MoP > Co2P ꢀ WP with values
around 90%, 91%, 53%, and 50%, respectively. Cu3P and Fe2P show
almost no activity and are excluded from our study. Subsequently,
Ni2P, MoP, Co2P and WP were further tested at different contact
times to explore the reaction mechanism of furfural HDO over
these catalysts (Fig. 7b,c,d,e).
According to previous research employed pure metal catalysts
[5,6,21,35], mainly two reaction pathways are proposed for fur-
fural HDO (Scheme 2). One is a hydrogenation pathway (pathway
1), in which the carbonyl group is first hydrogenated to an alco-
holic hydroxyl group, which is then removed by hydrogenolysis,
IR spectra of adsorbed furfural on silica supported Ni/SiO2 and
NiP(x)/SiO2 are shown in Fig. 6. The bands at 1675 and
1695 cmꢄ1 are assigned to the
m
(C@O) stretching vibration of
the carbonyl group; bands at 1570, 1465, and 1475 cmꢄ1 are due
to m(C@C) stretching vibration of the furan-ring [84–86]. The broad
bands between 1400 and 1300 cmꢄ1 are a result of the background
correction.
7

X. Lan, R. Pestman, Emiel J.M. Hensen et al.
Journal of Catalysis xxx (xxxx) xxx
Fig. 7. (a) HDO conversion of furfural over different metal phosphides at varying WHSV; Conversions (white squares in the bars) and selectivity of furfural HDO over (b) Ni2P,
(c) MoP, (d) Co2P and (e) WP at different WHSV. Catalyst loading = 60 mg, T = 200 °C, H2/furfural molar ratio = 74. Data collected after 2 h time on stream. MF: 2-methylfuran;
THFOL: tetrahydrofurfuryl alcohol; FOL: furfuryl alcohol; THMF: tetrahydro-2-methylfuran.
Scheme 2. Reaction pathways in the HDO of furfural according to literature [5,6,21,35]. FOL: furfuryl alcohol, MF: 2-methylfuran, THF: tetrahydrofuran, THMF: tetrahydro-2-
methylfuran, THFOL: tetrahydrofurfuryl alcohol.
8

X. Lan, R. Pestman, Emiel J.M. Hensen et al.
Journal of Catalysis xxx (xxxx) xxx
generating furfuryl alcohol (FOL) and subsequently 2-methylfuran
(MF) as products. The other one is decarbonylation (pathway 2)
yielding furan as a product. If applied catalysts (e.g., Ni, Co, Pd
(0.5), as well as the absence of THMF over NiP(1) and NiP(2) reveals
a suppressing effect of P on the furan-ring hydrogenation, i.e. a
weaker interaction between furan-ring and catalyst surface. High
reaction temperatures contribute to high production of undesired
decarbonylation and ring-opening reactions, consistent with the
effect of reaction temperature on furfural HDO over pure metal cat-
alysts [5,6,33]. The NiP(1) and NiP(2) catalysts show similar cat-
alytic performance, which can be attributed to the fact that these
catalysts contain the same active (Ni2P) phase (XRD). In line with
this, their surface characteristics as probed for XPS, CO-IR and
furfural-IR spectroscopy are also similar. The behaviors of the NiP
(0.33) and NiP(0.5) catalysts are clearly different from those of
NiP(1) and NiP(2), most likely due to the different active phases
of Ni3P and Ni12P5. Reasons of different catalytic behavior of the
Ni3P, Ni12P5 and Ni2P active phases are discussed below.
[5,33]) have
a high hydrogenation capacity, furfuryl alcohol
(FOL), 2-methylfuran (MF) and furan can be further hydrogenated
to tetrahydrofurfuryl alcohol (THFOL), tetrahydro-2-methylfuran
(THMF) and tetrahydrofuran (THF), respectively. If catalysts pos-
sess high hydrogenolysis capability, the formation of ring-
opening products (e.g. butanol, butane, pentanol, pentane, etc.)
would be obtained.
In our cases, metal phosphides show a promising performance
in the furfural HDO with activity orders of Ni2P ꢀ MoP > Co2P ꢀ
WP and a desirable product selectivity towards 2-methylfuran.
Product distributions differ across the series of metal phosphides
significantly. WP and MoP show highest selectivity towards MF
(>90%), whereas Co2P mainly produces the less-desired furan pro-
duct (>40%). This shows that each metal phosphide sample favors
the hydrogenation and decarbonylation routes in a different way.
At decreasing furfural conversions (attained by increasing
WHSV), all catalysts showed increasing furfural alcohol selectivity
at the expense of 2-methylfuran production, revealing the hydro-
genation mechanism (pathway 1) of metal phosphide catalysts.
The constant furan selectivity revealed a decarbonylation pathway
(pathway 2), which is independent from the hydrogenation path-
way. Small amounts of THFOL and THMF, and no trace of ring-
opening products were detected, indicating that furan-ring hydro-
genation and ring-opening reactions are successfully suppressed
over these catalysts. Accordingly, these catalysts can contribute
to lower hydrogen consumption and lower light-product formation
in comparison to traditional transition metal catalysts (i.e. Ni [5,6]
or Pd [5]).
4.2.1. Furan-ring/Ni(P) interaction
The interaction between the furan-ring and the catalyst surface
(labeled below as the furan-ring/Ni interaction) is essentially the
interaction between the d-band of Ni and the
p* bonding of the
furan-ring. As the Ni/SiO2 catalyst accommodates the highest elec-
tron density on Ni sites and the largest concentration of bridge-CO
sites among all of our catalysts, the furan-ring/Ni interaction is
expected to be the strongest here, contributing to ring-
hydrogenation and ring-opening reactions in the furfural conver-
sion. Therefore, it is plausible that substantial amounts of THFOL
are obtained on a Ni catalyst.
As the P content increases, the number of adjacent Ni sites,
required to form bridge-CO species, is substantially suppressed
(Fig. 5) and the electron density on Ni sites is gradually reduced
due to the electron withdrawing nature of phosphorus. From a geo-
metric viewpoint, the furan-ring/Ni interaction is expected to
decrease due to the decreasing amount of adjacent Ni sites. From
an electronic structural point of view, the interaction between
For MF production, indirect and direct reaction pathways have
been reported over metal and metal alloy surfaces (e.g., PtZn, NiFe,
and Mo2C) [6,7,20,37]. The indirect reaction pathway consists of
hydrogenation of
lowed by conversion into MF via FOL hydrogenolysis [6,20]. The
direct reaction pathway involves the conversion of the
2(C, O)
g
2(C, O) surface adsorbed species to FOL, fol-
the Ni d-band and the furan p-system is expected to decline at
higher P content due to the lower electron density on Nid+ sites.
Therefore, the furan-ring/Ni interaction as well as the ring-
hydrogenation capacity of the catalysts are expected to decline
as the P content increases, which is consistent with our activity
results that the product distribution shifts from THFOL for Ni/
SiO2 to THMF for NiP(0.33), then to MF for NiP(0.5), NiP(1), and
NiP(2).
g
species into C4H3O-CH2 or C4H3O-CH intermediates, which are
anticipated to directly produce MF in a H2-rich environment
[7,20,37]. The presence of FOL at high WHSV over Ni2P, MoP,
Co2P, and WP catalysts (Fig. 7) confirms the indirect MF formation
mechanism. Yet, the direct MF formation pathways cannot be
excluded, especially for MoP and WP, where high MF production
can be still obtained at low conversion (Fig. 7).
4.2.2. Carbonyl/Ni(P) interaction
The interaction between the carbonyl group and the catalyst
surface, which can be characterized by furfural-IR, is another key
factor influencing furfural HDO performance.
4.2. NiP(x)/SiO2
As shown above, Ni2P/SiO2 is practically the most active catalyst
among our metal phosphides with a promising product distribu-
tion. We therefore decided to study this catalyst in more detail:
Nickel phosphide samples were prepared by varying the P/Ni
molar ratios NiP(x) (with x = 0.33, 0.5, 1, 2) and were tested in
the furfural HDO (Fig. 8). Since the focus of our work is to optimize
selectivity and to develop a mechanistic explanation of the furfural
HDO reaction, instead of optimizing reaction rates, TOF investiga-
tions at low conversion are not included here. Nevertheless, upper
bound estimates of TOFs based on the given data are provided in
Table S3.
Significant yields of FOL and THFOL are produced over metallic
Ni, demonstrating the high hydrogenation ability, i.e. strong inter-
action between furfural and the Ni metal surface. Higher amounts
of MF and THMF are produced at the expense of FOL and THFOL
over NiP(0.33) and NiP(0.5), implying an enhanced C1AO1
(Scheme 1) hydrogenolysis ability of these catalysts, i.e. a stronger
interaction between the carbonyl group and the Ni-P catalyst sur-
face. The decreasing production of THMF over NiP(0.33) and NiP
According to literature, the carbonyl group adopts an
configuration on the Ni surface, which is unstable and tends to
rearrange into an
1(C) configuration at higher temperature
(Scheme 4c) [5]. As the
1(C) configuration is most likely the pre-
g
2(C, O)
g
g
cursor for decarbonylation, an increasing furan production is
observed as reaction temperature increases [5]. Our results are
consistent with those published in literature [5]: the
(C@C) ratio of the adsorbed species decreases at higher tempera-
ture (Fig. 6f) confirming the
2(C, O) ? 1(C) conversion at higher
temperature. In accordance with the formation of the
1(C) config-
uration, more furan is produced at higher temperature (Fig. 8).
After P addition, the (C@O) stretching vibration of adsorbed
m(C@O)/m
g
g
g
m
furfural shifts to lower frequency and the downward shift tends
to be more prominent when the P content increases (Fig. 6), indi-
cating a stronger carbonyl/catalyst interaction at higher P content.
It is likely that electron-deficient Nid+ binds to the lone pairs of the
carbonyl O, while electron-rich Pdꢄ donates electrons to the anti-
bonding orbitals of the C@O moiety, contributing to a stronger
carbonyl/catalyst interaction (i.e.,
a more stable g
2(C, O)
9

X. Lan, R. Pestman, Emiel J.M. Hensen et al.
Journal of Catalysis xxx (xxxx) xxx
Fig. 8. Activity results of (a) NiP(2)/SiO2, (b) NiP(1)/SiO2, (c) NiP(0.5)/SiO2, (d) NiP(0.33)/SiO2, (e) Ni/SiO2 for furfural HDO at different reaction temperature. Catalyst = 60 mg,
H2 = 20 ml/min, furfural = 0.001 ml/min in liquid. Data after 2 h on stream. MF: 2-methyl furan, FOL: furfuryl alcohol, THFOL: tetrahydrofurfuryl alcohol, THF:
tetrahydrofurfuryl alcohol, THMF: tetrahydro-2-methylfuran. The diamond in the bars indicates furfural conversion.
configuration, Scheme 4d). Such stronger carbonyl/catalyst inter-
action weakens the C1AO1 bond (Scheme 1) in the carbonyl group.
Consequently, the carbonyl hydrogenation and subsequent C1AO1
(Scheme 1) hydrogenolysis of FOL could be enhanced, which is
consistent with the product distribution shift from FOL and THFOL
of Ni catalyst to MF and THMF of NiP(0.33) catalyst.
C1AO1 (Scheme 1) conversion. Besides, phosphorus could act as
Brønsted acid which would facilitate the C1AO1 (Scheme 1)
hydrogenolysis in FOL and thus contribute to a higher production
of MF.
4.3. CoP(x) catalysts
Besides, the MF yields are relatively stable for NiP(1) and NiP(2)
at low or high furfural conversions. Likely that the Ni2P surface
exposes sites with the appropriate geometry and electronic prop-
erties that enable a direct MF formation from furfural deoxygena-
tion via C4H3OACH@ [7] or C4H3OACH2O@ [37] intermediates
(Scheme 3).
Another explanation for the changing catalytic performance by
P addition is based on the role of Brønsted acid sites: In the
hydrodeoxygenation of phenolic compounds, a Brønsted acid site
adjacent to a metal site is considered to facilitate CAO hydrogenol-
ysis by protonating the oxygen in AOH or AOCH3 groups [87,88].
Since the P species in NiP(x) catalysts could also act as Brønsted
acid sites adjacent to Ni sites, i.e. PAOH (Figs. 3 and 4), we infer
that the increasing P content in the catalyst could also offer a pro-
moting effect on C1AO1 (Scheme 1) hydrogenolysis in FOL and
THFOL, contributing to enhanced production of MF and THMF.
Scheme 5 illustrates above findings and possible explanation: in
the metal phosphides (in contrast to the pure metal) the weakened
C@C adsorption leads to less furan-ring hydrogenation products,
and the strengthened C@O adsorption contributes to easier
Since Co2P is the third active catalyst in furfural HDO (Fig. 6)
and the active phases of cobalt phosphide are adjustable by varying
the molar P/Co ratio, we investigated the effect of the P stoichiom-
etry in CoP(x) phases on the furfural HDO as well (Figure S6). The
conversion decreases with increasing P content, indicating the
decreasing activity in the order CoP < Co2P < Co.
Similar to NiP(x), the ring-opening products are suppressed
with increasing P content. This is understandable, since as in the
case of the nickel phosphides, P suppresses the formation of adja-
cent Co sites required for the planar adsorption of the furan ring,
and lowers the electron density of Cod+ sites leading to a weaker
interaction between furan ring and catalyst surface. In addition,
the MF production is enhanced at the expense of the FOL produc-
tion by increasing the P content as a result of Brønsted acid sites
facilitating the C1AO1 (Scheme 1) bond breaking.
A higher furan production is found on CoP(x) than on NiP(x); a
possible explanation for this observation is as follows. The
g
2(C, O)
configurations on the catalyst surface are stabilized by the car-
bonyl O donating electrons to electron-deficient Cod+ or Nid+ sites
10

X. Lan, R. Pestman, Emiel J.M. Hensen et al.
Journal of Catalysis xxx (xxxx) xxx
Scheme 3. Possible direct and indirect formation mechanism of methylfuran [6,7,20,37].
Scheme 4. Adsorption configurations of furfural on Ni and NiP(x) surfaces.
Scheme 5. Effect of P on furfural HDO over Ni and nickel phosphide catalysts (Catalyst = 60 mg, T = 180 °C, H2 = 20 ml/min, furfural = 0.001 ml/min in liquid.)
and the antibonding orbitals of C@O accepting electron density
from electron-rich Pdꢄ. Since the density of Cod+ sites on CoP(x)
surfaces is lower than that of Nid+ on NiP(x) (P/Co > P/Ni,
Table S2), the binding of carbonyl O to the catalyst surface will
O) ? g
1(C) transformation and, consequently to a higher furan for-
mation rate on CoP(x) catalysts.
Overall, we summarize that increasing the P/Co stoichiometry
in CoP(x) samples suppresses the ring-opening reactions but does
not suppress the undesired decarbonylation. In comparison with
be lower on CoP(x) surfaces, which contributes to an easier
g
2(C,
11

X. Lan, R. Pestman, Emiel J.M. Hensen et al.
Journal of Catalysis xxx (xxxx) xxx
[3] Z. He, X. Wang, Hydrodeoxygenation of model compounds and catalytic systems
for pyrolysis bio-oils upgrading, Cataly. Sustain. Energy 1 (2012) 28–52.
[4] W. Zhang, Y. Zhu, S. Niu, Y. Li, A study of furfural decarbonylation on K-doped
Pd/Al2O3 catalysts, J. Mol. Catal. A: Chem. 335 (2011) 71–81.
[5] S. Sitthisa, D.E. Resasco, Hydrodeoxygenation of Furfural Over Supported Metal
Catalysts: A Comparative Study of Cu, Pd and Ni, Cataly. Lett. 141 (2011) 784–
791.
[6] S. Sitthisa, W. An, D.E. Resasco, Selective conversion of furfural to methylfuran
over silica-supported NiFe bimetallic catalysts, J. Catal. 284 (2011) 90–101.
[7] D. Shi, J.M. Vohs, Deoxygenation of Biomass-Derived Oxygenates: Reaction of
Furfural on Zn-Modified Pt(111), ACS Catal. 5 (2015) 2177–2183.
[8] R. He, X.P. Ye, B.C. English, J.A. Satrio, Influence of pyrolysis condition on
switchgrass bio-oil yield and physicochemical properties, Bioresour. Technol.
100 (2009) 5305–5311.
NiP(x), CoP(x) catalysts are not desirable in furfural conversion due
to the higher furan production and the lower MF yield (furan has a
lower C yield than MF in furfural conversion). Our characterization
and activity results of CoP(x) and NiP(x) demonstrate that the type
of metal phosphide and the phosphorus/metal ratio plays a crucial
role in determining the furfural adsorption on the catalyst surface,
as well as the catalytic performance in furfural HDO.
5. Conclusions
A series of transition metal phosphides was evaluated in the
furfural HDO reaction, showing the following activity trend: Ni2P
ꢀ MoP > Co2P ꢀ WP ꢁ Cu3P > Fe2P. In comparison to traditional
transition metal catalysts (i.e. Ni or Pd), metal phosphides like
Ni2P, MoP, Co2P and WP are promising catalysts for furfural HDO
as they produce 2-methylfuran and furan as major products and
contribute to lower hydrogen consumption and lower light-
product formation.
[9] J. Adam, M. Blazsó, E. Mészáros, M. Stöcker, M.H. Nilsen, A. Bouzga, J.E. Hustad,
M. Grønli, G. Øye, Pyrolysis of biomass in the presence of Al-MCM-41 type
catalysts, Fuel 84 (2005) 1494–1502.
[10] L. Baijun, L. Lianhai, W. Bingchun, C. Tianxi, I. Katsuyoshi, Liquid phase
selective hydrogenation of furfural on Raney nickel modified by impregnation
of salts of heteropolyacids, Appl. Catal. A 171 (1998) 117–122.
[11] N. Lohitharn, B.H. Shanks, Upgrading of bio-oil: Effect of light aldehydes on
acetic acid removal via esterification, Catal. Commun. 11 (2009) 96–99.
[12] H.-Y. Zheng, Y.-L. Zhu, L. Huang, Z.-Y. Zeng, H.-J. Wan, Y.-W. Li, Study on Cu–
Mn–Si catalysts for synthesis of cyclohexanone and 2-methylfuran through
the coupling process, Catal. Commun. 9 (2008) 342–348.
By varying the P/Ni molar ratios in NiP(x) precursors, the effect
of P stoichiometry on catalyst properties and performance are
investigated in depth. A higher P content weakens the furan-ring/
catalyst interaction contributing to lower furan ring-
hydrogenation and ring-opening reaction. On the other hand, it
provides a stronger carbonyl/catalyst interaction by providing a
´
[13] J. Kijenski, P. Winiarek, T. Paryjczak, A. Lewicki, A. Mikołajska, Platinum
deposited on monolayer supports in selective hydrogenation of furfural to
furfuryl alcohol, Appl. Catal. A 233 (2002) 171–182.
[14] H. Li, H. Luo, L. Zhuang, W. Dai, M. Qiao, Liquid phase hydrogenation of furfural
to furfuryl alcohol over the Fe-promoted Ni-B amorphous alloy catalysts, J.
Mol. Catal. A: Chem. 203 (2003) 267–275.
[15] J. Wu, G. Gao, J. Li, P. Sun, X. Long, F. Li, Efficient and versatile CuNi alloy
nanocatalysts for the highly selective hydrogenation of furfural, Appl. Catal. B
203 (2017) 227–236.
[16] M.J. Taylor, L.J. Durndell, M.A. Isaacs, C.M.A. Parlett, K. Wilson, A.F. Lee, G.
Kyriakou, Highly selective hydrogenation of furfural over supported Pt
nanoparticles under mild conditions, Appl. Catal. B 180 (2016) 580–585.
[17] S. Sitthisa, T. Sooknoi, Y. Ma, P.B. Balbuena, D.E. Resasco, Kinetics and
mechanism of hydrogenation of furfural on Cu/SiO2 catalysts, J. Catal. 277
(2011) 1–13.
[18] B.M. Nagaraja, A.H. Padmasri, B. David Raju, K.S. Rama Rao, Vapor phase
selective hydrogenation of furfural to furfuryl alcohol over Cu–MgO
coprecipitated catalysts, J. Mol. Catal. A: Chem. 265 (2007) 90–97.
[19] X. Chen, H. Li, H. Luo, M. Qiao, Liquid phase hydrogenation of furfural to
furfuryl alcohol over Mo-doped Co-B amorphous alloy catalysts, Appl. Catal. A
233 (2002) 13–20.
[20] R. Mariscal, P. Maireles-Torres, M. Ojeda, I. Sádaba, M. López Granados,
Furfural: a renewable and versatile platform molecule for the synthesis of
chemicals and fuels, Energy Environ. Sci. 9 (2016) 1144–1189.
[21] H.-Y. Zheng, Y.-L. Zhu, B.-T. Teng, Z.-Q. Bai, C.-H. Zhang, H.-W. Xiang, Y.-W. Li,
Towards understanding the reaction pathway in vapour phase hydrogenation
of furfural to 2-methylfuran, J. Mol. Catal. A: Chem. 246 (2006) 18–23.
[22] H. Müller, Tetrahydrofuran, in: Ullmann’s Encyclopedia of Industrial
Chemistry, Wiley-VCH Verlag GmbH & Co. KGaA, 2000.
more stable
g
2(C, O) configuration with the electron-deficient
Nid+ binding to the lone pairs of carbonyl O and the electron-rich
Pdꢄ donating electrons to the antibonding orbitals of C@O. Such
enhanced carbonyl/catalyst interaction weakens the CAO bond in
the carbonyl group promoting its hydrogenation and further con-
version. Moreover, P species could also act as Brønsted acid sites
that facilitate hydrogenolysis of FOL and THFOL, contributing to a
higher production of MF and THMF. A comparable role of P was
observed in CoP(x) samples, except that the undesired decarbony-
lation is suppressed to a lesser extent. This is likely because the
density of available Cod+ sites on CoP(x) surfaces is lower than that
of Nid+ on NiP(x). This reduces the binding of the carbonyl O to the
catalyst surface, thereby contributing to an easier
g g
2(C, O) ? 1(C)
transformation and consequently a higher furan formation rate on
CoP catalysts.
Declaration of Competing Interest
´
[23] D. Artur Malinowski, WardzinskaCatalytic conversion of furfural towards fuel
biocomponents, CHEMIK nauka-technika-rynek 66 (2012) 982–990.
[24] J.N. Chheda, J.A. Dumesic, An overview of dehydration, aldol-condensation and
hydrogenation processes for production of liquid alkanes from biomass-
derived carbohydrates, Catal. Today 123 (2007) 59–70.
[25] R.M. West, E.L. Kunkes, D.A. Simonetti, J.A. Dumesic, Catalytic conversion of
biomass-derived carbohydrates to fuels and chemicals by formation and
upgrading of mono-functional hydrocarbon intermediates, Catal. Today 147
(2009) 115–125.
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. We consulted the
author’s guide when submitting the manuscript. Manuscripts are
prepared in accordance with publishing ethics policies described
in the Author’s Guide.
[26] C.J. Barrett, J.N. Chheda, G.W. Huber, J.A. Dumesic, Single-reactor process for
sequential aldol-condensation and hydrogenation of biomass-derived
compounds in water, Appl. Catal. B 66 (2006) 111–118.
Acknowledgements
[27] W. Yu, K. Xiong, N. Ji, M.D. Porosoff, J.G. Chen, Theoretical and experimental
studies of the adsorption geometry and reaction pathways of furfural over
FeNi bimetallic model surfaces and supported catalysts, J. Catal. 317 (2014)
253–262.
[28] S. Liu, Y. Amada, M. Tamura, Y. Nakagawa, K. Tomishige, One-pot selective
conversion of furfural into 1,5-pentanediol over a Pd-added Ir-ReOx/SiO2
bifunctional catalyst, Green Chem. 16 (2014) 617–626.
[29] F. Dong, G. Ding, H. Zheng, X. Xiang, L. Chen, Y. Zhu, Y. Li, Highly dispersed Cu
nanoparticles as an efficient catalyst for the synthesis of the biofuel 2-
methylfuran, Catal. Sci. Technol. 6 (2016) 767–779.
[30] F. Dong, Y. Zhu, H. Zheng, Y. Zhu, X. Li, Y. Li, Cr-free Cu-catalysts for the
selective hydrogenation of biomass-derived furfural to 2-methylfuran: The
synergistic effect of metal and acid sites, J. Mol. Catal. A: Chem. 398 (2015)
140–148.
The China Scholarship Council (CSC) is acknowledged for finan-
cial support. We thank Tiny Verhoeven for performing XPS mea-
surements and Adelheid Elemans for the ICP measurements.
Appendix A. Supplementary material
Supplementary data to this article can be found online at
https://doi.org/10.1016/j.jcat.2021.01.031.
References
[31] L.W. Burnett, I.B. Johns, R.F. Holdren, R.M. Hixon, Production of 2-Methylfuran
by Vapor-Phase Hydrogenation of Furfural, Ind. Eng. Chem. 40 (1948) 502–505.
[32] J. Yang, H.-Y. Zheng, Y.-L. Zhu, G.-W. Zhao, C.-H. Zhang, B.-T. Teng, H.-W. Xiang,
Y. Li, Effects of calcination temperature on performance of Cu–Zn–Al catalyst
[1] H.Y. Zhao, D. Li, P. Bui, S.T. Oyama, Hydrodeoxygenation of guaiacol as model
compound for pyrolysis oil on transition metal phosphide hydroprocessing
catalysts, Appl. Catal. A 391 (2011) 305–310.
[2] K. Li, R. Wang, J. Chen, Hydrodeoxygenation of Anisole over Silica-Supported
Ni2P, MoP, and NiMoP Catalysts, Energy Fuels 25 (2011) 854–863.
for synthesizing
c-butyrolactone and 2-methylfuran through the coupling of
dehydrogenation and hydrogenation, Catal. Commun. 5 (2004) 505–510.
12

X. Lan, R. Pestman, Emiel J.M. Hensen et al.
Journal of Catalysis xxx (xxxx) xxx
[33] C.L. Wilson, 18. Reactions of furan compounds. Part V. Formation of furan from
furfuraldehyde by the action of nickel or cobalt catalysts: importance of added
hydrogen, J. Chem. Soc. (Resumed) (1945) 61–63.
[34] J. Luo, M. Monai, H. Yun, L. Arroyo-Ramírez, C. Wang, C.B. Murray, P.
Fornasiero, R.J. Gorte, The H2 Pressure Dependence of Hydrodeoxygenation
Selectivities for Furfural Over Pt/C Catalysts, Catal. Lett. 146 (2016) 711–717.
[35] S. Sitthisa, T. Pham, T. Prasomsri, T. Sooknoi, R.G. Mallinson, D.E. Resasco,
Conversion of furfural and 2-methylpentanal on Pd/SiO2 and Pd–Cu/SiO2
catalysts, J. Catal. 280 (2011) 17–27.
[62] S.-K. Wu, P.-C. Lai, Y.-C. Lin, H.-P. Wan, H.-T. Lee, Y.-H. Chang, Atmospheric
Hydrodeoxygenation of Guaiacol over Alumina-, Zirconia-, and Silica-Supported
Nickel Phosphide Catalysts, ACS Sustain. Chem. Eng. 1 (2013) 349–358.
[63] C.P. Jiménez-Gómez, J.A. Cecilia, R. Moreno-Tost, P. Maireles-Torres, Nickel
Phosphide/Silica Catalysts for the Gas-Phase Hydrogenation of Furfural to
High–Added–Value Chemicals, ChemCatChem 9 (2017) 2881–2889.
[64] Y. Wang, F. Liu, H. Han, L. Xiao, W. Wu, Metal Phosphide: A Highly Efficient
Catalyst for the Selective Hydrodeoxygenation of Furfural to 2-Methylfuran,
ChemistrySelect 3 (2018) 7926–7933.
[36] J.R. McManus, J.M. Vohs, Deoxygenation of glycolaldehyde and furfural on
Mo2C/Mo(100), Surf. Sci. 630 (2014) 16–21.
[37] Y. Shi, Y. Yang, Y.-W. Li, H. Jiao, Mechanisms of Mo2C(101)-Catalyzed Furfural
Selective Hydrodeoxygenation to 2-Methylfuran from Computation, ACS Catal.
6 (2016) 6790–6803.
[38] K. Xiong, W.-S. Lee, A. Bhan, J.G. Chen, Molybdenum Carbide as a Highly
Selective Deoxygenation Catalyst for Converting Furfural to 2-Methylfuran,
ChemSusChem 7 (2014) 2146–2149.
[65] Y. Wang, X. Feng, S. Yang, L. Xiao, W. Wu, Influence of acidity on the catalytic
performance of Ni2P in liquid-phase hydrodeoxygenation of furfural to 2-
methylfuran, J. Nanopart. Res. 22 (2020) 67.
[66] R.W. Gosselink, D.R. Stellwagen, J.H. Bitter, Tungsten-Based Catalysts for
Selective Deoxygenation, Angew. Chem. Int. Ed. 52 (2013) 5089–5092.
[67] Y. Yang, C. Ochoa-Hernández, P. Pizarro, V.A. de la Peña O’Shea, J.M. Coronado,
D.P. Serrano, Synthesis of Nickel Phosphide Nanorods as Catalyst for the
Hydrotreating of Methyl Oleate, Top. Catal. 55 (2012) 991–998.
[39] Z. Lin, R. Chen, Z. Qu, J.G. Chen, Hydrodeoxygenation of biomass-derived
oxygenates over metal carbides: from model surfaces to powder catalysts,
Green Chem. 20 (2018) 2679–2696.
[40] B.-J. Liaw, S.-J. Chiang, S.-W. Chen, Y.-Z. Chen, Preparation and catalysis of
amorphous CoNiB and polymer-stabilized CoNiB catalysts for hydrogenation
of unsaturated aldehydes, Appl. Catal. A 346 (2008) 179–188.
[68] Y. Yang, J. Chen, H. Shi, Deoxygenation of Methyl Laurate as a Model
Compound to Hydrocarbons on Ni2P/SiO2, Ni2P/MCM-41, and Ni2P/SBA-15
Catalysts with Different Dispersions, Energy Fuels 27 (2013) 3400–3409.
[69] J. Chen, Y. Yang, H. Shi, M. Li, Y. Chu, Z. Pan, X. Yu, Regulating product
distribution in deoxygenation of methyl laurate on silica-supported Ni–Mo
phosphides: Effect of Ni/Mo ratio, Fuel 129 (2014) 1–10.
[41] S.-P. Lee, Y.-W. Chen, Selective hydrogenation of furfural on Ni-P-B
nanometals, in: A. Corma, F.V. Melo, S. Mendioroz, J.L.G. Fierro (Eds.), Studies
in Surface Science and Catalysis, Elsevier, 2000, pp. 3483–3488.
[42] X. Lan, E.J.M. Hensen, T. Weber, Silica-supported Ni2P: Effect of preparation
conditions on structure and catalytic performance in thiophene
hydrodesulfurization (HDS), Catal. Today 292 (2017) 121–132.
[70] S.T. Oyama, X. Wang, Y.K. Lee, K. Bando, F.G. Requejo, Effect of Phosphorus
Content in Nickel Phosphide Catalysts Studied by XAFS and Other Techniques,
J. Catal. 210 (2002) 207–217.
[71] H. Xin, K. Guo, D. Li, H. Yang, C. Hu, Production of high-grade diesel from
palmitic acid over activated carbon-supported nickel phosphide catalysts,
Appl. Catal. B 187 (2016) 375–385.
[43] S.T. Oyama, Novel catalysts for advanced hydroprocessing: transition metal
phosphides, J. Catal. 216 (2003) 343–352.
[44] S.T. Oyama, T. Gott, H. Zhao, Y.-K. Lee, Transition metal phosphide
hydroprocessing catalysts: A review, Catal. Today 143 (2009) 94–107.
[45] W. Li, B. Dhandapani, S.T. Oyama, Molybdenum phosphide: A novel catalyst for
hydrodenitrogenation, Chem. Lett. (1998) 207–208.
[46] R. Prins, M. Bussell, Metal Phosphides: Preparation, Characterization and
Catalytic Reactivity, Catal. Lett. 142 (2012) 1413–1436.
[47] C. Stinner, R. Prins, T. Weber, Formation, Structure, and HDN Activity of
Unsupported Molybdenum Phosphide, J. Catal. 191 (2000) 438–444.
[48] C. Stinner, R. Prins, T. Weber, Binary and Ternary Transition-Metal Phosphides
as HDN Catalysts, J. Catal. 202 (2001) 187–194.
[49] P. Bui, J.A. Cecilia, S.T. Oyama, A. Takagaki, A. Infantes-Molina, H. Zhao, D. Li, E.
Rodríguez-Castellón, A. Jiménez López, Studies of the synthesis of transition
metal phosphides and their activity in the hydrodeoxygenation of a biofuel
model compound, J. Catal. 294 (2012) 184–198.
[50] Y. Li, X. Yang, L. Zhu, H. Zhang, B. Chen, Hydrodeoxygenation of phenol as a
bio-oil model compound over intimate contact noble metal-Ni2P/SiO2
catalysts, RSC Adv. 5 (2015) 80388–80396.
[51] I.V. Shamanaev, I.V. Deliy, P.V. Aleksandrov, E.Y. Gerasimov, V.P. Pakharukova,
E.G. Kodenev, A.B. Ayupov, A.S. Andreev, O.B. Lapina, G.A. Bukhtiyarova, Effect
of precursor on the catalytic properties of Ni2P/SiO2 in methyl palmitate
hydrodeoxygenation, RSC Adv. 6 (2016) 30372–30383.
[72] J.A. Cecilia, A. Infantes-Molina, E. Rodríguez-Castellón, A. Jiménez-López, S.T.
Oyama, Oxygen-removal of dibenzofuran as a model compound in biomass
derived bio-oil on nickel phosphide catalysts: Role of phosphorus, Appl. Catal.
B 136–137 (2013) 140–149.
[73] Y.-K. Lee, Y. Shu, S.T. Oyama, Active phase of a nickel phosphide (Ni2P) catalyst
supported on KUSY zeolite for the hydrodesulfurization of 4,6-DMDBT, Appl.
Catal. A 322 (2007) 191–204.
[74] A. Iino, A. Cho, A. Takagaki, R. Kikuchi, S. Ted Oyama, Kinetic studies of
hydrodeoxygenation of 2-methyltetrahydrofuran on a Ni2P/SiO2 catalyst at
medium pressure, J. Catal. 311 (2014) 17–27.
[75] J.A. Cecilia, I. Jiménez-Morales, A. Infantes-Molina, E. Rodríguez-Castellón, A.
Jiménez-López, Influence of the silica support on the activity of Ni and Ni2P
based catalysts in the hydrodechlorination of chlorobenzene. Study of factors
governing catalyst deactivation, J. Mol. Catal. A: Chem. 368 (2013) 78–87.
[76] A.W. Burns, K.A. Layman, D.H. Bale, M.E. Bussell, Understanding the
relationship between composition and hydrodesulfurization properties for
cobalt phosphide catalysts, Appl. Catal. A 343 (2008) 68–76.
[77] K.A. Layman, M.E. Bussell, Infrared Spectroscopic Investigation of CO
Adsorption on Silica-Supported Nickel Phosphide Catalysts, J. Phys. Chem. B
108 (2004) 10930–10941.
[78] C.H. Rochester, R.J. Terrell, Infrared study of effects of sulphur-poisoning on
the adsorption of carbon monoxide by nickel, J. Chem. Soc., Faraday Trans. 1:
Phys. Chem. Condens. Phases 73 (1977) 609–621.
[52] M.E. Witzke, A. Almithn, C.L. Coonrod, D.D. Hibbitts, D.W. Flaherty,
Mechanisms and Active Sites for C-O Bond Rupture within 2-
[79] E.B. Pereira, G.-A. Martin, Alcohol synthesis from syngas over nickel catalysts:
Effect of copper and lithium addition, Appl. Catal. A 103 (1993) 291–309.
[80] T.I. Korányi, É. Pfeifer, J. Mihály, K. Föttinger, Infrared Spectroscopic
Investigation of CO Adsorption on SBA-15- and KIT-6-Supported Nickel
Phosphide Hydrotreating Catalysts, J. Phys. Chem. A 112 (2008) 5126–5130.
[81] S.J. Sawhill, K.A. Layman, D.R. Van Wyk, M.H. Engelhard, C. Wang, M.E. Bussell,
Thiophene hydrodesulfurization over nickel phosphide catalysts: effect of the
precursor composition and support, J. Catal. 231 (2005) 300–313.
[82] G. Rupprechter, T. Dellwig, H. Unterhalt, H.-J. Freund, CO adsorption on Ni
(100) and Pt(111) studied by infrared–visible sum frequency generation
spectroscopy: design and application of an SFG-compatible UHV–high-
pressure reaction cell, Top. Catal. 15 (2001) 19–26.
[83] W. Erley, H. Wagner, H. Mach, Adsorption sites and long range order —
vibrational spectra for CO on Ni(111), Surf. Sci. 80 (1979) 612–619.
[84] M. Baia, S. Astilean, T. Iliescu, Other Molecules with Pharmacological Activity,
in: Raman and SERS Investigations of Pharmaceuticals, Springer, Berlin
Heidelberg, Berlin, Heidelberg, 2008, pp. 143–185.
Methyltetrahydrofuran over Ni, Ni12P5, and Ni2P Catalysts, ACS Catal.
(2018) 7141–7157.
8
[53] M.A. Golubeva, E.M. Zakharyan, A.L. Maximov, Transition Metal Phosphides (Ni
Co, Mo, W) for Hydrodeoxygenation of Biorefinery Products (a Review), Pet.
Chem. 60 (2020) 1109–1128.
[54] S. Boullosa-Eiras, R. Lødeng, H. Bergem, M. Stöcker, L. Hannevold, E.A. Blekkan,
Catalytic hydrodeoxygenation (HDO) of phenol over supported molybdenum
carbide, nitride, phosphide and oxide catalysts, Catal. Today 223 (2014) 44–53.
[55] D.J. Rensel, J. Kim, Y. Bonita, J.C. Hicks, Investigating the multifunctional nature
of bimetallic FeMoP catalysts using dehydration and hydrogenolysis reactions,
Appl. Catal. A 524 (2016) 85–93.
[56] D.J. Rensel, S. Rouvimov, M.E. Gin, J.C. Hicks, Highly selective bimetallic FeMoP
catalyst for C-O bond cleavage of aryl ethers, J. Catal. 305 (2013) 256–263.
[57] H. Zhao, S.T. Oyama, H.-J. Freund, R. Włodarczyk, M. Sierka, Nature of active
sites in Ni2P hydrotreating catalysts as probed by iron substitution, Appl.
Catal. B 164 (2015) 204–216.
[58] J.-S. Moon, E.-G. Kim, Y.-K. Lee, Active sites of Ni2P/SiO2 catalyst for
hydrodeoxygenation of guaiacol: A joint XAFS and DFT study, J. Catal. 311
(2014) 144–152.
[85] M. Rogojerov, G. Keresztury, B. Jordanov, Vibrational spectra of partially
oriented molecules having two conformers in nematic and isotropic solutions:
furfural and 2-chlorobenzaldehyde, Spectrochim. Acta Part A Mol. Biomol.
Spectrosc. 61 (2005) 1661–1670.
[59] J.-S. Moon, Y.-K. Lee, Support Effects of Ni2P Catalysts on the
Hydrodeoxygenation of Guaiacol: In Situ XAFS Studies, Topics Cataly. 58
(2015) 211–218.
[86] A.A. El-Azhary, R.H. Hilal, Vibrational analysis of the spectra of furan and
thiophene, Spectrochim. Acta Part A Mol. Biomol. Spectrosc. 53 (1997) 1365–
1373.
[60] S. Ted Oyama, T. Onkawa, A. Takagaki, R. Kikuchi, S. Hosokai, Y. Suzuki, K.K.
Bando, Production of Phenol and Cresol from Guaiacol on Nickel Phosphide
Catalysts Supported on Acidic Supports, Top. Catal. 58 (2015) 201–210.
[61] S.-K. Wu, P.-C. Lai, Y.-C. Lin, Atmospheric Hydrodeoxygenation of Guaiacol
over Nickel Phosphide Catalysts: Effect of Phosphorus Composition, Catal. Lett.
144 (2014) 878–889.
[87] W. Song, Y. Liu, E. Barath, C. Zhao, J.A. Lercher, Synergistic effects of Ni and acid
sites for hydrogenation and C-O bond cleavage of substituted phenols, Green
Chem. 17 (2015) 1204–1218.
[88] X. Zhu, L.L. Lobban, R.G. Mallinson, D.E. Resasco, Bifunctional transalkylation
and hydrodeoxygenation of anisole over a Pt/HBeta catalyst, J. Catal. 281
(2011) 21–29.
13