Iron phosphides presenting different stoichiometry as nanocatalysts in the HDO of phenol

Article abstract of DOI:10.1016/j.cattod.2018.05.023

Iron phosphide catalysts supported on silica with an iron loading of 15 wtpercent were synthesized and studied in the hydrodeoxygenation (HDO) of phenol. The amount of phosphorus varied in order to obtain iron phosphides with different stoichiometry. Catalysts containing Fe₂P, FeP and FeP₂ phases were obtained. The textural and structural properties of the prepared catalysts were evaluated by using different experimental techniques such as N₂ adsorption-desorption at -196 °C, X-ray diffraction (XRD), M?ssbauer spectroscopy, high resolution transmission spectroscopy (HRTEM), infrared spectroscopy (IR) of adsorbed CO at low temperature, X-ray photoelectron microscopy (XPS) and NH₃ thermoprogrammed desorption (NH₃-TPD). The catalytic activity was studied at 275 °C and at 15 and 5 bar of hydrogen pressure in the hydrodeoxygenation reaction of phenol. Characterization results evidenced that the initial P/Fe ratio employed in the synthesis not only governed the stoichiometry of the iron phosphide, but also the particle size, metallic surface exposure and acidity. The catalysts presenting unique phases were those presenting better activity in the HDO reaction of phenol. Moreover, Fe₂P phase presented better results than FeP in terms of HDO conversion.

Full text of DOI:10.1016/j.cattod.2018.05.023

Catalysis Today xxx (xxxx) xxx–xxx
Contents lists available at ScienceDirect
Catalysis Today
journal homepage: www.elsevier.com/locate/cattod
Iron phosphides presenting different stoichiometry as nanocatalysts in the
HDO of phenol
a
a,⁎
a
b
c
E. Rodríguez-Aguado , A. Infantes-Molina , D. Ballesteros-Plata , J.F. Marco , E. Moretti ,
d
a
a,⁎
E. Finocchio , J.A. Cecilia , E. Rodríguez-Castellón
a
Departamento de Química Inorgánica, Cristalografía y Mineralogía (Unidad Asociada al ICP-CSIC), Facultad de Ciencias, Universidad de Málaga, Campus de Teatinos,
9071 Málaga, Spain
Instituto de Química Física Rocasolano, CSIC, Madrid, Spain
Dipartimento di Scienze Molecolari e Nanosistemi, Università Ca' Foscari Venezia, INSTM Venice Research Unit, Mestre Venezia, Italy
Dipartimento di Ingegneria Civile, Chimica e Ambientale, Universita degli Studi di Genova, Génova, Italy
2
b
c
d
A R T I C L E I N F O
A B S T R A C T
Keywords:
Iron phosphide
HDO
Heterogeneous catalysis
Silica
Phenol
Iron phosphide catalysts supported on silica with an iron loading of 15 wt% were synthesized and studied in the
hydrodeoxygenation (HDO) of phenol. The amount of phosphorus varied in order to obtain iron phosphides with
different stoichiometry. Catalysts containing Fe
structural properties of the prepared catalysts were evaluated by using different experimental techniques such as
adsorption-desorption at -196 °C, X-ray diffraction (XRD), Mössbauer spectroscopy, high resolution trans-
mission spectroscopy (HRTEM), infrared spectroscopy (IR) of adsorbed CO at low temperature, X-ray photo-
electron microscopy (XPS) and NH thermoprogrammed desorption (NH -TPD). The catalytic activity was stu-
2 2
P, FeP and FeP phases were obtained. The textural and
N
2
3
3
died at 275 °C and at 15 and 5 bar of hydrogen pressure in the hydrodeoxygenation reaction of phenol.
Characterization results evidenced that the initial P/Fe ratio employed in the synthesis not only governed the
stoichiometry of the iron phosphide, but also the particle size, metallic surface exposure and acidity. The cat-
alysts presenting unique phases were those presenting better activity in the HDO reaction of phenol. Moreover,
2
Fe P phase presented better results than FeP in terms of HDO conversion.
1. Introduction
challenges, which have to be solved to expand its use in the already
existing biorefineries [7,8]. In this respect, the consumption of low le-
vels of hydrogen along with the obtainment of a high degree of deox-
ygenation is one of the most important challenges. Therefore, the cat-
alyst design is a key factor to cope with that purpose. Typically,
catalysts used in the HDO reactions are constituted by a metal with
good hydrogen transfer properties supported on a high surface area
oxide support. Conventional sulfide catalysts such as CoMoS and
NiMoS, used in petroleum hydrodesulfurization, have been used in
HDO reactions; however, the oxidation of the sulfide species by the
Increasingly awareness on polluted atmospheric environment com-
bined with depletion of fossil fuels and increased reliance on un-
conventional crude oil resources which need to be highly purified, are
prompting many researchers to address this problematic. Thus, new
approaches on hydrotreating processes are being tackled as well as the
search of new catalytic materials for processing fossil fuel feedstocks
[
1]. In this sense, fast pyrolysis of solid lignocellulosic biomass has
drawn considerable attention for the production of liquid biofuels as
this type of biomass is renewable and easily available [2–4]. Indeed,
this process can be efficiently applied to commercial plants without
investing a high amount of capital. However, the obtained bio-oil must
be upgraded to ensure the removal of a wide variety of oxygenated
compounds. The removal of oxygen can be tackled by means of hy-
drodeoxygenation, which is regarded as an effective method to improve
the effective H/C ratio [5,6]. During HDO, oxygenated compounds are
converted to hydrocarbons under hydrogen pressure and in the pre-
sence of heterogeneous catalysts. Currently, HDO faces several
2
presence of H O, obtained as by-product in the HDO reaction, and the
formation of carbonaceous deposits cause a strong deactivation of the
active phase [9,10]. Catalysts based on noble metals have also been
widely studied but their high activity does not offset their high cost.
That is why alternative catalysts as those constituted by transition metal
phosphides have attracted great attention in recent years since they
have much lower costs and their electronic structure is close to that of
noble metals [11,12]. Concretely, transition metal phosphides have
resulted to be very active in hydrotreating reactions also showing high
⁎
Corresponding authors.
E-mail addresses: ainfantes@uma.es (A. Infantes-Molina), castellon@uma.es (E. Rodríguez-Castellón).
https://doi.org/10.1016/j.cattod.2018.05.023
Received 17 December 2017; Received in revised form 5 May 2018; Accepted 10 May 2018
0920-5861/ © 2018 Elsevier B.V. All rights reserved.
Please cite this article as: Rodríguez-Aguado, E., Catalysis Today (2018), https://doi.org/10.1016/j.cattod.2018.05.023

E. Rodríguez-Aguado et al.
Catalysis Today xxx (xxxx) xxx–xxx
stability and resistance to deactivation in the presence of water. Several
studies of the HDO of diluted oxygenated compound solutions with
transition metal phosphides have been reported, having shown re-
markable results in terms of activity, stability and durability [12–16].
Most of the studies have proved that nickel phosphide based catalysts
display the most suitable properties to carry out an excellent perfor-
mance on the HDO process [17–19]. However, although it is necessary
to search highly active catalysts it is also essential to employ cheaper,
abundant and non-toxic active phases, which is another challenge that
HDO must face to be implemented industrially. In that sense, the pre-
sent research has synthetized a series of phosphides containing iron as
active phase to cope with that challenge.
The use of iron phosphide in the hydrotreating field has been scarce.
It has been reported in the literature that Oyama et al. [12–14] carried
out preliminary analysis in hydrodesulfurization (HDS), hydro-
denitrogenation (HDN) and hydrodeoxygenation (HDO) reactions for
target molecules coming from the petrochemical industry and lig-
nocellulosic biomass. More recently, Huang et al. [20] have also studied
the catalytic activity of iron phosphide based catalysts in the HDS of
thiophene; while Yuan et al. [21] have evaluated the catalytic behavior
Fe molar ratios of 0.5, 1, 2 and 3 as described elsewhere [30], where it
is described the preparation of phosphides by using hydrogenopho-
sphite as a precursor salt. The impregnated supports were air-treated at
60 °C for 24 h and finally, the precursors were reduced ex-situ under a
−
1
hydrogen flow of 100 mL min in a tubular reactor by heating from rt
−
1
to 300 °C (over a temperature linear ramp of 10 °C min and from 300
−1
to 700 °C (3 °C min ) remaining 2 h at that temperature. After reduc-
tion, the iron phosphides based catalysts were cooled to room tem-
−
1
perature under He flow and then passivated under 100 mL min
of
0.5% O /N (Air Liquid) flow for 4 h at room temperature. The series of
2
2
iron phosphides were prepared by maintaining constant the iron con-
tent (15 wt%) and varying the initial P/Fe molar ratio. Thus, the syn-
thesized catalysts were denoted as P/Fe-x, where x is the P/Fe molar
ratio employed, x = 0.5, 1, 2 and 3.
2.3. Characterization of catalysts
The textural parameters of the passivated catalysts were evaluated
from N adsorption-desorption isotherms at -196 °C measured using an
2
automatic ASAP 2020 system (Micromeritics). Before measurements,
of FeP and FeP
2
in the HDS of dibenzothiophene, HDN of quinoline and
−4
samples were outgassed at 150 °C and 10 mbar overnight. Specific
hydrogenation of tetralin to decalin, obtaining the highest conversions
for the FeP based catalysts. Currently, iron phosphide based catalysts
are highly studied as efficient electrocatalyst and photocatalyst for the
oxygen and hydrogen evolution reactions [22–25]. The use of iron as
promoter of nickel phosphide based catalysts has also been evaluated
since the incorporation of iron can alter the electronic density of the
active phase, modifying the selectivity pattern [26–28].
surface areas (SBET) were determined by using the BET equation con-
2
sidering a N
2
cross section of 16.2 Å .
Metal Dispersion was calculated from CO-chemisorption measure-
ments. CO uptakes were measured in a Micromeritics ASAP 2020 ap-
paratus. Prior to measurement, the samples were re-reduced in situ in
2
H at 450 °C and then evacuated at 25 °C for 10 h. The chemisorption
isotherm was obtained by measuring the amount of CO adsorbed be-
tween 10 and 600 mmHg at 35 °C. After completing the initial analysis,
the reversibly adsorbed gas was evacuated and the analysis was re-
peated to determine only the chemisorbed amounts.
In summary, although there are some reports concerning the cata-
lytic behavior of iron based phosphides, none of them have focused on
the effects of the stoichiometry of the iron phosphide based catalysts in
HDO reactions. In previous works, it has been reported that phosphides
based on Ni and Co prepared from phosphite-type precursors displayed
a good catalytic performance in the HDO of phenol [18]. In addition,
the effect of the initial P/Metal ratio on their catalytic response was
studied, having proved that the stoichiometry of the phosphide formed
has direct effects on the catalytic process [15,29]. It is also known that
the effect of phosphorus strongly depends on its content, showing a
negative effect at high loadings. In the present work, the character-
ization and catalytic activity of iron phosphide supported catalysts
presenting different stoichiometry by changing the initial P/Fe ratio is
reported. Thus, the possible application of a cheap and non-toxic phase
such as iron in the form of iron phosphide is going to be evaluated. HDO
of phenol, as an oxygenated model compound present on fast-pyrolysis
bio-oil from lignocellulosic biomass, was the catalytic test chosen to
evaluate the activity.
CO uptake (mmoles/g)
Dispersion (%) =
mmoles
Metal loading (
)
g
Powder diffraction patterns were collected using a X´Pert Pro MPD
automated diffractometer (PANalytical) with Mo-Kα radiation
(λ = 0.70930 Å), scanning from 9 to 27° (2θ). The crystallographic
phases were identified by using Highscore Plus software from
PANalytical.
Mössbauer data were recorded at room temperature in the trans-
mission mode using a conventional constant acceleration spectrometer
57
equipped with a Co(Rh) source. An effective absorber thickness of
−
2
5 mg cm
of natural iron was used for all the samples. The velocity
scale was calibrated using a 6 μm thick α-Fe foil. All the spectra were
computer-fitted and the isomer shifts were referred to the centroid of
the spectrum of α-Fe at room temperature.
Size and morphology of the catalysts were studied by high resolu-
tion transmission electron microscopy (HR-TEM) using a TALOS F200x
instrument. TEM analysis was performed at 200 kV and 5.5 μA and
scanning transmission electron microscopy (STEM) with a HAADF de-
tector, at 200 kV and 200 nA.
2. Materials and methods
2.1. Materials
The catalytic support used in this study was a commercial fumed
silica from Sigma-Aldrich. The reagents used to prepare the precursor
solution were iron(III) nitrate nonahydrate (Fe(NO ·9H O, 98%, Alfa
Aesar) and phosphorous acid (H PO H, 99%, Analyticals). In the re-
activity study, phenol (98%, Amresco) was used as oxygenated com-
pound; octane (99%, Sigma-Aldrich) was used as solvent and nonane
The Fourier transform infrared (FT-IR) spectra were carried out with
a Nicolet Nexus instrument, using a conventional cell connected to a
gas-handling system. Prior to each adsorption, pressed disks of powder
samples with a diameter of 1.5 cm (ca. 20 mg) were thermally treated
3
)
3
2
2
3
within the cell at 400 °C under H
2
(40 kPa) for 1 h. Then, the samples
(
99%, Sigma-Aldrich) as internal standard. The gases employed were
He (Air Liquide 99.99%), H (Air Liquide 99.999%), N (Air Liquide
9.9999%), CO (Air Liquide 99.9%) and NH (Air Liquide 99.99%).
were outgassed at 400 °C and cooled to carry out the spectrum at rt
−1
2
2
between 4000 and 400 cm
using 200 scans and a resolution of
2 cm . CO adsorption (0.13 kPa) experiments were performed at
140 °C and the spectra were recorded in the range of -140 to 10 °C
−1
9
3
−
2.2. Preparation of catalysts
while degassing.
3
Temperature-programmed desorption of ammonia (NH -TPD) was
Iron based catalysts (15 wt%) were prepared by incipient wetness
impregnation of the silica support by adding different amounts of iron
III) nitrate nonahydrate and phosphorous acid (H PO H) to achieve P/
conducted to determine the acidity of the samples. For the experiments,
80 mg of catalytic precursors were reduced in-situ by flowing hydrogen
to 700 °C (same conditions as in 2.2 section). Subsequently, samples
(
2
3
2

E. Rodríguez-Aguado et al.
Catalysis Today xxx (xxxx) xxx–xxx
Fig. 1. A) HDO conversion for the different catalysts at ToS = 6 h and B) Evolution of the conversion with time on stream. Operating at 275 °C and WHSV = 0.4 h-1
.
were cooled to 100 °C and then exposed to flowing pure ammonia for
min. After ammonia adsorption, samples were cleaned with He flow
in order to remove the physisorbed ammonia. Desorption was per-
formed between 100 and 500 °C (10 °C·min ) and the desorbed am-
monia was analyzed online by a thermal conductivity detector (TCD).
In order to quantify the amount of ammonia desorbed, the equipment
was previously calibrated by measuring the corresponding signals of the
thermal decomposition of known amounts of hexaamminenickel(II)
2
activities were measured at 275 °C under 5 and 15 bar, with a H flow
−1
-1
5
rate of 30 ml·min
and weight hourly space velocity of 0.3 h .
The evolution of the reaction was monitored by collecting liquid
samples which were kept in sealed vials and subsequently analized by
gas chromatrography (Shimadzu GC-14B, equipped with a flame ioni-
zation detector and a capillary column, TBR-14, coupled to an auto-
matic Shimadzu AOC-20i injector).
−
1
The HDO conversion was calculated by using the following equa-
tion:
3 6 2
chloride, [Ni(NH ) ]Cl , supplied by Aldrich.
X-Ray Photoelectron spectroscopy (XPS) was conducted to study the
surface composition of samples. Spectra were collected using a Physical
Electronics PHI 5700 spectrometer with non-monochromatic Mg Kα
radiation (300 W, 15 kV, 1253.6 eV) with a multi-channel detector.
Spectra of passivated samples were recorded in constant pass energy
mode at 29.35 eV using a 720 μm diameter analysis area. Binding en-
ergy (BE) values were referenced to the C 1 s peak (284.8 eV) from the
adventitious contamination layer. The PHI ACCESS ESCA-V6.0 F soft-
ware package and Multipak v8.2b were used for acquisition and data
analysis, respectively. A Shirley-type background was subtracted from
the signals. Recorded spectra were fitted using Gauss-Lorentz curves, in
order to determine the binding energy of the different element core
levels more accurately. The error in BE was estimated to be ca. 0.1 eV.
CPh. Ini− CPh.Fin −CInt
HDOConv. (%) =
× 100
CPh. Ini
where C
Ph.Ini
is the phenol concentration in the feed, C
Ph.Fin
is the
phenol concentration in the liquid product and CInt is the concentration
of the O-containing intermediates.
Selectivity data were calculated considering all the reaction pro-
ducts by using the following equation:
Ci
S (%) =
i
× 100
n
∑
Ci
1
Where S
i
is the selectivity of i-compound, Ci is the concentration of i-
product and is the sum of concentration of all reaction products. The
turnover frequency (TOF) was calculated from the formula:
2
.4. Catalytic test
F/W
M
TOF =
× ln(1 − X)
The hydrodeoxygenation of phenol was performed in a high-pres-
where F is the molar rate of the reaction (mol min−1), W is the catalyst
weight (g), X is the conversion factor and M is the number of sites
loaded (mol g ). This formula was used because the conversion values
were close to 100% and were far from differential conditions making an
integral analysis necessary.
sure, fixed-bed, continuous flow, stainless steel catalytic reactor
9.1 mm in diameter, 230 mm in length), operated in the down-flow
(
−1
mode. The reaction temperature was measured with an interior ther-
mocouple in direct contact with the catalyst bed. The organic feed
consisted on a solution of phenol (1 wt%) in octane which was supplied
by means of a Gilson 307SC piston pump (model 10SC) with a flow rate
of 0.18 ml·min . For the activity test, 0.25 g of pelletized and passi-
vated catalyst were transferred into the reaction tube diluted with
carborundum to a total volume of 3 cm3 (particle size 0.85–1.18 mm).
−1
3. Results and discussion
The HDO reaction of phenol over the prepared materials was carried
out at 275 °C and under pressures of 5 and 15 bar. The detailed reaction
conditions are included in the experimental section. Fig. 1 represents a
comparison of the HDO conversion values working under 5 and 15 bar.
Prior to the activity test, samples were treated in-situ at atmospheric
pressure with a H
flow of 100 mL min- by heating from rt to 450 °C
1
2
(
2 h) to ensure total reduction of the surface passivation layer. Catalytic
3

E. Rodríguez-Aguado et al.
Catalysis Today xxx (xxxx) xxx–xxx
Scheme 1. Main reaction pathways [31].
The results reflect the strong dependency of the pressure on the activity
of the catalysts being far superior working at higher pressure. Under
cyclohexane, with selectivity values higher than 80%. Benzene and
cyclohexene are also obtained but in much lesser extent. Instead, at
5 bar, the selectivity to cyclohexene increases conspicuously being the
most important product in the case of P/Fe-3 sample. This suggests that
all tested catalysts react via hydrogenation of the aromatic ring fol-
lowed by dehydration of cyclohexanol to obtain cyclohexane, cyclo-
hexene and in a minority, benzene.
15 bar, all the catalysts result to be considerably actives, achieving
conversions values higher than 60% in all cases. At 15 bar, P/Fe-0.5 and
P/Fe-2 samples are those more active. Working under 5 bar, the con-
version values range from 40 to 60%. Again, P/Fe-0.5 and P/Fe-2
samples present better conversion values. Therefore, the evolution of
HDO conversion at 15 bar of these catalysts versus time on stream is
depicted in Fig. 1B. As observed in this graph, both catalysts present a
growing catalytic response during the first two hours on stream, fol-
lowed by a slight loss of catalytic activity which eventually remains
steady during last two hours of time on stream.
Considering the selectivity results, it has been reported in literature
that phenol hydrodeoxygenation can take place by direct hydro-
genolysis of the CeO bond, obtaining benzene or by hydrogenation of
the aromatic ring giving cyclohexanol and subsequent dehydratation to
obtain cyclohexene and later cyclohexane. As mentioned above, the
catalytic results seem to indicate that the reaction proceeds via this
second pathway, see Scheme 1, which is in agreement with previous
research with nickel and cobalt phosphide based catalysts [28,29].
The selectivity results presented by iron phosphide based samples
are included in Fig. 2. The selectivity values are strongly dependent on
the reaction pressure studied. Thus, at 15 bar, the majority product is
In the case of metallic phosphides, it has been suggested that co-
ordinatively unsaturated metal sites (CUS) can act as Lewis sites, while
surface oxygen behaves as Lewis basic sites [32]. Moreover, literature
data report that phenol is adsorbed by an heterolytic dissociation be-
tween the hydrogen of its eOH group on the oxygen sites in the oxide
surface layer leading to a phenoxide species [33] (Scheme 2), which are
stabilized by the existence of coordinatively unsaturated metal sites
(CUS) that act as Lewis sites [32]. Other authors have pointed out that
phenol adsorption can also proceed by the existence of Lewis sites
3+
(Fe ) or the presence of defects in the grain boundary that are favored
in small crystals by its interaction with the aromatic ring [34,35]. In the
next step, the metallic species (Fe
form cyclohexanone, being reduced to cyclohexanol easily. Finally,
cyclohexanol eOH group interacts with the metal sites (Fe ) favoring
x y
P ) hydrogenates the aromatic ring to
x y
P
its dehydration to cyclohexene and a subsequent hydrogenation to for
cyclohexane [35].
Fig. 2. Selectivity graphs for the studied catalysts under 5 and 15 bar, 275 °C, WHSV = 0.3 h-1, ToS = 6 h.
4

E. Rodríguez-Aguado et al.
Catalysis Today xxx (xxxx) xxx–xxx
x y
Scheme 2. Proposed mechanism of Phenol HDO over Fe P samples.
Previous researchers have evaluated the catalytic behavior of iron
catalysts supported on silica and carbon using guaiacol as target lignin
molecule, at atmospheric pressure and high contact times [36,37].
These works reported the formation of phenol and benzene, discarding
the hydrogenation of the aromatic ring. In the present research, the iron
phosphide based catalysts were studied at higher H pressure and lower
2
contact time favoring the formation of dearomatized products (cyclo-
hexene and cyclohexane). These data suggest that the hydrogenation of
consequence of the incorporation of the active phases. As reported in
previous works, this loss of specific surface area is attributed to the
blockage of the surface pores with phosphorus species in excess
[28,29]. Considering the pore size distribution, in all cases a bimodal
distribution is observed compared to pure silica where not a clear pore
size distribution is obtained. The sharp increase in the pore volume of
the iron containing catalysts at high relative pressures suggests the
formation of pores of bigger size in the range of macroporous such is
our case. In this regard, all samples present a maximum close to 50 nm;
however, pores of smaller size (22–27 nm) are also present and being
much more important in the case of samples with an intermediate P/Fe
molar ratio. In fact, P/Fe-1 is that presenting pores with smaller size
and their proportion being as important as that of higher size.
2
the aromatic ring is favored by the presence of H -pressure, but also the
iron phase plays an important role. Considering the functional groups of
phenol, the eOH group tends to interact with the active sites at milder
reaction conditions, while the interaction of the aromatic ring with the
metal sites is favored under more severe conditions. In this sense, it is
well reported in the literature the hydrogenation of benzene to form
cyclohexane under H -pressure using Fe-based catalysts [38,39].
2
In order to understand and correlate the catalytic behavior of these
catalysts with their properties a deep characterization was performed.
The crystalline phases identified by XRD are shown in Fig. 4. The
diffractograms were measured with monochromatic Mo Kα radiation in
order to diminish iron fluorescence. At first sight, the presence of dif-
ferent phosphide phases as a function of the initial P/Fe ratio is evident.
Hence, P/Fe-0.5 catalyst presents diffraction peaks at 18.2, 19.9, 21.3,
2
The textural characterization as a result of N adsorption-desorption
analysis of the samples is included in Table 1. The textural properties of
the bare support are also included for comparative purposes. Fig. 3
shows the corresponding isotherms (Fig. 3a) and pore size distribution
graphs (Fig. 3b). The isotherms are all of type II and characteristic of
solids with pores of diameters in the range of meso and macropores,
and generally attributed to interparticle voids. BET surface areas suffer
a remarkable decrease in comparison with the bare support as a
23.7, 24.2 and 24.4° which are all assigned to the Fe
01-085-1725). By increasing the amount of P present in the catalyst
formulation until P/Fe-1, the diffraction peaks ascribed to Fe P phase
diminish and new ones appear at 14.1, 14.9, 15.7, 16.1, 16.9, 20.8,
21.1, 21.7, 22.6, 24.7 and 24.9°, which correspond to an iron phosphide
phase richer in phosphorus such as FeP (PDF Nº: 01-078-1443).
2
P phase (PDF Nº:
2
2
As the P content increases, Fe P phase disappears and for P/Fe-2
sample, FeP phase becomes the dominant and unique phase. By in-
creasing the P/Fe ratio until 3, the catalyst with the greater content in
phosphorous, besides the presence of the FeP phase, new diffraction
lines emerge at 14.4, 16.6, 17.1, 17.9, 18.5, 21.9, 22.4, 23.2, 23.4, 25.8
and 26.4°. These reflection lines are ascribed to the presence of a new
Table 1
Textural and acidic properties of the catalysts.
BET (m2 g−1
Vp (cm3 g−1
Sample
S
)
)
Total acidity
mmolNH /g)
(
3
2
phase enriched in phosphorus, FeP (PDF Nº: 03-065-6088). So, from
these data it can be concluded that in so far as the P content increases
the formation of phosphorus richer phosphides takes place. Scherrer’s
equation was used to determine the average particle size of the various
phases present and the corresponding data included in Table 2. From
SiO
2
203
82
100
76
0.64
0.77
0.62
0.64
0.56
–
P/Fe-0.5
P/Fe-1
P/Fe-2
P/Fe-3
0.06
0.09
0.22
0.15
61
2
this table it can be observed that Fe P phase is more dispersed on P/Fe-
5

E. Rodríguez-Aguado et al.
Catalysis Today xxx (xxxx) xxx–xxx
2
Fig. 3. N -adsorption desorption isotherms (a) and pore size distribution by applying the BJH method to the desorption branch of the isotherm (b).
chemisorption measurements and the data are also included in Table 2.
It has been assumed that each CO molecule is adsorbed on one active
site according to previous results in literature [40]. The corresponding
calculated dispersion values are low in all cases, although the values are
close to those found in literature for iron phosphide [40]. From these
2
data, it seems that FeP phase chemisorbs more CO than Fe P if the data
corresponding to P/Fe-2 and P/Fe-0.5 are compared. Moreover, the
calculated dispersion values are in accordance to XRD results, where
samples with two different iron phosphide phases, P/Fe-1 and P/Fe3,
show lower particle size.
In order to get further insight of the iron phases present, Mössbauer
spectroscopy was used and the corresponding spectra are lumped to-
gether in Fig. 5. In line with this, Table 3 collects the Mössbauer
parameters obtained from the fit of the spectra as well as the assign-
ment of the various spectral components to the corresponding chemical
species. The spectrum recorded from the sample having a P/Fe-0.5 ratio
is composed by four quadrupole doublets. The most intense ones, and
accounting together for the 84% of the spectral area, correspond to the
Fig. 4. X-ray diffractograms of the different samples.
Table 2
Particle size, percentage of each phase, chemisorption results and surface
atomic ratios.
Crystal
phase
Size
Percentage CO uptakes Dispersion Surface
(nm) (%)a
(μmol g
−1
)
(%)
b
atomic ratio
P/
Fe
Fe/Si
P/Fe-0.5 Fe
2
P
P
41.2 100
23.6 24
26.6 74
42.8 100
36.5 66
52.5 34
4.6
12.2
0.17
0.45
0.26 0.05
0.58 0.037
P/Fe-1
Fe
2
FeP
FeP
FeP
FeP
P/Fe-2
P/Fe-3
6.6
4.8
0.25
0.82
0.73 0.044
0.79 0.037
2
a
Semiquantitative analysis obtained from HighScore Plus Software from
crystalline phases.
b
Calculated from CO uptakes.
1
sample than on P/Fe-0.5 one; the particle size of FeP phase follows
the order: P/Fe-1 < P/Fe-3 < P/Fe-2. In general, it can be stated that
samples containing different iron phosphides such is the case of P/Fe-1
and P/Fe-3 present lower particle sizes.
An estimation of the dispersion was evaluated from CO-
Fig. 5. Mössbauer spectra corresponding to the different phosphides.
6

E. Rodríguez-Aguado et al.
Catalysis Today xxx (xxxx) xxx–xxx
Table 3
bond, while the second band can be attributed to carbonyl species
weakly coordinated over Fe-species partially reduced, possibly super-
imposed to some liquid-like CO [18]. In all cases, the interaction with
CO is weak since both bands disappear when the temperature increases,
although the P/Fe-1 sample shows stronger interactions since the band
still remains after outgassing. The spectrum corresponding to P/Fe-3
sample did not show any noticeable band. The high amount of phos-
phorus present in this sample seems to be detrimental for IR analysis.
From this figure, it can be observed that for all three samples the
contribution due to OH groups is more important, and the contribution
due to partially reduced iron decreases by increasing the amount of P
present, indicating that CO coordination with metallic sites is favored
for iron-rich phosphides.
Mössbauer parameters obtained from the corresponding spectra fitting.
Sample
Component
δ
Δ
Area
(%)
Assignment
(
mm s−1)
(mm s−1)
P/Fe-0.5
A
B
C
D
0.52
0.23
1.11
0.88
0.62
0.30
2.58
2.41
46
38
7
2
Fe P
Fe(II) phosphate
Fe(II) phosphate
9
P/Fe-1.0
P/Fe-2.0
P/Fe-3.0
E
C
0.34
1.14
0.61
2.40
79
21
FeP
Fe(II) phosphate
E
C
0.31
1.22
0.71
2.44
93
7
FeP
Fe(II) phosphate
3
NH -TPD experiments were carried out in order to assess the total
E
F
C
0.31
0.07
1.13
0.69
2.04
2.40
63
27
10
FeP
acidity of the materials. Table 1 includes the total amount of ammonia
desorbed for each catalyst from 100 to 500 °C and Fig. 8 the desorption
profiles. The bare support desorbs a negligible amount of ammonia
indicating that the observed acidity of the catalysts is mainly due to
presence of the active phases. All catalysts exhibit a single desorption
peak in the range of 100 to 300 °C which is associated to a weak acidic
character. The total acidity increases remarkably with phosphorous
content achieving its maximum value for the P/Fe = 2 catalyst.
The acidity should be related to phosphorus content present on the
sample since it increases in so far as P/Fe ratio does and is expected to
be associated to eOH groups as reported in literature [45,46] and also
corroborated from previous IR analyses. The presence of acid sites is
necessary to favor the hydrogenolysis reaction of cyclohexanol to form
cyclohexene and cyclohexane [29].
The surface chemical composition of passivated catalysts was
evaluated by XPS. The Si 2p and O 1s signals are centered at 103.5 eV
and 532.9 eV, respectively, remaining unchanged in all catalysts. The
Fe 2p core-level spectra show the coexistence of various species on the
surface. The Fe 2p spectrum of the sample P/Fe = 0.5, see Fig. 9, shows
the majority presence of iron(III) phosphate (711 eV) with a weak
contribution owed to iron phosphide. The relative intensity of the signal
FeP
Fe(II) phosphate
2
two crystallographic sites existing in the structure of the iron phosphide
of composition Fe P [41]. The two minor doublets can be assigned to an
2
Fe(II) phosphate. Both the number of Fe(II) phosphates and the range of
their Mössbauer parameters is large [42], therefore, in the absence of
other complementary information, it is difficult to assign the obtained
parameters to a particular Fe(II) phosphate species.
The composition deduced from the spectra corresponding to the
samples with P/Fe-1.0 and P/Fe-2.0 is different from that observed in
the previous material. In both cases the major contribution is due to the
iron phosphide of composition FeP (Table 1,[41]). The proportion of
FeP is larger in the sample P/Fe = 2 (93%) than in the sample P/Fe-1.0
(
79%). In both samples the minor contribution corresponds, again, to
an unspecific Fe(II) phosphate. Finally, the material having P/Fe-3.0 is
also dominated by the presence of FeP (63%). A contribution amount of
27% of the spectral area corresponds to the phosphorus-rich FeP
2
phosphide (Table 3,[41]). The smaller contribution, again, is due to an
Fe(II) phosphate.
δ−
HRTEM micrographs for the different samples are included in Fig. 6.
In the case of the samples of lower P content, it was not possible to
elucidate the presence of particles. The micrographs were similar to
those of iron based catalysts reported by other authors [43]. Instead,
the presence of particles was more evident for P/Fe- 2 and P/Fe-3
samples, were FeP and FeP
ticles are dispersed on the SiO
local agglomeration can be detected where the particle size cannot be
calculated.
corresponding to P rises with increasing the molar ratio P/Fe in both
Fe 2p and P 2p core-level spectra. Beside this, the evolution of the Fe 2p
signal shows an increase of Fe(II) as the P/Fe rises, not only for the
presence of a shoulder in the region of low binding energy but also for
the modification of the signal owed to the presence of a shake-up sa-
tellite. The P 2p core-level spectra of all catalysts present two con-
tributions of the P 2p3/2 component centered at 133.2 eV and 128.2 eV
2
phases are present. In general, Fe
x y
P par-
2
matrix except in some zones where a
3
-
δ-
which respectively correspond to PO
species are present as Fe P, FeP or FeP
4
and P [30]. The phosphide
2
2
phases; and phosphate species
The FT-IR spectrum of the passivated sample after hydrogen acti-
vation is represented in Fig. 7a. The spectrum corresponding to FeP-0.5
sample was the only one represented in the OH-stretching region
are originated in the passivation process. Considering the surface
atomic ratios compiled in Table 2, it is clearly noticeable how the
phosphorus exposed on surface rises by increasing P/Fe and therefore,
the iron exposed on surface is lower.
(
Fig. 7a), since the other samples did not show representative con-
tributions. It should be noted that by increasing P content the spectra
became even noisier. FeP-0.5 catalyst displayed a main band located at
The manuscripts devoted to the use of iron in hydrotreating reac-
tions are scarce due to the low activity of iron in these reactions.
However its abundance, low toxicity and therefore its low cost make it a
good candidate to face the challenges in HDO catalyst design [47]. In
HDO reactions, Rezaei et al.[48] have reported that iron is an effective
metal for the selective conversion of lignin-derived phenolics into
aromatic hydrocarbons; when it is combined with acidic supports. Iron,
an oxophilic metal, may cause repulsion to the phenolic ring, and in
turn, a strong interaction with the carbonyl group, facilitating its se-
lective hydrogenation to the corresponding alcohol [49]. The tuning
among metallic and acidic sites could govern the reaction pathway
followed as reported by Rezaei et al [49] when using iron catalysts
supported on MCM-41 and promoted with Re and Zr to incorporate the
acidic function. This bifunctional requirement is accomplished in
transition metal phosphides. The use of iron phosphide in HDO reac-
tions is scarce, as stated before, and the change in the iron-phosphorous
ratio it is going to determine the stoichiometry of the iron phosphide
phase formed and therefore the metallic and phosphorus surface
−1
3
747 cm , which is attributed to the −OH stretching of isolated single
−
1
silanols. The shoulder with a maximum located about 3720 cm
is
assigned to the −OH stretching of silanol group that interacts with a
vicinal pair [44]. Finally, the broad and weak band located about
−1
3700 cm
could be attributed to the −OH stretching of surface hy-
droxyl groups with intermediate acid properties possibly interacting
with neighboring phosphate groups, clearly noticeable for Ni
based samples [18].
2
P and CoP
CO molecule was used as a target molecule to determine the kind of
metal sites. In all cases the spectra were carried out at low temperature
(
2
using N -liquid to cool the holder). This implies that the interaction
between the CO molecule and the metal centers must be weak. From
CO-adsorption data shown in Fig.7b, it can be observed two bands lo-
−1
cated at 2157 and 2137 cm . The first band is attributed to the weak
interaction of the CO with the eOH groups located on the surface by H-
7

E. Rodríguez-Aguado et al.
Catalysis Today xxx (xxxx) xxx–xxx
Fig. 6. HRTEM images for P/Fe-0.5, P/Fe-1, P/Fe-2 and P/Fe-3 samples.
exposure and therefore the acidity of the resultant material. From the
catalytic results presented here, it seems clear that the presence of
higher phosphorus content worsens the conversions results. In this
sense, it has been reported for nickel and cobalt phosphide based cat-
alysts that the incorporation of phosphorus species modifies the elec-
tronic density of the metal and diminishes the exposition of the metal
sites on the surface limiting its deactivation. However, the presence of
larger amount of phosphorus can limit the amount of available active
sites, which leads to poorer conversions [13,29,50]. In the present
paper, by increasing the amount of phosphorus iron phosphides richer
2 2
in phosphorus are formed: Fe P →FeP →FeP , iron surface exposure
decreases and acidity improves. Moreover, the P/Fe molar ratio em-
ployed determines the presence of unique or mixed phases in the cor-
responding catalyst. All these factors determine the catalytic response
Fig. 7. FT-IR spectra in the OH-stretching region a) and after CO adsorption at -140 °C (black) and the same spectra with the CO outgassed at -140 °C (dark grey) b).
8

E. Rodríguez-Aguado et al.
Catalysis Today xxx (xxxx) xxx–xxx
Table 4
Turnover frequency values.
TOF (s 1) x 103
−
Sample
1
5 bar
5 Bar
P/Fe- 0.5
P/Fe-1
P/Fe- 2
P/Fe-3
147
18
66
45
10
29
21
50
activity of the prepared samples. Again TOF values are correlated with
the catalyst acidity (Fig. 10b), the higher the better. Therefore iron
phosphide stoichiometry, surface exposure of the iron and the catalyst
acidity are directly correlated with the catalytic activity. These pre-
sented data are at odds with those shown for iron phosphide supported
on activated carbon in the hydrodesulfurization reaction of dibenzo-
tiophene, where FeP was more active than Fe
2
P [21]. In the case, of
HDO reactions the oxophilic character of iron could justify how Fe
2
P
phase presenting greater iron surface exposure provides more active
sites to interact with oxygen from the molecules.
Fig. 8. Ammonia temperature programmed desorption profiles.
Taking into account the selectivity data, in spite of the change in the
acidity values, all the samples follows the hydrogenation route, al-
though the sample presenting the highest acidity, P/Fe-2, is that where
the amount of benzene was noticeable, selectivity close to 15%.
However, the weak nature of those sites does not favor the hydro-
genolysis route, where after the first hydrogenation of the carbonyl
group forming cyclohexadienols, they must be dehydrated in acidic
sites to form aromatic hydrocarbons and water [51]. As suggested by
other authors [52], the strength of those acidic sites are clue in order to
favor the dehydration reaction to occur and to prevent catalyst deac-
tivation. Further research is required in this regard to better understand
the selectivity in HDO of iron phosphides and should be focused on the
of the presented samples. In this regard, samples containing unique
phases, such is the case of P/Fe-0.5 and P/Fe-2, presented the highest
conversion values at both reaction pressures studied. If both samples
are compared, Fe
2
P phase is that present in P/Fe-0.5 sample and FeP
that present in P/Fe-2 one. Therefore it can be concluded that Fe
2
P
phase is more active than FeP if conversion values are considered.
Moreover, if catalytic results of all the samples are lumped together the
catalytic activity of the iron phosphide phases follows the trend
2 2
Fe P > FeP > FeP . In addition, if TOF values are calculated to de-
termine the activity per active site (Table 4), the much higher activity
of the P/Fe-5 sample can be highlighted compared to others at both
reaction pressures studied. As previously observed from conversion
values, P/Fe-0.5 and P/Fe-2 were those more active, i.e., unique phases
are more appropriate to achieve better catalytic results.
change of acidity. The presented results evidences that Fe
most active one in HDO.
2
P phase is the
TOF data can also be correlated with catalysts properties such is the
case of acidity and metal surface exposure. These correlations are
plotted in Fig. 10.
From this figure it is observed a relationship between iron surface
exposure and TOF (Fig. 10a), thus the sample presenting the greatest
4
. Conclusions
In the present work the influence of the phosphorus content in the
properties of iron-based catalysts has been evaluated as well as its ac-
tivity in the phenol HDO reaction. The amount of phosphorus varied to
obtain iron phosphides with different stoichiometry. When the phos-
phorus content was increased, it was incorporated into the iron phos-
phide structure and Fe
iron-rich phosphide, stoichiometric iron phosphide and phosphorus-
2
iron surface exposure is that containing an iron rich phosphide, Fe P,
and being the most active in the studied reaction, P/Fe-0.5. The better
metal surface exposure was also observed from CO-IR experiments. In
addition, the acidity also plays an important role in the catalytic
2 2
P, FeP and FeP phases were obtained, that is,
Fig. 9. Fe 2p (A) and P 2p (B) core level spectra of the studied catalysts.
9

E. Rodríguez-Aguado et al.
Catalysis Today xxx (xxxx) xxx–xxx
Fig. 10. Dependency of TOF with iron surface ex-
posure and catalyst acidity. ( -15 bar and -5 bar).
rich phosphide, respectively. Furthermore, depending on the P / Fe
ratio used, the catalysts synthesized contained one or two phases. The
catalytic results showed that the samples were active in the HDO re-
action of phenol, being those with unique phases the most active ones.
In addition, the activity depended on the iron phosphide formed.
[13] X. Wang, P. Clark, S.T. Oyama, Synthesis, characterization, and hydrotreating ac-
tivity of several iron group transition metal phosphides, J. Catal. 208 (2002)
3
21–331.
[
14] H.Y. Zhao, D. Li, P. Bui, S.T. Oyama, Hydrodeoxygenation of guaiacol as model
compound for pyrolysis oil on transition metal phosphide hydroprocessing cata-
lysts, Appl. Catal. A 391 (2011) 305–310.
[
15] 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:
Environ. 136-137 (2013) 140–149.
2 2
Therefore, the observed trend was: Fe P > FeP > FeP . Finally, the
activity per active site depended on the surface exposure of iron and the
acidity of the catalyst. The results presented highlight the application of
a cheap and abundant phase, such as iron in HDO reactions.
[
16] S.T. Oyama, T. Gott, H. Zhao, Y.-K. Lee, Transition metal phosphide hydroproces-
sing catalysts: a review, Catal. Today 143 (2009) 94–107.
[
17] N. Koike, S. Hosokai, A. Takagaki, S. Nishimura, R. Kikuchi, K. Ebitani, Y. Suzuki,
S.T. Oyama, Upgrading of pyrolysis bio-oil using nickel phosphide catalysts, J.
Catal. 333 (2016) 115–126.
Aknowledgements
[
18] A. Infantes-Molina, E. Gralberg, J.A. Cecilia, E. Finocchiob, E. Rodríguez-Castellón,
Nickel and cobalt phosphides as effective catalysts for oxygen removal of di-
benzofuran: role ofcontact time, hydrogen pressure and hydrogen/feed molar ratio,
Catal. Sci. Technology 5 (2015) 3403–3415.
Thanks to project CTQ2015-68951-C3-3R (Ministerio de Economía
y Competitividad, Spain and FEDER Funds). A.I.M. thanks the Ministry
of Economy and Competitiveness for a Ramon y Cajal contract
[
19] J.A. Cecilia, A. Infantes-Molina, J. Sanmartín-Donoso, E. Rodríguez-Aguado,
(
RyC2015-17870).
2
D. Ballesteros-Plata, E. Rodríguez-Castellón, Enhanced HDO activity of Ni P pro-
moted with noble metals, Catal. Sci. Technol. 6 (2016) 7323–7333.
[
20] X. Huang, Q. Dong, H. Huang, L. Yue, Z. Zhu, J. Dai, Facile Synthesis of Iron
Phosphide Fe P Nanoparticle and Its Catalytic Performance in Thiophene
Hydrodesulfurization, (2014).
References
2
[
[
[
1] C. Xu, R.A. Arancon, J. Labidi, R. Luque, Lignin depolymerisation strategies: to-
wards valuable chemicals and fuels, Chem. Soc. Rev. 43 (2014) 7485–7500.
2] P.M. Mortensen, J.D. Grunwaldt, P.A. Jensen, K.G. Knudsen, A.D. Jensen, A review
of catalytic upgrading of bio-oil to engine fuels, Appl. Catal. A 407 (2011) 1–19.
3] M. Patel, A. Kumar, Production of renewable diesel through the hydroprocessing of
lignocellulosic biomass-derived bio-oil: a review, Renew. Sustain. Energy Rev. 58
[21] Y. Yuan, H. Pei, H. Chen, L. Zuo, J. Shen, Preparation of Fe P and FeP catalysts for
2
the hydrotreating reactions, Catal. Commun. 100 (2017) 202–205.
[22] H. Du, S. Gu, R. Liu, C.M. Li, Highly active and inexpensive iron phosphide na-
norods electrocatalyst towards hydrogen evolution reaction, Int. J. Hydrogen
Energy 40 (2015) 14272–14278.
[23] Z. Huang, C. Lv, Z. Chen, Z. Chen, F. Tian, C. Zhang, One-pot synthesis of diiron
phosphide/nitrogen-doped graphene nanocomposite for effective hydrogen gen-
eration, Nano Energy 12 (2015) 666–674.
[24] L. Tian, X. Yan, X. Chen, Electrochemical activity of iron phosphide nanoparticles in
hydrogen evolution reaction, ACS Catal. 6 (2016) 5441–5448.
[25] H. Zhao, J. Wang, Y. Dong, P. Jiang, Noble-Metal-Free iron phosphide cocatalyst
loaded graphitic carbon nitride as an efficient and robust photocatalyst for hy-
drogen evolution under visible light irradiation, ACS Sustain. Chem. Eng. 5 (2017)
8053–8060.
(2016) 1293–1307.
[
4] S. Xiu, A. Shahbazi, Bio-oil production and upgrading research: a review, Renew.
Sustain. Energy Rev. 16 (2012) 4406–4414.
[
5] N. Arun, R.V. Sharma, A.K. Dalai, Green diesel synthesis by hydrodeoxygenation of
bio-based feedstocks: strategies for catalyst design and development, Renew.
Sustain. Energy Rev. 48 (2015) 240–255.
[6] T.V. Choudhary, C.B. Phillips, Renewable fuels via catalytic hydrodeoxygenation,
Appl. Catal. A 397 (2011) 1–12.
[
7] V.K. Venkatakrishnan, W.N. Delgass, F.H. Ribeiro, R. Agrawal, Oxygen removal
from intact biomass to produce liquid fuel range hydrocarbons via fast-hydro-
pyrolysis and vapor-phase catalytic hydrodeoxygenation, Green. Chem. 17 (2015)
[26] A. Cho, J. Shin, A. Takagaki, R. Kikuchi, S.T. Oyama, Ligand and ensemble effects in
bimetallic NiFe phosphide catalysts for the hydrodeoxygenation of 2-
Methyltetrahydrofuran, Top. Catal. 55 (2012) 969–980.
1
78–183.
[27] H. Zhao, S.T. Oyama, H.-J. Freund, R. Włodarczyk, M. Sierka, Nature of active sites
[
[
8] H.-G. Chen, Y.H.P. Zhang, New biorefineries and sustainable agriculture: increased
in Ni P hydrotreating catalysts as probed by iron substitution, Appl. Catal. B:
2
food, biofuels, and ecosystem security, Renew. Sustain. Energy Rev. 47 (2015)
Environ. 164 (2015) 204–216.
1
17–132.
[28] E. Rodríguez-Aguado, A. Infantes-Molina, D. Ballesteros-Plata, J.A. Cecilia,
I. Barroso-Martín, E. Rodríguez-Castellón, Ni and Fe mixed phosphides catalysts for
o-removal of a bio-oil model molecule from lignocellulosic biomass, Mol. Catal. 437
(2017) 130–139.
[29] E. Rodríguez-Aguado, A. Infantes-Molina, J.A. Cecilia, D. Ballesteros-Plata,
R. López-Olmo, E. Rodríguez-Castellón, CoxPy catalysts in HDO of phenol and di-
benzofuran: effect of P content, Top. Catal. (2017).
[30] J.A. Cecilia, A. Infantes-Molina, E. Rodríguez-Castellón, A. Jiménez-López, A novel
method for preparing an active nickel phosphide catalyst for HDS of dibenzothio-
phene, J. Catal. 263 (2009) 4–15.
[31] 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
9] M. Toba, Y. Abe, H. Kuramochi, M. Osako, T. Mochizuki, Y. Yoshimura,
Hydrodeoxygenation of waste vegetable oil over sulfide catalysts, Catal. Today 164
(2011) 533–537.
[
10] V.N. Bui, D. Laurenti, P. Delichère, C. Geantet, Hydrodeoxygenation of guaiacol,
Appl. Catal. B: Environ. 101 (2011) 246–255.
[
11] S. Boullosa-Eiras, R. Lødeng, H. Bergem, M. Stöcker, L. Hannevold, E.A. Blekkan,
Catalytic hydrodeoxygenation (HDO) of phenol over supported molybdenum car-
bide, nitride, phosphide and oxide catalysts, Catal. Today 223 (2014) 44–53.
12] 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.
[
10

E. Rodríguez-Aguado et al.
Catalysis Today xxx (xxxx) xxx–xxx
minerals: implications for interpretation of martian data, Am. Mineral. 99 (2014)
914–942.
governing catalyst deactivation, J. Mol. Catal. A Chem. 368-369 (2013) 78–87.
[32] H.H. Kung, Chapter 4 surface coordinative unsaturation, Stud. Surf. Sci. Catal. 45
(
1989) 53–71.
[43] L. Song, S. Zhang, Q. Ma, Synthesis of an iron phosphide catalyst based on sulfides
and hydrodesulfurization property, Chem. Eng. J. 281 (2015) 281–285.
[44] A.J.V. Roosmalen, J.C. Mol, An infrared study of the silica gel surface. 1. Dry silica
gel, J. Phys. Chem. 82 (1978) 2748–2751.
[
33] A. Popov, E. Kondratieva, J.-P. Gilson, L. Mariey, A. Travert, F. Maugé, IR study of
the interaction of phenol with oxides and sulfided CoMo catalysts for bio-fuel hy-
drodeoxygenation, Catal. Today 172 (2011) 132–135.
[
[
[
[
34] 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.
35] P.M. Mortensen, J.-D. Grunwaldt, P.A. Jensen, A.D. Jensen, Screening of catalysts
for hydrodeoxygenation of phenol as a model compound for bio-oil, ACS Catal. 3
2
[45] K. Li, R. Wang, J. Chen, Hydrodeoxygenation of anisole over silica-supported Ni P,
MoP, and NiMoP catalysts, Energy Fuel 25 (2011) 854–863.
[46] Z. Pan, R. Wang, M. Li, Y. Chu, J. Chen, Deoxygenation of methyl laurate to hy-
drocarbons on silica-supported Ni-Mo phosphides: effect of calcination tempera-
tures of precursor, J. Energy Chem. 24 (2015) 77–86.
[47] D.A. Ruddy, J.A. Schaidle, J.R. Ferrell Iii, J. Wang, L. Moens, J.E. Hensley, Recent
advances in heterogeneous catalysts for bio-oil upgrading via “ex situ catalytic fast
pyrolysis”: catalyst development through the study of model compounds, Green.
Chem. 16 (2014) 454–490.
[48] P.S. Rezaei, H. Shafaghat, W.M.A.W. Daud, Aromatic hydrocarbon production by
catalytic pyrolysis of palm kernel shell waste using a bifunctional Fe/HBeta cata-
lyst: effect of lignin-derived phenolics on zeolite deactivation, Green Chem. 18
(2016) 1684–1693.
(2013) 1774–1785.
36] R.N. Olcese, M. Bettahar, D. Petitjean, B. Malaman, F. Giovanella, A. Dufour, Gas-
2
phase hydrodeoxygenation of guaiacol over Fe/SiO catalyst, Appl. Catal. B:
Environ. 115-116 (2012) 63–73.
37] R. Olcese, M.M. Bettahar, B. Malaman, J. Ghanbaja, L. Tibavizco, D. Petitjean,
A. Dufour, Gas-phase hydrodeoxygenation of guaiacol over iron-based catalysts.
Effect of gases composition, iron load and supports (silica and activated carbon,
Appl. Catal. B: Environ. 129 (2013) 528–538.
[38] K.J. Yoon, P.L. Walker, L.N. Mulay, M.A. Vannice, Benzene hydrogenation over
[49] P. Sirous-Rezaei, J. Jae, J.-M. Ha, C.H. Ko, J.M. Kim, J.-K. Jeon, Y.-K. Park, Mild
iron. 1. Specific activities and kinetic behavior over unsupported iron and iron
hydrodeoxygenation of phenolic lignin model compounds over a FeReOx/ZrO
2
dispersed on SiO
Dev. 22 (1983) 519–526.
2
, Al
2
O
3
, carbon, and doped carbon, Ind. Eng. Chem. Prod. Res.
catalyst: zirconia and rhenium oxide as efficient dehydration promoters, Green
Chem. 20 (2018) 1472–1483.
[
39] K.J. Yoon, M.A. Vannice, Benzene hydrogenation over iron. II. Reaction model over
unsupported and supported catalysts, J. Catal. 82 (1983) 457–468.
[50] J.A. Cecilia, A. Infantes-Molina, E. Rodriguez-Castellon, A. Jimenez-Lopez, Gas
phase catalytic hydrodechlorination of chlorobenzene over cobalt phosphide cata-
lysts with different P contents, J. Hazard. Mater. 260 (2013) 167–175.
[51] 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.
[52] X. Zhu, L. Nie, L.L. Lobban, R.G. Mallinson, D.E. Resasco, Efficient conversion of m-
Cresol to aromatics on a bifunctional Pt/HBeta catalyst, Energy Fuel 28 (2014)
4104–4111.
[
40] M.C. Alvarez-Galvan, G. Blanco-Brieva, M. Capel-Sanchez, S. Morales-delaRosa,
J.M. Campos-Martin, J.L.G. Fierro, Metal phosphide catalysts for the hydrotreat-
ment of non-edible vegetable oils, Catal. Today (2017).
[41] R. Wäppling, L. Häggström, S. Rundqvist, E. Karlsson, Mösbauer study of phos-
phides Containing iron, J. Solid State Chem. 3 (1971) 276–292.
[42] M.D. Dyar, E.R. Jawin, E. Breves, G. Marchand, M. Nelms, M.D. Lane,
S.A. Mertzman, D.L. Bish, J.L. Bishop, Mössbauer parameters of iron in phosphate
11