Nickel and cobalt phosphides as effective catalysts for oxygen removal of dibenzofuran: Role of contact time, hydrogen pressure and hydrogen/feed molar ratio

Article abstract of DOI:10.1039/c5cy00282f

The catalytic activity of nickel and cobalt phosphides, with a metal loading of 5 wt.%, supported on silica was investigated in the hydrodeoxygenation reaction (HDO) of dibenzofuran (DBF) as a model oxygenated compound at different contact times, H₂ pressures and H₂/DBF molar ratios. The aim of the study was to understand the mechanism of the reaction and to study the impact of H₂ pressure and H₂/DBF molar ratio on the reaction. The catalysts were characterized by N₂ adsorption-desorption isotherm measurement at -196°C, X-ray diffraction (XRD), X-ray photoelectron spectroscopy (XPS), CO chemisorption, NH₃ Temperature-Programmed Desorption (NH₃-TPD), IR spectroscopy and H₂ Temperature-Programmed Desorption (H₂-TPD). The prepared catalysts were tested in the HDO reaction of DBF in a continuous-flow fixed-bed stainless steel catalytic reactor at pressures ranging from 1-30 bar at 275°C. The results obtained indicate that the Ni₂P catalyst is more active than the CoP catalyst, converting more than 90% of DBF at the highest contact time into oxygen-free products. The activity of both catalysts increases with increased contact time. At low contact times, the intermediates tetrahydrodibenzofuran (THDBF) and hexahydrodibenzofuran (HHDBF) are observed as products, while an increment in the contact time led to the transformation of THDBF and HHDBF into O-free compounds, mainly bicyclohexane (BCH), indicating that the HDO of DBF follows the path: DBF → HHDBF → THDBF → 2-CHP → BCH. Further, both Ni₂P and CoP catalysts are active at medium pressures with HDO degrees similar to those obtained at 30 bar. Ni₂P is less affected by the changes in H₂/DBF ratio than CoP and the catalysts are more active at high H₂/DBF molar ratios.

Full text of DOI:10.1039/c5cy00282f

View Article Online
View Journal
Catalysis
Science &
Technology
Accepted Manuscript
This article can be cited before page numbers have been issued, to do this please use: A. Infantes-
Molina, E. Gralberg, J. A. Cecilia, E. Finocchio and E. Rodriguez-Castellón, Catal. Sci. Technol., 2015, DOI:
10.1039/C5CY00282F.
This is an Accepted Manuscript, which has been through the
Royal Society of Chemistry peer review process and has been
accepted for publication.
Accepted Manuscripts are published online shortly after
acceptance, before technical editing, formatting and proof reading.
Using this free service, authors can make their results available
to the community, in citable form, before we publish the edited
article. We will replace this Accepted Manuscript with the edited
and formatted Advance Article as soon as it is available.
You can find more information about Accepted Manuscripts in the
Information for Authors.
Please note that technical editing may introduce minor changes
to the text and/or graphics, which may alter content. The journal’s
standard Terms & Conditions and the Ethical guidelines still
apply. In no event shall the Royal Society of Chemistry be held
responsible for any errors or omissions in this Accepted Manuscript
or any consequences arising from the use of any information it
contains.
www.rsc.org/catalysis

Page 1 of 36
Catalysis Science & Technology
View Article Online
DOI: 10.1039/C5CY00282F
1
Nickel and Cobalt phosphide as effective catalysts for oxygen-removal of dibenzofuran:
Role of contact time, hydrogen pressure and hydrogen/feed molar ratio
A. Infantes Molinaa*, E. Gralberga, J.A. Ceciliaa, Elisabetta Finocchiob, E. RodríguezꢀCastellóna
aDepartamento de Química Inorgánica, Cristalografía y Mineralogía (Unidad Asociada al ICP-
CSIC), Facultad de Ciencias, Universidad de Málaga, Campus de Teatinos, 29071 Málaga
(Spain)
bDipartimento di Ingegneria Civile, Chimica ed Ambientale, Università di Genova, P.le
J.F.Kennedy 1, 16129 Genova (Italy)
Corresponding Author: ainfantes@uma.es
Abstract
The catalytic activity of nickelꢀ and cobalt phosphides, with a metallic loading of 5 wt.%, supꢀ
ported on silica was investigated in the hydrodeoxygenation reaction (HDO) of dibenzofuran
(DBF) as model oxygenated compound at different contact times, H2 pressures and H2/DBF moꢀ
lar ratios. The aim of the study was to understand the mechanism of the reaction and to study the
impact of H2 pressure and H2/DBF molar ratio on the reaction. The catalysts were characterized
by N2 adsorption–desorption isotherms at −196 °C, Xꢀray diffraction (XRD), Xꢀray photoelecꢀ
tron spectroscopy(XPS), COꢀchemisorption, NH3 TemperatureꢀProgrammed Desorption (NH3ꢀ
TPD), IRꢀ Spectroscopy and H2 TemperatureꢀProgrammed Desorption (H2ꢀTPD). The prepared
catalysts were tested in the HDO reaction of DBF in a continuousꢀflow fixedꢀbed stainless steel
catalytic reactor at pressures ranging from 1ꢀ30 bars at 275 °C. The results obtained indicate that
the Ni2P catalyst is more active than the CoP catalyst, converting more than 90% of the DBF at
the highest contact time into oxygenꢀfree products. The activity of both catalysts increases with
increased contact time. At low contact times the intermediates tetrahydrodibenzofuran (THDBF)
and hexahydrodibenzofuran (HHDBF) are observed as products, while an increment in the conꢀ
tact time led to the transformation of THDBF and HHDBF into Oꢀfree compounds, mainly bicyꢀ
clohexane (BCH), indicating that the HDO of DBF follows the path: DBF → HHDBF →
THDBF → 2ꢀCHP → BCH. Further both Ni2P and CoP catalysts are active at medium pressures
with HDO degrees similar to those obtained at 30 bar. Ni2P is less affected by changes in
H2/DBF ratio than CoP and the catalysts are more active at high H2/DBF molar ratios.
Keywords: HDO; Ni2P; CoP; Dibenzofuran; Silica; transition metal phosphides

Catalysis Science & Technology
Page 2 of 36
View Article Online
DOI: 10.1039/C5CY00282F
2

Page 3 of 36
Catalysis Science & Technology
View Article Online
DOI: 10.1039/C5CY00282F
3
1.
Introduction
The use of bioenergy as a renewable alternative to fossil fuels is nowadays attracting more and
more attention. The reasons are quite obvious: our world is running out of oil reserves; the raw
oil quality is decreasing;1 and the world petroleum consumption continues to rise as a conseꢀ
quence of the world’s high population growth, the industrialization and a big change in conꢀ
sumption pattern. Within this scenario, biofuels seem to be a potential energy substitute for fossil
fuels since it is a renewable resource that could contribute to sustainable development and global
environmental preservation2.
Bioꢀoil is formed from fast pyrolysis of biomass. It retains most of the original oxygen in the
feedstocks and the oxygen content varies in the range of 35ꢀ60 wt.% on a wet basis. First of all,
the oxygenated compounds make the bioꢀoil polar and for this reason it will be nonꢀmiscible
with nonꢀpolar petroleum fuels. The high oxygen content also gives the bioꢀoil other important
undesirable properties. Hence, upgrading is necessary to use the pyrolysisꢀoil as a competitive
substitute of petroleum fuels. Most important is a deꢀoxygenation step prior to coꢀprocessing3, 4.
HDO is a promising future process for the upꢀgrading of bioꢀoil. The HDO process requires the
presence of hydrogen and a catalyst, and it normally operates at moderate temperatures, up to
450 °C and at a high pressure, 10ꢀ300 bar. The high pressure increases the reaction rate and deꢀ
creases the formation of coke in the reactor. A good HDO catalyst should be highly active to
remove oxygen in the bioꢀoil with a high yield and it should not produce coke5.
HDO is a process very similar to HDS and commercial catalysts, often used for traditional hyꢀ
drotreating processes such as CoꢀMoS2 and NiꢀMoS2 supported on ꢀAl2O3, have been the most
frequently tested catalysts also for HDO reactions.6 However a problem with the above named
catalysts is that coꢀfeeding of H2S is needed in order to regenerate the sulphide sites. Further the
sulphidation agent must be separated from the products and the sulphur must be recovered from
the H2S. Another disadvantage is the corrosiveness of H2S7.
Noble metal supported catalysts as for example Ru/C and Pt/C have also shown good results in
HDO reactions. Noble metal catalysts in HDO are also active at low temperatures. However, the
main disadvantages with noble metal based catalysts are their high cost and additionally they
show a low resistance towards poisoning by the presence of sulphur8.
Alternative compositions that show good hydrogen transferring properties are metal carbides,
metal nitrides, metal borides and transition metal phosphides9. Transition metal phosphides have

Catalysis Science & Technology
Page 4 of 36
View Article Online
DOI: 10.1039/C5CY00282F
4
presented very interesting properties in hydrotreating reactions as they are bifunctional with acid
and metallic catalytic functions. The metallic function is responsible for hydrogenation, while on
the acidic sites, hydrogenolysis, dehydration, and isomerization reactions take place. The balance
among these functions determines the observed catalytic activity. From FTIR studies of CO adꢀ
sorption of the transition metal phosphide Ni2P/SiO2, it has been shown that there are enough
electrons on the metal site to allow πꢀback bonding. This πꢀback bonding enables molecular actiꢀ
vation as electrons fill antibonding orbitals and the hydrogenation activity is thereby enhanced.
The acidity arises from acidic POꢀH sites10.
In the last decade a series of studies have shown that MoP, WP, Fe2P, CoP and Ni2P are suitable
for hydrodesulfurization and hydrodenitrogenation of fossil crude oils11ꢀ15 . More recently, these
transition metal phosphides have also been applied as catalysts in HDO reactions of bioꢀoil 16, 17
as they show good hydrogen transfer properties. These kinds of catalysts are favored when supꢀ
ported on a less acidic support as silica, SBAꢀ15 or MCMꢀ41. Zhao et al.18 measured the activity
for the HDO reaction of guaiacol over Ni2P, Co2P, Fe2P, WP and MoP catalysts supported on
SiO2 in a fixed bed reactor with a molar ratio H2/guaiacol of 33. The results indicated that the
type of metal has a significant effect on product distribution and a different involvement of metꢀ
als during the hydrogen activation or the interaction of reactants with the catalyst surface were
suggested. The activity of the different catalyst followed the order: Ni2P >Co2P
>Fe2P≈WP≈MoP. The activities of the transition metal catalysts were compared to the catalytic
activity of the commercial catalyst CoMoS/Al2O3 and 5%Pd/Al2O3. The commercial 5%
Pd/Al2O3 catalyst turned out to be more active than the metal phosphides at lower contact time,
but the major product was catechol which is undesired. A. Iino et al.19 studied the HDO reaction
of 2ꢀmetyltetrahydrofuran (2ꢀMTHF) on a Ni2P/SiO2 catalyst with a continuousꢀflow quartz reꢀ
actor. The activity and selectivity of this transition metal phosphide catalyst were compared to
Pd/Al2O3 and Ru/C catalysts. The results showed that Ni2P/SiO2 had higher selectivity to nꢀ
pentane and lower selectivity to lower hydrocarbons with carbon numbers less than 5 compared
to Pd/Al2O3 and Ru/C. It was concluded that Ni2P/SiO2 catalyst has high HDO activity and low
carbon bond cracking propensity.
When it comes to the reaction parameters, Mortensen et al.20 describe in their review of catalytic
upgrading of bioꢀoil that a high pressure is beneficial in the HDO process. A high pressure inꢀ
creases the solubility of hydrogen in the oil and consequently there will be higher availability of
hydrogen in the proximity to the catalyst. This increases the reaction rate and counteracts carbon
deposition on the surface of the catalyst. However, a great amount of hydrogen interferes with

Page 5 of 36
Catalysis Science & Technology
View Article Online
DOI: 10.1039/C5CY00282F
5
the adsorption of reactant on the surface19. In the work of A. Iino et al.19 the results showed that
the conversion decreases as H2 partial pressure increases. These experiments were run under H2
partial pressures in the range 75ꢀ418 kPa. In a recent study of T.A. Le et al.21 the HDO process
of 2ꢀfuryl methyl ketone (FMK) catalyzed by CoP/γꢀAl2O3, the complete HDO of FMK into meꢀ
thyl cyclopentane and methane was achieved at 400 °C and at atmospheric pressure. The authors
urge that positive catalytic results can be achieved even at atmospheric pressure (low pressure).
In this work, the focus is on the activity of nickel phosphide catalyst (Ni2P) and cobalt phosphide
catalyst (CoP), both supported on silica and tested in the HDO of dibenzofuran (DBF). In a preꢀ
vious work the HDO of DBF on Ni2P with different metallic loadings and P/Ni molar ratios was
studied. In the present paper an evaluation of the impact of the experimental parameters on the
catalytic response of these catalysts is carried out (contact time, H2 pressure and H2/DBF molar
ratio) to better understand the reaction mechanism of the two phosphideꢀtype catalysts.
2. Experimental
2.1. Materials
The catalytic support used in this study was commercial fumed silica from SigmaꢀAldrich. The
reagents used for the preparation of the precursor solution were nickel(II) hydroxide (Ni(OH)2),
cobalt(II) hydroxide (Co(OH)2, 95%) and phosphorous acid (H2PO3H, 99%), all supplied by
Sigma Aldrich. In the catalytic reactivity study the chemicals used were dibenzofuran(C12H8O,
98%) and decahydronaphtalene, cis+trans (C10H18, 98%), both from Alfa Aesar. The gases used
in the experimentation, H2, N2, CO, NH3 had a purity of 99.9999% and were supplied by Air
Liquide.
2.2. Preparation of catalysts
Nickel phosphide and cobalt phosphide catalysts containing 5 wt.% of nickel and cobalt, respecꢀ
tively, supported on fumed commercial silica (SBET = 218 m2 g−1 and Vp = 0.52 cm3 g−1) were
prepared. Precursor solutions of nickel(II) dihydrogenphosphite (Ni(HPO3H)2) and cobalt(II)
dihydrogenphosphite (Co(HPO3H)2) were prepared as described elsewhere14, 15. The precursor
solutions were incorporated into the silica support by the incipient wetness impregnation method
to prepare catalysts containing 5 wt.% of metal. To convert the precursor phosphite into phosꢀ
phide the impregnated solids were dried at 60°C for 24 hours and treated in hydrogen by heating
at a linear temperature ramp (3 ºC min−1) in flowing H2 (100 ml min−1), from rt. to 620 ºC for
nickel phosphide catalysts and from rt. to 700 ºC for cobalt phosphide catalysts. The obtained
catalysts will hereafter be named as Ni2P and CoP.

Catalysis Science & Technology
Page 6 of 36
View Article Online
DOI: 10.1039/C5CY00282F
6
2.3. Characterization of catalysts
The textural properties (SBET, Vp, dp) of the fresh catalysts were obtained from the N2 adsorptionꢀ
desorption isotherms at ꢀ196 ºC measured with a Micromeritics ASAP 2020 apparatus. Prior to
the measurements the samples were outgassed at 200 ºC and 0.01 kPa overnight. Surface areas
were determined by using the BrunauerꢀEmmetꢀTeller equation and a nitrogen molecule cross
section of 16.2 Å2 (3). The pore size distribution was calculated by applying the BarrerꢀJoynerꢀ
Halenda method (BJH) to the desorption branch. The total pore volume was calculated from the
adsorption isotherm at P/P0 = 0.95.
Xꢀray diffraction patterns (XRD) of the samples were obtained with aX´Pert PRO MPD Philips
diffractometer (PANanalytical), equipped with a Ge (111) primary monochromator using CuKα1
(1.5406 Å) radiation. The Xꢀray tube was set at 45 kV and 40 mA and the measurements were
carried out from 10° to 70° (2theta) for 12 minutes.
CO chemisorption analyses were performed under static volumetric conditions in a Micromeritꢀ
ics ASAP 2020 apparatus. Samples were reduced ex situ and transferred into an inert atmosꢀ
phere. Prior to measurement samples were reꢀreduced in situ in H2 at 300 °C and evacuated at 25
°C for 10 h. The chemisorption isotherm was obtained by measuring the amount of CO adsorbed
between 10 and 600 mmHg at 35°C. After completing the initial analysis, the reversibly adꢀ
sorbed gas was evacuated and the analysis repeated to determine only the chemisorbed amounts.
Xꢀray photoelectron spectra were collected using a Physical Electronics PHI 5700 spectrometer
with nonꢀmonochromatic Al Kα radiation (300 W, 15 kV, and 1486.6 eV) with a multiꢀchannel
detector. Spectra of pelletized samples were recorded in the constant pass energy mode at 29.35
eV, using a 720 ꢁm diameter analysis area. Charge referencing was measured against adventiꢀ
tious carbon (C 1s at 284.8 eV). A PHIACCESS ESCAꢀV6.0 F software package was used for
acquisition and data analysis. A Shirleyꢀtype background was subtracted from the signals. Recꢀ
orded spectra were always fitted using GaussianꢀLorentzian curves in order to determine the
binding energy of the different element core levels more accurately. Reduced catalysts were
stored in sealed vials with an inert solvent. The sample were prepared in a dry box under a N2
flow and analyzed without previous treatment.
Temperatureꢀprogrammed desorption (TPD) experiments were carried out to determine the
acidity of the samples by using ammonia (NH3ꢀTPD) and the hydrogen activation capability was
determined by using Hydrogen (H2ꢀTPD). For NH3ꢀTPD experiments, 80 mg of catalyst precurꢀ

Page 7 of 36
Catalysis Science & Technology
View Article Online
DOI: 10.1039/C5CY00282F
7
sor was reduced at atmospheric pressure using the same procedure described in the section 2.2.
After flushing with helium and adsorption of ammonia at 100 °C, NH3ꢀTPD was performed usꢀ
ing a helium flow and raising the temperature from 100 to 800 °C at a heating rate of 10 °C minꢀ
1. Prior to raising the temperature, physisorbed ammonia was removed flushing with helium. For
the H2ꢀTPD measurements 300 mg of catalyst precursor were reduced using the same procedure
as described in section 2.2. The catalyst was cooled under flowing helium from the reduction
temperature down to 100 °C and then exposed to H2 gas with a flow of 60 mL minꢀ1 at 100 °C
for 30 minutes, followed by flushing in argon for another 30 minutes in order to remove phyꢀ
sisorbed hydrogen, and then heated in argon to 800 °C. The evolved gas was analyzed by online
gas chromatography (Shimadzu GCꢀ14A) equipped with a TCD.
Fourier transform infrared (FTIR) spectra of the samples after reduction and outgassing at 400°C
as well as FTIR spectra of adsorbed CO were recorded with a Nicolet Nexus instrument, using a
conventional cell connected to a gasꢀhandling system. All spectra were obtained after subtraction
of the spectrum corresponding to the degassed sample. Prior to adsorption, pressed disks of
powder samples with a diameter of 1.5 cm of sample powders (ca.20 mg) were thermally treated
within the cell at 400 ºC Under H2 (67 kPa) for 1 h. CO adsorption (0.13 kPa) experiments were
performed at ꢀ140 ºC and spectra were recorded in the range ꢀ140 to 10 ºC while degassing.
2.4. Catalytic Activity
Dibenzofuran (DBF) was chosen as a model compound for the HDO reaction of bioꢀoil. Furan
derivate compounds are common template molecules used to evaluate the catalyst performance
in HDO reactions including dibenzofuran which is a very low reactive compound. The reactions
were carried out in a highꢀpressure fixedꢀbed continuousꢀflow stainless steel catalytic reactor
(9.1 mm in diameter and 230 mm in length). The concentration of the organic feed was in the
range of 1ꢀ4 wt.% and decalin was used as solvent. The catalysts were reduced exꢀsitu and saved
in a close vial containing cyclohexane to avoid oxidation. For the activity tests, an amount of
0.25 g of catalyst were used (particle size 0.85–1.00 mm) diluted with quartz sand to a total volꢀ
ume of 3 cm3. Prior to the catalytic tests, the catalyst was left to dry overnight in a hydrogen flow
of 100 mL minꢀ1 with a temperature ramp starting at 40 °C and finishing at 400°C with a heating
rate of 3 °C minꢀ1. After drying and before catalytic testing, the catalysts were left in a nitrogen
flow of 20 mL minꢀ1 at 40 °C to avoid oxidation. Catalytic activities were measured at 275 °C
under a 0.1ꢀ3.0 MPa pressure, with a flow rate of H2 of 100 mL minꢀ1 and with hourly space veꢀ
locities (WHSV) in the range of 0.4 ꢀ 2.6 hꢀ1. The evolution of the reaction was monitored by
collecting liquid samples every 60 minutes for 6 hours. A volume of 0.33 ml of product was colꢀ

Catalysis Science & Technology
Page 8 of 36
View Article Online
DOI: 10.1039/C5CY00282F
8
lected and diluted up to 1 ml volume with pure decalin and 20 µl of nonane were added as an
internal standard. These samples were kept in sealed vials and subsequently analyzed by gas
chromatography (Shimadzu GCꢀ14B, equipped with a flame ionization detector and a capillary
column, TBRꢀ14, coupled to an automatic Shimadzu AOCꢀ20i injector). In order to identify inꢀ
termediates samples were also analyzed by CGꢀMS (MODEL) with a TBRꢀ14 capillarity colꢀ
umn.
The catalytic selectivity was calculated considering all the reaction products observed, biphenyl
(BP), cyclohexylbenzene (CHB), tetrahydrodibenzofuran (THDBF), hexahydrodibenzofuran
(HHDBF), 2ꢀcyclohexylphenol (2ꢀCHP), bicyclohexylꢀ2ꢀenꢀ3ꢀol (BCHꢀ2ꢀenꢀ3ꢀol) and cyclopenꢀ
tylmethylcyclohexane (CꢀPEꢀMEꢀCH). Carbon balances are in the range of 1.00 +/ꢀ 0.05
The HDO conversion was calculated by the following equation;
CDBF.Ini −CDBF.Fin −C
HDOConv.(%) =
Int ×100
CDBF.Ini
(1)
where CDBF.Ini is the DBF concentration in the feed, CDBF.Fin is the DBF concentration in the HDO
liquid product and CIntHyd is the concentration of Oꢀcontaining intermediates (THDBF, HHDBF,
2ꢀCHP, BCHꢀ3ꢀenꢀ2ꢀol).
The turnover frequency was calculated using equation (2):
F W
(2)
TOF = −
× ln(1− X )
M
where F is the molar rate of reactant (mol minꢀ1), W is the catalyst weight (g), X is the converꢀ
sion and M is the mole of sites loaded (mol gꢀ1) calculated from COꢀchemisorption. This formula
was used because the conversions were close to 100% so were far from differential conditions,
and an integral analysis was needed. This equation, in which ꢀln(1ꢀX) substitutes for X assumes a
pseudo firstꢀorder reaction which may be justified by the large excess of hydrogen.
In order to determine if internal diffusion is limiting the reaction the WeiszꢀPrater criterion can
be used. It uses measured values of the rate of reaction to determine if the reaction suffers from
problem of internal diffusion. Internal mass transfer effect can be neglected when the value of
CWP is less than 116. The detail of the calculations made is found as suplementary information
(S1). All the values of CWP obtained under the experimental conditions studied were in the order
of 10ꢀ2ꢀ10ꢀ3. This indicates that internal mass transfer effects can be neglected at the conditions
employed for this reactivity study.

Page 9 of 36
Catalysis Science & Technology
View Article Online
DOI: 10.1039/C5CY00282F
9
3. Results and discussion
3.1. Catalytic study
The HDO activity and selectivity of the reduced catalysts were investigated by using DBF as a
model molecule present in bioꢀoil. Several experimental parameters were evaluated in order to
better understand the reaction mechanism of the HDO reaction of DBF and to find the optimal
experimental conditions to transform DBF into Oꢀfree products. Several types of products can be
formed during HDO reaction, presented in Scheme 1, depending on the route followed. There is
a hydrogenation pathway 1, in which oxygen removal is preceded by hydrogenation of benzene
rings; a direct deoxygenation pathway 3, in which oxygen removal occurs by direct cleavage of
bonds linking oxygen to aromatic carbons; and an intermediate pathway 2, in which oxygen reꢀ
moval is preceded by hydrogenation of one benzene ring only.
In the present work a further study to investigate the reaction mechanism of HDO of DBF on
Ni2P and CoP catalysts is performed by changing several operational parameters; contact time,
reaction pressure and H2/DBF ratio feed.
3.1.1 Contact time analysis
Contactꢀtime experiments were carried out to improve the understanding of the reaction network
and therefore the hydrogen gas volume flow and the feed (DBF in decalin) volume flow were
adjusted. The hydrogen pressure, H2/DBF molar ratio and the reaction temperature maintained
constant. The catalytic results in terms of total conversion and HDO conversion are presented in
Figure 1 in a contact time range of 1.8ꢀ12 s. It is observed that both total (dotted lines) and HDO
conversion (solid lines) rise with increased contact time. Total conversion is always higher than
HDO conversion, since the former only considers DBF transformation; instead, the later considꢀ
ers DBF transformation into Oꢀfree compounds. It is clearly seen that the longer the catalyst is in
contact with the reactants the more of them can react and transform into oxygen free products.
The Ni2P catalyst turns out to be more active than CoP, whereas only at low contact times CoP is
slightly more active than Ni2P. Ni2P catalyst presents the highest HDO conversion values at a
contact time of 6 s, where the conversion values are reaching 98%. Instead, on CoP catalyst the
highest value (80%) is attained at a contact time of 12 s. At low contact times the catalysts exhibꢀ
it a low catalytic activity (less than 15% conversion). With regard of to the lower conversion
value observed at very high contact time for Ni2P catalyst, in a recent work of A. Lino et al19
about HDO of 2ꢀMTHF on Ni2P, these authors have observed a linear relationship in the plot of

Catalysis Science & Technology
Page 10 of 36
View Article Online
DOI: 10.1039/C5CY00282F
10
conversion versus contact time. These author have pointed that in this situation each reactant
molecule is reacting independently, meaning that each reactant molecule sees essentially a fresh
surface. On Ni2P, the linearity is maintained until a contact time of 6 s where the conversion is
close to 100%. By increasing the contact time the conversion lowers perhaps to surface coverage
with other organic species. If there were coverage, the rate of adsorption would be retarded by
those species and therefore decreasing the conversion.
The reaction network is studied by observing the trends of product selectivity at different contact
times. Figure 2 shows the product selectivity as a function of contact time for Ni2P catalyst and
CoP catalyst, respectively. The products are classified into three groups; saturated oxygenꢀfree
cyclic compounds, unsaturated oxygenꢀcontaining cyclic compounds and furans.
At low contact times the intermediates tetrahydrodibenzofuran (THDBF) and hexahydrodibenzoꢀ
furan (HHDBF) are observed as products along with 2ꢀcyclohexylphenol (2CHP) and bicycloꢀ
hexylꢀ2ꢀenꢀ3ꢀol (BCHꢀ2ꢀenꢀ3ꢀol). It should be noted that on Ni2P, the selectivity to 2ꢀCHP did
not change obviously, and the selectivity to BCHꢀ2ꢀenꢀ3ꢀol firstly increased and then decreased,
indicating that BCHꢀ2ꢀenꢀ3ꢀol is an intermediate for Ni2P catalyst. This is not the case for CoP
one: as the contact time increased, the selectivities to 2ꢀCHP and BCHꢀ2ꢀenꢀ3ꢀol did not change
monotonously.
By increasing the contact time the transformation of Oꢀcontaining intermediates into Oꢀfree
compounds occurs. Bicyclohexane (BCH) is the main Oꢀfree compound product formed in all
reactions.
Based on the observation of formed intermediates and products and reviews of earlier studies 9,22
,
the reaction mechanism of HDO on CoP and Ni2P catalysts follows the hydrogenation pathway
1. In the first step one of the benzene rings in DBF is hydrogenated and the furan ring opens up:
DBF→THDBF→HHDBF→2ꢀCHP
In the next step the second benzene ring is hydrogenated:
2ꢀCHP→BCHꢀ3ꢀenꢀ2ꢀol→BCHol
Then the oxygen removal takes place by dehydration and cycloalkane formation, which is the
last step when both aromatic rings are completely saturated.
BCHol→BCH + CꢀPEꢀCH + CꢀPEꢀMEꢀCH
From these results it can be concluded that the catalysts have both metallic and acidic properties.
They show good hydrogenation properties which can be explained by the high metallic disperꢀ

Page 11 of 36
Catalysis Science & Technology
View Article Online
DOI: 10.1039/C5CY00282F
11
sion and they also present acidic properties which shows in their hydroisomerization ability leadꢀ
ing to CꢀPEꢀCH and CꢀPEꢀMEꢀCH. Moreover a contact time of 6 s seems to be adequate to asꢀ
sess the transformation of DBF into Oꢀfree products by following the route 1 presented in
Scheme 1.
3.1.2 H2 Pressure study
Another parameter evaluated was the reaction H2 pressure, an important parameter in hydrogenaꢀ
tion reactions. For this aim the reaction was carried out at three different H2 pressures, 1, 15 and
30 bar, while the other parameters were kept constant; temperature 275 °C, H2/DBF molar ratio
69 and the contact time 6 s. Figure 3 shows the total conversion and the HDO conversion as a
function of the H2 pressure. The plots of Figure 3 indicate that the activity is very low when the
reaction is performed at atmospheric pressure (the conversion is below 5% with both catalysts).
Nonetheless, Ni2P is more active than CoP at medium and high pressures. Moreover, both Ni2P
and CoP are active at medium pressures with HDO degrees similar to those attained at 30 bar in
the case of CoP. These results indicate that the reaction can be carried out at moderate pressures
and that high pressures are not required to perform Oꢀremoval as in Sꢀ Nꢀ removal, where high
pressures are always required. A higher H2 pressure is beneficial also in hydrodeoxygenation as
by increasing the pressure the thermodynamics would shift further towards complete conversion.
However, too much H2 interferes with the adsorption of reactants on the surface19. This could
explain why the conversion is slightly higher at the 15 bar pressure compared to the 30 bar presꢀ
sure for CoP catalyst. Wang et al23 observed an improvement in DBF conversion via the hydroꢀ
genation pathway by increasing the pressure from 0.5 to 5.0 MPa on Pt/SBAꢀ15 catalysts. In our
case the excellent activity of metallic phosphides, even being highly active at medium pressures,
should be highlighted in this type of reaction.
Considering the Oꢀfree product yield (Figure 4) at the different pressures, it is possible to obꢀ
serve that there is no major changes in product yield distribution by increasing the pressure from
15 to 30 bar in the case of the CoP sample, however, the Ni2P catalyst yielded greater amounts of
BCH at high pressures. It is worth noting that the maximum yield of a certain intermediate was
<4% in all experiments. At pressures of 15 and 30 bar the yield of the final product, BCH, is
much higher for Ni2P. These observations are consistent with the mechanism and the conclusion
is that CoP catalyst is less active than Ni2P.

Catalysis Science & Technology
Page 12 of 36
View Article Online
DOI: 10.1039/C5CY00282F
12
3.1.3 H2/DBF molar ratio study
With the aim to study how the reaction is influenced by different H2/DBF molar ratios, the feed
was prepared with three different compositions (H2/DBF molar ratios of 34, 69 and 139). Figure
5 shows the total conversion and the HDO conversion as a function of H2/DBF molar ratio.
The first thing observed is that the Ni2P catalyst is much more active than CoP, and also that the
Ni2P sample is less affected by H2/DBF molar ratios than the CoP catalyst. In fact, at low
H2/DBF molar ratios the activity of the Ni2P sample is close to that observed at high molar ratiꢀ
os. On the contrary the CoP sample shows a sharp increase in activity by increasing the H2/DBF
molar ratio. The highest conversion of all the experiments, when keeping the contact time at 6 s,
is achieved at the highest H2/DBF molar ratio for both catalysts.
The yields of the different reaction products for this set of experiments are depicted in Figure 6.
It is observed that the Ni2P sample shows high yields to BCH (>80%) regardless the H2/DBF
molar ratio employed, which demonstrates its high efficiency to convert DBF into Oꢀfree comꢀ
pounds under poorer H2 conditions. The CoP samples only attain high yields of BCH at high
H2/DBF molar ratios.
In fact, when the HDO conversion evolution with time on stream is observed (Figure 7), at low
H2/DBF molar ratio (Figure 7A), cobalt catalyst suffers from a fast deactivation after the first
hour of reaction. Its initial HDO conversion is close to 80%, meanwhile after 6 h on stream the
HDO conversion attained is down at 35%. However HDO conversion for the Ni2P sample under
the same conditions slightly decreases and reaches the steady state after 6 hours with a converꢀ
sion of 89%.Contrary and as expected, at H2/DBF=139 (Figure 7.B) both catalysts are rather
stable, which confirms previous statement; higher H2 contents are required to improve the activiꢀ
ty of CoP.
3.1.4. Intrinsic activity of the prepared sample-TOF number
In order to evaluate the intrinsic activity of Ni2P and CoP catalysts at the different conditions
employed, TOF numbers were calculated according to Eq. 2 and is defined as the moles of reacꢀ
tant converted per mol of active sites per unit of time. This value gives information about the
activity per active site. Table1 compiles the calculated TOF numbers for contact time, pressure
effect and H2/DBF effect experiments, respectively.
From this table, it is observed how the TOF number is higher for the Ni2P catalyst except at low
contact time contact time. Moreover, the TOF number increases for the Ni2P catalyst when the
contact time is increased up to 6 s, for longer contact times the TOF number decreases. However,

Page 13 of 36
Catalysis Science & Technology
View Article Online
DOI: 10.1039/C5CY00282F
13
no great changes in TOF number are observed for the CoP sample, what could indicate that the
activity per active site is hardly affected by changes in contact time; it only increases at the highꢀ
est contact time studied. Contrary the Ni2P sample is much more sensitive to these changes with
the TOF number increasing with about 400% from 1.8 to 6.0 s, in accordance to the sharp inꢀ
crease in the conversion. From H2 pressure effect experiments it is observed that at low H2 presꢀ
sure, the activity per active site is very low for both samples. Nevertheless at 15 and 30 bar, the
Ni2P sites are much more active than the CoP ones. Finally, H2/DBF experiments also showed
that the Ni2P sites had the highest activity, presenting much higher TOF values than the CoP
samples. Wang et al.23 tested Pt, Pd and Ru catalysts in HDO of DBF and the most active catalyst
presented a TOF value, at 30% HDO conversion, lower than the TOF value presented by CoP
catalysts at the same conversion value (15 bar), 0.42 sꢀ1 vs. 2.55 sꢀ1, respectively.
It should be highlighted that the TOF value reached under H2ꢀpoor conditions (4 wt.%DBF) is
the highest 23.7 sꢀ1 and much higher than those found in the literature for nobleꢀmetal based cataꢀ
lysts22ꢀ24
.
3.2. Catalyst characterization
The textural properties of the prepared catalysts are included in Table 2. As showed in this table,
BET surface area of the SiO2 support was 218 m2 gꢀ1 and that of the fresh Ni2P and CoP catalysts
in pellets form were 109 and 126 m2 gꢀ1, respectively. The decrease in specific surface area indiꢀ
cates that pores are partly blocked due to the presence of phosphides particles on the support
which partially impede the access of gas molecules25, 26. The textural properties of Ni2P and CoP
prepared catalysts were similar.
The active phases formed after reduction were firstly evaluated from XRD experiments. Ni2P
and CoP catalysts do not show any wellꢀdefined diffraction lines (diffractograms not shown
here). This is due to the low metallic loading of the catalysts (5 wt.%) and because of the high
dispersion of the active phase as observed in earlier studies9 . The diffractogram of Ni2P shows
some weak diffraction signals centred at 2θ (º) = 40.7, 44.6 and 55.0 and assigned to the Ni2P
phase (PDF 01ꢀ089ꢀ4864). Meanwhile, CoP catalyst shows a very weak diffraction line located
at 2θ (º) = 48.3 and assigned to the CoP phase (JCPDS No. 29ꢀ0497).
CO chemisorption capacity of the reduced catalysts provides with information about the number
of surface metal atoms and gives an estimation of the number of active sites. Table 2 compiles
the CO uptake at room temperature for cobalt phosphide and nickel phosphide catalysts. Both

Catalysis Science & Technology
Page 14 of 36
View Article Online
DOI: 10.1039/C5CY00282F
14
samples chemisorb similar amounts of CO, i.e., the number of active sites is quite similar in both
catalysts.
A more detailed information about the phases formed on the catalyst surface was obtained by
XPS. The corresponding binding energies values are summarized in Table 3. Both, Ni 2p core
level spectrum and Co 2p core level spectrum are decomposed into three contributions, i.e., three
doublets containing the 2p3/2 and 2p1/2 components. The first contribution at lower binding enerꢀ
gy is due to Ni2P (853.1 eV) and CoP phase (777.3 eV); the second contribution centered at
856.6 eV for Ni2P catalyst and at 780.7 eV for CoP samples is due to Me2+ species; meanwhile,
the third contribution observed is due to the shakeꢀup satellite assigned in the literature to divaꢀ
lent species. The core level spectra of P 2p were also analyzed (See Fig S1, Supplementary Inꢀ
formation) and decomposed. Both catalysts show two doublets (P 2p3/2ꢀsolid lines and P 2p1/2
dash lines) were observed indicating that, at least, two phosphorus species are present on the
catalyst surface. The band close to 134.0 eV is due to phosphate species and the other band at
129.3 eV for Ni2P sample and at 129.0 eV for CoP one is due to the phosphide phase. No unreꢀ
duced HPO3Hꢀ species are observed, which could be a result of the high reduction temperature
employed31. A much more detailed discussion of XPS spectra for these phases can be found in
previous papers27, 28
.
3ꢀ
Table 3 includes the percentage of phosphorous as PO4 and as phosphide (Ni2P or CoP). It is
clearly seen the much greater amount of phosphate species on Ni2P rather than on CoP, indicatꢀ
ing a surface enrichment of P in the former. In line with this, the corresponding surface atomic
ratios indicate that the Ni/P molar atomic ratio is much lower than the expected values due to the
formation of the Ni2P phase. On the contrary, for CoP sample, the Co/P ratio is higher than the
expected stoichiometric value due to the presence of CoP which indicates that there is a greater
concentration of metallic atoms on the surface.
These results are explained considering that the same amount of Pꢀprecursor has been added in
both catalysts. Nonetheless Ni2P is a metallic rich phosphide and CoP is a stoichiometric phosꢀ
phide, therefore more P is included into the CoP structure. Moreover the reduction temperature
to form CoP is higher than that for Ni2P, 700 ºC vs 620 ºC. Therefore it is expected thatat greater
quantities of P are released during the reduction process of CoP. As a result, the remaining extraꢀ
P is expected to be higher in the Ni2P catalyst.
The acidic properties of the catalysts were measured by temperatureꢀprogrammed desorption of
ammonia (NH3ꢀTPD) between 100 and 800 ºC. Figure 8 shows the corresponding profiles while

Page 15 of 36
Catalysis Science & Technology
View Article Online
DOI: 10.1039/C5CY00282F
15
Table 4 compiles the amount of NH3 desorbed (ꢀmol of NH3 g−1) for the reduced catalysts. By
comparing both profiles and the data compiled in Table 3 one sees that the amount of NH3 reꢀ
leased is much higher on Ni2P than on CoP and Ni2P sample releases NH3 molecules at temperaꢀ
tures lower than 400 ºC (weak acidity) while CoP sample releases NH3 molecules at very high
temperatures (550ꢀ700 ºC), indicating the presence of strong acid sites on this CoP sample. Preꢀ
viously the presence of both Lewis and Brönsted acid sites has been reported for the Ni2P cataꢀ
lyst29. Lewis acidity was ascribed to unreduced Ni2+ and Brönsted acidity due to POꢀH species.
No much Co2+ or Ni2+ species are expected to be present here, due to both the high reduction
temperature applied. Therefore, the acidity would be expected to come from POꢀH groups. XPS
data show that the Ni2P sample presents a surface enrichment of P species which could account
for its higher acidity.
FTꢀIR spectra of the samples after reduction and outgassing at 400°C will also give information
about the presence of surface hydroxyl groups. The corresponding spectra are represented in
Figure 9. These clearly show a main band at 3742 cmꢀ1 due to stretching of isolated silanol speꢀ
cies together with a weaker band centered at 3670 cmꢀ1 due to POꢀH stretching modes, thus conꢀ
firming that the acidity detected over these samples arises mainly from the Phosphorousꢀ
containing species. The frequency of this band is completely consistent with the frequency valꢀ
ues reported for phosphoric acid supported on several oxides such as Al2O3, SiO2, ZrO2, forming
hydrogenphosphate species30, 31. A shoulder at 3700 cmꢀ1 also seems to be present, which could
be due to another kind of OH groups showing intermediate acidic properties, possibly some siꢀ
lanols affected by nearby phosphate species. Moreover, the intensity ratio I3670/I3742 for the two
silanol and PꢀOH bands is higher in the Ni2P sample spectrum, as a further evidence that this
catalyst shows an higher surface concentration of PꢀOH species, as it was previously suggested
by XPS data, and also justifying the greater acidity presented by this catalyst. In line with this,
IR analysis of adsorbed CO at liquid nitrogen temperature over Ni2P sample corroborates the
presence of PꢀOH groups on this catalyst. The corresponding subtraction spectra are reported in
Figure 10, in the CO stretching region the main bands of the spectrum are centered at 2187, 2158
and 2130 cmꢀ1, shoulder. The main band at 2158 cmꢀ1 is assigned to CO weakly interacting with
surface OH groups32 through Hꢀ bonds and this band is readily disappearing at decreasing CO
coverage. Its assignation is also confirmed by the negative band appearing in the OH stretching
region (inset) due to the consumption of isolated OH groups after interaction with CO. From
these data it seems that silanols are involved first (highest frequency band), but also PꢀOH are
interacting with CO which behaves as weak base (negative component at 2168 cmꢀ1). Decreasing

Catalysis Science & Technology
Page 16 of 36
View Article Online
DOI: 10.1039/C5CY00282F
16
CO coverage leads to a partial restoring of silanols, but on the other hand the negative band due
to PꢀOH group is still clearly detectable, thus the hydrogen bonding of CO with the surface hyꢀ
droxy groups of hydrogenphosphate is significantly stronger than that with silanol, in agreement
with PꢀOH higher acidic strength.
On the other hand the band detected at 2187 cmꢀ1 is due to carbonyls coordinated over Niꢀions,
acting as Lewis acidic sites at the catalyst surface29, 32. Upon outgassing this band decreases in
intensity and shifts towards higher frequencies, as expected. The weaker component at 2130 cmꢀ1
can be assigned to carbonyls coordinated over partially reduced Ni ions (possibly Ni+), whereas
only a very weak and broad absorption from 2050 to 2000 cmꢀ1 could be assigned to carbonyls
coordinated over a very dispersed nickel metal phase33. The presence of some metallic nickel
particles cannot be discarded, possibly arising from Ni ions strongly interacting with the support
and not involved with Pꢀspecies during reduction to form the phosphide phase.
As for the CoP sample, the strongest acidic sites detected by NH3ꢀTPD could be assigned to the
Co ions unreduced still exposed at the catalyst surface. In order to ascertain this hypothesis we
performed CO adsorption and indeed the IR spectrum recorded at room temperature, i.e. when
the strong band due to Hꢀbound CO is not detectable, shows a weak carbonyl band at 2200 cmꢀ1
which can be assigned to CO coordinated over Co ions (insert in Figure 10).
Finally, when testing a catalyst for hydrotreating reactions, it is important to consider its capabilꢀ
ity to activate hydrogen. To this end, H2ꢀTPD experiments on Ni2P and CoP reduced samples
were performed and their corresponding profiles are shown in Figure 11. In general, the temperaꢀ
ture at which Hꢀspecies are released helps to understand the nature of these species. The H2ꢀ
desorption below 250 °C is ascribed to the adsorbed H2 on metal Ni or Co sites, while the H2
desorption between 250°C and 450°C is ascribed to the H2 from the metalꢀsupport interface. Fiꢀ
nally, desorption at higher temperatures is usually ascribed to H2 spill over species. At first
glance, when comparing the H2ꢀTPD profiles of the two different catalysts it is clear that the
Ni2P catalyst desorbs greater amounts of hydrogen species, indicating a greater capacity to actiꢀ
vate H2 compared to CoP. Moreover, the Ni2P catalyst mainly desorbs hydrogen at low temperaꢀ
ture, that is, coming from the metallic surface. Besides, greater amounts of Hꢀspillover are also
present on this catalyst. Literature data indicate that H2ꢀdesorption at intermediate temperatures
on Ni2P catalysts could be related to PꢀOH groups at the metal support interface34 and catalysts
35
with greater amount of PꢀOH groups desorbs a greater amount of H2 . In fact, the Ni2P catalyst
desorbs greater amounts of H2 from both metal surface and metalꢀsupport interface, revealing the

Page 17 of 36
Catalysis Science & Technology
View Article Online
DOI: 10.1039/C5CY00282F
17
greater capability of Ni2P to activate hydrogen and also the greater amount of PꢀOH present,
contributing to H2ꢀdesorption at intermediate temperatures.
3.3. Discussion
Catalytic results have shown the greater activity of the Ni2P catalyst with regard to the CoP one.
The most important results are that both catalysts show important conversion levels at medium
pressure (15 bar) and that Ni2P catalyst is hardly affected by changes in the H2/DBF ratio,
meanwhile the CoP one requires greater amounts of hydrogen to be active.
The main cause for catalyst deactivation in HDO reactions is considered to be carbon deposition
on the surface20. Brønsted acid sites are responsible for coke formation as they act as proton doꢀ
nors that form carbocations, therefore coking increases with increasing acidity of the catalyst.
However the HDO reaction does require acidity and a balance should be obtained to have acid
sites capable to activate oxygen containing molecules but whose strength will not be too strong
to give rise to coke formation36. As stated before, the Ni2P sample, presents a higher amount of
acid sites of weak nature, probably related to its higher amount of PꢀOH groups on the surface.
The CoP sample presents an important concentration of very strong acid sites that could be reꢀ
sponsible for coke formation36, this could explain the low activity at low H2/DBF molar ratio
where the concentration of Cꢀspecies is expected to be higher. By increasing the H2 content in
the feed, hydrogen acts as a cleaner hydrogenating Cꢀspecies on the surface and the coke forꢀ
mation is minimized37.
Another reason for catalyst deactivation is the formation of water during reaction. The water
formed could interact with oxidized metal phosphide particles. It has been observed in a previous
study9 that Ni2P catalysts with an initial P/Ni atomic ratio of 2 or higher are more stable than
those with the P/Ni atomic ratio of 1. The oxidation resistance of Ni2P phase under water moleꢀ
cules has been explained in the literature38 by the dissociation of H2O on Ni2P surface (0 0 1)
which generates O atoms that interact with both P and Ni atoms, but preferably with P. This
leads to the formation of oxyꢀphosphide species with electron transfer from P to O. Hence P
shows a ligand effect and the P species may inhibit the interaction of Ni with O, improving the
oxidation resistance of Ni2P. In a previous work, we observed this fact on Ni2P catalysts, that is,
catalysts with greater P content were more resistant to deactivation due to H2O interacting preꢀ
sumably with Pꢀspecies9. By looking at XPS and IR data, it is clearly seen how the Ni2P sample

Catalysis Science & Technology
Page 18 of 36
View Article Online
DOI: 10.1039/C5CY00282F
18
has a higher Pꢀconcentration on the surface compared to CoP mainly in the form of PꢀOH. This
could be another reason why Ni2P shows the highest activity of the two catalysts.
The better results obtained with Ni2P vs CoP is not only related to the higher intrinsic activity of
Ni2P but also extra Pꢀspecies that play an important role in these systems. Characterization reꢀ
sults have evidenced the presence of much more Pꢀon the surface of Ni2P, mainly in the form of
PꢀOH groups and as a consequence this catalyst possesses better acidity and H2ꢀdesorption caꢀ
pacity, mainly at low temperature. Pꢀspecies provide PꢀOH groups that can act as acid site and as
a source of hydrogen species.
The role of P was also evaluated in CoP catalysts presenting different P/Co molar ratios and testꢀ
ed in the catalytic hydrodechlorination of chlorobenzene37. The most active catalysts were those
containing P/Co>2 and showing better acidity values and greater amount of Hꢀspecies from H2ꢀ
TPD measurements. In line with this, in a recent manuscript of Chen et al39, it has been evaluated
the promoting effect of water and oxygen pretreatment in the hydrodechlorination activity of
Ni2P/SiO2. H2O or O2 pretreatment increased the amount of PꢀOH groups on the surface and
improved the catalytic response of these systems. So the activity of Ni2P was explained in terms
of synergism between Ni atoms and PꢀOH groups.
4. Conclusions
The catalytic activity of Nickel and Cobalt phosphides on silica was tested in hydrodeoxygenaꢀ
tion of DBF as a model oxygenate compound under different experimental conditions. The reꢀ
sults obtained pointed out that the Ni2P catalyst is more active than the CoP one. Moreover it is
less affected by hydrogen pressure changes and presenting very high conversion values even at
low H2/DBF molar ratios. Both catalysts perform the HDO reaction through the hydrogenation
route: DBF→THDBF→HHDBF→2ꢀCHP→BCHꢀ3ꢀenꢀ2ꢀol→BCHol→BCH + CꢀPEꢀCH + Cꢀ
PEꢀMEꢀCH. Characterization results have indicated that the better activity of the Ni2P phase is
related to its greater intrinsic activity, and the greater presence of extraꢀP. Extra PꢀOH species
can prevent metal phase oxidation; act as weak acid sites that can interact with Oꢀspecies; but
they can also act as Hꢀdonors to perform reaction, meanwhile spillover species could act as surꢀ
face cleaners avoiding deactivation by coke formation.
The presented results seem to indicate that extra PꢀOH groups on Ni2P could be those govꢀ
erning the excellent results presented by the Ni2P phase. Ni2P sample should be considered a
good candidate to be employed in hydrotreating units to perform HDO reactions.

Page 19 of 36
Catalysis Science & Technology
View Article Online
DOI: 10.1039/C5CY00282F
19
Acknowledgements
The National Spanish Project CTQ2012ꢀ37925ꢀC03ꢀ03 of MINECO, FEDER funds and Junta de
Andalucía project P12ꢀRNMꢀ1565 are acknowledged for financial support.
References
1.
J. S. M. Khanna, and D. Zilberman, Handbook of Bioenergy Economics and Policy, Springer, New York,
NY, 2010.
2.
3.
A. Demirbas, Prog. Energy Combust. Sci., 2007, 33, 1ꢀ18.
Q. Bu, H. Lei, A. H. Zacher, L. Wang, S. Ren, J. Liang, Y. Wei, Y. Liu, J. Tang, Q. Zhang and R. Ruan,
Bioresour. Technol, 2012, 124, 470ꢀ477.
4.
5.
6.
Q. Lu, W.ꢀZ. Li and X.ꢀF. Zhu, Energy Convers. Manag, 2009, 50, 1376ꢀ1383.
T.ꢀS. Kim, S. Oh, J.ꢀY. Kim, I.ꢀG. Choi and J. W. Choi, Energy 2014, 68, 437ꢀ443.
M. L. Honkela, T. R. Viljava, A. Gutierrez and A. O. I. Krause, RSC Energy Environ. Ser., 2010, 2010,
288–306.
7.
8.
M. Krár, S. Kovács, D. Kalló and J. Hancsók, Bioresour. Technol., 2010, 101.
S. BoullosaꢀEiras, R. Lødeng, H. Bergem, M. Stöcker, L. Hannevold and E. A. Blekkan, Catal. Today,
2014, 223, 44ꢀ53.
9.
J. A. Cecilia, A. InfantesꢀMolina, E. RodríguezꢀCastellón, A. JiménezꢀLópez and S. T. Oyama, Applied
Catalysi B: Environmental, 2013, 136-137.
10.
11.
12.
13.
14.
Y. Lee and S. Oyama, J. Catal., 2006, 239, 376ꢀ389.
P. Clark, J. Catal., 2001, 200, 140ꢀ147.
P. Clark, J. Catal., 2002, 207, 256–265.
X. Wang, J. Catal., 2002, 208, 321–331.
J. A. Cecilia, A. InfantesꢀMolina, E. RodríguezꢀCastellón and A. JiménezꢀLópez, Journal of Catalysis,
2009, 263, 4ꢀ15.
15.
16.
J. A. Cecilia, A. InfantesꢀMolina, E. RodríguezꢀCastellón and A. JiménezꢀLópez, Appl. Catal. B Environ.,
2009, 92, 100ꢀ113.
P. Bui, J. A. Cecilia, S. T. Oyama, A. Takagaki, A. InfantesꢀMolina, H. Y. Zhao, D. Li, E. Rodríguezꢀ
Castellón and A. JiménezꢀLópez, Journal of Catalysis 2012, 294, 184ꢀ198.
S. T. Oyama, T. Gott, H. Zhao and Y.ꢀK. Lee, Catalysis Today, 2009, 143, 94ꢀ107.
H. Y. Zhao, D. Li, P. Bui and S. T. Oyama, Appl. Catal. A Gen., 2011, 391, 305ꢀ310.
A. Iino, A. Cho, A. Takagaki, R. Kikuchi and S. T. Oyama, J. Catal., 2014, 311, 17ꢀ27.
P. M. Mortensen, J.ꢀD. Grunwaldt, P. A. Jensen, K. G. Knudsen and A. D. Jensen, Appl. Catal. A Gen.,
2011, 407, 1ꢀ19.
17.
18.
19.
20.
21.
T. A. Le, H. V. Ly, J. Kim, S.ꢀS. Kim, J. H. Choi, H.ꢀC. Woo and M. R. Othman, Chem. Eng. J., 2014, 250,
157ꢀ163.
22.
23.
24.
25.
26.
27.
28.
Y. Wang, Y. Fang, T. He, H. Hu and J. Wu, Catal. Commun, 12, 1201–1205.
L. Wang, C. Li, S. Jin, W. Li and C. Liang, Catal. Letters, 2014, 144, 809ꢀ816.
L. Wang, M. Zhang, M. Zhang, G. Sha and C. Liang, Energy & Fuels, 2013, 27, 2209ꢀ2217.
S. Oyama, J. Catal., 2002, 210, 207ꢀ217.
T. Koranyi, Z. Vit, D. Poduval, R. Ryoo, H. Kim and E. Hensen, J. Catal, 2008, 253, 119–131.
J. N. Kuhn, N. Lakshminarayanan and U. S. Ozkan, J. Mol. Catal. A Chem, 2008, 282, 9–21.
D. ElicheꢀQuesada, J. MéridaꢀRobles, P. MairelesꢀTorres, E. RodríguezꢀCastellón and A. JiménezꢀLópez,
Langmuir, 2003, 19, 4985ꢀ4991.
29.
30.
J. A. Cecilia, A. InfantesꢀMolina, E. RodrıíguezꢀCastellón and A. JiménezꢀLópez, Journal of Physical
Chemistry C, 2009, 17032ꢀ17044.
T. Armaroli, G. Busca, C. Carlini, M. Giuttari, A. M. R. Galletti and G. Sbrana, J. Mol. Catal. A:Molecular
2000, 151, 233ꢀ243.
31.
32.
G. Busca, G. Ramis, V. Lorenzelli, P. F. Rossi, A. L. Ginestra and P. Patrono, Langmuir, 1989, 5, 911–916.
G. Garbarino, V. Sanchez Escribano, E. Finocchio and G. Busca, Applied Catalysis B: Environmental,
2012, 113-114, 281ꢀ289.
33.
34.
I. Rossetti, C. Biffi, C. L. Bianchi, V. Nichele, M. Signoretto, F. Menegazzo, E. Finocchio, G. Ramis and
A. D. Michele, Applied Catalysis B: Environmental, 2012, 117–118, 384ꢀ396.
F. Nozaki, T. Kitoh and T. Sodesawa, Journal of Catalysis, 1980, 62, 286ꢀ293.

Catalysis Science & Technology
Page 20 of 36
View Article Online
DOI: 10.1039/C5CY00282F
20
35.
36.
T. Guo, J. Chen and K. Li, Chinese Journal of Catalysis, 2012, 33, 1080ꢀ1085.
S. Echeandia, B. Pawelec, V. L. Barrio, P. L. Arias, J. F. Cambra, C. V. Loricera and J. L. G. Fierro, Fuel,
2014, 117, 1061–1073.
37.
J. A. Cecilia, A. InfantesꢀMolina, E. RodríguezꢀCastellón and A. JiménezꢀLópez, J. Hazard. Mater., 2013,
260, 167ꢀ175.
38.
39.
P. Liu, J. A. Rodriguez, Y. Takahashi and K. Nakamura, J. Catal., 2009, 262, 294ꢀ303.
J. Chen, T. Guo, K. Li and L. Sun, Catal. Sci. Technol., 2015.

Page 21 of 36
Catalysis Science & Technology
View Article Online
DOI: 10.1039/C5CY00282F
21
Scheme 1. Mechanism of HDO of DBF

Catalysis Science & Technology
Page 22 of 36
View Article Online
DOI: 10.1039/C5CY00282F
22
Table 1. TOF number for Ni2P/SiO2 and CoP/SiO2 catalysts at T = 275 ºC, T.o.S.=6 h under difꢀ
ferent experimental conditions
Ni2P/SiO2
CoP/SiO2
Contact time (s)
TOF x 102 (s-1) a
1.8
2.4
3.6
6.0
1.69
2.67
6.68
8.42
4.30
3.33
2.27
1.40
2.18
3.15
12.0
H2 Pressure
TOF x 102 (s-1)b
(Bar)
1
15
30
0.33
7.95
0.14
2.55
2.18
8.42
H2/DBF
TOF x 102 (s-1)c
34.3
69.2
139.0
23.7
8.42
6.09
3.01
2.18
2.93
a P= 30 Bar and H2/DBF=69
b Contact time= 6 s and H2/DBF=69
c P= 30 Bar and Contact time= 6 s

Page 23 of 36
Catalysis Science & Technology
View Article Online
DOI: 10.1039/C5CY00282F
23
Table 2. Textural Propertiesb and CO Chemisorptionc properties of the prepared samples
CO Chemisorptionc
b
b
SBET
Vp
(m2 gꢀ1)
(cm3 gꢀ1)
(
ꢀ
mol gꢀ1)
Support SiO2
Ni2P Pellets
CoP Pellets
218
0.52
109
0.57
32.9
36.7
126
0.67

Catalysis Science & Technology
Page 24 of 36
View Article Online
DOI: 10.1039/C5CY00282F
24
Table 3. Binding energy (eV) of core electrons and surface atomic ratios of fresh CoP and Ni2P
catalysts
Sample
Binding energy (eV)
Atomic ratio
Ni 2
p3/2/Co 2p3/2
P 2p3/2
Me/P Me/Si
3ꢀ
Co2+
780.7
Ni2+
CoP
PO4
CoP
CoP
Ni2P
1.65
0.41
0.04
0.01
134.1
129.0
777.3
(48.6)* (51.4)
*
3ꢀ
Ni2P
PO4
Ni2P
129.3
134.4
856.6
853.1
(25.5)* (74.5)*
*In parenthesis the percentage of each specie

Page 25 of 36
Catalysis Science & Technology
View Article Online
DOI: 10.1039/C5CY00282F
25
Table 4. Quantity of NH3 released (
ꢀ
mol gꢀ1) from NH3ꢀTPD experiments
ꢀmol NH3 gꢀ1
Temperature (ºC)
Catalyst
Total
481
100ꢀ200
155
200ꢀ300
202
300ꢀ400
98
400ꢀ500
500ꢀ800
Ni2P
CoP
20
8
6
211
57
71
35
41

Catalysis Science & Technology
Page 26 of 36
View Article Online
DOI: 10.1039/C5CY00282F
26
HDO conv. Ni2P
HDO conv. CoP
total conv. Ni2P
total conv. CoP
100
80
60
40
20
0
0
3
6
9
12
Contact time (s)
Figure 1. Total conversion and HDO conversion of DBF on Ni2P/SiO2 and CoP/SiO2 as a funcꢀ
tion of contact time. Experimental conditions: T = 275 ºC; P = 30 Bar; H2/DBF ratio 69; T.o.S=
6 h

Page 27 of 36
Catalysis Science & Technology
View Article Online
DOI: 10.1039/C5CY00282F
27
Ni2P
CoP
0
3
6
9
12
15
0
3
6
9
12
80
60
40
20
0
80
60
40
20
0
20
15
10
5
CꢀPEꢀCH
CꢀPEꢀMEꢀCH
BCH
CꢀPEꢀCH
CꢀPEꢀMEꢀCH
BCH
15
2ꢀCHP
BCHꢀ2ꢀenꢀ3ꢀol
2ꢀCHP
BCHꢀ2ꢀenꢀ3ꢀol
10
5
0
0
20
40
30
20
10
0
THDBF
HHDBF
THDBF
HHDBF
15
10
5
0
0
3
6
9
12
0
3
6
9
12
15
Contact time (s
)
Contact time (s)
Figure 2. Product selectivity on Ni2P/SiO2 and CoP/SiO2 as a function of contact time. Experiꢀ
mental conditions: T = 275 ºC; P = 30 Bar; H2/DBF molar ratio 69; T.o.S= 6 h

Catalysis Science & Technology
Page 28 of 36
View Article Online
DOI: 10.1039/C5CY00282F
28
HDO conv.Ni2P
HDO conv.CoP
total conv. Ni2P
total conv. CoP
100
80
60
40
20
0
0
10
20
30
Pressure (bar)
Figure 3. Total conversion and HDO conversion as a function of reaction pressure for Ni2P/SiO2
and CoP/SiO2 catalyst. Experimental conditions: T = 275 ºC; contact time 6 s; H2/DBF molar
ratio 69; T.o.S= 6 h

Page 29 of 36
Catalysis Science & Technology
View Article Online
DOI: 10.1039/C5CY00282F
29
CoP
Ni2P
BCH
BCH
C-PE-CH
C-PE-ME-CH
CHB
80
60
40
20
0
80
60
40
20
0
C-PE-CH
C-PE-ME-CH
CHB
0
5
10
15
20
25
30
0
5
10
15
20
25
30
Pressure (bar)
Pressure (bar)
Figure 4. Yield distribution of the products as a function of H2 pressure for Ni2P/SiO2 and
CoP/SiO2 catalyst. Experimental conditions: T = 275 ºC; contact time 6 s; H2/DBF molar ratio
69; T.o.S= 6 h