Hydrodeoxygenation of phenol using nickel phosphide catalysts. Study of the effect of the support

Article abstract of DOI:10.1016/j.cattod.2019.08.028

This work studied the performance of nickel phosphide phases supported on various supports (SiO₂, Al₂O₃, TiO₂, CeO₂ and CeZrO₂) for the hydrodeoxygenation of phenol in gas phase at 300 °C and 1 atm. The nature of the phosphide phase obtained by the temperature programmed reduction at 700 °C depended on the type of support. Only Ni₂P was formed on SiO₂, TiO₂, and CeZrO₂, whereas the Ni₁₂P₅ was the preferred phase on Al₂O₃. A mixture of both Ni₂P and Ni₁₂P₅ phases was obtained on CeO_2. Unsupported Ni₂P exhibited high selectivity to benzene (95%), indicating that the Ni₂P phase is responsible for the direct deoxygenation of phenol. Ni₁₂P₅ phase promoted the formation of cyclohexanone, cyclohexane and cyclohexene. However, the supported catalysts showed lower selectivity to benzene, even when the Ni₂P was the only phase present. The supports favored the formation of hydrogenation products via the tautomerization route. All catalysts only slightly deactivated with time on stream, which is likely due to the high activity of the phosphide phase.

Full text of DOI:10.1016/j.cattod.2019.08.028

Catalysis Today xxx (xxxx) xxx–xxx
Contents lists available at ScienceDirect
Catalysis Today
journal homepage: www.elsevier.com/locate/cattod
Hydrodeoxygenation of phenol using nickel phosphide catalysts. Study of
the effect of the support
Priscilla M. de Souza , Carlos V.M. Inocêncioa,b, Victoria I. Perezc,d
a
,
Raimundo Crisostomo Rabelo-Neto , Vinicius Ottonio O. Gonçalves , Gary Jacobsc,d,
Frédéric Richard , Victor Teixeira da Silva , Fabio B. Noronha
Federal University of Rio de Janeiro, Chemical Engineering Program – COPPE, 21941-972, Rio de Janeiro, RJ, Brazil
Catalysis Division, National Institute of Technology, Av. Venezuela 82, 20081-312, Rio de Janeiro, RJ, Brazil
University of Texas at San Antonio, Chemical Engineering Program, Department of Biomedical Engineering, 1 UTSA Circle, San Antonio, TX, 78249, USA
University of Texas at San Antonio, Department of Mechanical Engineering, 1 UTSA Circle, San Antonio, TX, 78249, USA
Universidade Federal do Rio de Janeiro, Instituto de Química, Núcleo de Desenvolvimento de Processos e Análises Químicas em Tempo Real - NQTR, Rua Hélio de Almeida
0, Cidade Universitária, CEP 21941-614, Rio de Janeiro, RJ, Brazil
b
e,f
g
a
b,h,⁎
a
b
c
d
e
4
f
Universidade Federal do Rio de Janeiro, Instituto de Química, Departamento de Físico-Química, Av. Athos da Silveira Ramos, n° 149, CEP 21949-900, Rio de Janeiro, RJ,
Brazil
Institut de Chimie des Milieux et Matériaux de Poitiers, UMR 7285 Université de Poitiers - CNRS, 4, rue Michel Brunet, BP633, 86022, Poitiers Cedex, France
g
h
Military Institute of Engineering, Chemical Engineering Department, Praça Gal. Tiburcio 80, Rio de Janeiro, 22290-270, Brazil
A R T I C L E I N F O
A B S T R A C T
Keywords:
2 2 3 2
This work studied the performance of nickel phosphide phases supported on various supports (SiO , Al O , TiO ,
Hydrodeoxygenation
Nickel phosphide: effect of support
Phenol
CeO2 and CeZrO ) for the hydrodeoxygenation of phenol in gas phase at 300 °C and 1 atm. The nature of the
phosphide phase obtained by the temperature programmed reduction at 700 °C depended on the type of support.
2
Only Ni
mixture of both Ni
benzene (95%), indicating that the Ni
phase promoted the formation of cyclohexanone, cyclohexane and cyclohexene. However, the supported cata-
lysts showed lower selectivity to benzene, even when the Ni P was the only phase present. The supports favored
2
P was formed on SiO
2
, TiO
phases was obtained on CeO2. Unsupported Ni
P phase is responsible for the direct deoxygenation of phenol. Ni12P
5
2
, and CeZrO
2
, whereas the Ni12
P
5
was the preferred phase on Al
2 3
O . A
2
P and Ni12
P
5
2
P exhibited high selectivity to
2
2
the formation of hydrogenation products via the tautomerization route. All catalysts only slightly deactivated
with time on stream, which is likely due to the high activity of the phosphide phase.
1. Introduction
The HDO process involves the deoxygenation of bio-oil molecules
using hydrogen and a catalyst. The studies about HDO of bio-oil model
The depletion of fossil fuels reserves and the increase in CO
2
compounds used different catalysts containing various active phases
such as sulfide [5,6], oxide [7,8] and metallic [9–11] phases. Currently,
phosphide catalysts have been extensively studied because they are
more economical than noble metal catalysts and less susceptible to
secondary reactions (e.g., hydrogenolysis of the CeC bond) than non-
noble metal catalysts [4,12,13]. Furthermore, it has been reported that
the phosphide phase exhibits higher acitivity to deoxygenation than
noble metals.
emissions have stimulated the search for cleaner and renewable solu-
tions. In this context, the use and valorization of lignocellulosic biomass
for the production of fuels and chemicals has great potential due to its
wide availability and carbon neutral emissions [1]. Among the biomass
conversion routes, fast pyrolysis is one of the most economically viable,
producing a liquid (so-called bio-oil) that can be used as fuel for
transportation and for industry [1–3]. However, pyrolytic bio-oil pre-
sents low calorific value, high corrosivity, high viscosity and thermal
and chemical instability due to the high concentration of oxygen
compounds in its composition [4]. The hydrodeoxygenation process
Regarding the comparison between metallic Ni and Ni phosphide,
Yang et al. [14] performed the HDO of methyl oleate under different
experimental conditions (pressure and temperature). According to the
(
HDO) can be used to upgrade bio-oil by decreasing its oxygen content
authors, Ni
2
P/SBA-15 catalyst favored deoxygenation and decarbox-
and making possible its utilization as a transportation fuel.
ylation reactions instead of decarboxylation and cracking over Ni/SBA-
⁎
Corresponding author at: Catalysis Division, National Institute of Technology, Av. Venezuela 82, 20081-312, Rio de Janeiro, RJ, Brazil.
E-mail address: fabio.bellot@int.gov.br (F.B. Noronha).
https://doi.org/10.1016/j.cattod.2019.08.028
Received 17 February 2019; Received in revised form 4 August 2019; Accepted 27 August 2019
0920-5861/ © 2019 Elsevier B.V. All rights reserved.
Please cite this article as: Priscilla M. de Souza, et al., Catalysis Today, https://doi.org/10.1016/j.cattod.2019.08.028

P.M. de Souza, et al.
Catalysis Today xxx (xxxx) xxx–xxx
1
5. Gonçalves et al. [15] studied the HDO of m-cresol under 4 MPa of
synthesized using the nickel nitrate hexahydrate (Ni(NO
Merck) and ammonium hydrogen phosphate (NH HPO , Vetec) salts.
First, both salts were solubilized separately and then, the solution of
(NH HPO was slowly dripped to the Ni(NO .6H O solution under
3 2 2
) .6H O,
total pressure and 340 °C over nickel and nickel phosphide catalysts
supported on silica and zirconia. The authors observed that Ni phos-
phide phase was much more active than the metallic Ni phase, re-
gardless of the support. Moreover, the deoxygenation activity also de-
pended on the type of support. The following order of deoxygenation
activity was observed: Ni
The selectivity to toluene was higher for Ni
catalysts than for metallic Ni supported on silica or zirconia.
4
)
2
4
4
)
2
4
3
)
2
2
constant stirring. Then, nitric acid was added to the mixture until
complete solubilization of the precipitate. The obtained solution was
placed in a silicone bath held at 105 °C under moderate agitation. After
4 h, a gel was formed, which was dried at 150 °C. The solid was calcined
2
P/ZrO
2
> Ni
2
P/SiO
2
> Ni/ZrO
2
> Ni/SiO
2
.
2
P/ZrO and Ni
2
2
P/SiO
2
−
1
at 500 °C for 6 h (10 °C min ). Before catalytic tests, this material was
Even though it is known that the nature of the support influences
product distribution and the extent of the HDO reaction [16,17], only
few studies have investigated the effect of the type of support on the
performance of nickel phosphide based catalysts, especially compared
to other active phases [15,18–21]. Gonçalves et al. [15] attributed the
higher activity to deoxygenated products for the HDO of m-cresol of
reduced ex situ using two heating steps with different rates: 30–350 °C
−1
−1
(10 °C min ), 350–650 °C (1 °C min ). H
with a flow rate of 1 mL min
2
was fed into the reactor
for each mg of phosphate. Then, the
reactor was cooled down to room temperature and fed with a 30 mL
−1
−1
min
2 2
of 0.5% O /N mixture during 16 h. For the catalytic experi-
−1
ments, in situ catalyst reactivation was done with a 30 mL min flow
−
1
Ni
2
P/ZrO
2
in comparison to Ni
2
P/SiO
Zr
2
to the oxophilic sites of the
cations. The same effect was
of pure H
For supported nickel phosphide catalysts, the following materials
were used as supports: SiO (Hi-Sil 915), Al (Puralox), TiO
(Aeroxide TiO P25), CeO and CeZrO . Silica (Hi-Sil 915, Aldrich),
alumina (Puralox, Sasol) and titania (Aeroxide TiO P25, Evonik
Industries) were commercial materials. Silica and alumina were pre-
2
at 300 °C for 1 h (10 °C min ).
3
+/ 4+
support represented by the Zr
observed for Pd catalysts in the phenol HDO reaction (300 °C, 1 atm), in
2
2
O
3
2
which the ones supported by oxophilic materials (ZrO
showed higher benzene selectivity [22,23].
2
, Nb
2
O
5
, TiO
2
)
2
2
2
2
The type of support affects the nature of phosphide phase formed
since the oxide may react with the phosphorous precursor salt leading
to the formation of phosphate species with the support during calci-
−1
−1
viously calcined under air flow (50 mL min at 5 °C min ) at dif-
2 3
ferent temperatures (800 °C: SiO , 1000 °C: Al O ) for 5 h, while titania
2
nation [24]. For example, Al
forming AlPO that inhibits Ni P synthesis and yields various Ni phases,
such as Ni12 [18]. Gonçalves et al. [25] prepared nickel phosphide
2
O
3
tends to interact strongly with P
was not pre-treated. Cerium oxide and mixed cerium-zirconium oxide
were synthesized by the precipitation method using cerium ammonium
nitrate, zirconium nitrate and aqueous ammonia [29]. For cerium
4
2
P
5
catalysts with different P/Ni ratios (0.8, 2 and 3) and observed that the
activity of the catalysts was dependent on the active phase formed on
the solid. The authors demonstrated that the highest P/Ni ratio pre-
oxide, a 4.0 mol/L solution of cerium ammonium nitrate ((NH
4 2
) Ce
(NO ) (Sigma-Aldrich) was slowly dripped with stirring in an aqueous
3 6
)
ammonia solution (7.85 mol/L). After the mixture was stirred for
30 min, the precipitate was filtered, washed with distilled water to
neutral pH, dried in a oven for 24 h and finally calcined at 500 °C (5 °C/
min) for 6 h. For the preparation of ceria-zirconia mixed oxide (Ce/Zr
sented higher activity because only a pure Ni
with P/Ni = 3, whereas the other samples containing either a mixture
of Ni P and Ni12 (NiP-2/ZrO ) or Ni12 (NiP-0.8/ZrO ) alone were
less active. The higher P/Ni ratio required to obtain the Ni P active
2
P phase was obtained
2
P
5
2
P
5
2
2
molar ratio = 0.5), the precursor salts of zirconia nitrate (ZrO(NO
3 2
) )
phase is clear evidence of the influence of the support during pre-
and ((NH Ce(NO were used. The salts were solubilized separately
4
)
2
3 6
)
paration of nickel phosphides. During the phosphide synthesis, phos-
phorous tends to simultaneously be released from the surface as PH by
3
temperature programmed reduction and/or to stay on the surface
forming POH groups, which imparts some acidity to the catalysts, as
in a suitable concentration and the precipitate was filtered and washed
with distilled water to neutral pH. The solid was then dried in an oven
for 24 h and finally calcined at 500 °C at a heating rate of 5 °C/min in
synthetic air stream (50 mL/min) for 6 h.
observed in NH
3
-TPD and IR-pyridine experiments [26,27].
The nickel phosphide phases were synthesized by the temperature
programmed reduction (TPR) method. This procedure can be divided
into three steps: preparation of the solution of the nickel phosphide
precursor, and impregnation of the precursor onto the support followed
by calcination and reduction. In the first step, appropriate amounts of
Deoxygenation activity and product distribution are also sig-
nificantly influenced by the nature of the nickel phosphide phase as
well as by the type of support. For example, Chen et al. [28] prepared
different nickel phosphide catalysts using SiO
SAPO-11 and HY as supports. The authors observed that the Ni
was obtained using SiO , CeO , TiO and SAPO-11 supports. The Ni
and Ni12 phases were formed using Al while Ni12 and Ni P were
2
, Al
2
O
3
, CeO
2
, TiO
P phase
2
,
2
nickel nitrate (Ni(NO
3 2 2
) .6H O) and dibasic ammonium phosphate
2
2
2
3
P
((NH HPO ) were solubilized in water and both solutions were mixed
4
)
2
4
P
5
2
O
3
P
5
2
and kept under constant stirring. The catalysts were prepared to obtain
10 wt.% of Ni using an excess of phosphorus (molar ratio P/Ni = 3).
Subsequently, a small amount of nitric acid (HNO ) was added into the
3
obtained on HY. From these results it is clear that the support displays
different interactions with the atoms of nickel and phosphorus giving
rise to the formation of different phases. The authors also showed that
resulting solution. At this stage, the precipitate formed from the mix-
ture of the two solutions was dissolved and the resulting solution was
used for the impregnation of the supports by the incipient wetness
impregnation method. After impregnation, the samples were dried in a
oven at 120 °C for 24 h and calcined under air flow (60 mL/min) at
500 °C (10 °C/min) for 6 h. The samples were ground and sieved in the
range of 250–315 μm. Finally, the reduction step was carried out in situ
before the catalytic tests at 700 °C (5 °C/min) for 1 h. The reduction
temperature before the catalytic tests was chosen based on the results of
in situ XRD and temperature programmed experiments. All reduced
catalysts were identified as NiP/support.
the activity for HDO methyl laurate followed the order: Ni
SiO > Ni P-Ni12 /Al > Ni P/TiO > Ni P/SAPO-11 > Ni P-
Ni12 /HY > Ni P/CeO
The aim of this work was to investigate the effect of the type of
support (SiO , Al , TiO , CeO and CeZrO ) on the structure of the
2
P/
2
3
P
5
2
O
3
2
2
2
2
P
5
2
2
.
2
2
O
3
2
2
2
nickel phosphide phases and on their catalytic properties for the HDO of
phenol reaction.
2. Experimental
2.1. Catalyst preparation
2.2. Catalyst characterization
The unsupported and supported nickel phosphides catalysts were
prepared by temperature programmed reduction (TPR) method. In
general, this methodology is based on two steps: (i) phosphate forma-
tion and; (ii) reduction of phosphate to phosphide.
The chemical composition was determined by inductively coupled
plasma optical emission spectrometry (ICP-OES) using an SPECTRO
ARCOS ICP-OES instrument. The specific area of the support and the
precursor of phosphide catalysts were measured by nitrogen adsorption
The unsupported NiP catalyst with a P/Ni molar ratio of 0.8 was
2

P.M. de Souza, et al.
Catalysis Today xxx (xxxx) xxx–xxx
at -196 °C on a Micromeritics ASAP 2000 analyzer.
2.3. Catalytic experiments
In situ XRD was carried out for unsupported nickel phosphide cat-
alyst in XRD1 light line of the Brazilian Synchrotron Light Laboratory
The hydrodeoxygenation of phenol was carried out in a gas phase
fixed bed quartz reactor, operating at atmospheric pressure of H and
(
1
LNLS). The sample to be analyzed was placed in a quartz capillary of
mm internal diameter between two pieces of wool of quartz. The
2
300 °C. First, the samples were diluted with inert material (mSiC/
mcatalyst = 3.0), except for bulk NiP. Prior to the reaction, the sample
was reduced in situ under pure hydrogen (60 mL/min) at 700 °C (5 °C/
min) for 1 h. In case of unsupported NiP, the catalyst was reduced ex
situ as mentioned previously in section 2.1. Its in situ reactivation oc-
capillary was added to a reaction cell connected to diffractometer of 3
circles of Newport (model N3050-P1). The temperature of analysis was
controlled with a robotic arm (Yaskawa) used to maintain a blower
(
FMB Oxford (modelo GSB1300) of hot air above the capillary. A Si
−1
−1
double crystal monochromator (111) positioned more than 10 m before
the experimental station selected the wavelength equal to 1.0332 Å.
curred under pure hydrogen (30 mL min ) at 300 °C (10 °C min ).
The reaction mixture was obtained by flowing H (at the same flow rate
2
The instrument was calibrated using Si (SRM 640d) and Al
40a), which are NIST standards (National Institute of Standards and
Technology). The measurements were performed with a flow rate of
2
O
3
(SRM
used in reduction/reactivation) through a saturator containing phenol,
which was kept at the specific temperature required to obtain the de-
6
sired H /phenol molar ratio (about 60). To avoid condensation, all lines
2
−1
10 mL min
of pure H
2
in the capillary and with two heating ramps.
were heated to 250 °C. The reaction products were analyzed by a gas
Initially, the samples were warmed from room temperature to 400 °C at
chromatograph (Shimadzu – GC-2014 model), using Innowax capillary
column and a flame-ionization detector (FID).
−
1
−1
10 °C min . Then, the samples were heated at 3 °C min from 400 °C
up to the final synthesis temperature, 650 °C. The XRD patterns were
obtained in the interval from 30° to 60°. The X-ray powder diffraction
The product phenol conversion and selectivity for each product
were calculated as follows:
(
XRD) patterns of supported nickel phosphides were obtained using a
Rigaku model DMax 2200 powder diffractometer with a copper tube (λ
1.54056 Å) with diffracted beam graphite monochromator and a
fed
(
mol
− molphenol )
phenol
Conversion (%) =
Selectivity(%) =
× 100
fed
=
mol
phenol
(1)
(2)
single channel scintillator detector. The diffraction data were obtained
in 0.02° steps in the range 15 < 2θ < 70°. An Anton Paar reaction
chamber (model XRK 900) was installed on the sample stage. In such
configuration, the gases flow through the catalyst bed. The sample was
molofproductproduced
molofphenolconsumed
× 100
reduced under pure H
2
flow (60 mL/min) from 25 to 750 °C at a heating
was
reduced ex situ at 700 °C (5 °C min ) and passivated under 0.5% O
mixture for overnight before the XRD measurements.
Temperature programmed reduction (TPR) was performed in a
multipurpose unit coupled to a Pfeiffer Vacuum mass spectrometer
MS) model QME 200. Prior to reduction, the sample (0.1 g) was pre-
treated under He (20 mL/min) at 200 °C (10 °C/min) for 30 min and
then cooled to 30 °C. The samples were reduced under pure H
100 mL/min H ) at a heating rate of 5 °C/min up to 1000 °C. For the
unsupported NiP catalyst, pretreatment was performed by He (50 mL
3
. Results and discussion
.1. Catalyst characterization
Table 1 summarizes the chemical composition and the surface area
of oxide precursors of the phosphide catalysts. Considering the Ni and P
content measured by ICP, the calcined samples showed a P/Ni molar
ratio close to the expected value (3.0). The specific surface areas of
precursors of nickel phosphides catalysts were lower than those of the
respective supports. This result can be attributed to the excess of
phosphorus used during the synthesis of the catalysts, which can cause
the blocking of the pores of the support. It was used an excess of
phosphorus in the synthesis (P/Ni = 3) of supported catalysts to ensure
rate of 5 °C/min, holding at each temperature for 0.5 h. NiP/TiO
2
−
1
2
/
3
N
2
(
2
(
2
−
1
−1
min ) at 450 °C (10 °C min ) for 30 min. After that, the catalyst was
cooled down to 30 °C. Finally, He was replaced by H , which had its
O signal (m/Z = 18) started
to be measured at the spectrometer. During reduction measurements,
2
−1
flow rate adjusted to 100 mL min , and H
2
the Ni
2
P phase formation and to prevent the presence of other phases
and Ni P [30,31].
−1
such as Ni12
P
5
3
temperature was increased linearly from 30 °C to 1000 °C at 1 °C min
.
In order to investigate the phase transitions of unsupported and
supported nickel phosphide catalysts during phosphate reduction, in
situ XRD was carried out. The diffractograms of unsupported NiP ob-
tained during reduction are shown in Fig. 1.
No lines were detected on the diffractogram of the calcined sample,
indicating that nickel phosphate is amorphous. The diffractograms
The nature of acid sites was determined by DRIFTS of adsorbed
pyridine using a Perkin Elmer model Spectrum 100 instrument with a
DTGS-TEC detector and a Thermo Spectra-Tech reaction chamber with
ZnSe windows. The samples were reduced at 300 °C for a period of 1 h,
and the temperature was decreased to 150 °C while He was flowed to
record a spectrum for use as a background. Pyridine adsorption in-
volved bubbling He (30 mL/min) through a saturator for 30 min, fol-
lowed by a He purge. The scan resolution was 4 cm , and 512 scans
were taken.
DRIFTS of adsorbed cyclohexanone was carried out using a Nicolet
S10 instrument with a DTGS-TEC detector and a Harrick Scientific
praying mantis reaction chamber with ZnSe windows. The samples
−1
Table 1
Chemical composition and surface area of oxide precursors of the phosphide
catalysts.
Samples
Ni
wt.
%
P wt.% P/Ni
molar
Specific surface area
XRD Ni
phases
d
xrd
2
−1
(m g
)
ratio
underwent the same H
2
reduction treatment previously described.
Support
Catalyst
(nm)
108a
27b
Then, the catalyst was cooled to 50 °C while He was flowed through a
bubbler containing cyclohexanone at 60 mL/ min, and spectra were
recorded at 100, 150, 300, and 200 °C. Peak fitting was performed on
the band corresponding to adsorbed cyclohexanone by using Gaussian
peaks and the nonlinear generalized reduced gradient (GRG) algorithm
for the spectra obtained at 50 °C.
Raman was used to characterize the nature of carbon formed on the
catalysts after 24 h of time on stream (TOS). The Raman spectra were
collected with a Horiba Jobin-Yvon LabRAM HR system equipped with
He − Ne laser (λ =632 nm), a CCD detector and an Olympus BX-41
microscope.
NiP
–
–
–
–
–
Ni
2
Ni
5
Ni
2
Ni
2
P,
P
P
P
4
NiP/SiO
2
6.2
7.8
7.5
8.0
9.3
2.8
2.6
2.7
2.6
208
48
207
49
53
19
78
5
b
NiP/TiO
NiP/Al O
2
10.7
10.7
11.0
35
19
31
c
2
3
Ni12P5
Ni P,
Ni12
Ni
b
NiP/CeO
2
2
P
5
NiP/CeZrO
a
2
7.5
9.9
2.5
112
–
2
P
39b
Crystallite size of Ni
Crystallite size of Ni
2
P (111) analyze carried out at LNLS.
P (111).
Crystallite size of Ni12P5 (312).
b
c
2
3

P.M. de Souza, et al.
Catalysis Today xxx (xxxx) xxx–xxx
transformation of initial amorphous nickel phosphate. In addition to
Ni
JCPDS 03-0953) appeared at 600 °C. After heating to 700 °C, only the
Ni P phase was detected. The diffractogram of reduced and passivated
NiP/TiO sample (Fig. 2B) displayed the lines characteristics of titania
support (anatase and rutile) and Ni P phase, indicating that the treat-
ment at 700 °C produced the desired nickel phosphide phase, main-
taining the crystallinity of the support. For NiP/Al catalyst, the
presence of the Ni12 phase (lines at 2θ = 38.3, 46.9 and 48.9°, JCPDS
2–1190) was observed at around 600 °C. Increasing the reduction
temperature to 750 °C increased the intensity of the lines corresponding
2 2 7 3 2 2
P O or Ni(PO ) phases, new lines corresponding to Ni P phase
(
2
2
2
2 3
O
P
5
2
to Ni12
NiP/CeO
P
5
phase but no other nickel phosphide phase was detected. For
sample, the reduction changed the crystal structure of the
2
supports. The lines characteristics of monoclinic CePO
199) phase start to appear at 550 °C and their intensities continuously
increased up to 750 °C. At this temperature, the diffraction lines char-
acteristic of Ce 11 (JCPD S 32-0196) phase were also observed, in-
dicating the partial reduction of ceria. Concerning the nickel phosphide
4
(JCPDS 32-
0
6
O
phase, Ni
2
P (JCPDS 03-0953) and Ni12
P
5
(JCPDS 22–1190) were ob-
at 650 °C. These diffractions
served on the diffractogram of NiP/CeO
2
lines became more evident at 700 °C. For NiP/CeZrO
duction step also changed the crystal structure of the support. At 700 °C,
the diffraction lines of CePO (JCPDS 32-0199), Ce75Zr25 (JCPDS 28-
271) and ZrO (JCPDS 37-0031) were identified, suggesting that the
reduction led to a phase segregation. Furthermore, NiP/CeZrO also
exhibited the line typical of Ni P phase, but the presence of Ni12P
5
2
sample, the re-
4
O
2
0
2
2
2
cannot be ruled out because of the overlapping of its main line with the
support.
2
Fig. 1. X-ray diffractograms of the bulk Ni P during its activation at specific
temperatures. •Ni
2
P, ♦ Ni
5
4
P , ♣ Ni12
P
5
.
Rodrigues et al. [32] used synchrotron-based time-resolved X-ray
diffraction to study in situ the crystalline phases present during the
preparation of bulk and silica-supported Ni
2
P by reduction of oxidic
/95% He.
remain unchanged up to 473 °C. After heating at 510 °C, it is observed
the appearance of the diffraction lines characteristic of Ni12 (JCPDS
2–1190) and Ni P (JCPDS 03-0953) phases. Further increases in
temperature to 533 °C decreased the intensities of the lines corre-
sponding to Ni12 phase. Meanwhile, the diffraction lines of the Ni
phase continued to grow, indicating the transformation of the Ni12P5
phase into Ni P. The diffraction lines corresponding to Ni12 were no
longer detected at 548 °C. New diffraction lines typical of the Ni
phase appeared at 593 °C and their intensities increased up to 631 °C.
Increasing the temperature to 650 °C significantly decreased the in-
2
tensities of these lines and the diffractograms exhibited mainly the Ni P
phase.
precursors (NH NiPO ‚nH O) under a gas mixture of 5% H
4
4
2
2
P
5
The diffractograms of bulk NiP exhibit only the lines characteristic of Ni
phosphate during heating from 25 to 400 °C. After 400 °C, the sample
becomes amorphous and all of the diffraction lines disappear. The
diffraction lines corresponding to Ni P phase appears in the tempera-
2
ture range of 650 to 800 °C. For the silica supported NiP sample, in-
itially is observed the reduction of NiO to Ni0 around 400 °C and then,
2
2
P
5
2
P
2
P
5
5 4
P
Ni12
well with our in situ diffractograms of NiP/SiO
Berhault et al. [33] investigated the reduction of the ammonium
nickel phosphate NiNH PO · H O precursor into nickel phosphide
Ni P) by using a combination of magnetic susceptibility and in situ X-
P
5
is formed. Ni
2
P begins to form above 600 °C, which agrees very
2
catalyst.
4
4
2
(
2
The XRD patterns of the supported Ni phosphate samples are shown
in Fig. 1S (Supplementary information). Comparing the intensities of
the lines of the support and supported Ni phosphate samples, it is ob-
served a decrease in the crystallinity since the calcined samples showed
ray diffraction and X-ray absorption spectroscopy (XAS) techniques.
The in situ XRD experiments revealed the presence of three different
zones: (1) from room temperature to 250 °C, NiNH PO ·H O was still
4 4 2
detected; (2) from 300 to 500 °C, only amorphous phases were ob-
served; (3) above 500 °C, a crystallization process occurred, leading to
lower line intensities than those of the supports. NiP/SiO
broad peak at 22.7° corresponding to an amorphous phase, whereas
NiP/TiO revealed the lines characteristic of both anatase (JCPDS
1–1272) and rutile (JCPDS 21–1276) titania phases. NiP/CeO and
NiP/CeZrO showed the diffraction pattern of the cubic fluorite struc-
ture (JCPDS 34-0394) typical of ceria. For NiP/Al , in addition to the
diffraction lines of the support (cubic Al , JCPDS 10-0425), a line at
1.8° corresponding to the formation of AlPO by migration of phos-
2
exhibited a
Ni
showed that the amorphous region corresponds to the nickel pyr-
ophosphate phase (Ni ). In our study, the formation of Ni
and Ni P phases was observed in the same temperature range of the
2
P. in situ XAS study and magnetic susceptibility measurements
2
2
2
2
P
2
O
7
2 2 7
P O
2
2
2 3
O
Berhault et al. [33] work.
2 3
O
The high phosphorous loading used in the synthesis of the catalysts
2
4
favors the formation of the Ni
Depending on the support, the migration of phosphorus to the support
can give rise to a deficit in phosphorus, insufficient to produce Ni
phase and then, leading to the Ni12 formation [24,34–36]. Table 1
2
P phase with high crystallinity.
phorus to the support structure was also observed. The absence of the
lines characteristic of nickel phosphate for all catalysts indicates that
this phase is amorphous, as also noted for unsupported NiP catalyst.
Fig. 2 shows the diffractograms of the supported nickel phosphide
catalysts. All the samples were reduced in situ at different temperatures,
2
P
P
5
reports the nickel phosphide phases detected by XRD and the crystallite
size of phosphide phases for all catalysts. Similar and large crystallite
with the exception of NiP/TiO
00 °C and passivated before the XRD analysis. For NiP/SiO
Fig. 2A), a diffraction line around 2θ = 30.2° was observed when the
sample was reduced from 450 °C to 650 °C. This diffraction line could be
attributed to the monoclinic Ni phase (JCPDS 33-0950) or to the
monoclinic Ni(PO phase (JCPDS 28-0708), produced by the
2
, which was previously reduced at
sizes were observed for Ni
2
P phase for all catalysts (NiP/SiO
2
(27 nm),
7
2
catalyst
NiP/TiO (35 nm), NiP/CeO
2
2
(31 nm) and NiP/CeZrO (39 nm).
2
(
Fig. 3 shows the water formation profiles of the calcined un-
supported and supported phosphate precursors. This analysis was per-
formed in order to determine the intermediate and final temperature of
synthesis of the phosphides during the reduction step. The TPR of
2 2 7
P O
3 2
)
4

P.M. de Souza, et al.
Catalysis Today xxx (xxxx) xxx–xxx
Fig. 2. X-ray diffractograms of the supported Ni
2
P catalysts: (A) NiP/SiO
2
2 2 3 2 2 2 2 7 3 2 2
, (B) NiP/TiO , (C) NiP/Al O , (D) NiP/CeO , (E)NiP/CeZrO . ∇ Ni P O or Ni(PO ) , •Ni P,
♣
5 4 6 11
Ni12P , ♠ CePO , ♦ Ce O .
5

P.M. de Souza, et al.
Catalysis Today xxx (xxxx) xxx–xxx
region in the water formation profile as observed for unsupported
phosphide catalyst, indicating that the transformation of nickel phos-
5
86
phate to Ni
60 and 700 °C confirm the presence of nickel oxy-phosphates since free
NiO species tend to reduce at lower temperatures. The reduction of
NiP/Al , NiP/CeO and NiP/CeZrO precursors catalysts start at
lower temperature than those for NiP/SiO and NiP/TiO , which sug-
gests that the nickel species are more easily reduced on NiP/Al , NiP/
CeO and NiP/CeZrO catalysts or that the reducibility characteristics
of CeO and CeZrO promote the reduction of phosphate.
2
P also occurs in two steps. Furthermore, the peaks between
4
(
F)
E)
2
O
3
2
2
2
2
2 3
O
(
2
2
2
2
The DRIFTS of adsorbed pyridine was carried out to get information
about the nature of acid sites on the catalysts. The spectra of adsorbed
pyridine at 150 °C are shown in Fig. S2. Only bands at around 1449 and
(D) 5x
6
79
−1
1489 cm corresponding to pyridine adsorbed on Lewis acid site were
observed on the spectra except for bulk NiP and NiP/SiO
not exbibited any detectable band.
2
, which did
(C)
DRIFTS experiments of adsorbed cyclohexanone followed by deso-
rption at diff ;erent temperatures was performed in order to to probe
the strength of the interaction of the carbonyl oxygen with the active
sites. The spectra obtained at 50 °C are presented because the S/N ratio
was unacceptable at higher temperatures. Spectra in the region between
(
B)
578
5
15
−
1
1600 and 1750 cm
are shown in Fig. S3. The spectra of cyclohex-
4
68
anone is quite complex and it can be decomposed in three regions: high
−
1
−1
−1
(
> 1695 cm ), medium (1660–1695 cm ) and low (< 1660 cm
)
(
A)
wavenumbers. NiP/TiO and NiP/CeZrO catalysts showed a higher
2
2
fraction of this band located in the medium/low wavenumber region
whereas the other catalysts exhibited mainly a band positioned at high
wavenumber region. The shift of the cyclohexanone band to lower
wavenumbers was attributed to the stronger interaction between the
oxygen from the carbonyl group and metal cations of the support [23].
1
00 200 300 400 500 600 700 800 900
Temperature (ºC)
Fig. 3. Water formation profiles during TPR of unsupported and supported
nickel phosphide precursors: (A) NiP, (B) NiP/SiO , (C) NiP/TiO , (D) NiP/
Al , (E) NiP/CeO , (F)NiP/CeZrO
2
2
3.2. HDO of phenol
2
O
3
2
2
.
The HDO of phenol was carried out over the bulk NiP catalyst to
unsupported NiP catalyst showed three peaks of formation of water at
68, 515 and 578 °C. The first peak corresponds to reduction of the
surface nickel phosphates due to the slow diffusion of hydrogen [37].
The second and the third peaks of the TPR profile correspond to the
shed light on the catalytic properties of this phosphide phase without
the interference of the support. The choice of Ni P among other nickel
phosphide phases (i.e. Ni12 , Ni P and Ni ) is justified by the the fact
4
2
P
5
3
5 4
P
that it is known that this phase exhibits the highest intrinsic activity for
HDO reactions [30,40].
The phenol conversion and product distribution as a function of W/
F for HDO of phenol at 300 °C over all supported catalysts are shown in
Fig. S4. Benzene and cyclohexane were the main products formed with
formation of Ni12
XRD analysis, which confirmed the presence of a mixture of Ni12
Ni P phases between 495 and 540 °C, and the disappearance of the
Ni12 phase and the predominance of the Ni P phase at 548 °C. Liu
et al. [38] demonstrated by Raman spectroscopy that the calcined Ni
precursors containing P/Ni molar ratio = 0.8 or 1.0 displayed the
P
5
and Ni
2
P species. This is in accordance with in situ
P
5
and
2
P
5
2
2
P
minor formation of cyclohexene for NiP/SiO
and NiP/CeZrO catalysts. Significant production of cyclohexene and
cyclohexanol were also observed for NiP/CeO catalyst.
2 2 2 3
, NiP/TiO , NiP/Al O ,
4
−
3-
2
[
P
2
O
7
]
and [PO
of nickel oxide by visible/ UV spectroscopy technical. Therefore, it can
be assumed that the calcined precursor is composed of a Ni and
Ni (PO mixture (amorphous species). The beginning of the reduction
4
] groups and the absence of Ni-O-Ni bonds typical
2
Table 2 reports the reaction rate and product distribution for HDO
of phenol at 300 °C and low phenol conversion over unsupported and
supported nickel phosphide catalysts. The unsupported Ni P phase ex-
2
2 2 7
P O
3
4 2
)
at 420 °C corresponds to the reduction of these phosphate species
through a mechanism known as shrinking-core. After the reduction of
hibited the activity for deoxygenation aproximately 5 to 22-fold lower
than the supported materials, which is likely due to the very low surface
surface species, the phosphate bulk is reduced forming Ni12
respectively. Although no other water formation peak was observed,
the transition from the Ni P phase to a Ni phase was detected at
93 °C (in situ XRD).
Stinner et al. [39] showed that by varying the reduction parameters
P
5
and Ni
2
P,
2 −1
area (lower than 5 m g ). The HDO reaction rate was approximately
the same as the direct deoxygenation rate because this catalyst pro-
duced mainly benzene (95%). The other minor products were cyclo-
hexanone, cyclohexane, and cyclohexene.
2
5 4
P
5
It is well established that bulk Ni
thorhombic crystalline structure that can present two types of termi-
nations with different kinds of active sites. The Ni presents the so-
called Ni(1) site and Ni P has Ni(2) sites, the latter featuring a lower
2
P catalyst is comprised of an or-
such as gas flow rate and heating rate, the type of nickel phosphide
phase can change. Other factors like loading of P and type of supports
can also influence the synthesis temperature of phosphide catalysts. In
general, higher P/Ni molar ratio shift the initial and final reduction
temperatures to higher values. Acidic supports interact strongly with
3 2
P
3
coordination number. Several studies about hydrotreatment have ex-
plained the product distribution obtained by the different interaction
that can occur with these active sites. For hydrodesulfurization (HDS)
of 4,6-dimethyldibenzothiophene, Oyama et al. [41,42] proposed that
Ni(1) was involved in the DDS route leading to dimethylbiphenyl,
whereas Ni(2) is likely involved in the hydrogenation route toward
methylcyclohexyltoluenes. Gonçalves et al. [25] suggested that Ni(1)
2
phosphorus and inhibits the formation of the Ni P phase.
In the present work, the temperature programmed reduction profile
for supported nickel phosphate precursors give rise to different reduc-
tion profiles depending on the type of support. However, the maximum
reduction temperature appeared within a similar temperature range
(
578–586 °C) and a shoulder was observed in the low temperature
3 2
sites present on the Ni P termination interact with the oxygen atom of
6

P.M. de Souza, et al.
Catalysis Today xxx (xxxx) xxx–xxx
Table 2
Product distribution from the HDO of phenol at 300 °C over unsupported and supported phosphide nickel catalysts.
Catalyst
W/F (h)
XPHE (%)
HYD ratea (mmol gcat−1 min-1
)
HDO rateb (mmol gcat−1 min-1
)
Selectivity (mol%)
Others
NiP
NiP/SiO
NiP/TiO
NiP/Al
NiP/CeO
0.84
0.03
0.11
0.13
0.03
0.03
4.5
5.4
5.5
3.8
5.2
2.2
0.00
0.14
0.02
0.01
0.10
0.02
0.01 (0.01)c
0.22 (0.18)
0.08 (0.05)
0.05 (0.02)
0.19 (0.14)
0.11 (0.08)
94.6
50.8
48.3
35.8
41.3
58.6
2.8
–
1.2
3.2
16.3
32.2
8.2
1.4
8.1
17.9
20.4
8.1
0
2
30.5
16.2
9.7
32.7
17.2
7.5
0.0
1.1
9.3
1.7
0.0
1.3
0.0
0.5
1.0
2
2 3
O
2
NiP/CeZrO
2
13.6
7.9
a
Reaction rate for hydrogenation (cyclohexanone, cyclohexanol).
Reaction rate for hydrodeoxygenation (benzene, cyclohexane, cyclohexene).
Value in parentheses corresponds to the reaction rate for direct deoxygenation to benzene.
b
c
cresol molecules in the same way that it interacts with CeSH, facil-
itating hydrogenolysis reactions and producing toluene. Furthermore, it
was proposed that Ni(2) sites would promote hydrogenation reactions,
promoting flat adsorption of ring-containing compound such as phe-
nolic groups [15,27,30].
such as ceria, ceria-zirconia and alumina. For NiP/CeO
new phases were created such as CePO and AlPO , respectively. In the
case of NiP/CeZrO , in addition to CePO , it was also observed a phase
segregation, producing Ce75Zr25 and ZrO phases.
Regarding product distribution, benzene was the main product
formed over NiP/SiO , NiP/TiO and NiP/CeZrO , whereas cyclohex-
anone was mainly obtained over NiP/CeO and NiP/SiO . Significant
formation of cyclohexane and cyclohexene was observed for all sup-
ported catalysts in comparison to bulk Ni P. NiP/Al exhibited the
2 2 3
and NiP/Al O ,
4
4
2
4
O
2
2
In our work, it seems that bulk Ni
Indeed, Density Functional Theory (DFT) calculations have demon-
strated that Ni termination is more stable than Ni P [43,44]. How-
ever, Hernandez et al. [45] performed dynamical LEED analysis for
Ni P (0 0 0 1) and observed that Ni P termination (Ni(2) sites) accounts
for about 80% of the total surface. The authors pointed out that the P
adsorbs on the three-fold sites present in Ni structure, which further
stabilize the Ni structure by filling the dangling bonds and do not
expose such termination. These findings are somehow conflicting and
the determination of Ni and Ni P proportions is not an easy task,
especially regarding supported phases.
2
P contains mainly Ni(1) sites.
2
2
2
2
2
P
3 2
3
2
2 3
O
2
3
highest selectivities to cyclohexane and cyclohexene. These results
suggest that the support significantly affects the product distribution
3
P
2
regardless of the nature of phosphide phase. For instance, NiP/SiO
2
,
P
3 2
NiP/TiO and NiP/CeZrO catalysts contains only the Ni P phase but
2
2
2
the selectivity to benzene for the bulk NiP catalyst was almost double
that of the supported catalysts. The higher selectivity to cyclohexane
and cyclohexene observed for the supported phosphide catalysts in
comparison to bulk phosphide catalyst is likely due to the presence of
Lewis acid sites on the supports, as shown by the DRIFTS of adsorbed
pyridine experiments.
3
P
2
3
3.3. HDO of phenol of supported NiP catalysts
Recently, we studied the effect of support on the reaction me-
chanism for HDO of phenol over supported metal catalysts. This me-
chanism is based on the tautomerization of phenol to a keto inter-
mediate that can be hydrogenated to cyclohexanone and then
cyclohexanol, which may be dehydrated to cyclohexene over supports
with enough acidity, followed by hydrogenation to cyclohexane. The
second reaction route involves the hydrogenation of the carbonyl group
of the keto tautomer, leading to the production of benzene after de-
hydration [23]. According to this reaction pathway, oxophilic supports
The influence of the support on the performance of nickel phosphide
phases was evaluated for the HDO of phenol at 300 °C under atmo-
spheric pressure. The hydrogenation rate was quite low for all the
supported phosphide catalysts, except for NiP/SiO
reaction rate for HDO followed the order: NiP/CeO
SiO > NiP/CeZrO > NiP/TiO > NiP/Al
tion rate for direct deoxygenation is the same: NiP/CeO
SiO > NiP/CeZrO > NiP/TiO > NiP/Al NiP/Al
2
and NiP/CeO
2
. The
NiP/
2
≈
2
2
2
2 3
O . The order for the reac-
2
≈ NiP/
2
2
2
2
O
3
.
2
O
3
exhibited
very low reaction rates, which could be attributed to the presence of
Ni12 phase that is less active than Ni P. For the hydrogenation rate,
the following order was observed: NiP/CeO ≈ NiP/SiO > > NiP/
CeZrO ≈ NiP/TiO ≈ NiP/Al
In comparison to the previous results obtained by our group using
Pd supported on SiO , Al , TiO , ZrO , CeO and CeZrO , the var-
such as TiO
deoxygenated products. In our work, the higher selectivity to cyclo-
hexanone and cyclohexanol over NiP/SiO and NiP/CeO suggestes that
2 2 5 2
, Nb O , and ZrO promote C]O hydrogenation leading to
P
5
2
2
2
2
2
2
2
2 3
O .
the hydrogenation of the ring pathway is favored over these supports.
This result agrees very well with the DRIFTS of adsorbed cyclohexanone
experiments that revealed a weaker interaction between the oxygen of
the carbonyl group and the cation of the support.
Furthermore, it has been suggested that more oxophilic metals such
2
as Ru as well as supports like TiO strongly interact with the oxygen
atom, reducing the energy barrier for the direct cleavage of the CeO
bond of the aromatic ring (Direct deoxygenation – DDO), leading to the
formation of benzene as well as hydrogenolysis products. A similar
reaction pathway has been proposed for phosphide catalysts. The
2
2
O
3
2
2
2
2
iation of the HDO rate for supported NiP catalysts was approximately
the same as that observed for supported Pd catalysts [29]. For Pd
supported materials, the support acts as a center for phenol adsorption,
whereas Pd activates hydrogen. In that case, the HDO activity followed
the order−1of support oxophilicity, with Pd/TiO
2
(0.72 mmol
(0.59 mmol gcat min ) displaying the
highest values. It is noteworthy to mention that the lowest HDO activity
−1
−1
−1
g
cat
min ) and Pd/ZrO
2
−
1
−1
was found for Pd/SiO
lower than that for Pd/TiO
catalysts, the HDO activity varied from 0.05 mmol gcat min
2
(0.09 mmol gcat min ), which was 8-fold
stronger adsorption of m-cresol over Ni
cleavage of CeO bond and the production of toluene in comparison to
metallic Ni particles [15]. In our work, NiP/SiO , NiP/TiO and NiP/
CeZrO catalysts contained only the Ni P phase and resulted in the
2
P phase favored the direct
2
. From Table 2, regarding NiP supported
−
1
−1
(NiP/
). The higher HDO reac-
suggests that the intrinsic activity
P phase is higher than that of Ni12
The addition of phosphorus changed the structure of some supports
−1
−1
2
2
Al
tion rate of NiP/SiO
of Ni
2
O
3
) to 0.22 mmol gcat min
(NiP/SiO
2
2
2
2
than NiP/Al O
2 3
highest selectivity to benzene, suggesting that this phosphide phase
promotes the direct deoxygenation pathway. The lowest formation of
2
5
P .
7

P.M. de Souza, et al.
Catalysis Today xxx (xxxx) xxx–xxx
2 2 2 3 2 2
Fig. 4. Conversion of phenol and selectivity to products as a function of TOS for: (A) NiP/SiO , (B) NiP/TiO , (C) NiP/Al O , (D) NiP/CeO , (E)NiP/CeZrO .
8

P.M. de Souza, et al.
Catalysis Today xxx (xxxx) xxx–xxx
benzene on NiP/Al
phase. Therefore, the nature of the phosphide phase affected the pro-
duct distribution. In addition, the use of zirconia as a support enhanced
2
O
3
catalyst is likely due to the presence of Ni12
P
5
the P/Ni molar ratio. The catalyst containing the higher P/Ni ratio
remained quite stable during the reaction, whereas the one with P/
Ni = 1 underwent gradual deactivation. According to the authors, the
the deoxygenation activity comparing to NiP/SiO
attributed to the oxophilic sites of the support (Zr
strongly adsorbed phenol and favored the cleavage of C]O bond. The
same result was observed in our work for TiO and CeZrO supported
catalysts. In the present work, the diffractograms detected the presence
of a ZrO segregated phase. The DRIFTS of adsorbed cyclohexanone
experiments showed a stronger interaction between the oxygen of the
2
[15]. This result was
excess of phosphorus for the P-rich catalysts protected the Ni
against oxidation with the water formed in the reaction.
2
P phase
3
+
4+
and Zr ) that
Comparing the deactivation of supported Pd catalysts [29] with that
observed for the phosphide catalysts supported on the same oxides, it is
clear that the deactivation rate is much lower on the phosphide cata-
lysts. This is likely due to the high activity of phosphide catalysts that
efficiently turnover the adsorbed phenol molecule, preventing the re-
tention of intermediate species; therefore, the catalyst only slightly
deactivated. Furthermore, phosphorus atoms located on the surface of
the nickel phosphide phase could inhibit carbon formation by the di-
lution effect on Ni atoms.
2
2
2
4
+
4+
carbonyl group and the Ti
and Zr
cations of the support.
P phase is
Then, the results of this work demonstrated that the Ni
2
responsible for the direct deoxygenation of phenol. Furthermore, the
support affected the nature of the phosphide phase and product dis-
tribution. For alumina and ceria supported catalysts, the Ni12
5
P phase
was also observed and promoted the hydrogenation reaction pathway.
4. Conclusion
3.4. Stability of supported NiP catalysts for HDO of phenol
This work investigated the effect of the type of support on the
The stability of all catalysts was compared at similar initial phenol
performance of supported nickel phosphide catalysts for HDO of
phenol.
Unsuppported NiP catalyst exhibited high selectivity to benzene
(95%), which demonstrate that the direct deoxygenation of phenol is
conversion (between 30–40%), as shown in Fig. 4. All catalysts only
slightly deactivated during time on stream (TOS). For instance, the
phenol conversion decreased from 42 to 32% after 20 h of TOS over
NiP/SiO
2
. NiP/CeZrO
2
exhibited the highest deactivation rate at the
2
promoted by the Ni P phase. The strong adsorption of the phenol on the
beginning of reaction and then remained quite stable as a function of
time. Concerning the product distribution, the selectivity to benzene
decreased while the formation of cyclohexanone, cyclohexane and cy-
clohexene increased depending on the catalyst. This result indicates
that the selectivity to deoxygenated products decreases as the se-
lectivity to hydrogenation products increases.
nickel phosphide reduces the energy barrier for the cleavage of the CeO
bond, leading to the formation of benzene.
The type of the support affected the nature of the phosphide phase
and, consequently the product distribution. For NiP/SiO
and NiP/CeZrO catalysts, Ni P was the only phosphide phase formed.
These catalysts exhibited the highest selectivity to benzene among the
supported catalysts (around 57%). Ni12 was the main phase on Al
whereas a mixture of both Ni P and Ni12 phases was obtained on
CeO . For these catalysts, hydrogenation products (cyclohexanone,
2 2
, NiP/TiO ,
2
2
Catalyst deactivation is one of the main challenges for the HDO of
model molecules and it has been attributed to carbon deposition, metal
sintering or strong adsorption of the model molecules. Carbon deposi-
tion on supported phosphide catalysts during HDO of phenol was in-
vestigated by Raman spectroscopy of the used catalysts. The spectra of
the used catalysts did not detect the presence of bands at around 1350
P
5
2 3
O ,
2
P
5
2
cyclohexane and cyclohexene) were mainly formed. However, all sup-
ported catalysts showed lower selectivity to benzene, despite the pre-
sence of only the Ni P phase. The supports favored the formation of
2
−
1
and 1580 cm
characteristic of carbonaceous materials, indicating
hydrogenation products via the tautomerization route.
that deposition of carbon does not occur to a significant degree over
these supported NiP catalysts under the conditions used in this work.
de Souza et al. [29] investigated the stability of Pd-based catalysts
All catalysts only slightly deactivated with time on stream. The
supported phosphide catalysts exhibited a lower deactivation rate than
supported metallic catalysts. This is likely due to the high activity of the
phosphide phase that turns over the adsorbed phenol molecule in an
efficient manner, preventing the retention and buildup of intermediate
species that result in catalyst deactivation.
Therefore, nickel phosphide catalysts are promising HDO catalysts
to selectively produce aromatics from phenol with a high degree of
deoxygenation and without significant deactivation.
for HDO of phenol. Pd/SiO
2
, Pd/Al
2
O
3
, Pd/TiO and Pd/ZrO
2
2
catalysts
strongly deactivated. For Pd/Al
2
O
3
and Pd/ZrO catalysts, the phenol
2
conversion after 20 h of reaction decreased from 58% to 10% and 65%
to 18%, respectively. They proposed that the main cause for catalyst
deactivation was Pd sintering, which resulted in an accumulation of
intermediate species, blocking the active sites. Furthermore, a stronger
interaction between the oxygen of the tautomer intermediate and the
oxophilic site could also contribute to catalyst deactivation by retaining
more strongly adsorbed intermediates formed. Temperature pro-
grammed oxidation (TPO) was used to investigate the presence of
carbonaceous species on the used catalysts after 22 h of TOS. The TPO
Acknowledgments
This contribution is dedicated to the memory of Victor Teixeira da
Silva. Priscilla M. de Souza and Carlos V. M. Inocêncio thanks
Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES)
and Fundação de Amparo à Pesquisa do Estado do Rio de Janeiro
(FAPERJ) for the scholarships. Vinicius Ottonio O. Gonçalves and Fabio
B. Noronha also acknowledge the financial support of the Conselho
Nacional de Desenvolvimento Científico e Tecnológico (CNPq -446066/
2014-1), Coordenação de Aperfeiçoamento de Pessoal de Nível Superior
2
profile of the of NiP/TiO spent catalyst after HDO of phenol (Fig. S5)
exhibited one broad peak between 250 and 500 °C. This is attributed to
the oxidation of phenol adsorbed on the catalyst surface.
In our work, the phenol is strongly adsorbed on the phosphide sites,
which promotes the direct deoxygenation. However, the strong ad-
sorption should cause the deactivation of supported phosphide cata-
lysts. This could also explain the changes in selectivity observed during
the reaction. The deactivation of the sites responsible for deoxygenation
would favor the hydrogenation route. The same result was reported
Moon et al. [46], who studied the HDO of guaiacol at different tem-
peratures and pressures. They proposed that the hydrogenation
pathway was less susceptible to coke deposition and catalyst deacti-
vation than the direct deoxygenation pathway. Cecilia et al. [47] in-
(CAPES
- CAPES−COFECUB program –88881.142911/2017-01),
Fundação de Amparo à Pesquisa do Estado do Rio de Janeiro (FAPERJ)
and Fundação de Amparo à Pesquisa do Estado de Minas Gerais
(FAPEMIG -2444946/2018). Vinicius Ottonio O. Gonçalves and
Frédéric Richard acknowledge financial support from the European
Union (ERDF) and “Region Nouvelle Aquitaine”. Research conducted at
UTSA was supported by a UTSA College of Engineering scholarship, the
State of Texas, and the STARs program.
vestigated the stability of NiP/SiO
2
catalysts with different P/Ni molar
ratio (1–3) for the HDO of dibenzofuran. Catalyst stability depended on
9

P.M. de Souza, et al.
Catalysis Today xxx (xxxx) xxx–xxx
Appendix A. Supplementary data
[22] P.M. de Souza, L. Nie, L.E. Borges, F.B. Noronha, D.E. Resasco, Catal. Lett. 144
(2014) 2005–2011.
[
23] P.M. de Souza, R.C. Rabelo-Neto, L.E. Borges, G. Jacobs, B.H. Davis, T. Sooknoi,
Supplementary material related to this article can be found, in the
online version, at doi:https://doi.org/10.1016/j.cattod.2019.08.028.
D.E. Resasco, F.B. Noronha, ACS Catal. 5 (2015) 1318–1329.
[24] A.I. d’Aquino, S.J. Danforth, T.R. Clinkingbeard, B. Ilic, L. Pullan, M.A. Reynolds,
B.D. Urray, M.E. Bussel, J. Catal. 335 (2016) 204–214.
[
25] V.O.O. Gonçalves, P.M. de Souza, T. Cabioc’h, V.T. da Silva, F.B. Noronha,
F. Richard, Catal. Commun. 119 (2019) 33–38.
References
[
26] S.T. Oyama, X. Wang, F. Requejo, T. Sato, Y. Yoshimura, J. Catal. 209 (2002) 1–5.
[
1] E. Furimsky, Appl. Catal. Gen. 199 (2000) 147–190.
[27] D. Lee, P. Bui, H.Y. Zao, S.T. Oyama, T. Dou, Z.H. Shen, J. Catal. 290 (2012) 1–12.
28] H. Shi, J. Chen, Y. Yang, S. Tian, Fuel Process. Technol. 118 (2014) 161–170.
[29] P.M. de Souza, R.C. Rabelo-Neto, L.E. Borges, G. Jacobs, B.H. Davis, D.E. Resasco,
F.B. Noronha, ACS Catal. 7 (2017) 2058–2073.
[30] J. Chen, H. Shi, L. Li, K. Li, Appl. Catal. B: Environ. 144 (2014) 870–884.
[31] S.-K. Wu, P.-C. Lai, Y.-C. Lin, Catal. Lett. 144 (2014) 878–889, https://doi.org/10.
[
[2] S. Fernando, S. Adhikari, C. Chandrapal, N. Murali, Energy Fuels 20 (2006)
1
727–1737.
[
[
[
[
[
3] E. Furimsky, Catal. Today 217 (2013) 13–56.
4] D. Mohan, C.U. Pittman, P.H. Steele, Energy Fuels 20 (2006) 848–889.
5] Y. Romero, F. Richard, S. Brunet, Appl. Catal. B Environ. 98 (2010) 213–223.
6] V.O. Gonçalves, S. Brunet, F. Richard, Catal. Lett. 146 (2016) 1562–1573.
7] V.O. Gonçalves, C. Ciotonea, S. Arrii-Clacens, N. Guignard, C. Roudaut, J. Rousseau,
J.-M. Clacens, S. Royer, F. Richard, Appl. Catal. B Environ. 214 (2017) 57–66.
8] C. Ranga, R. Lødeng, V.I. Alexiadis, T. Rajkhowa, H. Bjørkan, S. Chytil, I.H. Svenum,
J. Walmsley, C. Detavernier, H. Poelman, Chem. Eng. J. 335 (2018) 120–132.
9] P.M. Mortensen, J.-D. Grunwaldt, P.A. Jensen, A.D. Jensen, Catal. Today 259
1
007/s10562-014-1231-7.
[32] J.A. Rodriguez, J.-Y. Kim, J.C. Hanson, S.J. Sawhill, M.E. Bussell, J. Phys. Chem. B
07 (26) (2003) 6276–6285.
1
[
[
[33] G. Berhault, P. Afanasiev, H. Loboué, C. Geantet, T. Cseri, C. Pichon, G.-
D. Catherine, A. Lafond, Inorg. Chem. 48 (7) (2009) 2985–2992.
[
34] M. Peroni, G. Mancino, E. Baráth, O.Y. Gutiérrez, J.A. Lercher, Appl. Catal. B
(2016) 277–284.
Environ. 180 (2016) 301–311.
[
35] Y.K. Lee, S.T. Oyama, Stud. Surf. Sci. Catal. 159 (2006) 357–360.
[
[
[
[
[
[
10] C.A. Teles, R.C. Rabelo-Neto, J.R. de Lima, L.V. Mattos, D.E. Resasco, F.B. Noronha,
Catal. Lett. 146 (2016) 1848–1857.
[36] J. Chen, S. Zhou, D. Ci, J. Zhang, R. Wang, J. Zhang, Ind. Eng. Chem. Res. 48 (2009)
11] L. Nie, P.M. de Souza, F.B. Noronha, W. An, T. Sooknoi, D.E. Resasco, J. Mol. Catal.
Chem. 388 (2014) 47–55.
12] A. Berenguer, T. Sankaranarayanan, G. Gómez, I. Moreno, J. Coronado, P. Pizarro,
D. Serrano, Green Chem. 18 (2016) 1938–1951.
3812–3819.
[37] L.F. Feitosa, G. Berhault, D. Laurenti, T.E. Davies, V.T. da Silva, J. Catal. 340 (2016)
154–165.
[
38] X. Liu, L. Xu, B. Zhang, J. Solid State Chem. 212 (2014) 13–22.
13] V.O. Gonçalves, P.M. de Souza, V.T. da Silva, F.B. Noronha, F. Richard, Appl. Catal.
[39] C. Stinner, R. Prins, T. Weber, J. Catal. 202 (2001) 187–194.
[
40] T.I. Korányi, A.E. Coumans, E.J. Hensen, R. Ryoo, H.S. Kim, É. Pfeifer,
B Environ. 205 (2017) 357–367.
14] Y. Yang, C. Ochoa-Hernández, V.A. de la Peña O’Shea, J.M. Coronado, D.P. Serrano,
ACS Catal. 2 (2012) 592–598.
15] V.O. Gonçalves, P.M. de Souza, T. Cabioc’h, V.T. da Silva, F.B. Noronha, F. Richard,
Appl. Catal. B Environ. 219 (2017) 619–628.
16] H. Shafaghat, P.S. Rezaei, W.M.A.W. Daud, RSC Adv. 5 (2015) 103999–104042.
17] N. Arun, R.V. Sharma, A.K. Dalai, Renewable Sustainable Energy Rev. 48 (2015)
Z. Kasztovszky, Appl. Catal. Gen. 365 (2009) 48–54.
[41] S.T. Oyama, Y.-K. Lee, J. Catal. 258 (2008) 393–400.
[
42] H. Zhao, S.T. Oyama, H.-J. Freund, R. Włodarczyk, M. Sierka, Appl. Catal. B
Environ. 164 (2015) 204–216.
[
43] P. Liu, J.A. Rodriguez, T. Asakura, J. Gomes, K. Nakamura, J. Phys. Chem. B 109
(2005) 4575–4583.
[
[
2
40–255.
[44] Q. Li, X. Hu, Phys. Rev. B 74 (2006) 035414.
[
45] A.B. Hernandez, H. Ariga, S. Takakusagi, K. Kinoshita, S. Suzuki, S. Otani,
[
18] S.-K. Wu, P.-C. Lai, Y.-C. Lin, H.-P. Wan, H.-T. Lee, Y.-H. Chang, ACS Sustain. Chem.
Eng. 1 (2013) 349–358.
19] S.T. Oyama, T. Onkawa, A. Takagaki, R. Kikuchi, S. Hosokai, Y. Suzuki, K.K. Bando,
Top. Catal. 58 (2015) 201–210.
20] Y. Yang, J. Chen, H. Shi, Energy Fuels 27 (2013) 3400–3409.
21] J.-S. Moon, Y.-K. Lee, Top. Catal. 58 (2015) 211–218.
S.T. Oyama, K. Asakura, Chem. Phys. Lett. 513 (2011) 48–52.
[
46] J.S. Moon, E.G. Kim, Y.K. Lee, J. Catal. 311 (2014) 144–152.
[
[47] J.A. Cecilia, A. Infantes-Molina, E.R. Castellón, A. Jiménez-Lopez, S.T. Oyama,
[
[
Appl. Catal. B 136-137 (2013) 140–149.
10