Ni2P and CoP catalysts prepared from phosphite-type precursors for HDS-HDN competitive reactions

Article abstract of DOI:10.1016/j.apcata.2010.10.019

Silica-supported nickel and cobalt phosphide catalysts were synthesized with low metal-loading (5 wt%). The formation of Ni₂P and CoP phases was achieved by temperature-programmed reduction of phosphite-based precursors, Ni(HPO₃H)₂ and Co(HPO₃H)₂, respectively. The catalysts were characterized by X-ray fluorescence (XRFS), N₂ adsorption-desorption, X-ray diffraction (XRD), transmission electron microscopy (TEM), X-ray photoelectron spectroscopy (XPS) and elemental analysis (CNHS). These catalysts were tested in the individual reactions of hydrodesulfurization (HDS) of dibenzothiophene (DBT) and hydrodenitrogenation (HDN) of quinoline (Q), and also in the simultaneous HDS and HDN reactions. These tests were conducted by feeding 3000 ppm of DBT or 3000 ppm of Q. Additional tests were performed with the aim to evaluate S-removal capability by keeping constant Q concentration in the feed (3000 ppm) and varying DBT concentration (200 and 2000 ppm). The catalytic results show that these phosphide catalysts achieve high DBT and Q conversion values at high temperatures though the rate of HDS reaction is higher than that of HDN. In all cases Ni₂P is more active than CoP. Moreover, the presence of N-containing molecules in the feed does not affect the S-removal as sulfur is almost completely removed after 48 h on-stream.

Full text of DOI:10.1016/j.apcata.2010.10.019

Applied Catalysis A: General 390 (2010) 253–263
Contents lists available at ScienceDirect
Applied Catalysis A: General
journal homepage: www.elsevier.com/locate/apcata
Ni2P and CoP catalysts prepared from phosphite-type precursors for HDS–HDN
competitive reactions
A. Infantes-Molinaa, J.A. Ceciliab, B. Paweleca, J.L.G. Fierroa,
E. Rodríguez-Castellónb, A. Jiménez-Lópezb,∗
a Instituto de Catálisis y Petroleoquímica, CSIC, Cantoblanco, 28049 Madrid, Spain
b Departamento de Química Inorgánica, Cristalografía y Mineralogía (Unidad Asociada al ICP-CSIC), Facultad de Ciencias, Universidad de Málaga,
Campus de Teatinos, 29071 Málaga, Spain
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 19 July 2010
Received in revised form 18 October 2010
Accepted 19 October 2010
Available online 23 October 2010
Silica-supported nickel and cobalt phosphide catalysts were synthesized with low metal-loading (5 wt%).
The formation of Ni2P and CoP phases was achieved by temperature-programmed reduction of phosphite-
based precursors, Ni(HPO3H)2 and Co(HPO3H)2, respectively. The catalysts were characterized by
X-ray fluorescence (XRFS), N2 adsorption–desorption, X-ray diffraction (XRD), transmission electron
microscopy (TEM), X-ray photoelectron spectroscopy (XPS) and elemental analysis (CNHS). These cata-
lysts were tested in the individual reactions of hydrodesulfurization (HDS) of dibenzothiophene (DBT)
and hydrodenitrogenation (HDN) of quinoline (Q), and also in the simultaneous HDS and HDN reactions.
These tests were conducted by feeding 3000 ppm of DBT or 3000 ppm of Q. Additional tests were per-
formed with the aim to evaluate S-removal capability by keeping constant Q concentration in the feed
(3000 ppm) and varying DBT concentration (200 and 2000 ppm). The catalytic results show that these
phosphide catalysts achieve high DBT and Q conversion values at high temperatures though the rate of
HDS reaction is higher than that of HDN. In all cases Ni2P is more active than CoP. Moreover, the pres-
ence of N-containing molecules in the feed does not affect the S-removal as sulfur is almost completely
removed after 48 h on-stream.
Keywords:
Commercial silica
Nickel phosphide (Ni2P)
Cobalt phosphide (CoP)
Dibenzothiophene
HDS
Quinoline
HDN
© 2010 Elsevier B.V. All rights reserved.
1. Introduction
S, N, O, and metals from different petroleum streams in a refin-
ery.
Demands for a cleaner environment have led to a global
tightening in the allowed sulfur content in fuels and increas-
ing restrictions in the release of nitrogen oxides are also being
imposed. The environmental protection agency (EPA) has legis-
lated a reduction of sulfur content from 500 to 10 parts per
million by weight (wppm) at the beginning of 2011 [1]. The
restrictive regulations in the European union are also similar,
and a maximum sulfur content of 10 wppm is imposed [2].
Moreover, feedstocks of lower quality (typical crude contains
1.5 wt% S) are being treated due to the reduction in petroleum
reserves. Accordingly, considerable efforts are being devoted to
develop new technologies for the production of clean fuels such as
biodesulfurization (BDS) [3,4], oxidative desulfurization (ODS) [5],
ultrasound-assisted oxidative desulfurization (UAOD) [6]. How-
ever, hydroprocessing appears to be the technologically preferred
solution. Hydroprocessing refers to a variety of catalytic hydro-
genation processes that saturate heteroatomatic rings and remove
The environmental legislation demands in terms of fuel sulfur
content have been attained with HDS commercial catalysts such
sulfides based on Mo or W and doped with Ni or Co [3] that have
been modified using several promoters or changing the material
support [7–9]; and non-traditional catalytic materials such car-
bides and nitrides [10,11]. Recently, transition metal phosphides,
MoP [12,13], WP [14,15], CoP [16,17] and Ni2P [18–20] have been
reported as very active phases in hydrodesulfurization (HDS) and
hydrodenitrogenation (HDN) reactions. Among these phosphides,
the Ni2P phase has been reported the most active one for HDS reac-
tions, even presenting a higher HDS activity than that found for
traditional sulfides [21,22]. Several researches have also pointed
out a high activity and stability of the CoP phase in HDS reactions
[17,23]. It has been shown that during the HDS reaction there is
a transformation on the phosphide surface due to sulfur incor-
poration and formation of a phosphosulfide phase, considered as
the real active phase in HDS reactions catalysed by phosphides
[24,25]. Moreover, it should be kept in mind that sulfur and nitrogen
are always present in crude oil. Accordingly, and in order to fulfil
the environmental laws, the strong suppression of sulfur removal
by nitrogen compounds [26] is a fundamental problem which is
∗
Corresponding author. Tel.: +34 952131876; fax: +34 952137534.
E-mail address: ajimenezl@uma.es (A. Jiménez-López).
0926-860X/$ – see front matter © 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.apcata.2010.10.019

254
A. Infantes-Molina et al. / Applied Catalysis A: General 390 (2010) 253–263
exacerbated by the decline in the overall quality of crudes. Thus,
nitrogen removal is a key problem in HDS, in particular deep HDS
to levels of 50 ppm S or less. It is well recognized that HDN is
more difficult than HDS, requiring more severe conditions of pres-
sure and temperature. With this premises, it turns out interesting
to find a catalytic system capable of eliminating both N- and S-
containing molecules. Ni2P has been considered one of the most
active phases for both HDS and HDN reactions [16]. The effect of
phosphorus content in the hydroprocessing performance of phos-
phides is considered a key factor. Oyama et al. [27] reported that
the HDS reaction hardly changes with the initial Ni/P ratio while
the HDN conversion goes through a maximum when a Ni/P molar
ratio of 0.5 was employed, as in the case of this study. On contrary,
Bussell and co-workers [25] have reported the opposite effect, i.e.,
they found a strong variation of the HDS activity with the initial
Ni/P ratio, as long as when high Ni/P ratios are used, Ni12P5 impu-
rities are present that deleteriously affected the HDS properties of
the catalysts.
On the other hand, the role of the support has shown to be com-
plex due to the interaction active phase precursor-support. In the
case of acid supports such as MCM-Zr or ꢀ-Al2O3, a strong inter-
action with the active phase precursors and support is developed
[17,20] thus requiring a higher reduction temperature. Other sup-
ports based on hexagonal mesoporous silica such as MCM-41 [28]
or SBA-15 [29] have been widely used but the phosphorus excess
necessary to allow the formation of phosphide causes a blockage
of the mesoporous structure. Silica has been often tested as a sup-
port for phosphide catalysts [16] because the support–precursor
interaction is weak due to the low acidity of the silica.
In this paper, the formation, activity and catalytic stability of
nickel and cobalt phosphide catalysts in HDS, HDN and HDS–HDN
competitive reactions are described. These catalysts have been
synthesized by impregnation of the support with nickel(II) dihy-
drogenphosphite, Ni(HPO3H)2, or cobalt(II) dihydrogenphosphite,
Co(HPO3H)2. Subsequently, the precursor was reduced in H2 flow
(TPR) leading to the Ni2P [30] or CoP [17] phase, respectively. As
regard to the support, commercial silica has been chosen due to its
low cost and the interesting results found previously in hydrotreat-
ing reactions [17,20]. The catalysts prepared here only possess a
5 wt% of metal and they will be studied in the DBT HDS, Q HDN reac-
tions and in HDS–HDN competitive reactions by varying the ratio
of N- and S-compounds, i.e., studying the possible suppression of
sulfur removal by the presence of nitrogen compounds.
nickel(II) dihydrogenphosphite (Ni(HPO3H)2) and cobalt(II) dihy-
drogenphosphite (Co(HPO3H)2) were prepared by mixing the
stoichiometric amounts of phosphorus acid (H2PO3H) and nickel(II)
hydroxide (Ni(OH)2) or cobalt(II) hydroxide (Co(OH)2) [30], as
accordingly, to the incipient volume. Subsequently, each solution
was added to the pelletized silica (0.85–1.00 mm), and air dried at
40 ◦C. Finally, a temperature programmed reduction was used to
convert the phosphite into the desired phosphide according to the
method previously described [30], i.e., the sample was placed in
a tubular reactor, heating at linear temperature ramp (3 ◦C min−1
)
in flowing H2 (100 mL min−1) from 100 ◦C to 500 ◦C as it was opti-
mised in previous works [17,20].
2.3. Characterization of catalysts
Elemental bulk composition of the catalysts in terms of Ni, Co
and P loading was achieved by X-ray fluorescence spectrometry
(XRFS) of dispersive energy using a XGT-500 equipment with a
probe diameter of 1.2 mm. The X-ray tube was set at 50 kV with
an acquisition time of 500 s and a current intensity between: 0.26
and 0.32 mA.
The textural parameters were evaluated from the nitrogen
adsorption–desorption isotherms at −196 ◦C as determined by an
automatic ASAP 2020 system from Micromeritics.
X ray diffraction patterns (XRD) of the precursor, reduced and
spent catalysts were obtained with a X’Pert PRO MPD Philips
diffractometer (PANanalytical), using monochromatic CuK␣ radi-
˚
ation (ꢁ = 1.5406 A). The K␣ radiation was selected with a Ge (1 1 1)
primary monochromator. The X-ray tube was set at 45 kV and
40 mA.
Transmission electron micrographs of the precursor and
reduced catalysts were obtained by using a Philips CM 100
Supertwin-DX4 microscope. Samples were dispersed in ethanol
and a drop of the suspension was put on a Cu grid (300 mesh).
Elemental chemical analysis was performed for spent catalysts
with a LECO CHNS 932 analyser to determine the sulfur and the
nitrogen content present after the catalytic test through the com-
bustion of the samples at 1100 ◦C in pure oxygen to form NO and
SO2.
X-ray photoelectron spectra were collected using a physical
electronics PHI 5700 spectrometer with non monochromatic AlK␣
radiation (300 W, 15 kV, and 1486.6 eV) with a multi-channel detec-
tor. 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 adventitious car-
bon (C 1s at 284.8 eV). A PHI ACCESS ESCA-V6.0 F software package
was used for acquisition and data analysis. A Shirley-type back-
ground was subtracted from the signals. Recorded spectra were
always fitted using Gaussian–Lorentzian curves in order to deter-
mine the binding energies of the different element core levels more
accurately. Reduced and spent catalysts were stored in sealed vials
with an inert solvent. The sample preparation was done in a dry
box under a N2 flow, where the solvent was evaporated prior to
its introduction into the analysis chamber, and directly analyzed
without previous treatment.
2. Experimental
2.1. Materials
The support used in this study was commercial silica (Cab-
osil® M-5, Riedel-de Haën, Sigma–Aldrich). The reactants used to
prepare the catalyst precursors were phosphorus acid (H2PO3H,
Aldrich 99%), cobalt(II) hydroxide (Co(OH)2, Aldrich 95%) and
nickel(II) hydroxide (Ni(OH)2, Aldrich). The chemical products uti-
lized in the reactivity study were dibenzothiophene (Aldrich 98%)
and quinoline (Aldrich 98%) in cis-, trans-decahydronaphthalene
(Sigma–Aldrich 98%). The gases employed were He (Air Liquide
99.99%), H2 (Air Liquide 99.999%) and N2 (Air Liquide 99.9999%).
2.4. Catalytic test
For the catalytic tests, the HDS of DBT and the HDN of quinoline
were chosen, which were performed in a high-pressure fixed-bed
continuous-flow stainless steel catalytic reactor (9.1 mm in diam-
eter, and 230 mm in length), operated in the down-flow mode. The
reaction temperature was measured with an interior placed ther-
mocouple in direct contact with the catalyst bed. The organic feed
was adjusted to solutions in cis- trans-decaline of: DBT (3000 ppm)
for the study of HDS reaction, quinoline (Q) (3000 ppm) for the
2.2. Preparation of catalysts
Nickel phosphide and cobalt phosphide catalysts with 5 wt%
of nickel and cobalt, respectively, were prepared. The sup-
port used was a commercial silica Cab-osil (SBET = 257 m2 g−1
and Vp = 0.72 cm3 g−1). The catalysts were prepared follow-
ing the incipient wetness impregnation method. Solutions of

A. Infantes-Molina et al. / Applied Catalysis A: General 390 (2010) 253–263
255
study of HDN reaction and finally 3000 ppm Q-2000 ppm DBT and
Ni2P (Solid)
CoP (Open)
3000 ppm Q-200 ppm DBT for the study of HDS–HDN competitive
reactions and the possible suppression of S-removal by nitrogen
compounds.
100
80
60
40
20
0
Each solution was supplied by means of a Gilson 307SC piston
pump (model 10SC). For the activity tests, 0.5 g of catalyst were
used (particle size 0.85–1.00 mm) and were diluted with silicon car-
bide to 3 cm3. Prior to the activity test, the catalysts were reduced
in situ at atmospheric pressure with a H2 flow of 100 mL min−1
by heating from room temperature to 500 ◦C at a heating rate of
3 ◦C min−1. Catalytic activities were measured at different tem-
peratures (300–475 ◦C), under 3.0 MPa of H2, with a flow rate of
3000DBT
3000Q 2000DBT
3000Q 200DBT
100 mL min−1 and with hourly space velocities (WHSV) of 32 h−1
.
Moreover, the stabilities of the catalysts were studied at a constant
temperature reaction for 48 h. The evolution of the reaction was
monitored by collecting liquid samples after 60 min at the desired
reaction temperature. These liquid samples were kept in sealed
vials and subsequently analyzed by gas chromatography (Shimadzu
GC-14B, equipped with a flame ionization detector and a capil-
lary column, TBR-14, coupled to an automatic Shimadzu AOC-20i
injector).
For these catalysts, the main products of the HDS reac-
tion were biphenyl (BP), cyclohexylbenzene (CHB), bicyclohexyl
(BCH), benzene (B) and cyclohexane (CH). In the case of HDN
(Scheme 1), the main products were 1,2,3,4-tetrahydroquinoline
(THQ-1), 5,6,7,8-tetrahydroquinoline (THQ-5), propylcyclohexane
(PCH), propylbenzene (PB) and decahydroquinoline (DHQ). The
total conversion of HDS reaction was calculated from the ratio of
converted DBT/initial DBT. The total HDN conversion was calcu-
lated according to Eq. (1):
300
325
350
375
400
425
450
475
Temperature (ºC)
100
95
90
85
80
75
70
65
CQ.Init. − CQ.Final − CInt.Hyd.
HDNConv. (%) =
× 100
(1)
CQ.Init.
where CQ.Init. represents the quinoline concentration in the feed,
CQ.Final is the quinoline concentration in the HDN liquid product and
CInt.Hyd. is the concentration of hydrogenated intermediates (THQ-1,
THQ-5 and DHQ). The selectivity of the different reaction products
was calculated considering BP, CHB, B and CH as the only products
obtained for HDS. In the case of HDN, the selectivity was calcu-
lated considering PCH, PB, DHQ, THQ-1 and THQ-5 as products of
HDN reaction. Moreover slight traces of some unknown compounds
were observed in few cases.
300
325
350
375
400
425
450
475
Temperature (ºC)
30
25
20
15
10
5
3. Catalytic results
Two catalysts with 5 wt% of Ni and Co, respectively, were tested
in the HDS reaction of DBT and in the HDN reaction of Q. The mech-
anisms of HDN and HDS and the products obtained are shown in
Scheme 1.
0
3.1. Role of the reaction temperature
300
325
350
375
400
425
450
475
Temperature (ºC)
The starting point was to study the role of reaction temperature
on both catalytic reactions, whose results are shown in Fig. 1 for
the HDS reaction and Fig. 2 for the HDN reaction. In the case of
the HDS reaction (Fig. 1), we can observe that Ni2P and CoP reach
high conversion values when the temperature increases. Compar-
ing both catalysts we can conclude that Ni2P is more active than CoP
in HDS reaction, because Ni2P shows higher conversion values than
CoP, as reported by others [16]. On the other hand, when mixing
sulfur and nitrogen containing molecules i.e., the HDS–HDN com-
petitive reactions, the HDS reaction is improved, especially when
only 200 ppm of DBT is in the feed. Some authors have pointed out
a strong suppression of sulfur removal in the presence of nitrogen
compounds due to the deactivation of the active phase for the HDS
reaction [26]. In our case, we discard this suppression because at
Fig. 1. HDS of DBT in the absence and presence of Q. Evolution of conversion and
selectivity versus reaction temperature for Ni2P and CoP catalysts.
325 ◦C, both catalysts remove 90% of sulfur when the feed contains
3000 ppm of Q and 200 ppm of DBT, i.e., both catalysts perform deep
HDS.
With regards to the selectivity of HDS reaction (Fig. 1), this
proceeds through two different pathways; a hydrogenation (HYD)
route and a direct desulfurization (DDS) route (Scheme 1) [31]. BP
and CHB are the main products of the DDS and HYD routes, respec-
tively. These two main products were detected here: BP formed
in greater proportions in all cases, and CHB in a lesser extend, as

256
A. Infantes-Molina et al. / Applied Catalysis A: General 390 (2010) 253–263
Scheme 1. Reaction mechanism of HDS of DBT (top) and HDN of Q (bottom).
reported in the literature for phosphide catalysts [25]. The tendency
of nickel phosphide and cobalt phosphide with the temperature is
similar. Both of them show a higher selectivity toward BP, reaching
even values of 90–95% at 475 ◦C. The reaction with the lowest sulfur
content (200 ppm) both catalysts presented the lowest selectivity
toward BP reaching values of about 80% at 475 ◦C. The selectivity
toward CHB decreases at high temperatures due to the hydro-
genation reaction is not favoured at high temperatures because
of thermodynamic considerations. Moreover, the lower the sulfur
content in the feed the higher the selectivity to CHB, i.e., at 350 ◦C
with only 200 ppm of DBT, a selectivity to CHB of 27% is achieved
for both catalysts.
Considering the HDN activity of these catalysts versus the tem-
perature (Fig. 2) it is observed that Ni2P is also more active than
CoP, in agreement with a previous report by Oyama et al. [16].
The HDN conversion of quinoline is low at moderate temperatures
increasing in so far as the temperature does. When 200 ppm of DBT
is present in the feed a slight increase in the HDN conversion is
observed with respect to the feed with only quinoline, meanwhile
the opposite effect is obtained with a feed containing 2000 ppm of
DBT, i.e., a slight decrease in the HDN conversion is obtained for
Ni2P. Nevertheless, some authors have established that the HDN
conversion catalysed by phosphides diminishes dramatically in the
presence of H2S produced in the HDS reaction [28], indicating that
H2S has an inhibiting effect on HDN and it also alters the HDN path-
ways as it happens in conventional hydrotreating catalysts [32,33].
In our case, no loss of activity is observed. CoP catalysts exhibit
conversions below 20% in all range of studied temperatures.
The hydrogenated intermediates and the products obtained in
the HDN reaction are shown in Fig. 2. HDN has been suggested
to be a structure-sensitive reaction due to the fact that nitrogen
is removed indirectly [19] as consequence of the greater strength
of the C–N bonds. The HDN reaction of quinoline can proceed
via fully hydrogenated intermediates, DHQ-PCHA, or via partially
hydrogenated intermediates, THQ-1-OPA (Scheme 1). In a first step,
there is a hydrogenation of one of the aromatic rings (HYD). The
hydrogenation of the N-containing ring is easily achieved, and
a thermodynamic equilibrium between quinoline and THQ-1 is
readily established under typical HDN conditions and favoured
at low temperatures. At moderate temperatures, it has been sug-
gested that HDN of quinoline mainly takes place through the fully
satured intermediates (DHQ-PCHA) that results in PCHE which
undergoes further (de)hydrogenation to PCH and PB. Nonetheless,
the mechanism can also take place through an intermediate par-
tially hydrogenated as THQ-1 [27] and the hydrogenolysis of the
aliphatic C–N bond (CNH) to obtain orto-propylaniline (OPA). Sev-
eral authors have established that the reaction of OPA is greatly
inhibited by quinoline and THQ-1, so the partially hydrogenated
pathway should be unfavourable by self-inhibiting effect [34].
However, at high temperatures and with the presence of phos-
phorus on NiMo sulfide catalysts, Jian and Prins [35] reported that
the reaction of Q → THQ-1 → OPA → PB is favoured and that PB can
hardly be hydrogenated under the presence of quinoline-type com-
pounds.
The amounts of DHQ and OPA detected, in all cases, are minimal.
The low detected concentrations of OPA and DHQ are due to their
transformation under the corresponding conditions. With regard
to the partially hydrogenated intermediate THQ-1, it is favoured at
low temperatures and the selectivity decreases when the temper-
ature is higher as observed in Fig. 2. The selectivity toward THQ-1
shows a similar pattern for both catalysts reaching selectivity val-
ues from 70 to 80% at 300 ◦C to 5–20% at 475 ◦C. The presence
of THQ-1 at low temperatures is lower for both catalysts when
200 ppm of DBT is present in the feed. Lu et al. established a relation
between Ni/P ratio and the dominant hydrogenated intermedi-
ate. The main intermediates THQ-1 and DHQ are favoured for Ni/P
ratio close to 2 and for lower Ni/P ratios the intermediate THQ-1 is
favoured at moderate temperatures [28]. In our case, with an initial
Ni/P close to 0.5, the hydrogenated intermediate THQ-1 is favoured.
At high temperatures THQ-1 suffers ring opening and loss of NH3,
being this effect maximum at 450 ◦C as can be seen in the formation
of PB curves mainly for Ni2P catalyst.

A. Infantes-Molina et al. / Applied Catalysis A: General 390 (2010) 253–263
257
Ni2P (solid)
3000Q
3000Q 200DBT
3000Q 2000DBT
CoP(Open)
100
80
60
40
20
0
300
325
350
375
400
425
450
475
Temperature (ºC)
80
80
60
40
20
0
60
40
20
0
300 325 350 375 400 425 450 475
Temperature (ºC)
300 325 350 375 400 425 450
Temperature (ºC)
475
40
30
20
10
0
60
45
30
15
0
300 325 350 375 400 425 450 475
Temperature (ºC)
300 325 350 375 400 425 450 475
Temperature (ºC)
Fig. 2. HDN of Q in the absence and presence of DBT. Evolution of conversion and selectivity versus reaction temperature for Ni2P and CoP catalysts.
Kinetic analysis have revealed that when the temperature is
increased, THQ-1 is hydrogenated to the fully hydrogenated inter-
mediate DHQ and in the next step DHQ leads to THQ-5 (Scheme 1),
more favoured than PCHA [36]. This is the case of CoP catalyst,
where values of THQ-5 selectivity of about 50–60% at 425–475 ◦C
are attained. On the other hand, Ni2P catalyst shows a higher
selectivity values at intermediate temperatures, 30% at 375 ◦C,
decreasing with the temperature. On the other hand, the higher
selectivity toward THQ-5 for CoP catalyst reveals the lack of activ-
ity to C–N bond cleavage of this catalyst, so that the reaction is
limited to the intermediate THQ-5.
The formation of the denitrogenated product PCH occurs as can
be seen in Fig. 2. The path that is very favoured kinetically is DHQ-
PCHA at intermediates temperatures, reaching the highest values
at about 375–425 ◦C for Ni2P catalyst (20–28%). All the tests shown
that Ni2P, being more active than CoP, possesses higher PB and
PCH selectivity values. Moreover it cannot be discarded a contri-
bution to PB formation from the dehydrogenation of PCHA, that
is favoured at higher temperatures [37]. Similar results have been
published by Oyama et al where PB and PCH selectivities of about
39.8% and 37.1%, respectively were obtained with a Ni2P/SiO2 cat-
alyst at 340 ◦C [27]. The phosphorus excess present on the catalyst
surface prepared from phosphite based precursor seems to be an
important factor to consider in terms of PB selectivity as it has been
reported that it leads the Q HDN reaction through the partially
hydrogenated route [35].
With regards to the effect of sulfur content in PB and PCH selec-
tivity, the presence of 200 ppm of DBT in the feed improves the
selectivity to these products. In contrast, higher sulfur content,
2000 ppm of DBT in the feed, gives rise to lower selectivity due to
a possible competitiveness of Q or DBT for the active sites and the
poisoning of the catalysts due to the presence of H2S in the process
[38].
3.2. Stability with time
The stability of these catalysts was tested for 48 h on-stream at
a reaction temperature of 400 ◦C in the case of Ni2P and 425 ◦C for

258
A. Infantes-Molina et al. / Applied Catalysis A: General 390 (2010) 253–263
Table 1
Conversion and selectivity of HDN and HDS reactions after 48 h of reaction.
Simples
Ni2P
CoP
HDS
HDN
HDS
HDN
Conv. (%)
BP (%)
CHB (%)
Conv. (%)
PCH (%)
PB (%)
Conv. (%)
BP (%)
CHB (%)
Conv. (%)
PCH (%)
PB (%)
3000Q
–
–
–
17.8
9.1
57.7
49.5
49.7
–
26.9
31.1
24.6
–
25.1
22.3
28.6
–
–
–
–
20.3
16.3
14.5
–
8.0
6.4
5.6
–
14.0
13.9
11.8
–
3000Q 200DBT
3000Q 2000DBT
3000DBT
95.1
98.5
99.6
78.8
89.0
81.5
81.6
78.8
78.6
89.5
90.6
90.3
7.3
8.2
8.4
13.7
CoP. Table 1 compiles the selectivity values to the different products
in the HDS and HDN reactions and the conversion values on each
case after 48 h of reaction. Fig. 3 depicts the evolution of conversion
and selectivity for Ni2P catalyst with a feed containing 3000 ppm
Q and 200 ppm DBT, considered as representative. The HDS cat-
alytic results show that both catalysts, Ni2P and CoP, present a high
activity during the first hours on-stream. In this sense, Ni2P cata-
lysts possess a conversion close to 97–99% after 6 h on stream that
is maintained for 48 h. Meanwhile, CoP ones suffers a slight loss in
the conversion after 48 h of reaction, in spite of its initial high activ-
ity. In both cases, the same pattern of selectivity are observed and
the inhibiting effect of nitrogen molecules is hardly noticeable after
48 h of reaction, possibly due to the formation of more active phase
as previously reported [30] and the formation of phosphosulfide
species [39].
The selectivity values after 48 h on-stream for the HDS reac-
tion are also compiled in Table 1. BP is the favoured product in
all cases. The selectivity to this product is higher for CoP catalyst
reaching selectivity values close to 89–92% after 48 h on-stream. In
the case of Ni2P, the selectivity toward BP is slightly lower, show-
ing values of 79% after 48 h for the reaction with the lowest sulfur
content in the feed, as found in the temperature test. The product
of the HYD pathway, such as CHB, is less favoured, as expected.
Moreover, benzene was also formed in some cases with selectiv-
ity values lower than 10%. We can conclude that the presence of
N-containing compounds does not affect the HDS conversion after
48 h on-stream. The main pattern of selectivity hardly changes with
the presence of quinoline, although for Ni2P catalyst, when the pro-
portion Q/DBT is high, the HYD route is more favoured with regards
to feeds containing more DBT molecules.
With regard to the stability with the time on-stream in the HDN
reaction, Table 1 compiles the percentages of HDN removed for
Ni2P and CoP catalysts. In all cases the conversion increases during
the first hours on-stream, as shown in Fig. 3 for Ni2P catalyst with
a feed containing 3000 ppm Q and 200 ppm DBT, as representative.
This is due to the formation of more metal transition phosphide as
active phase for HDN reaction. Thereafter catalysts are deactivated
with the time on-stream. CoP catalyst shows lower activity than
Ni2P one, so Ni2P catalysts reach nitrogen removed values of about
50% after 48 h on stream. In the case of CoP one, it reaches nitrogen
removed values close to 14–20% after 48 h reaction time. Moreover,
conversion values are lower when sulfur content is present in the
feed. Contrary to what was obtained in the temperature test, the
stability of these catalysts in the HDN reaction is influenced by the
presence of S-containing molecules. From this data it can be con-
cluded that the catalysts are less active in HDN than in HDS, and
the slight deactivation in the HDN reaction after 48 h on-stream
is more important when DBT molecules are present in the feed.
Authors have pointed out a decrease in the stability due to H2S
molecules coming from HDS reaction, that can adsorb on the active
sites and poisoning them [38].
HDS reaction
100
80
60
Conv
BP
40
20
CHB
0
8
16
24
Time (h)
32
40
48
HDN reaction
With regards to the denitrogenated products obtained in the
HDN reaction, the main products are PCH and PB (Table 1). In the
case of the PB selectivity, Ni2P has a higher selectivity toward PB
(22–29% at 48 h) than CoP (11–14% at 48 h). On the other hand, PCH
reaches selectivity values of 24–33% for Ni2P catalyst and 5–8% for
CoP at 48 h.
70
60
50
40
30
20
Conv
PB
PCH
4. Characterization and evolution of catalysts
4.1. Elemental analysis
Table 2 reports the nickel, cobalt and phosphorus contents of the
reduced and spent catalysts as determined by X-ray fluorescence
spectroscopy. As seen in this table, both the Ni/P and Co/P molar
ratios for reduced catalysts are lower to the nominal ones (2 and 1,
respectively), indicating that only a partial reduction has happened
and that a phosphorus excess is present in the catalysts. After the
catalytic tests the Ni/P molar ratio increases, indicating that the
0
8
16
24
Time (h)
32
40
48
Fig. 3. Evolution of HDS and HDN conversion and selectivities versus reaction time,
for a Ni2P catalyst with a feed containing 3000Q200DBT.

A. Infantes-Molina et al. / Applied Catalysis A: General 390 (2010) 253–263
259
Table 2
Chemical analysisa, XPS atomic ratiosb and sulfur content on spent catalystsc.
∗
Ni2P
CoP
∗
∗
∗
Samples
Nickel catalysts
Cobalt catalysts
CoP
∗
∗
Ni/Pa
Ni/Pb
S (%)c
Co/Pa
0.41
Co/Pb
0.58
S (%)c
∗
∗
Reduced catalysts
0.36
0.34
–
–
Spent catalysts after temperature test
3000Q
0.52
0.52
0.52
0.50
–
–
0.16
0.20
–
0.41
0.46
0.60
0.44
0.53
0.65
0.72
0.65
–
3000Q 200DBT
3000Q 2000DBT
3000DBT
0.002
0.013
0.040
0.001
0.009
0.012
Co(HPO H)
3
2
Spent catalysts after time test
Ni2P
3000Q
0.32
0.54
0.49
0.54
0.34
0.34
0.42
0.44
–
0.52
0.81
0.54
0.52
0.60
0.46
0.68
0.58
–
3000Q 200DBT
3000Q 2000DBT
3000DBT
0.024
0.086
0.094
0.009
0.035
0.037
Ni(HPO H)
3
2
a
Bulk molar ratios determined by XRFS.
Superficial atomic ratio calculated by XPS.
Sulfur content on spent catalysts as determined by CNHS.
30
40
50
θ (º)
60
70
b
c
2
Fig. 4. X-ray diffractograms of precursors and reduced catalysts at 500 ◦C.
formation of more active phase has happened. However, the Co/P
ratio hardly increases. These results reveal that a phosphorus excess
is always present on the catalysts surface.
35.4, 36.4, 46.7, 48.5, 55.3 and 56.8 (PDF 00-029-0497) have been
observed.
XRD analyses of the spent catalysts after temperature and time
tests (not shown) are similar to those obtained in previous studies
[17,20]. Ni2P catalyst presents the same diffraction pattern, while
CoP one suffers an increase in the diffraction domains, indicating
the appearance of a larger amount of CoP phase during the catalytic
test. On the other hand, with this technique, we do not observe
evidence of the formation of new sulfided and nitrogenated phases
or the alteration of metal phosphide during the catalysis, indicating
a high stability of the transition metal phosphides prepared here.
4.2. Textural properties
The textural properties of support, precursor and reduced cat-
alysts are listed in Table 3. As a preliminary observation, both
catalysts show similar textural properties. The precursors suffer
a loss of both specific surface area and pore volume. These values
increase after reduction, due to the loss of water and phosphine in
the process of the reduction of phosphite to phosphide [17,30]. Tex-
tural properties were also studied after the catalytic tests. Firstly,
no noticeable change in textural properties was observed, when
the composition of the organic feed was modified. Hence, average
values after temperature and stability tests are compiled in Table 3.
4.4. TEM
Transmission electron microscopy was performed in order to
elucidate the distribution of the active phase on the support. Fig. 5
shows the micrograph of Ni2P (Micrograpgh 5.A) and that of CoP
(Micrograpgh 5.B). It can be clearly seen how Ni2P presents a bet-
ter dispersion of the active phase with smaller particles than that
observed in CoP. These results confirm the statement suggested in
XRD section. Ni2P is better dispersed than CoP.
4.3. XRD
The identification of the crystalline phases was carried out by
XRD. The diffraction profiles of reduced catalysts do not show
well defined diffraction lines (Fig. 4) due to the low metal load-
ing employed (5 wt% of nickel or cobalt). Moreover as previously
observed [17,20], Cab-osil favours the formation of transition metal
phosphide particles with small particle size and high dispersion.
In both cases, the diffraction profiles show a unique phase. In the
case of Ni2P catalyst, the main diffraction lines of the Ni2P phase
appear, which are located at 2ꢂ (◦) = 40.7, 44.6, 47.4, 54.2 and 55.0
(PDF 01-089-4864). On the other hand, for the cobalt catalyst,
only the diffraction lines of the CoP phase located at 2ꢂ (◦) = 31.7,
4.5. XPS
The surface composition of catalyst precursors, reduced and
spent catalysts was evaluated by X-ray photoelectron spectroscopy.
Table 4 compiles the Ni 2p3/2 and Co 2p3/2 and P 2p3/2 binding
energy values. In the case of the Ni-based catalyst, the same spec-
tral characteristics previously found for Ni-based catalysts were
observed [30]. In general, the Ni 2p3/2 core-level spectra for the
precursor catalyst possess a main component at 856.8 eV due to
Ni2+ ions forming Ni(HPO3H)2. Similarly the P 2p core-level spec-
tra, present the P 2p3/2 (solid) component centered at ca. 134.1 eV
assigned to phosphorus in the form of HPO3H− [40]. The dotted
lines are the P 2p1/2 components coming from the spin-orbit split-
ting. The P 2p1/2 and P 2p3/2 doublet are separated 0.8 eV. In the case
of reduced catalysts (Fig. 6), the Ni 2p3/2 core-level spectra show a
new contribution located at ca. 852.6 eV due to the presence of Ni2P
[41]. In the same way, the formation of Ni2P is also observed from
the P 2p core level spectra (Fig. 6) with the appearance of a new
P 2p3/2 band at 129.1 eV. The P 2p3/2 signal at 133.1 eV is due to
unreduced (HPO3H−) species [30] and finally the band at 134.4 eV
arises from phosphate species (PO43−) due to a superficial oxidation
of phosphide to phosphate due to air contact [42].
Table 3
Textural properties of the support, catalyst precursors, reduced and spent catalysts.
Sample
SBET (m2 g−1
)
Vp (Cm3 g−1
)
dp (av) nm
Support
Cab-osil
257
0.72
11.2
Catalyst precursors
Ni(HPO3H)2
Co(HPO3H)2
Catalysts
Ni2P
103
130
0.36
0.43
13.9
13.4
125
146
0.43
0.50
14.2
13.6
CoP
Spent catalysts (after temperature test)
Ni2Pa
93
128
0.43
0.47
18.4
14.6
CoPa
Spent catalysts (after time test)
Ni2Pa
CoPa
120
146
0.57
0.50
19.8
13.8
With regards to the cobalt compound based catalyst, as reported
before [17], the Co 2p3/2 core-level spectra for the precursor present
a
Average values for spent catalysts.

260
A. Infantes-Molina et al. / Applied Catalysis A: General 390 (2010) 253–263
Fig. 5. TEM microsgraphs of reduced catalysts: (A) Ni2P and (B) CoP.
the main contribution at 781.6 eV assigned to Co2+ ions forming
Co(HPO3H)2. For the reduced catalysts (Fig. 6), the Co 2p3/2 core-
level spectra show a new contribution centred at ca. 777.9 eV due to
the presence of CoP. The P 2p3/2 core-level spectra (Fig. 6) present
the same features as those of Ni2P catalysts, being the P 2p signal
of CoP compound located at ca. 128.9 eV.
From Fig. 6, it is evident that a complete formation of Ni2P and
CoP does not occur after the reduction. Moreover, the Ni 2p and Co
2p signals are not well defined due to both, the low metal loading
and some surface passivation.
The Ni 2p, Co 2p and P 2p spectra for spent catalysts are also
depicted in Fig. 6, corresponding to the catalysts tested with a feed
containing 3000 ppm of quinoline and 2000 ppm of DBT. Moreover,
the corresponding binding energy values are lumped together in
Table 4. In all cases the binding energies hardly change. Nonethe-
less from the spectra plotted in Fig. 6 the intensity of Ni 2p, Co
2p and P 2p signals along with the relative intensity of each band
present important changes. The signals are better defined in CoP
catalysts than in Ni2P ones. There is greater cobalt and phosphorus
surface enrichment according to the higher particle size of cobalt
based catalysts observed from TEM. The general trend is that after
temperature test, a considerable increase of the intensity of the Ni
2p3/2, Co 2p3/2 and P 2p3/2 signals assigned to Ni2P and CoP phases
with regard to the component assigned to the precursor is observed,
revealing the formation of more active phase during the hydrotreat-
ing process, as previously reported [27]. However, after the stability
test, the intensity of the Ni 2p, Co 2p and P 2p signals decreases, indi-
cating a lower concentration of these elements on the surface by
a redispersion or by a superficial covering with organic molecules.
Moreover the relative intensity of the Ni 2p3/2 and Co 2p3/2 signals
arising from the Ni2P and CoP phases also decreases. The S 2p signal
is not detected in any case. According to literature data, the incor-
poration of sulfur into the phosphide particles to form the active
phosphosulfide phase is only superficial. Therefore the low metal
loading employed here leads to a lower superficial sulfur incor-
poration that makes the S 2p signal detection not possible. In this
regard, a similar catalyst with a twice higher metallic loading [30],
presented a noisy S 2p signal.
Table 4
Spectral parameters obtained by XPS analysis.
Sample
Binding energy (eV)
Nickel catalysts
Ni 2p3/2
Cobalt catalysts
P 2p3/2
Co 2p3/2
P 2p3/2
Ni2+
Ni2P
HPO3H−
PO34−
–
Ni2P
–
Co2+
CoP
HPO3H−
PO34−
CoP
Precursors
Ni(HPO3H)2
Co(HPO3H)2
Catalysts
Ni2P
856.8
–
134.1
133.1
781.6
781.7
–
134.0
133.1
–
–
856.2
852.6
134.3
129.2
CoP
777.9
134.2
128.9
Spent catalysts (after temperature test)
3000Q
–
–
–
–
133.1
133.1
133.1
133.1
134.4
134.7
134.6
134.5
129.1
129.2
129.2
129.0
781.9
781.5
781.5
782.0
778.2
778.0
778.3
778.0
133.1
–
–
134.2
134.1
133.4
134.3
128.6
128.9
128.6
128.2
3000Q200DBT
3000Q2000DBT
3000DBT
855.9
856.3
852.7
852.8
133.1
Spent catalysts (after stability test)
3000Q
856.0
–
856.1
856.5
852.7
–
853.1
853.0
133.1
133.1
133.0
133.1
134.5
134.5
134.5
134.5
129.3
128.9
129.1
129.0
781.8
781.7
782.0
782.1
777.9
777.9
778.2
777.9
133.1
–
133.1
133.1
134.3
133.7
134.3
134.2
128.0
128.1
128.6
128.6
3000Q200DBT
3000Q2000DBT
3000DBT

A. Infantes-Molina et al. / Applied Catalysis A: General 390 (2010) 253–263
261
Ni2P
Ni2p
Ni2 P
P2p
c
c
b
b
a
a
888
876
864
852
140
136
132
128
Binding Energy (eV)
Binding Energy (eV)
P2p
CoP
Co 2p
CoP
c
c
b
a
b
a
136
132
128
810
800
790
780
Binding Energy (eV)
Binding Energy (eV)
Fig. 6. XPS spectra for Ni2P and CoP reduced catalysts (a) and spent catalysts after temperature (b) and time (c) tests, with a feed containing 3000Q2000DBT.
Table 2 compiles the Ni/P and Co/P atomic ratios for reduced
and spent catalysts. Both Ni/P and Co/P ratios increases after the
catalytic reaction due to the formation of more active phase during
the test. Only nickel catalysts after temperature test, present a very
low value or even in some cases the Ni 2p signal is not detected.
Ni2P phase is more dispersed than CoP one and a redispersion of
such a phase could have happened during the temperature test. This
fact along with the covering with carbonaceous residues makes its
detection by XPS difficult.
of the sample to DBT molecules. Burns et al. have reported that
CoxPy plays a critical role in the resistance to sulfur incorporation
[23]. Moreover, some authors have observed that sulfur content in
the catalysts after the catalytic reaction decreases with increasing
the surface P/metal molar ratio, indicating that the surface phos-
phorus blocks sulfur incorporation into the catalysts [23], thus it
has been reported that the P-rich metal phosphide phase (i.e., CoP)
retains less sulfur in its structure than the more metal-rich phos-
phide phase (i.e., Co2P or Ni2P) [18,23] and the same conclusion
are extracted here. Although the sulfur retained does not affect the
HDS and HDN activities versus temperature reaction, the higher
percentages of sulfur present in the catalysts after the stability test
could be the responsible of the loss of activity found in the HDN
reaction, especially in the case of CoP catalysts (Table 1).
On the other hand, the percentages of nitrogen obtained by
CHNS elemental analysis do not follow a trend, the obtained values
ranged between 0.06 and 0.12%, higher than sulfur ones. Moreover
the nitrogen percentages obtained with cobalt catalysts are higher
than with nickel ones. Both catalysts treated with quinoline present
higher nitrogen contents, while by varying the amount the DBT
molecules in the feed, they all retained less nitrogen.
4.6. Chemical analysis (CNHS)
Sulfur content of spent catalysts was measured from CHNS ele-
mental analysis (Table 2). In all cases, Ni2P catalysts possess higher
sulfur levels than that obtained for CoP suggesting that a higher
formation of phosphosulfide species could be taking place on Ni2P
due to its higher activity; however we cannot determine what sul-
fur is forming such a phase or what sulfur is in adsorbed sulfided
molecule. The percentage of retained sulfur increases in so far as
the amount of DBT present in the feed does and is also higher after
the stability test, which is related to the higher time of exposure

262
A. Infantes-Molina et al. / Applied Catalysis A: General 390 (2010) 253–263
5. Discussion
sites where C–N bond cleavage occurs, and therefore diminishes
the HDN activity of these catalysts. The found percentages of nitro-
gen are lower when DBT molecules are present, just when catalysts
are less active in the HDN reaction, i.e., nitrogen retained on the
catalysts surface does not alter either HDS or HDN reaction.
Ni2P and CoP catalysts, with only 5 wt% of metal (Ni or Co),
supported on commercial silica and prepared from dihydrogen-
phosphite precursors, were very active in the HDS of DBT in the
presence and absence of quinoline molecules, where the HDS activ-
ity is similar to that found when the metal loading is twice [17,20].
In the HDN reaction, Ni2P is much more active than CoP, improving
the nitrogen removal when 200 ppm of DBT is present in the feed,
while a slight sulfur inhibiting effect is observed when increasing
the concentration of DBT in the feed (2000 ppm).
The characterization results reveal that nickel catalyst possesses
lower particle size than CoP one and these particles are more dis-
persed, as can be seen by TEM. XPS results confirm such a statement,
in such a way that the Ni 2p signal is much more noisy than Co 2p
one. A surface enrichment of cobalt phosphide is clearly observed
in Fig. 6, especially in the spent catalyst after the temperature test.
It has been reported in the literature the higher activity of Ni2P
with regards to CoP. At this point it can be stated that besides Ni2P
inherently activity in HDS and HDN reactions, the higher disper-
sion of Ni2P with regards to CoP also contributes to the excellent
behaviour of the Ni2P catalyst.
6. Conclusions
From the above results it can be concluded that: (i) Ni2P shows
a higher activity than CoP in the HDS and HDN reaction; (ii) the
catalytic activity and stability is higher in the HDS reaction than in
the HDN one; (iii) the competitive HDS–HDN reactions reveal that
Ni2P catalyst is highly active and there is no suppression of sulfur
removal by competitive effects; (iv) nitrogen removal is affected by
the presence of sulfur containing molecules; and (v) HDN and HDS
reaction must occur on different sites in a way that one does not
affect the other in term of selectivity.
Acknowledgements
We gratefully acknowledge the support from the Ministry of
Science and Innovation, Spain (MICINN, Spain) through the project
MAT2009-10481, the Regional Government (JA) through the Excel-
lence Projects (P06-FQM-01661 and P07-FQM-5070) and FEDER
funds. J.A.C.B. thanks the Ministry of Science and Innovation, Spain
for a fellowship (BES-2007-15735). One of us (AIM) also thanks the
MICINN, Spain, for a Juan de la Cierva contract.
The HDS reaction (Fig. 1) shows that at low T (below 450 ◦C) the
activity of Ni2P is higher than that observed for CoP. The activity
found for both catalysts with 200 ppm of DBT in the feed is similar.
From XPS data it can be clearly seen that during the catalytic reac-
tion more active phase is formed, and therefore by increasing the
temperature the activity of both catalysts does. The fact that with a
small concentration of DBT (200 ppm) both catalysts possess 90% of
conversion at 325 ◦C is ascribed to the presence of enough Ni2P or
CoP active sites as to convert only 200 ppm of DBT, in spite of being
present 3000 ppm of quinoline. From these data it can be said that
these catalysts are very active for deep HDS in the presence of a big
excess of nitrogen containing molecules.
References
[1] C. Song, Catal. Today 86 (2003) 211–263.
[2] P. Greening, Top. Catal 16/17 (2001) 5–13.
[3] G. Shan, H. Zhang, J. Xing, G. Chen, W. Li, H. Liu, Biochem. Eng. J. 27 (2006)
305–309.
[4] M. Rashtchi, G.H. Mohebali, M.M. Akbarnejad, J. Towfighi, B. Rasekh, A. Keytash,
Biochem. Eng. J. 29 (2006) 169–173.
The HDN reaction results turn out much more complex. In gen-
eral Ni2P is more active than CoP. The results found here present
changes in the selectivity patterns with regards to results previ-
ously reported by others. It has been stated that the relative rate of
each reaction depends on the reaction conditions and the relative
rates of the two hydrogenolysis reactions can influence the equi-
librium state. The majority product found for Ni2P catalyst is PB,
in agreement with the results of Oyama et al. [27], who reported
high PB selectivity values with similar conditions of pressure, tem-
perature and Ni/P atomic ratio. Independently of the amount of
DBT present in the feed, PB is the majority HDN product found
for Ni2P catalyst, i.e., we do not observe an alteration of the HDN
mechanism under the presence of DBT molecules. Nonetheless,
other authors underlined that the presence of H2S alters the HDN
mechanism [32,38]. Moreover it has also been reported that the
presence of phosphorus on a NiMo catalysts promotes the reac-
tion via THQ1-OPA [35], so the presence a phosphorus excess on
the catalyst surface seems to lead the reaction toward the partially
hydrogenated route.
The general trend observed here is that no change in the selectiv-
ity data is observed in HDS and HDN reactions with different DBT/Q
ratios. The behaviour of the catalysts prepared here suggests that
the active sites for C–N and C–S bond cleavage are different in a
way that one does not infer in the activity of the other. From CNHS
analysis it is obtained higher sulfur retention at higher DBT load-
ings especially after stability test, while the amount of nitrogen
retained does not follow a pattern. It is not possible to determine
the amount of phosphosulfide phase formed on each catalyst from
this sulfur percentages data, but it is clear that with 200, 2000 and
3000 ppm of DBT, the amount of such a phase should be the nec-
essary to remove DBT molecules almost its completion in the case
of Ni2P phase. However part of this sulfur could be blocking active
[5] S. Liu, B. Wang, B. Cui, L. Sun, Fuel 87 (2008) 422–428.
[6] L.H. Thompson, L.K. Doraiswamy, Ind. Eng. Chem. Res. 38 (1999) 1215–1249.
[7] H. Topsøe, B.S. Clausen, F.E. Massoth, Hydrotreating Catalysis – Science and
Technology, Springer, Berlin, 1996.
[8] S.E. Skrabalak, K.S. Suslick, J. Am. Chem. Soc. 127 (2005) 9990–9991.
[9] M.J.B. Souza, B.A. Marinkovic, P.M. Jardim, A.S. Araujo, A.M.G. Pedrosa, R.R.
Souza, Appl. Catal. A: Gen. 316 (2007) 212–218.
[10] Furimsky, Appl. Catal. A: Gen. 240 (2003) 1–28.
[11] M. Nagai, Appl. Catal. A: Gen. 322 (2007) 178–190.
[12] S. Yang, C. Liang, R. Prins, Stud. Surf. Sci. Catal. 162 (2006) 307–314.
[13] Y. Teng, A. Wang, X. Li, J. Xie, Y. Wang, Y. Hu, J. Catal. 266 (2009) 369–379.
[14] S.T. Oyama, P. Clark, X. Wang, T. Shido, Y. Iwasawa, S. Hayashi, J.M. Ramallo-
Lopez, F.G. Requejo, J. Phys. Chem. B 106 (2002) 1913–1920.
[15] V. Zuzaniuk, R. Prins, J. Catal. 219 (2003) 85–96.
[16] S.T. Oyama, J. Catal. 216 (2003) 343–352.
[17] J.A. Cecilia, A. Infantes-Molina, E. Rodríguez-Castellón, A. Jiménez-López, Appl.
Catal. B: Environ. 92 (2009) 100–113.
[18] S.J. Sawhill, D.C. Phillips, M.E. Bussell, J. Catal. 215 (2003) 208–219.
[19] S.T. Oyama, X. Wang, Y.-K. Lee, W.-J. Chun, J. Catal. 221 (2004) 263–273.
[20] J.A. Cecilia, A. Infantes-Molina, E. Rodríguez-Castellón, A. Jiménez-López, J.
Phys. Chem. C 113 (2009) 17032–17044.
[21] J.M. Lewis, R.A. Kydd, R.M. Boorman, P.H. Van Rhyn, Appl. Catal. A: Gen. 84
(1992) 103–121.
[22] C. Stinner, R. Prins, T. Weber, J. Catal. 191 (2000) 438–444.
[23] A.W. Burns, K.A. Layman, D.H. Bale, M.E. Bussell, Appl. Catal. A: Gen. 343 (2008)
68–76.
[24] S.T. Oyama, T. Gott, H. Zhao, Y.-K. Lee, Catal. Today 143 (2009) 94–107.
[25] S.J. Sawhill, K.A. Layman, D.R. Van Wyk, M.H. Engelhard, C. Wang, M.E. Bussell,
J. Catal. 231 (2005) 300–313.
[26] F. van Looij, P. van der Laan, W.H.J. Stork, D.J. DiCamillo, J. Swain, Appl. Catal.
A: Gen. 170 (1998) 1–12.
[27] S.T. Oyama, X. Wang, K.Y.-K. Lee Bando, F.G. Requejo, J. Catal. 210 (2002)
207–217.
[28] M. Lu, A. Wang, X. Li, X. Duan, Y. Teng, Y. Wang, C. Song, Y. Hu, Energy & Fuels
21 (2007) 554–560.
[29] T.I. Korányi, Z. Vít, D.G. Poduval, R. Ryoo, H.S. Kim, E.J.M. Hensen, J. Catal. 253
(2008) 119–131.
[30] J.A. Cecilia, A. Infantes-Molina, E. Rodríguez-Castellón, A. Jiménez-López, J.
Catal. 263 (2009) 4–15.
[31] F. Bataille, J.-L. Lemberton, P. Michaud, G. Pérot, M. Vrinat, M. Lemaire, E. Schulz,
M. Breysse, S. Kasztelan, J. Catal. 191 (2000) 409–422.

A. Infantes-Molina et al. / Applied Catalysis A: General 390 (2010) 253–263
263
[32] C.N. Satterfield, S. Gültekin, Ind. Eng. Chem. Process. Des. Dev. 20 (1981) 62–68.
[33] S. Yang, C.N. Satterfield, J. Catal. 81 (1983) 168–178.
[34] T. Kawai, K.K. Bando, Y.-K. Lee, S.T. Oyama, W.-J. Chun, K. Asakura, J. Catal. 241
(2006) 20–24.
[37] C. Stinner, Z. Tang, M. Haouas, Th. Weber, R. Prins, J. Catal. 208 (2002) 456–466.
[38] M. Egorova, R. Prins, J. Catal. 221 (2004) 11–19.
[39] A.E. Nelson, M. Sun, A.S.M. Junaid, J. Catal. 241 (2006) 180–188.
[40] A. Bertrand, J. Vac. Sci. Technol. 18 (1981) 28–33.
[35] M. Jian, R. Prins, J. Catal. 179 (1998) 18–27.
[41] T. Korányi, Appl. Catal. A: Gen. 239 (2003) 253–267.
[36] M. Jian, R. Prins, Stud. Surf. Sci. Catal. 113 (1998) 111–123.
[42] D.J. Sicinolfi, R.P. Frankenthal, J. Electrochem. Soc. 136 (1989) 2475–2480.