Hydrodesulfurization enhancement of heavy and light S-hydrocarbons on NiMo/HMS catalysts modified with Al and P

Article abstract of DOI:10.1016/j.apcata.2014.06.033

Deep hydrodesulfurization (HDS) of oil fractions is a key process in petroleum refineries due to the increasing demand for S-free diesel and gasoline fractions. In this work, a series of NiMo catalysts supported on Al- and P-modified HMS mesoporous substrate were tested in the HDS of thiophene and 4,6-dimethyldibenzothiophene (4,6-DMDBT) to evaluate the effect of Al and P additives on the catalyst response in HDS reactions (thiophene at 1 bar and 4,6-DMDBT at 5.5 MPa). The catalysts were characterized by a variety of techniques (XRD, N₂ adsorption-desorption, TPR, TPS, FT-IR of adsorbed pyridine, UV-vis and H₂ chemisorption). NiMo/Al-HMS-P catalyst containing 1.0 wt.% of P exhibited the best performance in both HDS reactions as a consequence of the proper balance between the active phase dispersion and the largest hydrogenation ability among the studied catalysts. The 4,6-DMDBT HDS reaction proceeded toward direct desulfurization (DDS) and hydrogenation (HYD) reaction routes. Concerning the HDS of thiophene, the increase of the catalyst acidity led to the formation of butadiene and diminished the hydrogenation of olefins (formation of butane). At the reaction temperature of 320 °C, the NiMo/Al-HMS-P1.0 catalyst exhibited the highest activity and the lowest butane formation among the catalysts studied. The isomerization of olefins in the thiophene HDS reaction did not occur, in line with the observed absence of isomerization in the 4,6-DMDBT HDS reaction.

Full text of DOI:10.1016/j.apcata.2014.06.033

Applied Catalysis A: General 484 (2014) 108–121
Contents lists available at ScienceDirect
Applied Catalysis A: General
journal homepage: www.elsevier.com/locate/apcata
Hydrodesulfurization enhancement of heavy and light
S-hydrocarbons on NiMo/HMS catalysts modified with Al and P
T.A. Zepedaa,∗, A. Infantes-Molinab, J.N. Díaz de Leóna, S. Fuentesa, G. Alonso-Nún˜eza,
G. Torres-Otan˜eza, B. Pawelecb
a Centro de Nanociencias y Nanotecnología, UNAM, Km. 107 Carretera Tijuana-Ensenada, CP, 22800 Ensenada, BC, Mexico
b Instituto de Catálisis y Petroleoquímica, CSIC, c/Marie Curie 2, Cantoblanco, 28049 Madrid, Spain
a r t i c l e i n f o
a b s t r a c t
Article history:
Deep hydrodesulfurization (HDS) of oil fractions is a key process in petroleum refineries due to the
increasing demand for S-free diesel and gasoline fractions. In this work, a series of NiMo catalysts sup-
ported on Al- and P-modified HMS mesoporous substrate were tested in the HDS of thiophene and
4,6-dimethyldibenzothiophene (4,6-DMDBT) to evaluate the effect of Al and P additives on the catalyst
response in HDS reactions (thiophene at 1 bar and 4,6-DMDBT at 5.5 MPa). The catalysts were charac-
terized by a variety of techniques (XRD, N2 adsorption–desorption, TPR, TPS, FT-IR of adsorbed pyridine,
UV–vis and H2 chemisorption). NiMo/Al-HMS-P catalyst containing 1.0 wt.% of P exhibited the best perfor-
mance in both HDS reactions as a consequence of the proper balance between the active phase dispersion
and the largest hydrogenation ability among the studied catalysts. The 4,6-DMDBT HDS reaction pro-
ceeded toward direct desulfurization (DDS) and hydrogenation (HYD) reaction routes. Concerning the
HDS of thiophene, the increase of the catalyst acidity led to the formation of butadiene and diminished
the hydrogenation of olefins (formation of butane). At the reaction temperature of 320 ◦C, the NiMo/Al-
HMS-P1.0 catalyst exhibited the highest activity and the lowest butane formation among the catalysts
studied. The isomerization of olefins in the thiophene HDS reaction did not occur, in line with the observed
absence of isomerization in the 4,6-DMDBT HDS reaction.
Received 6 March 2014
Received in revised form 29 June 2014
Accepted 30 June 2014
Available online 6 July 2014
Keywords:
Hydrodesulfurization
Thiophene
4,6-Dimethyldibenzothiophene
Catalysts
Mesoporous materials
Heavy and light S-hydrocarbons
© 2014 Elsevier B.V. All rights reserved.
1. Introduction
former compounds are very difficult to desulfurize because of
the steric hindrance of their alkyl groups, especially when both
During the past years the challenge for producing enough hydro-
carbon fuels to fulfill the worldwide increasing transportation
needs are accentuated by the environmental regulations regarding
the sulfur specifications for diesel and gasoline, which have become
stricter, and will continue to do so [1–3]. In order to fulfill such
requirements, one of the strategies is hydrodesulfurization (HDS)
of petroleum feedstock’s [4]. This cannot achieved with the classical
Ni(Co)Mo(W)/Al2O3 hydrotreating catalysts due to their thermo-
dynamic limitations for simultaneous deep hydrodesulfurization
(HDS) and hydrodearomatization (HDA) at high temperature oper-
ation [5]. Thus, it is urgent to develop more efficient catalysts with
significantly higher HDS activities than the classical ones [6].
To achieve deep desulfurization of diesel and gasoline, the
refractory alkyl-substituted dibenzothiophenes and thiophenic
compounds, respectively, should be removed. In particular the
are in the 4- and 6-positions, for the adequate interaction of the
S-atom with the catalytic active site [6]. Since the transforma-
tion of those compounds via direct desulfurization reaction route
is limited, alternative strategies for elimination of planar config-
uration of alkyl-substituted DBT involve hydrogenation of their
aromatic rings, dealkylation, isomerization and/or C C bond scis-
sion reactions [7]. Contrary to diesel fractions, deep HDS of FCC
gasoline is usually accomplished by the use of Ni(Co)Mo(W)/Al2O3
sulfide catalysts. Unfortunately, as the degree of desulfurization
increases, the research octane number (RON) of gasoline products
decreases due to the hydrogenation of olefins as well as of aromat-
ics present in the feed [8]. Thus, on one hand, sulfur removal from
gasoline range (C4–C10) streams without hydrogenation of olefins
remains a significant challenge [8,9]. On the other hand, to enhance
the octane number of the gasoline pool, the olefin chain-branching
(isomerization) is needed.
It is known that olefins isomerization during HDS of thiophene
(through classical or non-classical carbenium chemistry) requires
support’s acid sites [10]. Thus, to improve the catalyst acid function,
∗
Corresponding author. Tel.: +52 6461750650; fax: +52 6461750640.
E-mail addresses: trino@cnyn.unam.mx, tzepeda@hotmail.com (T.A. Zepeda).
http://dx.doi.org/10.1016/j.apcata.2014.06.033
0926-860X/© 2014 Elsevier B.V. All rights reserved.

T.A. Zepeda et al. / Applied Catalysis A: General 484 (2014) 108–121
109
Table 1
research activities are directed toward the use of strongly acidic
Labeling and nominal composition of the synthesized samples.
supports, such as medium pore ZSM-5 [8] or large pore USY zeolites
[9], which demonstrated to be effective for acid-catalyzed reactions
such as isomerisation, cracking, hydrotreating and/or alkylation.
Concerning the novel supports, the potential application of
highly ordered mesoporous materials, such as MCM-41, SBA-15,
HMS or SBA-16, etc., for supporting hydrotreating catalysts has
been recently intensively studied [11–13]. This is because their
interesting textural properties, such as well-defined uniform meso-
pores and high surface area, which are particularly attractive for
reactions involving large organic molecules [11,12]. Unfortunately,
those purely siliceous materials have no Brønsted acidity and
limited capacity for ion-change [13]. Thus, considerableefforts have
been applied to create acid sites by the substitution of Si4+ atoms by
Al3+ ions in the pore walls of HMS [14–21], SBA-16 [20,21], MCM-
41 [22–28], and SBA-15 [29–35] materials. Among those materials,
the hexagonal mesoporous silica (HMS) is of particular interest
because it was found that HMS doped with Al, Ti, Zr, etc., exhib-
ited a larger HDS activity than MCM-41 ones in the HDS of DBT
[28]. Indeed, the NiMo and CoMo sulfide catalysts supported on Al-
HMS demonstrated to be more active in the HDS reaction than their
Al-free counterparts [15]. The activity enhancement was explained
in terms of the structure and specific electronic properties of the
supported active species [15].
Catalysts
Catalyst labeling
Mo (wt.%)
Ni (wt.%)
P (wt.%)
NiMo/Al-HMS 0.0
NiMo/Al-HMS 0.5
NiMo/Al-HMS 1.0
NiMo/Al-HMS 1.5
NiMo/Al-HMS 2.0
NiMo-0.0
NiMo-0.5
NiMo-1.0
NiMo-1.5
NiMo-2.0
9.0
9.0
9.0
9.0
9.0
3.0
3.0
3.0
3.0
3.0
0.0
0.5
1.0
1.5
2.0
Al-HMS solid was air-dried at room temperature for 24 h and sub-
sequently at 110 ◦C for 2 h. Finally the samples were calcined in air
at 500 ◦C for 4.5 h (heating rate 2.5 ◦C min−1).
Phosphorus was incorporated into Al-HMS by impregnation
with an aqueous solutions of H3PO4 using the pore filling method.
Thus, an aqueous solution of phosphoric acid (H3PO4, Fluka 85 wt.%
in water) with the corresponding concentration of phosphoric acid
was added to the support. These impregnates were dried at room
temperature for 16 h, and subsequently at 110 ◦C for 2 h. Then they
were calcined at 500 ◦C for 4.5 h (heating rate 2.5 ◦C min−1).
2.2. Catalyst preparation
The catalysts were prepared by simultaneous impregnation
using the pore filling method. An aqueous solution of ammonium
heptamolybdate ((NH4)6Mo7O24, Aldrich 99%) and nickel nitrate
(Ni(NO3)2, Aldrich 98%), with an appropriate concentration of Ni
and Mo, was added to the support. The impregnates were dried at
room temperature for 18 h, then at 110 ◦C for 2 h and finally calcined
at 500 ◦C for 4.5 h (heating rate 2.5 ◦C min−1). Nominal composition
and catalysts labeling are given in Table 1.
Similarly to the support modification with Al3+, the increase
of the support acidity could be achieved also by support graft-
ing with phosphate [36–41]. Indeed, in the past, this method
was the most frequently employed to improve HDS activity of
Co(Ni)–Mo(W)/Al2O3 sulfide catalysts [36–41]. However, com-
pared to alumina [36], the effect of P incorporation into the
mesoporous siliceous substrate was much less studied and the final
conclusions are contradictory [42–49]. In addition, there are also
some works reporting no effect, or even negative effect, of phos-
phorus in the HDS reaction [43,44]. This is probably because the
final effect of P-doping is the result of various cumulative effects
such as its loading, catalyst preparation method, precursor of active
phases and reaction conditions employed. Moreover, the increase
of support acidity strongly depends on P loading being the largest
acidity achieved for P loadings close to 1.0 wt.% [36–42].
2.3. Characterization methods
2.3.1. X-ray diffraction (XRD)
The X-ray patterns of the calcined and sulfided catalysts were
recorded on a Rigaku 2100 diffractometer, using monochromatic
CuK␣ radiation (ꢀ = 0.1541 nm). The diffractograms were recorded
in the 2ꢁ range of 0.15–80◦ at a step of 0.02◦.
As compared with the support modification with one element
or group, the effect of support modification with two additives
is scarcely reported [45,46,48]. In this sense, our previous study
demonstrated that the post-synthesis modification of Ti-HMS sup-
port with phosphorous led to enhancement of dealkylation and
isomerization routes of 4,6-DMDBT transformation over CoMo/Ti-
HMS-P sulfide catalysts [45]. This prompted us to study the effect
of Al-HMS support modification with P on the HDS activity of sul-
fided NiMo catalysts in deep HDS reactions of model compounds
(thiophene and 4,6-DMDBT). To examine the effect of phosphorus
concentration on the dispersion of nickel and molybdenum species,
and the type of structures formed in the oxide and sulfide form of
the P-containing catalysts, the supports and catalysts were char-
acterized by different techniques (XRD, N2 adsorption–desorption
isotherms, DRS UV–vis, TPS, H2-TPR, H2-chemisorption capability,
FTIR spectroscopy of adsorbed pyridine) and tested in the reac-
tions of HDS thiophene and 4,6-DMDBT reactions performed at
atmospheric and high hydrogen pressures.
2.3.2. N2 adsorption–desorption isotherms
The textural properties of the supports and oxidic catalysts were
determined by N2 adsorption–desorption isotherms recorded at
−196 ◦C with an ASAP 2000 Micromeritics equipment. The cata-
lysts were sulfided ex-situ in a U-shape glass reactor under the
same conditions described in the catalytic activity measurement
section. After sulfidation, the sample was cooled to room tem-
perature under N2 flow and transferred into the Micromeritics
apparatus in an Ar atmosphere without exposing the sample to air.
Prior to the experiments, the supports and sulfided catalysts were
degassed at 270 ◦C for 5 h. The volume of adsorbed N2 was normal-
ized to standard temperature and pressure. Specific surface area
(SBET) was calculated by applying the BET equation to the range of
relative pressures 0.05 < P/P0 < 0.30. The average pore diameter was
calculated following the Barret–Joyner–Halenda method (BJH) and
using the adsorption branch of the N2 isotherm. The cumulative
pore volume was obtained from the isotherms at P/P0 = 0.99. The
normalized BET area was calculated using the equation:
SBET catalyst
(1 − y) × SBET support
2. Experimental
NSBET
=
(1)
2.1. Synthesis of the supports
where NSBET is the normalized SBET and y is the weight fraction of
the guest phases.
The Al-HMS support (Si/Al molar ratio of 40) was prepared fol-
lowing the procedure described by Gontier and Tuel [50] using
dodecylamine (C12H25NH2, Aldrich 98%) as surfactant and alu-
minum isopropoxide (98%, Aldrich) as aluminium precursor. The
2.3.3. UV–vis diffuse reflectance spectroscopy
The UV–vis diffuse reflectance spectra (DRS UV–vis) of the bare
supports and oxide catalyst precursors were recorded using a Cary

110
T.A. Zepeda et al. / Applied Catalysis A: General 484 (2014) 108–121
5E spectrophotometer with a specially designed Praying Mantis dif-
fuse reflection attachment (Harrick) for in situ measurements. All
spectra were recorded after heating the samples at 250 ◦C in He
flow for 1 h. The spectrum of the corresponding support was sub-
tracted from the spectrum of the catalyst. Decomposition of each
spectrum was performed by non-linear fitting of multiple Gaussian
peak functions sharing a common baseline.
catalyst activation by sulfidation, the catalyst was dried under a N2
flow of 100 mL min−1 at 150 ◦C for 0.5 h. Then, the sample was sul-
fided with heating up to 400 ◦C, at a heating rate of 4 ◦C min−1, in a
H2/H2S gas mixture (15% v/v of H2S) and kept at this temperature
for 2 h. After sulfidation, the catalyst was purged with N2 at 150 ◦C
for 0.5 h to eliminate H2S, which could be adsorbed on the catalyst
surface. The reaction was carried out at atmospheric pressure and
at different temperatures: 300, 320 and 340 ◦C. Before the experi-
mental run, the catalytic bed was heated to the desired reaction
temperature. Once the test was performed at a given tempera-
ture, then the temperature was increased to the next temperature
while maintaining the catalyst in a flow of inert gas. Meanwhile, the
saturation of hydrogen with thiophene was obtained by bubbling
hydrogen (50 mL min−1) through a saturator containing thiophene
liquid at 0 ◦C. For each catalyst studied, steady state conditions were
reached after 1 h of time on-stream reaction. Reaction products
were analyzed (by) with an online gas chromatograph (HP-7820,
FID) equipped with a CP-Sil 5 CB column. A conventional indus-
trial NiMo/Al2O3 catalyst was tested under the same experimental
conditions. This industrial reference sample has a chemical compo-
sition of 12, 4 and 2.4 wt.% of Mo, Ni and P, respectively. The textural
2.3.4. Temperature-programmed reduction (H2-TPR)
TPR experiments of the oxide catalyst precursors were con-
ducted in a Micromeritics 2900 equipment. Prior to reduction, the
catalysts (ca. 50 mg) were heated at a rate of 20 ◦C min−1 up to
400 ◦C, and kept for 2 h under a flow of He to remove water and
other contaminants. The catalysts were cooled to room tempera-
ture in the same He flow; then reduced in flowing 10% H2/Ar gas
mixture (50 mL min−1) from r.t. until 1000 ◦C at a heating rate of
15 ◦C min−1
.
2.3.5. Temperature programmed sulfidation (TPS)
The TPS experiments of the oxide precursors were carried out
on Micromeritics ChemiSorb 2720 apparatus, equipped with TCD
and UV detectors. Prior to sulfidation, the catalyst (ca. 50 mg) was
heated from r.t. until 300 ◦C at a heating rate of 20 ◦C min−1 and
kept at that temperature for 2 h under a flow of He to remove water
and other contaminants. Subsequently, the catalysts were cooled to
room temperature in the same He flow, and sulfided with heating
at a linear temperature ramp (15 ◦C min−1) in flowing 5% H2S/H2
gas mixture (50 mL min−1) from r.t. 600 ◦C.
properties for the reference sample are the follows: 215 m2 g−1
,
0.45 cm3 g−1 and 7.6 nm of SBET, cumulative pore volume and aver-
age pore diameter, respectively.
2.4.2. 4,6-Dimethyldibenzothiophene hydrodesulfurization
The catalytic activity was evaluated in the reaction of HDS of 4,6-
DMDBT (300 ppm of S) carried out in a batch Parr reactor (300 mL
capacity) charged with 0.2 g of catalyst (particle size between
−80/+100 mesh) and 0.3 g of 4,6-DMDBT dissolved in 100 mL of
n-dodecane. The reaction was carried out at 320 ◦C under a total
H2 pressure of 5.5 MPa for 6 h. Before the activity test, the cata-
lyst was sulfided in a U-shape glass flow reactor. First the sample
was flushed in a nitrogen flow gradually increasing the tempera-
ture from room temperature up to 150 ◦C for 0.5 h. Then, the sample
was sulfided with a 15% v/v of H2S gas mixture (60 mL min−1) from
150 ◦C up to 400 ◦C (heating rate of 4 ◦C min−1), and kept at this
temperature for 2 h. After sulfidation, the catalyst was purged with
2.3.6. H2-Chemisorption measurements
The dispersion of the active phase was determined from the
amount of chemisorbed H2 measured with a pulse method using
a Micrometrics ChemiSorb 2720 apparatus. Prior to chemisorption
measurements the oxide samples (ca. 50 mg) were reduced in situ
in an H2/Ar stream at 400 ◦C for 2 h, and then cooled in an Ar flow
in order to remove physisorbed hydrogen. Pulses (0.113 mL) of 10%
H2/Ar were injected into a stream of Ar carrier gas (50 mL min−1
)
and contacted with the catalyst at room temperature. Metal disper-
sion was calculated assuming that one Ni and Mo atom chemisorb
one hydrogen atom.
N
2 at 150 ◦C for 0.5 h to eliminate H2S, which could be adsorbed on
the catalyst surface. After cooling down to room temperature, the
sulfided sample was transferred to the batch reactor in an argon
atmosphere with the aim to avoid contact with air. The reaction
products were analyzed by GC on a Perkin-Elmer XL equipment
using 30 m capillary column coated with a non-polar methyl sil-
icone phase (DB-1, J & W). For comparison purpose, the activity
of a conventional industrial NiMo/Al2O3 catalyst was tested under
the same experimental conditions. This industrial reference sam-
ple has a chemical composition of 12, 4 and 2.4 wt.% of Mo, Ni
and P, respectively. The textural properties for the reference sam-
2.3.7. FTIR spectra of adsorbed pyridine
The FT-IR measurements were performance in a Tensor 27
Bruker spectrophotometer. Self-supporting wafers of the oxide cat-
alyst precursor with a thickness of 12 mg cm−2 were prepared by
pressing (7 × 103 kg cm−2) the powdered sample during 10 min.
The wafer was introduced into a special IR cell having greaseless
stopcocks and KBr windows. Then the sample was sulfided in situ
under the same conditions described in the catalytic activity mea-
surement section. The sulfided samples were outgassed at 450 ◦C
for 2 h and cooled down to 120 ◦C prior to contact with ca. 2 mbar
of pyridine. Then, IR spectrum was recorded after evacuation of
physically adsorbed pyridine (10−5 mbar) at 120 ◦C for 0.5 h.
ple are the follows; 215 m2 g−1, 0.45 cm3 g−1 and 7.6 nm of SBET
,
cumulative pore volume and average pore diameter, respectively.
3. Results and discussion
2.3.8. High-resolution transmission electron microscopy (HRTEM)
HRTEM micrographs were collected on a JEOL-2010F instru-
ment. The samples were suspended in heptane as solvent in order
to be deposited on lacey carbon (440 mesh) Cu grid holders.
3.1. Textural properties
The values of specific BET surface area (SBET), total pore volume
and pore diameter of bare supports and oxide catalyst precursors
are listed in Table 2. Fig. 1 shows the influence of P loading on the
specific BET surface area and total pore volume of the bare supports
and oxide catalyst precursors. For the bare substrates, Al-HMS P-
loaded samples suffer a decrease in the specific BET area (SBET) and
pore volume suggesting a partial blockage of pores by P species.
As expected, the NiMo-2.0 sample shows the largest decrease of
SBET and total pore volume among the studied catalysts. A further
2.4. Catalytic activity measurements
2.4.1. Gas-phase thiophene hydrodesulfurization
The HDS of thiophene was carried out in a vapor phase using
a fixed bed micro flow reactor (15 mm ID) housed in a furnace. A
quartz reactor was loaded with 100 mg of catalyst (particle size
between the 80 and 120 mesh) diluted with 1 g of SiC. Prior to the

T.A. Zepeda et al. / Applied Catalysis A: General 484 (2014) 108–121
111
Table 2
Textural properties of pure supports and oxide precursors as determined by N2 adsorption desorption isothermsa at −196 ◦C and low-angle XRD diffraction patternsb.
Sample
SBET (m2 g−1
)
dp (nm)
Vp (cm3 g−1
)
NSBET
d100 (nm)
a0 (nm)
wt (nm)
Supports
Al-HMS-P 0.0
Al-HMS-P 0.5
Al-HMS-P 1.0
Al-HMS-P 1.5
Al-HMS-P 2.0
Oxide precursors
NiMo-0.0
NiMo-0.5
NiMo-1.0
NiMo-1.5
NiMo-2.0
1013
1021
994
972
901
3.1
3.1
3.0
3.0
2.9
1.17
1.01
0.98
0.93
0.88
–
7.18
6.54
6.40
6.74
6.13
8.29
7.55
7.39
7.78
7.08
5.19
4.45
4.39
4.78
4.18
1.00
0.99
0.97
0.90
796
723
691
668
604
3.1
3.0
3.0
3.1
2.7
1.01
0.94
0.89
0.83
0.75
0.89
0.80
0.79
0.78
0.76
6.35
6.54
6.09
5.77
5.42
7.33
7.55
7.03
6.66
6.25
4.23
4.55
4.03
3.56
3.55
a
SBET: specific surface area; dp: pore diameter; Vp: total pore volume; NSBET: normalized BET surface area calculated using Eq. (1).
b
d100, d-Spacing as calculated from the X-ray patterns; a0, unit cell parameter estimated from the position of the (100) diffraction line and calculated employing equation
√
ao = d100 × 2/ 3; wt, wall thickness was obtained by subtracting pore diameter (dp) from the unit cell parameter.
1100
of pure supports and oxide precursors are shown in Fig. 2(A) and
120
110
100
90
(B), respectively. In all cases, the diffractograms present a broad
reflection peak at 2ꢁ = 1.0–2.0◦ that can be indexed to the d100
reflection of a 2D hexagonal structure that is characteristic of the
HMS systems and therefore indicative of the maintenance of the
mesoporous structure after the impregnation and calcination pro-
cesses. The comparison of patterns of oxide catalyst precursors
reveals that the diffraction peak becomes less intense and broader
and shifts to higher angles, from 1.37◦ in NiMo-0.0 to 1.57◦ in NiMo-
2.0. These data indicate a partial collapse of the HMS mesoporous
structure upon P, Ni and Mo incorporation. In this regard, by cal-
culating the d-spacing values of the samples (Table 2), it decreases
with the P content and in a greater extent after the incorporation
of Ni and Mo species. Thus, the d100 values change from 7.18 nm
for the support without P (Al-HMS-P0.0) to 6.13 nm for the support
containing the highest P loading (Al-HMS-P2.0). A similar trend is
observed for the oxide catalyst precursors. The a0 and wall thick-
ness present the same tendency. For pure supports, a decrease of
the cell parameter and wall thickness with an increase of P load-
ing suggest some collapse of the support structure upon support’s
treatment with phosphoric acid (Table 2). Similarly, the incorpo-
ration of the metal oxides followed by calcination led to a further
decrease of both unit cell parameter and wall thickness with respect
to the original support values indicating some collapse of pore
structure.
1000
900
800
700
600
support
support
80
oxide precursor
oxide precursor
70
60
50
0,0
0,5
1,0
1,5
2,0
2,5
P loading (wt.%)
Fig. 1. Influence of P loading on the specific BET surface area and total pore volume
of the pure Al-HMS-P supports and oxide NiMo/Al-HMS-P catalyst precursors.
decrease in the SBET is observed after Ni and Mo incorporation due
to a progressive blockage of the support’s pores by the metal oxide
particles formed during calcination.
To get an idea about the location of the guest phases in the
support structure, the normalized NSBET of the supports and the
corresponding catalysts have been calculated (Eq. (1)), and the
corresponding values are included in Table 2. The values corre-
sponding to pure supports ranged from 0.9 to 1.0 pointing to a
homogeneous location of P species on both the support surface
and within the inner porous structure. After Ni and Mo addition,
a further decrease is observed. The catalyst without P, NiMo-0.0
presents the highest value 0.89, indicating the main location of
the active phase in the outer surface. By increasing the P loading,
the catalysts containing 0.5, 1.0, 1.5 and 2.0 possess lower values
although the differences between them are very low. These data
suggest that P favors the incorporation of the active phase into the
mesoporous structure. The lowest NSBET value recorded for the cat-
alyst containing the highest P-loading suggests that Ni–Mo species
are located not only inside the pores but also at the entrance of the
mesopores.
3.3. DRS UV–vis spectra
Information about the distribution of Mo oxide species on the
supports, and therefore their dispersion, was also evaluated by
UV–vis DRS spectra (Fig. 3). The spectrum of each support was
subtracted from the spectrum of the corresponding NiMo cata-
lyst. The bands in the 260–280 nm range are due to O2−–Mo6+
ligand-to-metal charge transfer transition (LMCT) in tetrahedral
coordination of isolated molybdate species (MoTd); the bands in
the 300–370 range are due to polymolybdate octahedral (Oh) Mo
species (MoOc); moreover, both types of Mo6+ species show a sec-
ond absorption band at about 230 cm−1 [51]. As seen in Fig. 2, all the
catalysts show a similar profile, with a strong absorption band at ca.
230–240 cm−1 indicative of the coexistence of a mixture of Mo6+
species in octahedral and tetrahedral coordination. Deconvolution
of all these spectra gives rise to two absorption bands located at ca.
244 and 302 cm−1 whose relative intensity depends on the P con-
tent. In all cases, mixtures of MoTh and MoOh are present, although
the proportion of tetrahedral Mo species is much more important
than octahedral ones. In the NiMo-0.0 sample the contribution of
MoTh species is the most important while a tiny absorption band
due to MoOh is observed. Upon P incorporation, the band due to
3.2. XRD
As the HMS substrate shows wormhole hexagonal ordered
mesoporous structure [50] low angle XRD analyses were per-
formed to reveal whether or not this structure is preserved after
support treatment with phosphoric acid and its further impreg-
nation with metal oxide precursors. The low-angle XRD patterns

112
T.A. Zepeda et al. / Applied Catalysis A: General 484 (2014) 108–121
Fig. 2. Low-angles X-ray diffraction patterns of the pure supports (A) and oxide catalyst precursors (B).
MoTh species becomes narrower and shifts to lower wavelengths,
while the contribution of MoOh ones increases in intensity, becom-
ing a distinguishable shoulder of the MoTh absorption for NiMo-1.0,
NiMo-1.5 and NiMo-2.0 samples. As the position of the absorption
bands is related to the coordination and agglomeration degree of
Mo species, it can be inferred that MoTh species are better dispersed
over these samples, besides they also present a greater proportion
MoOh ones. The exception is NiMo-2.0 which in spite of presenting
a greater amount of MoOh species, the peak positions are similar to
that of P-free sample, indicating that the dispersion is worse than
for the other P-containing catalysts.
3.4. H2-TPR experiments
231 nm
5
5
.
1
-
244 nm
o
The influence of P on the reducibility of nickel and molybde-
num species was studied by H2-TPR. The corresponding profiles are
depicted in Fig. 4. The TPR profiles of all NiMo/Al-HMS-P(x) samples
show two main reduction peaks. The first one located at 527 ◦C in
all the samples, and a second one at higher temperatures, whose
position depends on P content. Thus, for the P-free sample this sec-
ond peak is located at 745 ◦C but it shifts to higher temperature
upon increasing P-loading. In addition, another band is observed
at intermediate temperatures, ca. 600 ◦C, and seen as a shoulder of
the low temperature peak. This band becomes better defined for
the P-free sample (NiMo-0.0). The peak located at 527 ◦C can be
ascribed to the reduction, Mo6+ to Mo4+, of polymeric octahedral
Mo species, MoOh, weakly bounded to the support surface [51,52].
The position of this peak does not change under phosphorous pres-
ence but becomes narrower and more intense, except for NiMo-2.0
sample. These data suggest that the interaction of MoOh species
does not depend on the P presence but their amount increases, as
previously observed by UV–vis DRS spectra. The shoulder at 600 ◦C
has been previously ascribed to reduction of Ni2+ ions in tetrahe-
dral coordination (␣-NiMoO4 phase) [15] taking into account that
the reduction peak of Ni2+ ions in octahedral coordination overlaps
with that of Mo6+. This contribution is less important when P is
present, suggesting that P inhibits to a some extent the formation
M
Ni
.
0
-
o
M
Ni
Ni
0
.
2
-
o
M
0
.
0
-
o
M
Ni
0
.
1
-
o
M
Ni
302 nm
200
250
300
350
400
Wavelength (nm)
Fig. 3. DRS UV–vis spectra of the NiMo/Al-HMS and NiMo/Al-HMS-P(x) oxide cata-
lyst precursors.

T.A. Zepeda et al. / Applied Catalysis A: General 484 (2014) 108–121
113
Fig. 4. (a) TPR profiles of the NiMo/Al-HMS-P(x) oxide precursors. (b) Gaussian deconvolution of the TPR profiles of NiMo-2.0 and NiMo-0.0 samples.
of Ni2+ in tetrahedral coordination. Finally, the reduction peak at
higher temperature (750 ◦C) corresponds to the second reduction
step of polymeric octahedral molybdenum species (Mo4+ to Mo0)
and to the first step of reduction of isolated tetrahedral molybdates
(MoTh) strongly interacting with the support [53]. This reduction
process is less important and shifts to higher temperatures under
the presence of P, from 750 ◦C for NiMo-0.0 to 790 ◦C for NiMo-2.0.
These data indicate that P favors the formation of a higher amount
of MoOh species, and therefore it reduces the proportion of unre-
ducible Mo species (MoTh), even though they become reduced at
higher temperatures.
and overlaps with the sulfidation process. The second H2S con-
sumption peak at around 430 ◦C corresponds to the sulfidation
process of MoO2 to produce the MoS2 active phase. The TPS patterns
of the catalysts containing P showed more or less the same general
behavior as the described for P-free catalyst, nevertheless some
differences aroused as the P loading varied. In the case of the NiMo-
0.5 catalyst three peaks were observed, suggesting the presence of
MoO3 particles that exhibit different interaction strengths with the
support. When P was increased to 1.0 wt.% both an increase in the
amount of H2S consumption and a decrease in the temperature of
the first consumption peak were observed. This indicates a larger
amount of MoS2 sulfided species with weaker active phase-support
interactions. For the NiMo-1.5 catalyst the observed TPS pattern
was very similar to that of the NiMo-1.0 sample but a consider-
ably lower H2S consumption was observed in the low temperature
peak. Finally, the NiMo-2.0 catalyst showed two main peaks located
around 270 ◦C and 400 ◦C. Again, this shift to lower temperatures
indicates a weaker metal-support interaction than that observed
for the P-free catalyst.
3.5. TPS experiments
Fig. 5 shows the H2S consumption profiles observed during TPS
experiments of supports (dashed lines) and oxide precursors (con-
tinuous lines). All supports presented more or less the same general
behavior. Physically adsorbed H2S was released at temperatures
above 100 ◦C. The maximum of this peak was observed at 150 ◦C
for P-free, NiMo-0.5 and NiMo-2.0 catalysts but it shifted to higher
temperatures for NiMo-1.0 and NiMo-1.5 ones. All the supports
presented a broad H2S consumption band at around 400 ◦C.
3.6. Characterization of sulfided catalysts
The TPS of the oxide precursors also presented a desorption peak
similar to that observed for the P-free sample. The total amount of
consumed H2S increased with P-content catalysts until 1.0% and
then it decreased for NiMo-1.5 catalyst after which the amount of
consumed H2S again increased for NiMo-2.0 sample. All catalysts
exhibited at least two peak maxima. For the P-free material the first
consumption peak at around 346 ◦C is attributed to the exchange of
sulfur for oxygen in MoO3 to produce MoO2S and water [54]. After
this process a decomposition–reduction process of the oxysulfide
species formed has been reported to produce MoO2. However it
cannot be observed directly from the depicted TPS patterns since it
should be accompanied by a negative H2S consumption peak aris-
ing from the hydrogenation of elemental sulfur and the resulting
H2S production, thus indicating that this band is not well resolved
3.6.1. XRD
XRD measurements were used to study the crystalline phases
of sulfided catalysts. The diffraction patterns of sulfide catalysts in
Fig. 6 shows the same general features and reveal that the structure
of HMS substrate apparently is not modified by P introduction.
Moreover, no diffraction peaks corresponding to any phosphate
phase are observed. All samples present a broad peak between
18 and 32◦ (2ꢁ) which is attributed to the amorphous part of the
Al-HMS supports. Diffraction peaks located at 14.4, 33.1, 36.0,
39.5, 49.8 and 58.3◦ (2ꢁ) belong to MoS2 crystallites (JCPDS card
37-1492). Upon increasing P content until 1 wt.% the diffraction
peaks of MoS2 are less intense than in the P-free sample suggesting
a loss of crystallinity and a better dispersion of the MoS2 crystallites
formed. The crystallite sizes were estimated using the Scherrer

114
T.A. Zepeda et al. / Applied Catalysis A: General 484 (2014) 108–121
Fig. 5. TPS profiles of the oxide catalyst precursors.
equation for the highest intensity diffraction peak located at 14.4◦
corresponding to the (0 0 2) plane. Crystallite sizes increased in
the order: NiMo-1.0 (12.8 nm) < NiMo-2.0 (17.1 nm) < NiMo-0.5
(20.8 nm) < NiMo-0.0 (25.7 nm). No diffraction peaks of other
crystalline than MoS2 phase were recorded.
Table 3
Chemisorption capacity and metal dispersion of the NiMo-P(x) freshly sulfided cat-
alysts as determined by H2 chemisorption.
−1
Sample
H2 uptakea (␮mol gcat
)
Dispersiona (%)
NiMo-0.0
NiMo-0.5
NiMo-1.0
NiMo-1.5
NiMo-2.0
73
96
112
105
66
13.1
13.8
17.2
16.8
8.7
3.6.2. H2-chemisorption measurements
The dispersion of the oxide Ni-Mo species exposure was deter-
mined by H2 chemisorption measurements. The obtained values
are included in Table 3. As observed from XRD measurements, the
presence of phosphorous improves the Ni–Mo dispersion. Thus an
increment of P content until 1.0–1.5 wt.% increases the metallic dis-
persion from 13% to 17%. However, further addition of P (2.0 wt.%)
provokes a drastic decrease of metallic dispersion to 8.7%. These
results are in agreement with the tendency observed in the textural
properties, where no important changes in dispersion are observed
until 1.0 wt.% of P, on the contrary dispersion drops severely upon
increasing P-loading.
a
Chemisorption capacity as determined by pulse method of H2 chemisorption in
the oxidized catalysts.
3.6.3. FT-IR spectra of adsorbed pyridine
FT-IR spectra of adsorbed pyridine on sulfide catalysts were
recorded with the objective to reveal acidity differences between P-
free and P-containing samples. Fig. 7 shows the infrared spectra of
the pyridine adsorbed on P-free and P-containing sulfide catalysts.
All catalysts show similar spectra in the range 1400 to 1650 cm−1
.

T.A. Zepeda et al. / Applied Catalysis A: General 484 (2014) 108–121
115
Table 4
Effect of support functionalization with PO3H2 groups of NiMo-P(x) sulfided cat-
alysts on the amount of Brønsted and Lewis acid sitesa.
−1
−1
)
Catalyst
Brønsted (␮mol gcat
)
Lewis (␮mol gcat
NiMo-0.0
NiMo-0.5
NiMo-1.0
NiMo-1.5
NiMo-2.0
0.4
0.2
0.2
0.2
1.1
2.8
11.0
13.7
13.4
15.9
a
As determined by FTIR spectroscopy of adsorbed pyridine.
Lewis acid sites. Thus, from FT-IR spectra the quantitative anal-
ysis of Brønsted and Lewis acidity was done as reported in [57]
and the corresponding amount of each acid site-type Lewis (L) and
Brønsted (B)) are reported in Table 4. The amount of Brønsted acid
sites was much lower than the estimated concentration of Lewis
sites, regardless the P content in the catalyst and always lower
under the presence of P, with the exception of NiMo-2.0 sample.
Instead, a considerable increase in Lewis acidity occurs after the
addition of 0.5 wt.% of P to the support. Further addition of P also
increases the population of Lewis acid sites, but to a lesser extent.
As seen in Table 4, the Brønsted acidity trend is: NiMo-2.0 ꢀ NiMo-
0.0 > NiMo-0.5 = NiMo-1.0 = NiMo-1.5. Thus, the NiMo-2.0 sample
displays the largest amount of both Brønsted and Lewis acid sites
among the catalysts studied.
Summarizing the catalyst characterization data, we conclude
that support modification with phosphorous has a detrimental
effect on the textural properties because it decreases specific sur-
face area and collapses partially the Al-HMS mesoporous structure,
as confirmed by N2 physisorption and low-angle XRD, respectively.
Moreover, it was found that the Lewis acidity depends strongly on
the amount of phosphorous incorporated over the support. On the
contrary, the phosphorus addition on the support did not influence
on the Brønsted acid sites. For the NiMo-0.5, NiMo-1.0 and NiMo-
1.5 samples, the Brønsted acidity deceases with respect to P-free
counterpart. This is suggesting a multiple bonding of phosphoric
acid to three different hydroxide groups of the support [38]. How-
ever, an important increment in the Brønsted acidity was observed
in the NiMo-2.0 sample. This could related that the hydroxyl groups
bonded to tetrahedral Si4+ ions acts as Brønsted acid sites, consid-
ering one point bonding of the phosphoric acid to one hydroxyl
group of the carrier, as it was suggested by Fitz and Rase [38].
On the contrary to acidity, the presence of P species on the sup-
port surface enhances both the dispersion of metal oxide species
and their reducibility, as deduced from XRD and TPR data, respec-
tively. The calculation of the normalized SBET (expressed as m2 g−1
of support) of the catalyst indicated the homogeneous distribution
of metal oxide species within the inner pores of the support and
on their surface. Upon sulfidation, all catalysts show the formation
of MoS2 phase being nickel species so small to be detected by XRD
(Fig. 6). This technique also shows that a large amount of phosphate
(2.0 wt.%) favors the particles growth due to excessive weakening
of the metal-support interaction deduced from TPS measurements
(Fig. 5).
Fig. 6. X-ray diffraction patterns of fresh sulfided catalysts: (a) NiMo-0.0; (b) NiMo-
0.5; (c) NiMo-1.0, and (d) NiMo-2.0.
Two strong bands assigned to the pyridine chemisorbed on Lewis
sites and hydrogen-bonded pyridine is observed at 1445 and
1600 cm−1, respectively [55]. A small band located at 1542 cm−1
is assigned to the vibration mode of pyridinium ion adsorbed on
Brønsted sites while a band observed at 1486 cm−1 is assigned
to pyridine associated with both Lewis and Brønsted sites [56].
The P-free sample presents the bands at 1571 and 1445 cm−1 of
weak and strong adsorbed pyridine on Lewis sites, respectively,
and a small one at 1542 cm−1 due to Brønsted acid sites, indicating
that the main acid sites on sulfided NiMo-0.0 catalyst are of Lewis
nature. Upon P addition, the intensity of these bands increases
suggesting that the incorporation of P increases the population of
In this sense, the morphology for the some sulfided samples
was studied by statistical calculation of ca. 500 particles from
12 HRTEM micrographs. HRTEM images of sulfided NiMo-0.0 and
NiMo-1.0 catalyst are shown in Fig. 8(a) and (b), respectively.
One can observe that the stacking in both samples is closed at 4.
Nonetheless, an important effect on the dispersion and sulfide slabs
size were enhancement with the phosphorus addition. The cata-
lyst prepared on the phosphate support modification has the lower
average size of the slabs value of 2.2 0.25 nm, which represents
2.9 nm obtained for the free phosphorus containing sample. The
surface density of MoS2 slabs was also calculated. The surface den-
sity of the slabs for the 1 wt.% P-containing sample was of 1 MoS2
Fig. 7. FT-IR spectra of adsorbed pyridine on the NiMo/Al-HMS-P(x) sulfide cat-
alysts: (a) NiMo-0.0; (b) NiMo-0.5; (c) NiMo-1.0; (d) NiMo-1.5; (e) NiMo-2.0. B,
Brønsted acid sites; L, Lewis acid sites.

116
T.A. Zepeda et al. / Applied Catalysis A: General 484 (2014) 108–121
Fig. 8. HRTEM images of the sulfide NiMo-0.0 and NiMo-1.0 samples.
slabs (particle)/10 nm2, while for the P-free catalyst was of 1 MoS2
slabs (particle)/25 nm2. These displayed values confirm a high dis-
persion of the active phase in the sample with 1 wt.% of phosphorus.
These results are in agreement with the discussed previously.
higher than the P-free catalyst, probably related with the higher
acidity measure for these samples. This is an important finding
because our objective in this work is to diminish olefins hydro-
genation. At a reaction temperature of 300 ◦C, both NiMo-2.0 and
NiMo/Al2O3 samples showed very low HYD capability. However,
an increase of the reaction temperature led to the undesirable
enhancement of olefin hydrogenation for all the catalysts studied.
Besides this, it is also noted that the most active NiMo-1.0 catalyst
displayed larger olefin/butane ratio than its NiMo-0.0 and NiMo-
2.0 counterparts. However, regardless the reaction temperature,
its selectivity toward butane formation is larger than that of the
3.7. Catalytic results
In the present study, NiMo/Al-HMS-P(x) catalysts were tested
in the HDS of thiophene (TF) and 4,6-dimethyldibenzothiophene
(4,6-DMDBT); the later sulfur-containing molecule is considered
as representative of the refractory sulfur compounds contained in
the gas oil fraction. For comparative purposes a reference NiMo
catalyst was also tested (NiMo/Al2O3).
3.7.1. HDS of thiophene
The activity of NiMo-0.0, NiMo-1.0 and NiMo-2.0 sulfide cata-
lysts was evaluated in the HDS of thiophene at 300, 320 and 340 ◦C
at atmospheric pressure. The influence of the reaction temperature
and P loading on the thiophene conversion are shown in Fig. 9(a).
From this results we the activation energies were estimated for
each sample, the Ea values were 24.7, 30.4, 28.7 and 27.9 kJ mol−1
for the NiMo-0.0, NiMo-1.0, NiMo-2.0 and NiMo/Al2O3 samples,
respectively. The activation energies values follow the trend: NiMo-
1.0 > NiMo-2.0 > NiMo/Al2O3 > NiMo-0.0. This trend indicates that
activity increases upon P incorporation up to 1.0 wt.% of P and then
declines at higher P-loading. It can be noted that the commercial
NiMo/Al2O3 catalysts is less active than NiMo-1.0 one.
The influence of the reaction temperature on product selectiv-
ity, taken to a thiophene conversion of 20–25%, is shown in Fig. 10.
Under the reaction conditions employed, the reaction products
identified were n-butane, 1-butene and cis and trans-butene for
P-free catalyst (NiMo-0.0) and also 1,3 butadiene for the other cat-
alysts. As expected, the product selectivity strongly depends on
the reaction temperature and catalyst composition. The general
trend observed is the increase of butane and 1-butene selectivity
and a decrease of 2-butenes and 1,3 butadiene with temperature.
The proportion of each one depends on catalyst composition. Thus,
NiMo-0.0 produces the highest amount of butane and 1-butene.
However, 1,3-butadiene is formed upon P-adding to the catalyst,
mainly at low temperatures because it becomes hydrogenated at
higher temperatures. Besides, a considerable increase of 2-butenes
selectivity is also observed upon P addition (mainly over NiMo-1.0).
Phosphorus it also has an important effect on selectivity, con-
sidering the olefins/paraffin ratio ((butenes + butadiene)/butane)
(Fig. 9(b)), in both P- containing samples the olefins selectivity was
Fig. 9. HDS of thiophene over NiMo/Al-HMS and NiMoAl-HMS-P(x) sulfide catalysts
(a flow reactor, a steady-state; atmospheric pressure, T = 300, 320, 340 ◦C and con-
versions between 20 and 25%). Influence of the reaction temperature and P loading
on catalyst activity (a) and the comparison of olefins/butane ratio (b).

T.A. Zepeda et al. / Applied Catalysis A: General 484 (2014) 108–121
117
Fig. 10. Influence of the reaction temperature on the selectivity of products in the HDS of thiophene over sulfide catalysts (atmospheric pressure; T = 300, 320 340 ◦C;
steady-state, a flow reactor).
Scheme 1. Reaction scheme for the hydrodesulfurization of thiophene over NiMo/Al-HMS-P(x) sulfide catalysts.

118
T.A. Zepeda et al. / Applied Catalysis A: General 484 (2014) 108–121
[44]. The observed activity trend is: NiMo-1.0 > NiMo-1.5 > NiMo-
0.5 > NiMo-0.0 > NiMo-2.0. Thus, the NiMo-1.0 catalyst turns out to
be also the most active one, with an increment in activity of 72%.
Compared with the commercial catalyst with higher metal loading,
there are two catalysts (NiMo-1.0 and NiMo-1.5) showing higher
activity than the reference one. Thus, the optimum performance is
observed for the catalyst whose P-loading is close 1 wt.%, which is
in good agreement with that reported for NiMo/P-MCM-41 cata-
lysts [42]. As observed in the present work, the NiMo/P-MCM-41
sulfide catalysts demonstrated higher activity than the commercial
NiMo/Al2O3 catalyst when tested in HDS of 4,6-DMDBT in a batch
reactor [42].
Besides, a linear dependence was observed between cata-
lyst activity and the active phase dispersion derived from H2
chemisorption (Fig. 11(b)). Thus, the highest activity of the NiMo-
1.0 catalyst appears to be related to the highest dispersion of the
active phase among the catalysts studied and therefore to the for-
mation of the smallest MoS2 crystallites on the support surface
(from HRTEM). Moreover, one might expect that the type of sym-
metry adopted by molybdenum and nickel ions in the Al-HMS-P(x)
support surface might have some influence on the final catalytic
response. In this sense, the oxide precursor of NiMo-1.0 sample
showed the lowest amount of the Mo in tetrahedral coordination
(from UV–vis). It is well known that tetrahedral molybdate (MoO4)
species are more difficult to reduce and sulfide than a supported
bulk MoO3 acidic species. Additionally, a volcano-curve trend of
the catalyst activity could also be explained considering a possi-
ble increase in the amount of nickel available to form the NiMoS
active phase, as it was proposed by López Cordero et al. [52] for
NiMo/P/Al2O3 catalysts.
Finally, the detrimental effect of a large P-loading on the HDS
activity of both NiMo-1.5 and NiMo-2.0 catalysts could be explained
considering the sharp decrease in their specific area with respect
to P-free sample (Table 2) and considering the large concentration
of tetrahedral molybdenum species bound to the support surface
(from DRS UV-vi). For those samples, the decrease of the active
phase dispersion is consistent with a more difficult sulfidation
of the support-linked tetrahedral Mo species (from TPS measure-
ments, Fig. 5). This situation is similar to that observed by Anderson
et al. [59] where the lower sulfidation degree has been achieved
with Mo/USY sulfide catalyst containing a high concentration of
tetrahedral molybdenum bound to the support. In addition, Her-
rera et al. [42] showed that an increase of P loading from 1.0 to
5.0 wt.% resulted in the formation of crystalline MoO3 species, with
subsequent formation of larger MoS2 crystals during sulfidation
of the NiMo/P-MCM-41 catalyst and a parallel activity drop in the
4,6-DMDBT HDS reaction.
Fig. 11. HDS of 4,6-DMDBT DMBT (P = 5.5 MPa; T = 320 ◦C, reaction time of 6 h; a
batch reactor) over NiMo/Al-HMS and NiMo/Al-HMS-P(x) sulfide catalysts. Influence
of the P loading on the activity and selectivity (a). Influence of the active phase
dispersion (from H2 chemisorption) on the HDS activity of catalysts (b).
reference catalyst. It is noteworthy that for all catalysts studied
the olefin chain-branching did not occur. Scheme 1 shows the
possible HDS reaction mechanism over P-containing catalysts and
the reference one. 1,3-butadiene has been suggested as the most
likely product formed directly following C S bond breaking, which
is lately hydrogenated to form butene and butane. The reaction
mechanism proposed by Schultz and Rahman [58] involves the per-
pendicular adsorption of thiophene molecule through the S atom
on a S anion vacancy, followed by attack of the S C bond by H from
adjacent OH groups, thus forming butadiene as primary product.
3.7.2. HDS of 4,6-DMDBT reaction
3.7.2.1. Factors influencing the catalyst activity. Alkyl-substituted
DBTs are known to be more difficult to desulfurize than thiophene
because of the steric hindrance of their alkyl groups hindering the
S-atom interaction with the catalyst surface. Thus, our challenge
in this work was to investigate if the catalyst modification with P
could eliminate the planar configuration of alkyl-substituted DBT
through the enhancement of hydrogenation, isomerization, dealky-
lation, and/or C C bond scission reaction routes [45]. The effect
of P incorporation on the Al-HMS substrate on the activity of the
sulfided NiMo catalysts in the HDS of 4,6-DMDBT reaction was eval-
uated in a batch reactor at T = 320 ◦C and P = 5.5 MPa of hydrogen
pressure. A commercial NiMo/␥-Al2O3 catalyst with a metal load-
ing higher than that of the catalysts studied (16 vs. 12 wt.% Mo + Ni)
was used as a reference.
Fig. 11(a) shows the specific reaction rates versus P-loading
of the catalysts at a reaction time of 6 h. In good agreement
with the HDS of thiophene results, it is noted that P-loading on
the HMS substrate results in an increase of activity. The HDS
activity of P-loaded sulfide catalysts follows a volcano-shaped
curve, in line with the study of NiMoW/P/SBA-16 sulfide catalysts
3.7.2.2. Factors influencing the selectivity. Additional information
on the P effect was obtained by analyzing selectivity data (cal-
culated at the same 4,6-DMDBT conversion of ca. 30%) given
in Table 5. For all the studied catalysts, including a commercial
NiMo/Al2O3 one, the products detected were dimethylbiphenyl
(DMBP), methylcyclohexyltoluene (MCHT) and dimethylbicy-
clohexyl (DMBCH) being MCHT the main reaction product.
Additionally, different isomers of the principal reaction prod-
ucts together with traces of the intermediate products such
as tetrahydro- and hexahydro-dimethyldibenzothiophene (THD-
MDBT and HHDMDBT, respectively) were detected. Based on the
products detected, Scheme 2 shows a simplified reaction mecha-
nism of 4,6-DMDBT transformation over sulfided NiMo/Al-HMS-P
catalysts. Under the reaction conditions employed, the HDS reac-
tion of 4,6-DMDBT proceeds through two main reaction routes:
direct desulfurization (DDS) leading to formation of 3,3ꢁ-DMBP and
hydrogenation (HYD) reaction route producing tetrahydro- and
hexahydro-DMDBT intermediates, and after S-elimination leading

T.A. Zepeda et al. / Applied Catalysis A: General 484 (2014) 108–121
119
Table 5
HDS of 4,6-DMDBTa specific rate, selectivityb and HYD/DDS selectivities ratio for NiMo-P(x) catalysts and NiMo/Al2O3 reference catalyst.
Selectivity (%)c
DMBP
HYD/DDSc
−1
]
Rate ×107 [mol g−1
s
Catalyst
cat .
DMDCHCHc
MCHT
NiMo-0.0
NiMo-0.5
NiMo-1.0
NiMo-1.5
NiMo-2.0
NiMo/Al2O3
4.0
4.6
6.9
5.8
3.8
4.7
29
28
21
23
37
25
5
7
14
11
4
66
65
65
66
59
63
2.44
2.56
3.70
3.33
1.69
3.03
12
a
The reaction conditions were: T = 320 ◦C; P = 5.5 MPa; 6 h of reaction time; a batch reactor.
Achieved at 30% of DBT conversion: DMBP, dimethylbiphenyl; MCHT, methylcyclohexyltoluene; DMDCH, dimethyldicyclohexyl.
DDS: direct desulfurization; HYD: hydrogenation pathway. The HYD/DDS selectivities ratio as calculated from selectivity of the products [(MCHT + DMDCH)/DMBP].
b
c
to the formation of MCHT and DMDCH. This is in good agreement
with the CoMo/Ti-HMS-P(x) catalysts studied by us previously
[45]. The isomerisation of 4,6-DMDBT also occurs but only small
amounts of different isomers of the principal reaction products
were detected for all the studied catalysts. Indeed, with the excep-
tion of NiMo-2.0, all the catalysts showed the same amount of
Brønsted acid sites, as derived from FT-IR spectra of adsorbed pyri-
dine (Table 4). Similarly to the NiMo-00 catalyst, the formation
of different isomers in the main products was also observed over
NiMo catalysts supported on Al-SBA-15 [33] and Al-SBA-16 sulfide
catalysts [21].
It is generally accepted the hypothesis that ␲-adsorption favors
hydrogenation reaction, whereas the sulfur adsorption is related
to desulfurization reactions [60]. In general, it was found that the
hidrodesulfurization via hydrogenation way over the catalyst was
enhanced by the support modification with phosphorous, being
in good agreement with the work of Herrera et al. [42]. Fig. 6
shows that (MCHT + DMDCH)/DMBP selectivity ratio (HYD/DDS
ratio) increases with P loading up to 1.0 wt.%, which also presents
better activity results, but then levels off. For all the studied cat-
alysts, MCHT was the main reaction product (selectivity ca. 66%).
Although the NiMo-2.0 catalyst shows the lowest selectivity toward
this product (59%), at the same time, it can be noted that in all cases
the HYD/DDS ratio was much higher than 1 (Table 5), indicating
the predominance of the HYD reaction route over the HDS one. If
the HYD/DDS ratios are plotted against the specific activity a linear
increase is again observed (Fig. 11(a)), i.e., the most active cata-
lyst presents better hydrogenating properties suggesting that the
MoS2 crystallites of this sample contain (a) the largest amount of
promoted sites among the catalysts [61].
the enhancement of their hydrogenation function is needed. For
all the studied catalysts, the HYD/DDS selectivity ratio follows
the trend: NiMo-1.0 > NiMo-1.5 > NiMo/Al2O3 > NiMo-0.5 > NiMo-
0.0 > NiMo-2.0 indicating that the two former catalysts show larger
hydrogenation capability than the commercial one. The promotion
of HYD route was explained in terms of the increase in catalyst acid-
ity induced by H2S adsorption on the catalyst surface [62]. However,
the reverse trend observed in this work between the HYD/DDS
ratio and the catalyst acidity suggests a compromise between the
enhancement of the catalyst HYD function and the decrease of the
catalyst deactivation by decreasing the catalyst acidity. Indeed, the
high acidity of NiMo-2.0 confirms the hypothesis that the intro-
duction of high amounts of phosphate created medium strength
acid sites, which in turn increases the total acidity of the support
[37]. As a consequence, one might expect that coking could also
increase.
The influence of the active phase dispersion on the specific reac-
tion rate (expressed as activity per the catalyst surface area) is
shown in Fig. 11(b). Thus, the decrease of the active phase disper-
sion may also explain the lowest HYD/DDS selectivity ratio of the
NiMo-2.0 sample in the HDS of 4,6-DMDBT reaction (Fig. 12(b)).
This is because the samples showing an inferior dispersion possess
a lower amount of promoted active sites, which are widely accepted
to decreases the formation of active sites for HYD reaction [61].
In this sense, phosphorous poisoning was explained by Bouwens
et al. [63] in terms of the adsorption of phosphide (formed dur-
ing the presulfiding treatment) on the anion vacancies. Although
this explanation is also valid in this work, considering our previ-
ous HRTEM study on the NiMoW/SBA-16 sulfide catalysts with a
high amount of P (2.0 wt.%) [44]. We strongly believe that the high
presence of P2O5 species on the support surface might force the
formation of less active “onion-type” Mo(W)S2 structures, which
possess a low amount of rim/edge sites.
Interestingly, with the exception of NiMo-2.0 catalyst, there is a
linear increase between 4,6-DMDBT transformation and HYD/DDS
selectivity ratio (Fig. 12(a)) confirming that, for deep HDS catalysts,
Scheme 2. Reaction scheme for the hydrodesulfurization of 4,6-DMDBT over sulfide NiMo/Al-HMS and NiMo/Al-HMS-P catalysts.

120
T.A. Zepeda et al. / Applied Catalysis A: General 484 (2014) 108–121
[3] J.D.K. Bishop, C.J. Axon, M. Tran, D. Bonilla, D. Banister, M.D. McCulloch, Fuel 99
HDS of 4,6-DMDBT
(2012) 88–105.
[4] H. Topsøe, B.S. Clausen, F.E. Massoth, J.R. Anderson, M. Boudart (Eds.),
Hydrotreating Catalysis, Springer Verland, Berlin Heidelberg, 1996, pp. 1–269.
[5] M. Grings, B.C. Gates, Ind. Eng. Chem. Res. 30 (1991) 2021–2058.
[6] D. Laurenti, B. Phung-Ngoc, C. Roukoss, E. Devers, K. Marchand, L. Massin, L.
Lemaitre, C. Legens, A.A. Quoineaud, M. Vrinat, J. Catal. 297 (2013) 165–175.
[7] S.Y. Yu, W. Li, E. Iglesia, J. Catal. 187 (1999) 257–261.
[8] B. Pawelec, R. Mariscal, R.M. Navarro, J.M. Campos-Martin, J.L.G. Fierro, Appl.
Catal., A: Gen. 262 (2004) 155–166.
[9] R. Zhao, C. Yin, H. Zhao, C. Liu, Fuel Process. Technol. 81 (2003) 201–209.
[10] C. Sepulveda, V. Bellière, D. Laurenti, N. Escalona, R. García, C. Geantet, M. Vrinat,
Appl. Catal., A: Gen. 393 (2011) 288–293.
[11] D. Trong On, D. Desplantier-Giscard, C. Danumah, S. Kaliaguine, Appl. Catal., A:
Gen. 253 (2003) 545–602.
NiMo-1.0
7
6
5
4
3
( a )
NiMo-1.5
NiMo/Al2O3
NiMo-0.5
NiMo-2.0
NiMo-0.0
[12] A. Taguchi, F. Schu˝ th, Microporous Mesoporous Mater. 77 (2005) 1–45.
[13] S. Kawi, S.C. Shen, P.L. Chew, J. Mater. Chem. 12 (2002) 1582–1586.
[14] A. Tuel, Microporous Mesoporous Mater. 27 (1999) 151–169.
[15] T.A. Zepeda, B. Pawelec, J.L.F. Fierro, A. Olivas, S. Fuentes, T. Halachev, Micro-
porous Mesoporous Mater. 111 (2008) 157–170.
1.5
2.0
2.5
3.0
3.5
4.0
[16] T. Chiranjeevi, P. Kumar, M.S. Rana, S.K. Maity, G. Murali Dhar, T.S.R. Prasada
Rao, Microporous Mesoporous Mater. 44–45 (2001) 547–556.
[17] T. Chiranjeevi, P. Kumar, M.S. Rana, G. Murali Dhar, T.S.R. Prasada Rao, J. Mol.
Catal. A: Chem. 181 (2002) 109–117.
[18] M. Alibouri, S.M. Ghoreishi, H.R. Aghabozorg, Ind. Eng. Chem. Res. 48 (2009)
4283–4292.
HYD/DDS ratio
20
16
12
8
( b )
NiMo-1.5
NiMo-1.0
[19] M. Alibouri, S.M. Ghoreishi, H.R. Aghabozorg, J. Supercrit. Fluids 49 (2009)
239–248.
[20] R. Huirache-Acun˜a, B. Pawelec, C.V. Loricera, E.M. Rivera-Mun˜oz, R. Nava, B.
Torres, J.L.G. Fierro, Appl. Catal., B: Environ. 125 (2012) 473–485.
[21] T. Klimova, L. Lizama, J.C. Amezcua, P. Roquero, E. Terrés, J. Navarrete, J.M.
Domínguez, Catal. Today 98 (2004) 141–150.
[22] M.V. Landau, L. Vradman, M. Herskowitz, Y. Koltypin, A. Gedanken, J. Catal. 201
(2001) 22–36.
[23] K.C. Park, D.J. Yim, S.K. Ihm, Catal. Today 74 (2002) 281–290.
[24] T. Klimova, M. Calderón, J. Ramírez, Appl. Catal., A: Gen. 240 (2003) 29–40.
[25] 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.
[26] Y. Wang, S.K. Ihm, G. Lu, Catal. Lett. 92 (2007) 165–173.
[27] A.M. Venezia, R. Murania, V. La Parola, B. Pawelec, J.L.G. Fierro, Appl. Catal., A:
Gen. 383 (2010) 211–216.
[28] K. Segawa, K. Takahashi, S. Satoh, Catal. Today 63 (2000) 123–132.
[29] S. Zeng, J. Blanchard, M. Breysse, Y. Shi, X. Su, H. Nie, D. Li, Appl. Catal., A: Gen.
294 (2005) 59–67.
[30] G. Muthu Kumaran, S. Garg, K. Soni, M. Kumar, L.D. Sharma, G. Murali Dhar, K.S.
Rama Rao, Appl. Catal., A: Gen. 305 (2006) 123–129.
[31] G. Muthu Kumaran, S. Garg, K. Soni, M. Kumar, L.D. Sharma, K.S. Rama Rao, G.
Murali Dhar, Ind. Eng. Chem. Res. 46 (2007) 4747–4754.
[32] G. Macías Esquivel, J. Ramírez, A. Gutiérrez-Alejandre, Catal. Today 148 (2009)
36–41.
[33] T. Klimova, J. Reyes, O. Gutiérrez, L. Lizama, Appl. Catal., A: Gen. 335 (2008)
159–171.
0.979
=
R
NiMo-0.5
NiMo-0.0
NiMo-2.0
1.5
2.0
2.5
3.0
3.5
4.0
HYD/DDS ratio
Fig. 12. HDS of 4,6-DMDBT DMBT (P = 5.5 MPa; T = 320 ◦C, reaction time of 6 h; a
batch reactor) over NiMo/Al-HMS and NiMo/Al-HMS-P(x) sulfide catalysts. Catalytic
activity against HYD/DDS selectivities ratio (a) and the influence of the active phase
dispersion (from H2 chemisorption) on the HYD/DDS selectivities ratio (b).
4. Conclusions
All catalysts showed changes in morphology and acidity induced
by the presence of phosphate species on the support surface. The
catalysts were tested in the HDS of thiophene and 4,6-DMDBT. The
catalyst with nominal 1.0 wt.% P content exhibited the best per-
formance in both HDS reactions. This was related to an increase
in the active phase dispersion, as demonstrated by H2 chemisorp-
tion data, and to the enhancement amount of promoted sites in the
active NiMoS phase. The HDS of 4,6-DMDBT reaction over all cat-
alysts proceeds toward direct desulfurization, hydrogenation and
isomerization reaction routes being inhibited dealkylation and/or
[34] M. Gómez-Cazalilla, A. Infantes-Molina, R. Moreno-Tost, P.J. Maireles-Torres,
J. Mérida-Robles, E. Rodríguez-Castellón, A. Jiménez-López, Catal. Today 143
(2009) 137–144.
[35] R. Huirache-Acun˜a, B. Pawelec, E. Rivera-Mun˜oz, R. Nava, J. Espino, J.L.G. Fierro,
Appl. Catal., B: Environ. 92 (2009) 168–184.
[36] R. Iwamoto, J. Grimblot, Adv. Catal. 44 (2000) 417–503.
[37] A. Stanislaus, M. Absi-Halabi, A. Al-Doloma, Appl. Catal. 39 (1988) 239–253.
[38] C.W. Fitz, H.F. Rase, Ind. Eng. Chem. Prod. Dev. 22 (1983) 40–44.
[39] S. Eijsbouts, J.N.M. van Gestel, J.A.R. van Veen, V.H.J. de Beer, R. Prins, J. Catal.
131 (1991) 412–432.
[40] J.M. Lewis, R.A. Kydd, P.M. Boorman, P.H. Van Rhyn, Appl. Catal., A: Gen. 84
(1992) 103–121.
[41] J.M. Lewis, R.A. Kydd, J. Catal. 136 (1992) 478–486.
[42] J.M. Herrera, J. Reyes, P. Roquero, T. Klimova, Microporous Mesoporous Mater.
83 (2005) 283–291.
C
C bond scission reaction routes. Concerning the HDS of thio-
phene, the increase of the catalyst acidity led to the formation of
butadiene and a diminished hydrogenation of olefins (formation of
butane) while no olefin chain-branching occurred.
[43] R. Nava, B. Pawelec, J. Morales, R.A. Ortega, J.L.G. Fierro, Microporous Meso-
porous Mater. 118 (2009) 189–201.
[44] M.A. Guzmán, R. Huirache-Acun˜a, C.V. Loricera, J.R. Hernández, N. Díaz de León,
J.A. de los Reyes, B. Pawelec, Fuel 103 (2013) 321–333.
[45] B. Pawelec, J.L.G. Fierro, A. Montesinos, T.A. Zepeda, Appl. Catal., B: Environ. 80
(2008) 1–14.
Acknowledgments
[46] T. Halachev, J.A. de los Reyes, C. Araujo, L. Dimitrov, G. Cordoba, in: B.
Delmon, G.F. Froment, P. Grange (Eds.), Hydrotreatment and Hydrocrack-
ing of Oil Fractions, Studies in Surface Science and Catalysis, Elsevier, 1999,
p. 401.
The authors are grateful to Eric Flores, Eloisa Aparicio, David
Dominguez, Jose Antonio Díaz, Francisco Ruiz and Israel Gradilla for
their technical assistances and CONACYT projects 152012, 155388
and 117373 for the financial supports.
[47] R. Nava, A. Infantes-Molina, P. Castan˜o, R. Guil-López, B. Pawelec, Fuel 90 (2011)
2726–2737.
[48] R. Nava, J. Morales, G. Alonso, C. Ornelas, B. Pawelec, J.L.G. Fierro, Appl. Catal.,
A: Gen. 321 (2007) 58–70.
[49] B. Pawelec, S. Damyanova, R. Mariscal, J.L.G. Fierro, I. Sobrados, J. Sanz, L. Petrov,
J. Catal. 223 (2004) 86–97.
References
[50] S. Gontier, A. Tuel, Zeolites 15 (1995) 601–610.
[51] O.Y. Gutiérrez, T. Klimova, J. Catal. 281 (2011) 50–62.
[52] R. López Cordero, A. López Agudo, Appl. Catal., A: Gen. 202 (2000) 23–35.
[1] U.S. EnvironmentalProtectionAgency, ꢂhttp://www.epa.gov/otaq/gasoline.htmꢃ,
ꢂhttp://www.epa.gov./otaq/diesel.htmꢃ.
[2] EU Regulations, ꢂhttp://www.dieselnet.com/standards/eu/fuel.phpꢃ.

T.A. Zepeda et al. / Applied Catalysis A: General 484 (2014) 108–121
[53] R. López Cordero, F.J. Gil Llambias, A. López Agudo, Appl. Catal. 74 (1991) [58] H. Schultz, N.M. Rahman, Stud. Surf. Sci. Catal. 75 (1993) 585–595.
121
125–136.
[59] J.A. Anderson, B. Pawelec, J.L.G. Fierro, Appl. Catal., A: Gen. 99 (1993) 55–70.
[60] V. Meille, E. Schultz, M. Lemaire, M. Vrinat, J. Catal. 170 (1997) 29–36.
[61] M. Daage, R.R. Chianelli, J. Catal. 149 (1994) 414–427.
[62] E. Lecrenay, K. Sakanishi, I. Mochida, Catal. Today 39 (1997) 13–20.
[63] S.M.A.M. Bouwens, J.P.R. Vissers, V.H.J. de Beer, R. Prins, J. Catal. 112 (1988)
401–410.
[54] J.W. Niemantsverdriet, Spectroscopy in Catalysis. An Introduction, third ed.,
Wiley-VCH, 2007, pp. 217–249 (Chapter 8).
[55] B. Chakraborty, B. Viswanathan, Catal. Today 49 (1999) 253–260.
[56] G. Busca, Catal. Today 41 (1–3) (1998) 191–206.
[57] C.A. Emeis, J. Catal. 141 (1993) 347–354.