Effect of the type of precursor and the synthesis method on thiophene hydrodesulfurization activity of activated carbon supported Fe-Mo, Co-Mo and Ni-Mo carbides

Article abstract of DOI:10.1016/j.molcata.2007.09.015

In this work it is studied the effect of the type of precursor (sulfate vs. nitrate of promotor) and the synthesis method (conventional vs. carbothermal carbiding) on the thiophene hydrodesulfurization (HDS) activity of activated carbon supported Fe-Mo, C

Full text of DOI:10.1016/j.molcata.2007.09.015

Available online at www.sciencedirect.com
Journal of Molecular Catalysis A: Chemical 281 (2008) 85–92
Effect of the type of precursor and the synthesis method
on thiophene hydrodesulfurization activity of
activated carbon supported Fe-Mo,
Co-Mo and Ni-Mo carbides
∗
Esneyder Puello-Polo, Joaqu ´ı n L. Brito
Laboratorio de Fisicoqu ´ı mica de Superficies, Centro de Qu ´ı mica, Instituto Venezolano de Investigaciones Cient ´ı ficas,
IVIC, Apartado 20632, Caracas 1020-A, Venezuela
Available online 18 September 2007
Abstract
In this work it is studied the effect of the type of precursor (sulfate vs. nitrate of promotor) and the synthesis method (conventional vs. carbothermal
carbiding) on the thiophene hydrodesulfurization (HDS) activity of activated carbon supported Fe-Mo, Co-Mo and Ni-Mo carbides. Catalytic
precursors were prepared by co-impregnation of the support with solutions of ammonium heptamolybdate and the promotor salt. Conventional
carbiding consists of a temperature-programmed treatment under a CH
The passivated carbided solids were characterized by XRD and XPS. The presence of metals and Fe
samples from sulfate precursors and of ␤-Mo C and metals in those from nitrate precursors was verified by XRD, whereas XPS showed the presence
4
/H
2
(1:4) atmosphere, while the carbothermal method employs pure H
2
.
3
Mo C or M Mo phases (M = Co or Ni) in
3
6
6 2
C
2
6+
δ+
4+
3+
2+
0
2+
at the surface of Mo (0 ≤ δ ≤ 2), Mo , Mo , Fe , Co , Ni and Ni species. The bimetallic carbides obtained from sulfate precursors retained
2
−
2−
4
sulfur on the surface, as shown by signals in the S 2p region, at 169 and 162 eV, assigned to S and SO , respectively. Prior to the catalytic
reaction, the passivated carburized forms of carbon supported catalysts were presulfided in situ. The catalytic activity of carbides was strongly
influenced by the type of precursor and slightly by the synthesis method. The activity of carbides obtained by the carbothermal method and with
sulfate precursors was greater than that of the solids obtained by the conventional method and with nitrate precursors. The carbides obtained from
nitrates showed increased catalytic activity when presulfided, suggesting that the carbides with sulfided surfaces or mixed carbo-sulfide species
could be the active phase in HDS on carbide catalysts. The catalysts derived from sulfates showed higher activity than those obtained from nitrates,
even without presulfiding, suggesting that sulfide from the precursor sulfate results in more active species than those obtained after presulfiding.
©
2007 Elsevier B.V. All rights reserved.
Keywords: Activated carbon; Carbides; Precursor effect; Sulfided surface; Synthesis method effect; Thiophene hydrodesulfurization
1
. Introduction
tory sulfur compounds [2]. The quality of fuels prescribed by
the new environmental regulations can be attained using the cur-
rent catalysts, but only after significant modifications of refining
operations. At certain point, these modifications would cause
negative effects on the catalyst life and would increase the costs
of the processes [2]. In the last decades, efforts have been made
to develop novel catalysts that can fulfill these regulations, and
the transition metals carbides and nitrides have been identified
as potential catalysts for such applications [3]. In this regard,
the Mo based solids have been receiving most of the attention
The high demand of hydrocarbons in the world during the
last century has resulted in the exhaustion of most of the
light oil reserves; therefore the petroleum used nowadays is
increasingly contaminated with sulfur, which upon combus-
tion of fuels produces sulfur oxides [1]. The current generation
of hydrodesulfurization catalysts are sulfides of Co(Ni)-Mo
supported on alumina, and although they are very active for
conventional oil, they exhibit low activity towards highly refrac-
[3–14]. The catalytic properties of carbide materials strongly
depend on their surface structure and composition, which are
closely associated with the preparation methods [15]. When
employing bimetallic precursors, there exist the possibility that
the resulting carbides (nitrides) consist either of the separated
∗
Corresponding author.
E-mail addresses: snypollqco@yahoo.com (E. Puello-Polo),
joabrito@ivic.ve (J.L. Brito).
1
381-1169/$ – see front matter © 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.molcata.2007.09.015

8
6
E. Puello-Polo, J.L. Brito / Journal of Molecular Catalysis A: Chemical 281 (2008) 85–92
monometallic phases or of the mixed Co(Ni)-Mo-C(N) ones.
While in unsupported solids the presence of such compounds
as Co3Mo3C or Ni2Mo3N has been shown by XRD, in sup-
ported systems it is not easy to observe such mixed phases, and
only the monometallic carbides are generally detected [16]. In
a few works, bimetallic phases supported on alumina have been
observed [17,18]. Nevertheless, it is clear that regardless of the
observation or not of the bimetallic phases, the mixed metal cat-
alysts are more active than the catalysts consisting only of Mo
carbide/nitride [5,16,18–21].
best knowledge, this is the first report on the preparation of
bimetallic carbides from sulfate precursors, showing an effect
of sulfur derived from the precursor salt in the HDS of thiophene.
2. Experimental
2.1. Preparation of precursors and carbides
The bimetallic precursors were supported on a com-
2
mercial activated carbon (Merck, 899 m /g) by one-step
The transition metals carbides as catalysts are still in devel-
opment and the type of support is a variable to be studied in
order to improve the hydrotreating process efficiency. Activated
carbon has received attention as a carrier for HDS catalysts for
several reasons (variable amount of surface functional groups,
large specific surface area with an easily controlled pore volume
and relatively lower coking activity) [22–25].
In the preparation of carbides of catalytic interest, the con-
ventional preparation method is the temperature-programmed
reaction between oxide precursors and a flowing mixture of
H2 and carbon-containing gases [26]. However, some problems
exist in this method of preparation, namely, the resultant car-
bide surface is usually contaminated by polymeric carbon from
the pyrolysis of the carbon-containing gases [27,28]. Recently,
Mordenti and co-workers developed a novel synthesis route to
preparing high surface area carbides by employing the carboth-
ermal method using H2, which can avoid the formation of carbon
residues on the surface [29]. Liang et al. tried to optimize the car-
bothermal hydrogen reduction varying the H2 flow and the load
of precursor in order to obtain supported Mo2C and bimetallic
carbides [30]. Sayag, using the carbothermal H2 method pre-
pared supported Mo2C, modifying the chemical surface of the
support and obtained excellent results in HDS [31].
co-precipitation. Aqueous solutions of ammonium heptamolyb-
date (Merck, 99%) and iron, cobalt or nickel nitrates (Riedel
de Haen, typically 98%) or sulfates (Aldrich, typically 98%)
◦
were prepared with M/Mo molar ratio 1:1 and kept at 80 C,
and only then were simultaneously added dropwise to a flask
containing the support, under stirring. The impregnation was
performed at pH 7. A slight excess of the solution was used to
provide an uniform coating of the activated carbon surface. The
impregnation step lasted until removal of the solvent by evapo-
◦
ration. The mass obtained was further dried at 120 C for 12 h
overnight, andthenthesolidsweretransferredtoaquartzreactor,
deposited on a quartz fiber bed. A tubular electrical furnace con-
trolled by a temperature programmer (Thermolyne 21100) was
used to heat the reactor. The amount of the sample was about
2 g/batch. Either pure H2 or a CH4/H2 (1:4, v:v) mixture was
3
−1
.
passed through the sample at a total flow rate of 100 cm min
◦
−1
The temperature was increased at a linear rate of 5 C min
◦
to the final temperature (700 C), which was held for 1 h. The
samples were quenched to room temperature and then passi-
vated by a 1% O2/Ar mixture for 1 h. The activated carbon
supported carbides will be identified as MMo-P-m, where M,
P, and m are, respectively, the type of metal promotor (Fe, Co,
Ni), type of precursor (S = sulfate, N = nitrate), and synthesis
method (CM = conventional; CTH = carbothermal with H2).
The exact nature of the active sites in these systems is still
a matter of debate. It is believed that the carbide particles may
simply serve as a template on which a strained or highly dis-
persed, sulfided phase is formed. In this sense, carbon could be
considered as a textural promoters, acting to increase the number
of active sites instead of their activity [14]. It has been suggested
that during hydrotreating, these catalysts form a new active
site (MoSxCy), which can enhance the HDS activity [32,33].
Aegerter et al. [13] observed sulfidation of Mo2C during HDS
of thiophene and proposed that a thin layer of highly dispersed
MoS2 is present on the surfaces of the carbide. This model is
further supported by Bussell and co-workers [14,18]. Oyama et
al. found an increase in the HDS activity of molybdenum car-
bides during hydrogenation reactions in presence of thiophenic
compounds [5,34]. Szymanska-Kolasa et al. showed in recent
studies, the sulfidation at the surfaces of Mo2C during the HDS
process [35]. Sundaramurthy et al. confirmed by DRIFTS of
adsorbed CO that the surface of NiMo carbides are partially
sulfided during the hydrotreating reaction as a result of the
sulfidation [33].
2.2. Catalyst characterization
The activated carbon supported Fe-Mo, Co-Mo and Ni-Mo
carbides were characterized by X-ray diffraction (XRD), X-
ray photoelectron spectroscopy (XPS) and elemental CHNS
analysis. XRD analysis of the samples was carried out using
˚
a SIEMENS D-5005 diffractometer with a Cu K␣ (1.5456 A)
◦
◦
radiation source, within a range 10 ≤ 2θ ≤ 90 . The identifica-
tion of the different phases was made using the JCPDS library
[36] for Mo2C (card No. 11-0680), Fe (card No. 06-0696), Co
(card No. 15-0806), Ni (card No. 04-0850), Fe3Mo3C (card No.
47-1191), Co Mo C2 (card No. 80-0339) and Ni Mo C2 (card
6
6
6
6
No. 80-0337). The surface composition (XPS) of the carbides
was determined with a VG ESCALAB 220 i-XL photoelectron
spectrometer equipped with a dual (non monochromatic) Mg/Al
anode, operated at 400 W. The Al K␣ radiation (1486.6 eV) was
employed for the experiments reported here. All measurements
were performed under UHV, better than 10 Torr. Calibra-
tion of the instrument was done employing the Au 4f7/2 line
at 83.9 eV. Quantitation of the XPS signals and curve fitting
of the spectra was carried out with the XPSPEAK 4.1 and
XPS GRAPH routines after baseline subtraction by the Shirley
−
8
In the present work we have studied the effect of the type
of precursor salt and the carbiding method (conventional vs.
carbothermal) on the thiophene hydrodesulfurization activity of
activated carbon supported Fe-Mo, Co-Mo and Ni-Mo. To our

E. Puello-Polo, J.L. Brito / Journal of Molecular Catalysis A: Chemical 281 (2008) 85–92
87
method, employing typically a 80% Gausian-20% Lorentzian
combination and tabulated atomic sensitivity factors. Due to the
relatively insulating character of samples, internal referencing
of spectrometer energies was made using the dominating C 1s
peak of the support at 284.6 eV. Binding energies reported in the
current study were accurate to within 0.2 eV. Analysis of sulfur
content in selected catalyst samples was carried out by oxida-
and (d)) show predominant peaks of Mo2C and metals, the pres-
ence of small amounts of the mixed Fe3Mo3C and Co Mo C2
phases is evident from the less intense peaks at 42.11 and 42.30,
respectively.
The XRD patterns of activated carbon supported Ni-Mo car-
bides are shown in the Fig. 1(e) and (f). Fig. 1(e) shows peaks
6
6
of Ni Mo C2 at 2θ = 43.11, 40.53, 47.16, 35.91, 73.74, 75.52,
6
6
◦
tion at 1000 C and chromatographic separation of the resulting
gases using an EA 1108 Carlo Erba CHNS-O instrument.
60.66, 32.86, 68.90, 55.25, and metallic Ni at 2θ = 44.51, 51.85,
and 76.37, while Fig. 1(f) shows diffraction peaks corresponding
to Mo2C and nickel.
2
.3. Catalytic test
3.2. XPS analysis
Prior to the catalytic reaction, the carburized and passivated
forms of activated carbon supported catalysts were activated in
situ under a 1 vol% CS2/H2 mixture at 400 C for 2 h, in order
X-ray photoelectron spectroscopy was used to obtain the sur-
◦
face composition of the carbides and the chemical environment
of Mo Fe, Co and Ni, which was estimated by curve fitting
of the Mo 3d, Fe 2p, Co 2p and Ni 2p spectra. Fig. 2(a)–(d)
show the Mo 3d spectra of the activated carbon supported Fe-
Mo carbides obtained with different synthesis method and type
of precursor. These figures are typical of those obtained with
the rest of Mo-containing catalysts. Regardless of the passi-
vation procedure, they show the presence at the surface of a
signal between 228.0 and 228.5 eV that can be attributed to a
Mo 3d5/2 peak from a carbide phase. Previously, a similar peak
to remove the passivation layer and attain a stable and repro-
ducible state at the surface. Tests of thiophene HDS were carried
◦
out in a fixed bed, continuous flow reactor, at 400 C and atmo-
spheric pressure. The test conditions were: 250 mg of catalyst,
3
−1
flow 100 cm min of the thiophene (2.27 mol%)/H2 mixture.
Reaction products were analyzed by means of gas chromatog-
raphy, with sampling of the gas effluent occurring at 10 min
intervals. After stabilization (≈3 h), the catalytic activities of
Fe-Mo, Co-Mo and Ni-Mo catalysts were determined.
δ+
denoted as Mo has been assigned to low oxidation state species
2+
3
. Results and discussion
(0 ≤ δ ≤ 2) [37], and also to Mo species derived from allyl
precursor on silica support [38]. Table 1 presents the B.E. of
the Mo 3d5/2 signals and the percent distribution of Mo oxida-
tion states that were obtained by deconvolution of the Mo 3d
3
.1. XRD analysis
envelopes. The presence of Mo4 (229.3–230.4 eV), and Mo
+
6+
Fig. 1 show selected results of the analysis by XRD. Indif-
ferently from the used synthesis method, the supported carbides
obtained from sulfate precursors showed diffraction peaks due to
mixed Fe3Mo3C or M Mo C2 phases (M = Co or Ni) plus peaks
due to the metals. On the other hand, the carbides obtained from
nitrate precursors showed the presence of ␤-Mo2C and metallic
phases. All figures show narrow and defined diffraction peaks
suggesting that these supported phases have good crystallinity.
Fig. 1(a) and (b) show XRD patterns of FeMo samples
obtained from different precursors. In Fig. 1(a), the peaks at
(232.0–232.5 eV) species can be identified, respectively, with
Mo (IV) (oxy-) carbide and Mo (VI) oxide [37,39]. The use of
nitrate precursors, indifferently from the used synthesis method,
generated a greater proportion of high oxidation state species,
6
6
4+
6+
Mo and Mo . The same is true of the use of the CTH method
with sulfate precursors, but the CM with sulfates generated a
δ+
larger proportion of Mo species. This may be due to a higher
Table 1
2
4
θ = 42.11, 39.60, 46.05, 71.85, 35.09, 73.58, 32.12, 64.24,
9.06, 54.60 are due to the Fe3Mo3C phase, and those at 2θ
Distribution of Mo oxidation states in activated carbon supported Fe-Mo, Co-
Mo, and Ni-Mo carbides obtained with different synthesis method and type of
precursor
of 44.67, 65.02, and 82.33 correspond to metallic iron. Inter-
estingly, the peaks due to Fe are more intense that those of
similarly prepared samples of the Co and Ni promoted carbides
Catalyst*
Mo 3d5/2
Mo␦+/eV (%)
Mo4+/eV (%)
Mo6+/eV (%)
(
Fig. 1(c)and (e)). Nopeaks due to sulfurcontainingphases were
observed. Fig. 1(b) shows peaks at 2θ = 39.49, 37.93, 34.47,
2.23, 69.64, 74.74, 61.65 and 44.67, 65.02, 82.33 due to ␤-
FeMo-S-CM
FeMo-N-CM
FeMo-S-CTH
FeMo-N-CTH
CoMo-S-CM
CoMo-N-CM
CoMo-S-CTH
CoMo-N-CTH
NiMo-S-CM
NiMo-N-CM
NiMo-S-CTH
NiMo-N-CTH
228.4 (15.07)
228.4 (32.51)
228.4 (12.13)
228.4 (5.68)
228.3 (14.96)
228.2 (25.33)
228.3 (14.23)
228.2 (19.63)
228.1 (61.70)
228.4 (78.15)
228.0 (28.51)
228.0 (34.90)
229.5 (45.49)
230.4 (10.66)
229.5 (33.05)
230.2 (7.14)
229.3 (47.76)
229.3 (18.59)
229.5 (54.98)
229.3 (22.19)
229.3 (10.85)
230.4 (8.96)
229.3 (16.36)
229.3 (17.57)
232.5 (39.44)
232.5 (56.83)
232.5 (54.82)
232.4 (87.19)
232.5 (37.28)
232.4 (56.08)
232.5 (30.79)
232.4 (58.18)
232.2 (27.45)
232.3 (12.89)
232.4 (55.13)
232.2 (47.53)
5
Mo2C, and also the signals due to iron.
The XRD patterns of CoMo carburized with different syn-
thesis method and precursors are shown in the Fig. 1(c) and
(
3
d). Fig. 1(c) shows peaks at 2θ = 42.30, 39.89, 46.39, 72.44,
5.34, 32.35, 87.71, 64.67, 69.49, 59.64, 49.39, 74.11, 78.75
due to the Co Mo C2 phase, while the other diffraction peaks
6
6
at 2θ of 44.22, 51.52, and 75.85 correspond to cobalt metal.
Fig. 1(d) shows diffraction peaks corresponding to Mo2C and
cobalt metal. Although, the diffraction patterns of FeMo and
CoMo carbides obtained by the carbothermal method (Fig. 1(b)
*
S = sulfate; N = nitrate; CM = conventional method; CTH = carbothermal
method.

8
8
E. Puello-Polo, J.L. Brito / Journal of Molecular Catalysis A: Chemical 281 (2008) 85–92
Fig. 1. XRD of activated carbon supported M-Mo carbides synthetized by the conventional (CM) and carbothermal (CTH) methods: (a) Fe-Mo from CTH and
sulfate precursor—Fe3Mo3C (᭹) and Fe (*); (b) Fe-Mo from CTH and nitrate precursor—␤-Mo2C (+), Fe (*) and Fe3Mo3C (ꢀ); (c) Co-Mo from CM and sulfate
precursor—Co6Mo6C2 (᭹) and Co (*); (d) Co-Mo from CTH and nitrate precursor—␤-Mo2C (+), Co (*) and Co6Mo6C2 (ꢁ); (e) Ni-Mo from CTH and sulfate
precursor—Ni6Mo6C2 (᭹) and Ni (*); (f) Ni-Mo from CTH and nitrate precursor—␤-Mo2C (+) and Ni (*).
reducibily induced by the conventional method that consists of a
temperature-programmed treatment under a CH4/H2 (1:4) mix-
ture together with the support, both acting as sources of carbon
atoms, while the carbothermal method employs only the support
as source of carbon atoms.
amounts of sulfur, expressed as atomic %, obtained from XPS
data and from elemental analysis for several catalyst samples,
including in the case of the latter technique samples submitted to
a sulfiding pretreatment. In can be seen that for samples derived
from sulfate precursors (not presulfided) the amount of sulfur
at the surface (%SXPS) is consistently higher than that in the
bulk (%SCNHS). This suggests that the sulfide species tend to
segregate towards the surface of the carbided phases.
The bimetallic carbides obtained from sulfate precursors
retained sulfur on the surface, as suggested by a quite visible
shoulderat226.5 eV(Fig. 2(a)and(b))closetotheMo3dregion,
and confirmed by the presence of two bands in the S 2p region, at
Fig. 4(a) shows the Fe 2p spectrum of the activated carbon
supported Fe-Mo carbide carburized by conventional method
and sulfate precursor. Independently from the used synthesis
method and type of precursor, the XPS analysis in the Fe 2p
region shows a peak between 710.9 and 711.5 eV, with a strong
2
− 2−
1
69 and 162 eV (Fig. 3), assigned to S and SO4 [39]. This
suggests that the carbides possess sulfided surfaces or mixed
carbo-sulfide species present on the surface of the catalysts.
Fig. 3 is typical of those obtained with the rest of Mo-containing
catalysts synthesized from sulfate precursors. Table 2 shows the
3
+
shake-up line at 717 eV. These signals correspond to Fe typi-

E. Puello-Polo, J.L. Brito / Journal of Molecular Catalysis A: Chemical 281 (2008) 85–92
89
Fig. 2. X-ray photoelectron spectra (Mo 3d region) of the activated carbon supported Fe-Mo carbides: (a) synthetized by CM from sulfate precursor; (b) synthetized
by CTH from sulfate precursor; (c) synthetized by CM from nitrate precursor; (d) synthetized by the CTH from nitrate precursor.
cal of a Fe3O4 structure [40], suggesting that the Fe species in
the surface of the carbide are easily oxidized. On the order hand,
the XPS analysis of the activated carbon supported Co-Mo car-
bide obtained by CM and sulfate precursor in the Co 2p region
(
Fig. 4(b)) shows a peak between 781.5 and 781.8 eV, together
2
+
with a strong shake-up line at 786 eV, characteristic of Co
2
+
species. The Co 2p3/2 band corresponding to Co is typical,
e.g., of a CoMoO4 structure [39], a possible precursor for the
formation of “Co-Mo-S” species involved in the catalytic activ-
ity of hydrodesulfirization reactions [41]. The Co-Mo carbides
have shown similar species on the surface indifferently from the
used synthesis method and type of precursor.
The major XPS features for the activated carbon supported
Ni-Mo carbides (Fig. 4(c) and (d)) are the peaks between
Table 2
Sulfur content (as atomic %) measured by XPS (SXPS) and by chemical analysis
(
SCHNS) of selected catalyst samples
Catalyst*
Non presulfided
Presulfided
% SCHNS
%
SXPS
% SCHNS
FeMo-S-CM
FeMo-S-CTH
CoMo-S-CM
CoMo-S-CTH
NiMo-S-CM
NiMo-S-CTH
CoMo-N-CM
CoMo-N-CTH
6.3
6.7
13.4
13.7
5.7
5.2
–
3.2
–
–
3.1
–
3.5
–
15.4
–
–
13.9
–
14.4
12.0
11.0
–
–
*
Fig. 3. X-ray photoelectron spectra (S 2p region) of the activated carbon sup-
ported Fe-Mo carbides synthetized from sulfate precursors: (a), CM; (b), CTH.
S = sulfate; N = nitrate; CM = conventional method; CTH = carbothermal
method.

9
0
E. Puello-Polo, J.L. Brito / Journal of Molecular Catalysis A: Chemical 281 (2008) 85–92
8
52.7–852.8 and 856.1–856.7 eV, which correspond to metal-
2+
lic Ni and Ni , respectively [42]. The strong shake-up line that
appears at 862.0 eV in these spectra is associated to the Ni
2
+
species. The nitrate precursor generated a greater proportion of
metallic Ni, while the sulfate precursor generated a larger abun-
dance of Ni2 species, which may be due to the lower thermal
stability of the nitrate precursors, as compared to the sulfates,
that results in generation of metallic Ni at low temperature due
to the reducing effect of the carbon support. Additionally, this
precludes the union between the nickel and molybdenum atoms
in a mixed phase such as NiMoO4 in the precursor and NiMo
carbide or NiMoS in the working catalyst. The sulfate moiety,
being more thermally stable, possibly allows maintaining the
bonding between Mo and Ni in a more extended range of tem-
perature. It is important to mention that Fig. 4 is similar to those
obtained with the rest of Fe, Co and Ni-containing catalysts.
+
3.3. Catalytic HDS activity
Thiophene HDS activities are reported as pseudo-first-order
rate constants for thiophene disappearance in units of moles
of thiophene converted to products per gram of catalyst per
minute (mol Th/g cat min) after ≈3 h of reaction time (steady
state). The samples were presulfided before the activity tests in
order to remove the passivation oxidic layer and to attain a repro-
ducible and stable initial state of the catalytic surface, assuming
that in the absence of any pretreatment oxygen surface species
would eventually be exchanged for sulfur in the feed. It has been
reported that passivated carbide or nitride catalysts submitted to
presulfidation attain higher HDS activities than those submit-
ted to either degassing in inert atmosphere or prereduction [18].
Table 2 shows that the amounts of sulfur in presulfided samples
are four to five times higher than those originally present in the
non presulfided, sulfate-derived samples. However, the amount
of sulfur is lower for the samples obtained from nitrates than for
those synthesized from sulfate precursors.
Figs. 5 and 6 show the effect of the type of precursor and the
synthesis method on the thiophene hydrodesulfurization activity
of activated carbon supported Fe-Mo, Co-Mo and Ni-Mo car-
bides. These figures show that the activity of carbides obtained
Fig. 4. X-ray photoelectron spectra (M 2p regions for M = Fe, Co and Ni) of
activated carbon supported bimetallic carbides: (a) Fe-Mo carbide synthetized
by CM from sulfate precursor; (b) Co-Mo carbide synthetized by CM from
sulfate precursor; (c) Ni-Mo carbide synthetized by CM from nitrate precursor;
(
d) Ni-Mo cabide synthetized by CM from sulfate precursor.
Fig. 5. Thiophene hydrodesulfurization activity at steady state of activated car-
bon supported bimetallic carbides prepared by the conventional method: (A)
Fe-Mo, (B) Co-Mo, and (C) Ni-Mo.

E. Puello-Polo, J.L. Brito / Journal of Molecular Catalysis A: Chemical 281 (2008) 85–92
91
Fig. 7. Thiophene hydrodesulfurization activity as a function of time of activated
carbon supported bimetallic carbides synthesized by the carbothermal method:
Fe-Mo (from sulfate) and Co-Mo (from nitrate), presulfided and non presulfided.
Fig. 6. Thiophene hydrodesulfurization activity at steady state of activated car-
bon supported bimetallic carbides prepared by the carbothermal method: (A)
Fe-Mo, (B) Co-Mo, and (C) Ni-Mo.
to the increased HDS activities of the nitrate catalysts when
they were presulfided (Fig. 7), showing that there is a change
in the nature of surface active sites of these catalysts during
HDS reaction. This points to the formation of a different active
site, which can enhance the HDS activities of carbide catalysts,
suggesting that the carbides with sulfided surfaces or mixed
carbo-sulfide species could be the most active ones in HDS on
carbide catalysts. On the other hand, the activity differences
between carbides prepared with sulfate precursor (either presul-
fided or not) were not significative (Fig. 7).
from sulfate precursor was greater than that of solids obtained
from nitrates. The differences in activity between conventional
and carbothermal hydrogen reduction methods were less signif-
icant. The Fe-Mo and Co-Mo carbide catalysts obtained from
sulfates showed thiophene HDS activities around five and two
times higher, respectively, than those of Fe-Mo and Co-Mo
carbides obtained from nitrate precursor, while both Ni-Mo car-
bides have similar catalytic activities, independently from the
used synthesis method and type of precursor.
The results of catalytic activity obtained in this work may
be correlated with the presence at the surface of molybdenum
species, such as Mo4 (Table 1), which could be identi-
fied with molybdenum in an oxo-carbide (or sulfo-carbide)
environment or coordinatively unsaturated (anionic vacancy)
sulfide centers; therefore the Mo4 species could be the
active centers [14]. The overall rate of the HDS of thiophene
was found to increase as follows: NiMo-N-CM < FeMo-N-
CTH < NiMo-S-CM < FeMo-N-CM < NiMo-S-CTH < NiMo-
N-CTH < CoMo-N-CM < CoMo-N-CTH < FeMo-S-
The partial sulfidation of these carbide catalysts has been evi-
denced in a variety of studies [32–36,45]. These works showed
that, it is possible that the carbide catalysts react with the H2S
gas produced during the HDS reaction and get sulfided. As a
result of sulfidation, there is a possibility that “MMoS” phase
(M = Fe, Co, or Ni) could be formed on the surface of carbide
catalysts. The “MMoS” are the most active species in sulfide cat-
alyst for HDS reactions [46,47]. The “MMoS” phases formed
on the carbide surface are beneficial for hydrotreating and would
enhance the HDS activities of carbide catalysts. Gaseous H2S,
produced during the pretreatment with CS2/H2, carries out the
sulfidation on the surface, and that way this treatment allows
to have mixed carbo-sulfide species at the beginning of HDS
reaction.
Fig. 7 shows that non presulfided, sulfate-derived sam-
ples present identical behavior in HDS of thiophene than the
corresponding presulfided samples. Thus, for sulfate-derived
samples, the relatively low sulfur content already present after
carbiding (∼3 atom %) seems to be the optimum content. The
large excess observed after presulfiding (∼15 atom %) is prob-
ably being retained in the microporous structure of the support,
by means of a “sink effect” as proposed by Laine and co-workers
[48]. Thus, sulfur added by means of presulfiding does not nec-
essarily result in an optimum sulfo-carbide active phase in the
case of nitrate precursors. In line with this, the amount of sulfur
in catalysts derived from nitrates is slightly lower than that in
catalysts obtained from sulfates (Table 2). These effects might
be related to the observation by several workers (see, e.g., [15])
of irreversible loss of activity of carbide catalysts after passi-
vation as compared with catalysts not submitted to passivation
procedures.
+
+
CTH < FeMo-S-CM < CoMo-S-CM < CoMo-S-CTH;
a
sequence that almost parallels the increase of Mo4 species
content as shown in Table 1. The same molybdenum oxidation
state was suggested by Manoli [43] to be the active site for
HDS, his work showed the presence of molybdenum atoms
+
4+
mainly as Mo , assigned to molybdenum oxycarbide; these
catalysts displayed a high desulfurization rate in the 4,6-
DMDBT hydrodesulfurization that was assigned to the Mo4+
abundance.
The results shown above demonstrate that the Ni-Mo car-
bides exhibited the lowest catalytic activities; this observation is
contradictory with the results obtained for conventional sulfide
catalysts. These results suggest that the nickel metal present on
the surface of Ni-Mo carbides caused deactivation during the
reaction, possibly due to the deposition of coke on the catalyst
surface, since this coke covers the active sites and deactivates
the catalysts [44].
The activated carbon supported Fe-Mo, Co-Mo and Ni-Mo
carbides obtained from the sulfate precursors retained sulfur
on the surface, as shown by XPS analysis (Figs. 2 (a and b)
and 3(a and b) and Table 2). This behavior may be related

9
2
E. Puello-Polo, J.L. Brito / Journal of Molecular Catalysis A: Chemical 281 (2008) 85–92
[5] S. Ramanathan, S.T. Oyama, J. Phys. Chem. 99 (1995) 16365.
4
. Conclusions
[
[
6] G.M. Dolce, P.E. Savage, L.T. Thompson, Energy Fuels 11 (1997) 3.
7] P. Da Costa, C. Potvin, J.M. Manoli, J.L. Lemberton, G. Perot, G. Mari-
adassou, J. Mol. Catal. A: Chem. 184 (2002) 323.
Theinfluenceofthetypeofprecursor(sulfatevs. nitrateofFe,
Co, or Ni promotor) and the synthesis method (conventional vs
carbothermal carbiding) of activated carbon supported Fe-Mo,
Co-Mo and Ni-Mo carbides was determined by XRD and XPS,
and also evidenced in thiophene hydrodesulfurization (HDS)
activity. The XRD analysis showed that the carbides obtained
[
8] J.C. Schlatter, S.T. Oyama, J.E. Metcalfe, J.M. Lambert, Ind. Eng. Chem.
Res. 27 (1988) 1648.
[9] P. Da Costa, C. Potvin, J.M. Manoli, B. Genin, G. Mariadassou, Fuel 83
2004) 1717.
10] V. Schwartz, S.T. Oyama, J.G. Chen, J. Phys. Chem. B 104 (2000) 8800.
11] B. Dhandapani, S. Ramanathan, C.C. Yu, B. Fruhberger, J.G. Chen, S.T.
Oyama, J. Catal. 176 (1998) 61.
(
[
[
with sulfate precursors generated Fe3Mo3C or M Mo C2 phases
6
6
(
M = Co or Ni), while the solids obtained from nitrate precursors
[
12] D.J. Sajkowski, S.T. Oyama, Appl. Catal. A: Gen. 134 (1996) 339.
formed ␤-Mo2C and metals. XPS showed the presence at the sur-
[13] P.A. Aegerter, W.W.C. Quigley, G.J. Simpson, D.D. Ziegler, J.W. Logan,
K.R. McCrea, S. Glazier, M.E. Bussell, J. Catal. 164 (1996) 109.
δ+
4+
6+
3+
2+
0
2+
face of Mo , Mo , Mo , Fe , Co , Ni and Ni species.
All the carbides show similar species of Fe, Co, and Ni on the
surface, however the relative proportions of Mo species depends
on the synthesis method and/or type of precursor. Thus, the cat-
alytic activity was influenced by the proportion of molybdenum
species at the surface. The carbothermal method and nitrate pre-
cursor indifferently from the used synthesis method generated
[
14] K.R. McCrea, J.W. Logan, T.L. Tarbuck, J.L. Heiser, M.E. Bussell, J. Catal.
71 (1997) 255.
1
[
15] L. Leclercq, M. Provost, H. Pastor, J. Grimblot, A. Hardy, L. Gengembre,
G. Leclercq, J. Catal. 117 (1989) 371.
[16] L.A. Santill a´ n-Vallejo, J.A. Melo-Banda, A.I. Reyes de la Torre, G.
Sandoval-Robles, J.M. Dom ´ı nguez, A. Montesinos-Castellanos, J.A. de
los Reyes-Heredia, Catal. Today 109 (2005) 33.
[
[
17] S. Korlann, B. D ´ı az, M.E. Bussell, Chem. Mater. 14 (2002) 4049.
18] B. D ´ı az, S.J. Sawhill, D.H. Bale, R. Main, D.C. Phillips, S. Korlann, R.
Self, M.E. Bussell, Catal. Today 86 (2003) 191.
4+
a greater proportion of high oxidation state species (Mo and
6
+
Mo ), whiletheconventionalmethodgeneratedalargerpropor-
δ+
tion of reduced species (Mo ). The results of catalytic activity
[19] H.K. Park, J.K. Lee, J.K. Yoo, E.S. Ko, D.S. Kim, K.L. Kim, Appl. Catal.
A: Gen. 150 (1997) 21.
4+
could be correlated with the presence at the surface of Mo ,
which shows an increase in the proportion of these species on
carbides prepared from the sulfate precursors. In consequence,
the activity of carbides obtained by the carbothermal method and
withsulfateprecursorswashigherthanthatofthesolidsobtained
by the conventional method and with nitrate precursors. How-
ever, no significant differences in catalytic activity were found
for the Ni-Mo carbides. From these results, it appears possible
to correlate the Mo4 content to HDS activity. The bimetallic
carbides obtained from sulfate precursors retained sulfur on the
surface, as shown by XPS and chemical analysis. Generally, the
carbides showed greater catalytic activity when they were pre-
sulfided, suggesting that the carbides with sulfided surfaces or
mixed carbo-sulfide species could be the active phase in HDS
on carbide catalysts. In the present work it is shown that the use
of sulfate precursors provides such mixed carbo-sulfide species
prior to the catalytic reaction, even without the need of presulfi-
dation, and that these phases are more active than those obtained
by presulfidation.
[
20] J.W. Logan, J.L. Heiser, K.R. McCrea, B.D. Gates, M.E. Bussell, Catal.
Lett. 56 (1998) 165.
21] J. Trawcynski, Appl. Catal. A: Gen. 197 (2000) 289.
22] G.M.K. Abotsi, A.W. Scaroni, Fuel Process. Technol. 22 (1989) 107.
[
[
[23] H. Farag, D.D. Whitehurst, K. Sakanishi, I. Mochida, Catal. Today 50
1999) 9.
24] B. Pawelec, R. Mariscal, J.L.G. Fierro, Appl. Catal. A: Gen. 206 (2001)
95.
25] M. Ferrari, S. Bosmans, R. Maggi, B. Delmon, P. Grange, Catal. Today 65
2001) 257.
[26] L. Volpe, M. Boudart, J. Solid St. Chem. 59 (1985) 348.
(
[
[
2
+
(
[
[
27] J.S. Lee, L. Volpe, H. Ribeiro, M. Boudart, J. Catal. 112 (1988) 44.
28] C. Pham-Huu, A.P.E. York, M. Ledoux, Ind. Eng. Chem. Res. 34 (1995)
1
107.
29] D. Mordenti, D. Brodzki, G. Dj e´ ga-Mariadassou, J. Solid St. Chem. 141
1998) 114.
[30] C. Liang, W. Ma, Z. Feng, C. Li, Carbon 41 (2003) 1833.
[
(
[
[
31] C. Sayag, G. Dj e´ ga-Mariadassou, Appl. Catal. A: Gen. 275 (2004) 15.
32] P. Liu, J.A. Rodriguez, J.T. Muckerman, J. Mol. Catal. A: Chem. 239 (2005)
116.
[
[
33] V. Sundaramurthy, A.K. Dalai, J. Adjaye, Catal. Today 125 (2007) 239.
34] B. Dhandapani, T. St. Clair, S.T. Oyama, Appl. Catal. A: Gen. 168 (1998)
219.
Acknowledgements
[35] A. Szymanska-Kolasa, M. Lewandowsky, C. Sayag, G. Dj e´ ga-
Mariadassou, Catal. Today 119 (2006) 7.
[
36] Power Dittraction File, International Center for Diffraction Data, Newtown
Square, PA, 1995.
37] M. Ledoux, G.P. Huu, J. Guille, H. Dunlop, J. Catal. 134 (1992) 383.
The authors would like to acknowledge financial support by
FONACIT (through Project G-2000001537) and the fine techni-
cal assistance of Carlos Bastidas (XRD) and Alberto Albornoz
[
[38] J. Aigler, J.L. Brito, P.A. Leach, M. Houalla, A. Proctor, N.J. Cooper, W.K.
Hall, D.M. Hercules, J. Phys. Chem. 97 (1993) 5699.
(XPS). E. P.-P. personally thanks the Surface Physicochemistry
[
[
[
39] L. Portela, J. Catal. 156 (1995) 243.
Laboratory and Dr. J.L. Brito, who have contributed in my scien-
tific education, the IVIC “Centro de Estudios Avanzados”, and
Julio A. Puello Polo.
40] M. Muhler, R. Schlogl, G. Ertl, J. Catal. 138 (1992) 413.
41] V. La Parola, G. Deganello, C.R. Tewell, A.M. Venezia, Appl. Catal. A:
Gen. 235 (2002) 171.
[42] X. Wang, U.S. Ozkan, J. Phys. Chem. B 109 (2005) 1882.
References
[43] J.-M. Manoli, P. Da Costa, M. Brun, M. Vrinat, F. Maug e´ , C. Potvin, J.
Catal. 221 (2004) 365.
[
[
[
[
[
44] D. Ferdous, A.K. Dalai, J. Adjaye, Fuel 85 (2006) 1286.
45] R.R. Chianelli, G. Berhault, Catal. Today 53 (1999) 357.
46] N.Y. Topsoe, H. Topsoe, J. Catal. 84 (1983) 386.
47] S.P.A. Louwers, R. Prins, J. Catal. 139 (1993) 525.
48] J. Laine, M. Labady, F. Severino, S. Yunes, J. Catal. 166 (1997) 384.
[
[
[
[
1] V. Babich, Fuel 82 (2003) 607.
2] E. Furimsky, Appl. Catal. A: Gen. 240 (2003) 1.
3] S.T. Oyama, Catal. Today 15 (1992) 179.
4] C.C. Yu, S. Ramanathan, B. Dhandapani, J.G. Chen, S.T. Oyama, J. Phys.
Chem. B 101 (1997) 512.