Efficient hydrodesulfurization catalysts based on Keggin polyoxometalates

Article abstract of DOI:10.1016/j.apcata.2015.10.001

Bulk and supported hydrodesulfurization catalysts based on Mo and W and containing Co or Ni as promoters and phosphorus as a modifier are prepared through the polyoxometalate route using Keggin type phosphomolybdates and phosphotungstates and tested in the HDS of thiophene at 350-400°C and 1 bar pressure. The corresponding oxidic pre-catalysts retain intact Keggin structure of the parent polyoxometalates and possess Br?nsted and Lewis acidity. In the course of sulfidation, the oxidic pre-catalysts transform into an active sulfidic phase with the loss of Keggin structure and catalyst acidity. Catalyst activity increases in the order of supports: SiO₂ < TiO₂ < γ-Al₂O₃. CoMoP/γ-Al₂O₃ catalyst prepared through the polyoxometalate route shows higher HDS activity and butene selectivity than industrial catalyst of comparable composition. The results indicate that polyoxometalate catalyst preparation route can be considered a performance enhancement methodology for HDS catalysis.

Full text of DOI:10.1016/j.apcata.2015.10.001

Accepted Manuscript
Title: Efficient hydrodesulfurization catalysts based on Keggin
polyoxometalates
Author: Jocelyn North Olivia Poole Abdullah Alotaibi
Hossein Bayahia Elena F. Kozhevnikova Ali Alsalme M.
Rafiq H. Siddiqui Ivan V. Kozhevnikov
PII:
S0926-860X(15)30178-2
DOI:
Reference:
http://dx.doi.org/doi:10.1016/j.apcata.2015.10.001
APCATA 15577
To appear in:
Applied Catalysis A: General
Received date:
Revised date:
Accepted date:
24-8-2015
23-9-2015
1-10-2015
Please cite this article as: Jocelyn North, Olivia Poole, Abdullah
Alotaibi, Hossein Bayahia, Elena F.Kozhevnikova, Ali Alsalme, M.Rafiq
H.Siddiqui, Ivan V.Kozhevnikov, Efficient hydrodesulfurization catalysts
based on Keggin polyoxometalates, Applied Catalysis A, General
http://dx.doi.org/10.1016/j.apcata.2015.10.001
This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.

Efficient hydrodesulfurization catalysts based on Keggin
polyoxometalates
Jocelyn North1, Olivia Poole1, Abdullah Alotaibi1, Hossein Bayahia1, Elena F.
Kozhevnikova1, Ali Alsalme2, M. Rafiq H. Siddiqui2, Ivan V. Kozhevnikov1*
1Department of Chemistry, University of Liverpool, Liverpool L69 7ZD, United Kingdom
2Department of Chemistry, King Saud University, P.O. Box 2455, Riyadh, Kingdom of
Saudi Arabia
* Corresponding author. Tel.: +44(0)1517942938
E-mail address: kozhev@liverpool.ac.uk (I. V. Kozhevnikov)
1

Graphical abstract
2

Highlights
 HDS pre-catalysts prepared via polyoxometalate (POM) route retain the POM
structure and acidity.
 The structure and acidity lost upon sulfidation to form HDS active phase(s).
 HDS activity increases in the series of supports: SiO2<TiO2<γ-Al2O3.
 POM-based CoMo catalyst more active and selective in thiophene HDS than
industrial counterpart.
3

Abstract
Bulk and supported hydrodesulfurization catalysts based on Mo and W and
containing Co or Ni as promoters and phosphorus as a modifier are prepared through the
polyoxometalate route using Keggin type phosphomolybdates and phosphotungstates and
o
tested in the HDS of thiophene at 350-400 C and 1 bar pressure. The corresponding
oxidic pre-catalysts retain intact Keggin structure of the parent polyoxometalates and
possess Brønsted and Lewis acidity. In the course of sulfidation, the oxidic pre-catalysts
transform into an active sulfidic phase with the loss of Keggin structure and catalyst
acidity. Catalyst activity increases in the order of supports: SiO2 < TiO2 < γ-Al2O3.
CoMoP/γ-Al2O3 catalyst prepared through the polyoxometalate route shows higher HDS
activity and butene selectivity than industrial catalyst of comparable composition. The
results indicate that polyoxometalate catalyst preparation route can be considered a
performance enhancement methodology for HDS catalysis.
Key words: desulfurization, thiophene, heterogeneous catalysis, polyoxometalate,
molybdenum, cobalt.
4

1. Introduction
Hydrodesulfurization (HDS) driven by ever stringent environmental legislation is
one of the most important processes of the petroleum industry. It removes sulfur from oil
fractions by catalytic hydrotreatment to upgrade the quality of fuels [1,2]. Present day
HDS technology largely employs sulfided Co(Ni)Mo/γ-Al2O3 catalysts [1-3]. These
catalysts, although durable, are not sufficiently active and selective to achieve the
required very low sulfur content in transportation fuels. Consequently, research into
improvement of HDS catalysts continues all over the world. Recent developments
include new methods of preparation of Co(Ni)Mo catalysts such as chemical vapor
deposition improving the dispersion of active phases [4], addition of phosphorus [4] and
citric acid [5], new active phases for HDS catalysts, e.g., noble metals [6], transition
metal carbides [7], nitrides [8] and phosphides [9,10].
The Co(Ni)Mo/Al2O3 HDS catalysts are obtained by in-situ sulfidation of an
oxidic precursor (pre-catalyst), which is usually prepared by impregnation of γ-alumina
support with appropriate Mo and Co(Ni) compounds from aqueous solution, most
typically using ammonium heptamolybdate and adding Co(II) or Ni(II) nitrate as a
promoter, often with addition of phosphoric acid as a modifier [1-3]. Depending on the
pH and P and Mo concentrations, such impregnation solutions contain P-Mo heteropoly
6-
3-
6-
anions such as P2Mo5O236-, PMo9O31(OH)3 , PMo11O397-, PMo12O40 and P2Mo18O62
[2]. On the other hand, heteropoly compounds, especially molybdophosphates
incorporating Co(II) and/or Ni(II) promoters as counter cations or addenda atoms, can be
used directly for the preparation of oxidic precursors for HDS catalysts [11-23]. In the
past decades, heteropoly compounds, also known as polyoxometalates (POM) [24], have
5

attracted much interest as catalysts of various reactions and as such found application in
several large-scale industrial processes [25-27]. Various types of heteropoly compounds
have been used for the preparation of HDS pre-catalysts, the common Keggin type
heteropoly compounds comprising heteropoly anions of the composition [XM12O40]n- (M
= Mo6+, W6+; X = P5+, Si4+) most frequently applied. Catalyst preparation methodology
through the polyoxometalate route has important advantages such as (i) incorporation of
all the elements required for HDS catalyst within heteropoly compound thus allowing the
preparation of oxidic precursor in a single impregnation step, (ii) close interaction
between the key elements in the pre-catalyst and sulfided catalyst and (iii) extraneous
counter ions can be excluded from the impregnation solution to reduce waste [16,17].
In this work, which follows up our recent study on polyoxometalate NiMo/SiO2
HDS catalysts [23], we prepared bulk and supported Co and Ni phosphomolybdate and
phosphotungstate catalysts via the POM route from Co(II) and Ni(II) salts of
phosphomolybdic H3[PMo12O40] and phosphotungstic H3[PW12O40] heteropoly acids
(HPA) and tested for their activity in the HDS of thiophene, which is relevant to the
hydrotreating of FCC naphtha [1,2]. Our aim was to investigate the effect of POM
composition and catalyst support on the catalytic activity, the role of Brønsted and Lewis
acidity of the pre-catalyst and to follow the evolution of pre-catalyst in the course of HDS
reaction. It is demonstrated that CoMo/Al2O3 catalyst prepared through the POM route
has the potential of reducing hydrogen consumption in HDS process and the loss of high
octane alkenes in the final gasoline pool. Therefore, the POM catalyst preparation route
can be considered a performance enhancement methodology for HDS catalysis.
6

2. Materials and methods
2.1. Chemicals and materials
Phosphomolybdic and phosphotungstic acid hydrates (H3[PMo12O40]∙12H2O and
H3[PW12O40]∙28H2O), Ni(NO3)2∙6H2O, Co(NO3)2∙6H2O and thiophene (99%) were from
Sigma-Aldrich. Catalyst supports Aerosil 300 silica (SBET = 300 m2g-1), P25 titania
(anatase/rutile = 3:1, SBET = 51 m2g-1) and Aluminiumoxid C γ-alumina (SBET = 120 m2g-
1) were from Degussa.
2.2. Preparation of HDS pre-catalysts
Bulk Ni(II) and Co(II) salts of H3[PMo12O40] and H3[PW12O40] with M(II)/HPA
molar ratios of 1:1 and 1.5:1 (i.e., M(II)/Mo(VI) or M(II)/W(VI) ratio of 1:12 and 1:8) were
prepared by dissolving the HPA in a minimum amount of deionized water at room
temperature followed by addition of the required amount of the corresponding metal nitrate
(Co(NO3)2 or Ni(NO3)2) with stirring. Excess water was distilled off using a rotary
o
evaporator and the resulting residue oven dried at 100 C overnight. The solids were
o
calcined in air at 350 oC for 2 h with a heating rate of 10 C min-1and finally ground into a
powder of 45-180 μm particle size.
Supported H3[PMo12O40] and H3[PW12O40] pre-catalysts with 30 wt% HPA loading
based on anhydrous HPA were prepared by wet impregnation of support (γ-Al2O3, SiO2 or
TiO2) with an aqueous solution of HPA at room temperature with stirring for 2 h. Solid
o
residues were isolated by rotary evaporation, oven dried at 100 C and air calcined at 350
o
oC for 2 h at a heating rate of 10 C min-1. The pre-catalysts were then ground into a
7

powder with a particle size of 45-180 μm. γ-Al2O3-supported Ni(II) and Co(II)
phosphomolybdate and phosphotungstate pre-catalysts with M(II)/HPA molar ratios of 1:1
and 1.5:1 and 30 wt% POM loading were prepared similarly by impregnation of γ-Al2O3
with HPA and Ni(NO3)2 or Co(NO3)2 in required quantities from aqueous solution. The
slurry was stirred for 2 h at room temperature followed by rotary evaporation to remove
o
o
water. The residue was dried overnight at 100 C and air calcined at 350 C for 2 h
resulting in lightly colored powders, which were ground into 45-180 μm particle size.
Industrial CoMo/Al2O3 pre-catalyst (ICI Catalyst 41-6, Co/Mo = 0.62 mol/mol),
containing CoO (4.0%), MoO3 (12.0%), SiO2 (1.0%) and γ-Al2O3 (balance) [3], was
crushed and sieved to a powder with a particle size of 45-180 μm. It had a surface area of
220 m2g-1 and a pore volume of 0.6 cm3g-1 [3]. For comparison with the industrial ICI 41-6
catalyst, 15%Co1.5[PMo12O40]/Al2O3 pre-catalyst containing similar Mo loading (13%
MoO3) but a lower Co/Mo molar ratio of 0.125 was prepared by grinding
30%Co1.5[PMo12O40]/Al2O3 with equal quantity of γ-Al2O3 support. The powder was
pressed into a pellet then crushed to a 45-180 μm catalyst powder.
2.3. Techniques
Catalyst surface area, pore volume and pore size were determined on a
o
Micromeritics ASAP 2010 instrument by measuring nitrogen adsorption at -196 C.
o
Before measurement the catalysts were pre-treated at 240 C in vacuum. Powder X-ray
diffractograms were collected on a PANalytical Xpert diffractometer using Cu Kα
radiation (λ = 0.1542 nm). ICP-AES analysis of sulfur was carried out on a Spectro Ciros
emission spectrometer. Fourier transform infrared (FTIR) spectra of pre-catalysts and
8

spent catalysts were recorded on a Nicolet Nexus FTIR spectrometer using powdered
catalyst mixtures with KBr. In FTIR spectra of γ-Al2O3-supported catalysts, the
absorption of support was subtracted using a γ-Al2O3-KBr mixture as the background.
DRIFT (diffuse reflectance infrared Fourier transform) spectra of adsorbed pyridine were
taken on the same spectrometer. Catalyst samples were ground with KBr (1-10 wt% in
KBr) and pre-treated at 150 oC/10-3 kPa for 1 h. The samples were then exposed to
pyridine vapor at room temperature for 1 h, followed by pumping out at 150 oC/10-3 kPa
for 1 h to remove physisorbed pyridine. The DRIFT spectra of adsorbed pyridine were
recorded at room temperature at a 4 cm−1 resolution. Temperature programmed reduction
(H2-TPR) of catalysts was carried out on a Micromeritics TPD/TPR 2900 apparatus
equipped with a thermal conductivity detector. Catalyst samples (20-30 mg) were heated
up to 950 oC at a rate of 10 oC min-1 in a H2-N2 (5:95) gas flow (60 mL min-1).
2.4. Catalyst testing
o
The HDS of thiophene was carried out in flowing H2 at 350-400 C under
atmospheric pressure in a down-flow quartz fixed-bed reactor (9 mm i.d.) with online GC
analysis (Varian 3800 instrument with a 30 m × 0.25 mm × 0.5 μm HP-INNOWAX
capillary column and a flame ionization detector) (Supplementary material, Fig. S1). For
more accurate offline GC analysis of C1-C4 hydrocarbon products, a 60 m x 0.32 mm GS-
GasPro capillary column was used, which allowed for complete separation of these
hydrocarbons. The temperature in the reactor was controlled by a Eurotherm controller
using a thermocouple placed at the top of the catalyst bed. Thiophene was fed by passing
hydrogen flow controlled by a Brooks mass flow controller through a stainless steel
9

o
saturator which held liquid thiophene at 0 C to keep the chosen thiophene partial
pressure of 2.63 kPa in hydrogen flow [28]. The reactor was packed with 0.2-0.4 g of pre-
catalyst powder of 45–180 μm particle size. The gas feed entered the reactor at the top at
a flow rate of 10-20 mL min-1 (space time W/F = 152-608 g h mol-1, where W (g) is the
catalyst weight and F (mol h-1) is the molar flow rate of thiophene). Prior to reaction
monitoring, the pre-catalysts were sulfided in situ in the feed flow containing thiophene
(2.63 kPa) in H2 for 1 h at the reaction temperature. The downstream flow was analyzed
by the on-line GC to continuously monitor thiophene conversion. Product selectivity was
determined at selected times on stream by offline GC analysis using the GS-GasPro
capillary column. The selectivity was defined as the percentage of thiophene converted
into a particular product and quoted in mol. %. The mean absolute percentage error in
conversion and selectivity was ≤ 5% and the carbon balance was maintained within 95%.
3. Results and discussion
3.1. Pre-catalyst characterization
Bulk HPA and M(II)-HPA (M = Co, Ni) pre-catalysts prepared had a very low
surface area (2.3-5.7 m2g-1) and pore volume (0.004-0.012 cm3g-1), which is typical of
these compounds [25-27]; their pore diameter was in the range of 58-114 Å (Table S1).
The M(II)-HPA pre-catalysts with the sub-stoichiometric M(II)/HPA molar ratio of 1:1
can be viewed as the partially substituted heteropoly salts MIIH[PMo12O40] and
MIIH[PW12O40]. For the stoichiometric M(II)/HPA ratio of 1.5:1, the pre-catalysts could
be attributed to neutral salts MII1.5[PMo12O40] and MII1.5[PW12O40], although HPA salts
10

prepared to be neutral often contain protons due to incomplete substitution and Keggin
ion degradation [27]. It should be noted that Ni(II) and Co(II) nitrates decompose to form
o
metal oxides upon calcination above 300 C: M(NO3)2 → MO + 2NO2 + O2 [29]. From
powder X-ray diffraction (XRD), bulk H3[PMo12O40] and H3[PW12O40] calcined at 350
oC were crystalline materials with close packed cubic structure. Bulk metal salts of
H3[PW12O40] were crystalline too, whereas the salts of H3[PMo12O40] were amorphous as
exemplified for Co(II) salts in Fig. 1.
Supported HPA and M(II)-HPA pre-catalysts (30 wt% loading) prepared through
Keggin POMs were mesoporous materials with a BET surface area of 29-172 m2g-1, 0.14-
0.86 cm3g-1 pore volume and 119-249 Å pore diameter depending on support (γ-Al2O3,
SiO2 or TiO2) (Table S2). The Co1.5[PMo12O40]/Al2O3 pre-catalyst, which showed the
highest activity among the catalysts studied (see below), had a surface area of 69 m2g-1,
0.27 cm3g-1 pore volume and 157 Å pore diameter. For this catalyst, nitrogen adsorption
isotherm and pore size distribution are shown in Fig. S2 and Fig. S3.
H3[PMo12O40] exhibited a crystal phase on SiO2 and Al2O3, but was amorphous
on TiO2 (Fig. S4). H3[PW12O40] was crystalline on SiO2 and TiO2, but amorphous on
Al2O3 (Fig. S4). Co(II) and Ni(II) salts of H3[PMo12O40] were crystalline on Al2O3 (Fig.
2), whereas the salts of H3[PW12O40] did not show POM crystal phase on Al2O3 (Fig. S5).
Regardless of crystallinity, the Keggin primary structure of PMo and PW
heteropoly anions remained intact in all these materials, bulk and supported, as
demonstrated by infrared spectroscopy. Fig. 3 and Fig. 4 show the DRIFT spectra of bulk
Co1.5[PMo12O40] and NiH[PW12O40] pre-catalysts (powdered with KBr), which match
perfectly the literature spectra of H3[PMo12O40] and H3[PW12O40] and their salts [30]. Fig.
11

5 shows the DRIFT spectrum of Co1.5[PMo12O40]/Al2O3 pre-catalyst, with the strong
absorption band of alumina at 900 cm-1 largely subtracted. It also matches quite well the
literature spectra of bulk H3[PMo12O40] and its salts (cm-1): 1062-1068 (P-O), 954-963
(Mo=O), 869-880 (Mo-O-Mo corner-sharing) and 785-805 (Mo-O-Mo edge-sharing)
[30]. In the spectrum of Co1.5[PMo12O40]/Al2O3, the weak Mo-O-Mo bands at 865 and
803 cm-1 are partly obscured by the γ-Al2O3 band at 900 cm-1.
M(II)-HPA pre-catalysts possess Brønsted and Lewis acidity owing to the presence
of heteropoly acid protons and M(II). These materials showed the characteristic peaks of
Brønsted (~1540 cm-1) and Lewis (~1450 cm-1) acid sites in their DRIFT spectra of
adsorbed pyridine, together with the band at 1490 cm-1 attributed to both types of acid sites
[31] as exemplified for bulk NiH[PMo12O40] in Fig. 6. This acidity, however, was lost
during the HDS reaction as a result of catalyst sulfidation (see below).
Fig.
7
shows the H2-TPR (temperature programmed reduction) of
H3[PMo12O40]/Al2O3 and Co1.5[PMo12O40]/Al2O3 pre-catalysts. Both exhibit two
reduction peaks at about 550 and 800 oC. The Keggin structure of HPMo would probably
collapse to form MoO3 and P2O5 before the reduction took place. These peaks can be
attributed to reduction of Mo(VI) to Mo(IV). As Mo(VI) is in large excess over Co(II) in
the Co1.5[PMo12O40]/Al2O3 pre-catalyst, reduction of Co(II) is probably masked by the
Mo peaks. Similar H2-TPR results have been reported previously for Ni(II)-
[PMo12O40]/SiO2 [23] and non-promoted HPMo/TiO2 [17].
12

3.2. Catalyst performance in HDS of thiophene
3.2.1. Bulk catalysts
Bulk M(II)-HPA pre-catalysts have low surface areas and porosities (Table S1)
and therefore could be expected to have low HDS activities. Nevertheless, it is interesting
to test these catalysts because their evolution in the course of HDS reaction is easier to
follow with spectroscopic techniques as compared to supported catalysts.
Table 1 shows the performance of the bulk HPA and M(II)-HPA catalysts in the
HDS of thiophene at 400 oC and a space time W/F = 152 g h mol-1. It is clearly seen that
tungsten catalysts possessing more stable POM structure [27] are less active (<1%
thiophene conversion) than more labile molybdenum catalysts (2-3.5% conversion). Ni-
Mo catalysts show marginally higher activities than Co-Mo ones, the latter being
practically equal in activity to the non-promoted H3[PMo12O40]. The main reaction
products include n-butane (10-21% selectivity) and n-butenes (1-butene, trans-2-butene
and cis-2-butene) (30-59%), as represented by eq. (1). On top of that, significant amount
of C1-C3 cracking by-products (25-60%) and a little of isobutane was also formed, which
may be explained by strong acidity of the bulk HPA pre-catalysts (see above).
All bulk catalysts, initially lightly colored, turned black after reaction. This was
also the case for supported M(II)-HPA catalysts (see below). The change of color
indicates catalyst sulfidation. Indeed, from post-reaction characterization, sulfur content
13

in spent Ni1.5[PMo12O40] and Co1.5[PMo12O40] catalysts was found to be 7.5 and 4.5 wt%,
which corresponds to 18 and 10% catalyst sulfidation, respectively, as estimated
assuming formation of MoS2 and NiS or CoS. This implies transformation of M(II)-
H3[PMo12O40] pre-catalysts into active metal sulfide phases. A small amount of coke was
also found in spent catalysts (0.32% carbon content in spent Co1.5[PMo12O40]).
Amorphous bulk Co(II)-H3[PMo12O40] and Ni(II)-H3[PMo12O40] pre-catalysts
remained amorphous after HDS reaction (Fig. 1). It was found that they lost their Keggin
structure during the reaction as can be seen from Fig. 3, which shows the DRIFT spectra
for Co1.5[PMo12O40] (pre-catalyst versus spent catalyst). In contrast, the more stable and
less active M(II)-H3[PW12O40] catalysts retained the Keggin structure after reaction as
clearly seen from the DRIFT spectra for NiH[PW12O40] (Fig. 4). As to the M(II)-
H3[PMo12O40] catalysts, not only their Keggin structure but also their acidity, both
Brønsted and Lewis, was lost in the course of HDS reaction, as seen from the DRIFT
spectra of adsorbed pyridine for the NiH[PMo12O40] pre-catalyst and spent catalyst (Fig.
6).
The overall picture emerging from these results is that the M(II)-H3[PMo12O40]
oxidic pre-catalysts with labile POM structure readily undergo sulfidation with the loss of
Keggin structure and acidity to form an HDS active sulfidic phase. In contrast, the more
stable M(II)-H3[PW12O40] pre-catalysts are rather resistant to sulfidation hence having
difficulty to form active HDS catalysts, which results in their low catalytic activity.
14

3.2.2. Supported catalysts
Table 2 shows the HDS of thiophene over non-promoted supported HPA catalysts
at 400 oC. These catalysts had considerably higher activities than their bulk counterparts,
which can be attributed to higher HPA dispersion in supported catalysts. Again, for the
reasons given above, Mo catalysts showed higher HDS activities in comparison to W
catalysts. The type of support played an important role; the HDS activity (thiophene
conversion) of Mo catalysts increased in the order of supports: SiO2 < TiO2 < γ-Al2O3,
which is in agreement with previous reports [1,2]. This activity order does not correlate
with support surface area (Table S2), but is in line with increasing HPA-support
interaction in this series [27,32], which might increase dispersion of active species on the
surface of support. It should be noted that the supported HPA catalysts were much more
selective than the bulk ones, giving butane-butene product mixtures with only traces of
C1-C3 cracking by-products (cf. results in Table 1 and Table 2). Therefore, as the
alumina-supported Mo catalysts showed better HDS activities, our further studies were
focused on these catalysts.
o
Table 3 shows the HDS of thiophene at 350-400 C over γ-Al2O3-supported
catalysts promoted by Co and Ni prepared through the POM route. It is evident that Mo
catalysts are again much more active than their W counterparts. As expected, addition of
Co and Ni promoters significantly increased thiophene conversion. The catalyst activity
also increased with increasing the M(II)/HPA molar ratio. In all cases the catalyst activity
predictably increased with increasing the temperature from 350 to 400 oC.
Among the catalysts studied, 30%Co1.5[PMo12O40]/Al2O3 was found to be the
most active one in the HDS of thiophene, closely followed by 30%Ni1.5[PMo12O40]/Al2O3
15

(Table 3). The 30%Co1.5[PMo12O40]/Al2O3 catalyst was tested for a longer time on stream
o
(132 h in total) at different space times; the results are shown in Fig. 8. At 350 C and
W/F = 152 g h mol-1, thiophene conversion gradually increased with the time on stream
from 47 to 55% in 4 h. This can be explained by increasing catalyst sulfidation to form an
active sulfidic phase. At 400 oC and the same space time, the sulfided catalyst performed
with a steady conversion of 83%. A notable increase in butene selectivity occurred along
the time on stream (from 38 to 47%) at the expense of n-butane formation. This may be
explained by inhibition of catalyst hydrogenation activity due to coke formation. The
conversion further scaled with the W/F to 96% at W/F = 304 g h mol-1 and 99% at 608 g
h mol-1 without evidence of catalyst deactivation. The selectivity to n-butane increased
with thiophene conversion at the expense of butenes reaching 75% at 99% conversion
(Fig. 8). Only traces of C1-C3 cracking by-products were observed (Fig. S6), which can
be explained by the loss of catalyst acidity upon sulfidation.
After reaction, the lightly colored 30%Co1.5[PMo12O40]/Al2O3 pre-catalyst turned
black, however its texture practically did not change (Table S2). Post-reaction
characterization revealed only a small amount of coke in the spent catalyst (0.69 wt%
carbon content). The sulfur content was found to be 8.2 wt%. This corresponds to 64%
catalyst sulfidation assuming formation of MoS2 and CoS and implies transformation of
the pre-catalyst into an active metal sulfide phase on the alumina surface (hence the black
color). Upon sulfidation, the Keggin structure of P-Mo polyoxometalate was completely
destroyed as clearly seen from the DRIFT spectra of the pre-catalyst and spent catalyst
(Fig. 5). Previously, non-promoted H3[PW12O40] has been found to lose its Keggin
structure upon the HDS of dibenzothiophene as evidenced by 31P MAS NMR [15].
16

No correlation was found between the activity of the M(II)-H3[PMo12O40]
catalysts in the HDS of thiophene and the redox behavior of the corresponding pre-
catalysts in H2-TPR (Fig. 7). It is obvious that the HDS proceeds at much lower
o
o
temperatures (350-400 C) compared to the H2-TPR (550–800 C). This implies that
sulfidation of these pre-catalysts to form catalytically active phases occurs at lower
temperatures than their reduction with H2. It should be noted, however, that the H2-TPR
was carried out under transient conditions. Slow isothermal catalyst reduction with H2
may be possible at lower temperatures below 550 oC.
Thermodynamic analysis of reaction products provides further evidence about the
HDS process. The equilibrium ratio of n-butene isomers has been found to be 1-
o
butene/trans-2-butene/cis-2-butene = 1:1.8:1.2 at 400 C and 1 bar [33]. This is close to
the ratio obtained here with M(II)-H3[PMo12O40] catalysts (Table 3), which shows that
the butenes produced were at equilibrium. On the other hand, hydrogenation of butenes to
n-butane was far from equilibrium, with butane/butene molar ratios 2-3 orders of
magnitude lower than the equilibrium values. This indicates that butene-to-butane
hydrogenation was a slow process, and as a result, n-butane selectivity predictably
increased with thiophene conversion.
For comparison with industrial oxide pre-catalyst CoMo/γ-Al2O3 (ICI Catalyst 41-
6) with 12% MoO3 content and a Co/Mo molar ratio of 0.62 [3], the
30%Co1.5[PMo12O40]/Al2O3
pre-catalyst
was
diluted
with
γ-Al2O3
to
15%Co1.5[PMo12O40]/Al2O3 with a similar MoO3 content (13%) but a lower Co/Mo molar
o
ratio of 0.125. This catalyst was tested at 400 C to give 54% thiophene conversion with
77% total butene selectivity at W/F = 152 g h mol-1 and 74% conversion with 64% butene
17

selectivity at W/F = 304 g h mol-1 (Fig. 9, Table 3). The industrial catalyst ICI 41-6 under
the same conditions gave 37% conversion with 74% butene selectivity at W/F = 152 g h
mol-1 and 54% conversion with 62% butene selectivity at W/F = 304 g h mol-1 (Fig. 10,
Table 3). The spent ICI 41-6 catalyst contained 0.26% of coke and its sulfidation was
65% based on the formation of MoS2 and CoS, which is close to the sulfidation of the
30%Co1.5[PMo12O40]/Al2O3 catalyst (see above). Both 15%Co1.5[PMo12O40]/Al2O3 and
ICI 41-6 catalysts showed a slight decrease in n-butane selectivity with the time on
stream (Fig. 9 and 10) probably due to inhibition of their hydrogenation activity by
coking, similar to that displayed in Fig. 8. In the latter case, however, this trend is more
notable due to much longer time on stream.
Therefore, the Co1.5[PMo12O40]/Al2O3 catalyst, despite its lower Co/Mo ratio,
gave higher thiophene conversion compared to the industrial catalyst per equal catalyst
weight. It also gave higher butene selectivity thus exhibiting lower hydrogenation activity
compared to the industrial catalyst, which can be explained by the lower Co content.
Overall, these results demonstrate that the Co1.5[PMo12O40]/Al2O3 catalyst prepared
through the POM route has enhanced HDS performance compared to the conventional
industrial CoMo/Al2O3 catalyst. This can be attributed to close interaction between
oxomolybdenum species with Co promoter and to the presence of phosphorus modifier in
the Co1.5[PMo12O40]/Al2O3 catalyst.
18

4. Conclusions
It has been demonstrated that the Co1.5[PMo12O40]/Al2O3 catalyst (Co/Mo = 0.125
mol/mol) prepared through the POM route using Keggin type phosphomolybdate has
high activity in the HDS of thiophene. Compared to the industrial CoMo/Al2O3 catalyst
with similar Mo loading and a higher Co/Mo ratio of 0.62, the new catalyst shows higher
thiophene conversion and higher butene selectivity. Therefore Co1.5[PMo12O40]/Al2O3
catalyst has the potential of reducing hydrogen consumption in HDS process and the loss
of high octane alkenes in the final gasoline pool. On the basis of these results, the POM
catalyst preparation route is considered to be a performance enhancement methodology
for HDS catalysis.
Acknowledgements
This research was partly funded by the National Plan for Science, Technology and
Innovation (MAARIFAH), King Abdulaziz City for Science and Technology, Saudi
Arabia, Award No. 11-NAN-1852-02. We thank Shaqra University, Shaqra, Saudi Arabia
for PhD studentship (A. Alotaibi).
19

References
[1] I. V. Babich, J. A. Moulijn, Fuel 82 (2003) 607-31.
[2] R. Prins, in Handbook of Heterogeneous Catalysis, ed. G. Ertl, H. Knözinger, F.
Schüth, J. Weitkamp, Vol. 6, Wiley-VCH, Weinheim, 2008, p. 2695-2718.
[3] P. J. H. Carnell, in Catalyst Handbook, ed. M. V. Twigg, Wolfe Publishing,
London, 1989, p. 204-208.
[4] Y. Okamoto, Catal. Today 132 (2008) 9–17.
[5] N. Rinaldi, K. Usman, T. Al-Dalama, Y. Kubota, Okamoto, Appl. Catal. A 360
(2009) 130–136.
[6] A. Niquille-Röthlisberger, R. Prins, Catal. Today 123 (2007) 198–207.
[7] M. Lewandowski, A. Szyma´nska-Kolasa, P. Da Costa, C. Sayag, Catal. Today 119
(2007) 31–34.
[8] M. Nagai, Appl. Catal. A 322 (2007) 178–190.
[9] Y. Kanda, C. Temma, A. Sawada, M. Sugioka, Y. Uemichi, Appl. Catal. A 475
(2014) 410–419.
[10] R. Prins, M. E. Bussell, Catal. Lett. 142 (2012) 1413–1436.
[11] S. Damyanova, A. Spojakina, D. Shopov, Appl. Catal. 48 (1989) 177-186.
[12] A. M. Maitra, N. W. Cant, D. L. Trimm, Appl. Catal. 48 (1989) 187-197.
[13] A. Griboval, P. Blanchard, L. Gengembre, E. Payen, M. Fournier, J. L. Dubois, J.
R. Bernard, J. Catal. 188 (1999) 102–110.
[14] C. I. Cabello, I. L. Botto, H. J. Thomas, Appl. Catal. A 197 (2000) 79-86.
[15] R. Shafi, M. R. H. Siddiqui, G. J. Hutchings, E. G. Derouane, I. V. Kozhenikov,
Appl. Catal. A 204 (2000) 251-256.
20

[16] A. Griboval, P. Blanchard, E. Payen, M. Fournier, J. L. Dubois, J. R. Bernard, Appl.
Catal. A 217 (2001) 173–183.
[17] E. Kraleva, A. Spojakina, K. Jiratova, L. Petrov, Catal. Lett. 112 (2006) 203-212.
[18] R. Palcheva, A. Spojakina, K. Jiratova, L. Kaluza, Catal. Lett. 137 (2010) 216-223.
[19] K. Ben Tayeb, C. Lamonier, C. Lancelot, M. Fournier, A. Bonduelle-Skrzypczak,
F. Bertoncini, Appl. Catal. B 126 (2012) 55–63.
[20] P. A. Nikul’shin, A. V. Mozhaev, D. I. Ishutenko, P. P. Minaev, A. I. Lyashenko,
A. A. Pimerzin, Kinet. Catal. 53 (2012) 620-631.
[21] F. J. Méndez, A. Llanos, M. Echeverría, R. Jáuregui, Y. Villasana, Y. Díaz, G.
Liendo-Polanco, M. A. Ramos-García, T. Zoltan, J. L. Brito, Fuel 110 (2013) 249–
258.
[22] B. Pawelec, R. Mariscal, J. L. G. Fierro, A. Greenwood, P. T. Vasudevan, Appl.
Catal. A 206 (2001) 295–307.
[23] A. Alsalme, N. Alzaqri, A. Alsaleh, M. R. H. Siddiqui, A. Alotaibi, E. F.
Kozhevnikova, I. V. Kozhevnikov, Appl. Catal. B 182 (2016) 102-108.
[24] M. T. Pope, Heteropoly and Isopoly Oxometalates, Springer, Berlin, 1983.
[25] T. Okuhara, N. Mizuno, M. Misono, Adv. Catal. 41 (1996) 113-252.
[26] I. V. Kozhevnikov, Chem. Rev. 98 (1998) 171-198.
[27] I. V. Kozhevnikov, Catalysis by Polyoxometalates, Wiley & Sons, Chichester,
2002.
[28] D. R. Stull, Ind. Eng. Chem. 39 (1947) 517-550.
[29] S. Yuvaraj, L. Fan-Yuan, C. Tsong-Huei, Y. Chuin-Tih, S. Yuvaraj, L. Fan-Yuan,
C. Tsong-Huei and Y. Chuin-Tih, J. Phys. Chem. B 107 (2003) 1044-1047.
21

[30] C. Rocchiccioli-Deltcheff, M. Fournier, R. Franck, R. Thouvenot, Inorg. Chem. 22
(1983) 207-216.
[31] H. Knözinger, in Handbook of Heterogeneous Catalysis, ed. G. Ertl, H. Knözinger,
F. Schüth, J. Weitkamp, Wiley-VCH, Weinheim, 2008, p. 1154.
[32] A. M. Alsalme, P. V. Wiper, Y. Z. Khimyak, E. F. Kozhevnikova, I. V.
Kozhevnikov, J. Catal. 276 (2010) 181-189.
[33] J. Happel, M. A. Hnatow, R. Mezaki, J. Chem. Eng. Data 16 (1971) 206-209.
22

4
3
2
1
10
20
30
40
50
60
Cu Kα 2Θ (degrees)
o
Fig. 1. XRD for bulk catalysts (air calcined at 350 C for 2 h): (1) H3[PW12O40], (2)
Co1.5[PW12O40], (3) Co1.5[PMo12O40] and (4) Co1.5[PMo12O40] spent catalyst.
23

1
2
3
4
5
10
20
30
40
50
60
Cu Kα 2Θ (degrees)
Fig. 2. XRD for supported catalysts (air calcined at 350 oC for 2 h): (1)
30%H3[PMo12O40]/SiO2, (2) 30%Ni1.5[PMo12O40]/Al2O3, (3) 30%Co1.5[PMo12O40]/Al2O3,
(4) 30%Co1.5[PMo12O40]/Al2O3 spent catalyst, (5) γ-alumina.
24

1063
871
814
966
1200
1100
1000
900
800
700
600
500
Wavenumbers (cm-1)
Fig. 3. DRIFT spectra of bulk Co1.5[PMo12O40] pre-catalyst (solid line) and spent catalyst
(dashed line).
25

1080
891
891
983
984
816
826
1081
1100
1200
1000
900
800
700
600
500
Wavenumbers (cm-1)
Fig. 4. DRIFT spectra of bulk NiH[PW12O40] pre-catalyst (solid line) and spent catalyst
(dashed line).
26

803
865
964
1064
1200 1150 1100 1050 1000 950
900
850
800
750
700
Wavenumbers (cm-1)
Fig. 5. DRIFT spectra of 30%Co1.5[PMo12O40]/Al2O3 pre-catalyst (solid line) and spent
catalyst (dashed line).
27

1535
1446
1486
1580 1560 1540 1520 1500 1480 1460 1440 1420 1400 1380 1360
Wavenumbers (cm-1)
Fig. 6. DRIFT spectra of adsorbed pyridine on bulk NiH[PMo12O40]: pre-catalyst (solid
line) and spent catalyst (dashed line).
28

55
50
45
40
35
1
2
0
200
400
600
800
1000
Temperature (°C)
Fig.
7.
H2-TPR
for
pre-catalysts:
(1)
30%H3[PMo12O40]/Al2O3,
(2)
30%Co1.5[PMo12O40]/Al2O3.
29