Hierarchization of Mordenite as NiW Sulfide Catalysts Support: Towards Efficient Hydrodesulfurization

Article abstract of DOI:10.1002/cctc.201500686

NiW-based catalysts were supported on Al₂O₃, commercial mordenite (HM), and hierarchical mordenite prepared by acid-base-acid treatment (HM-M). Their catalytic performance was evaluated in the hydrodesulfurization (HDS) of dibenzothiophene (DBT) and 4,6-dimethyldibenzothiophene (4,6-DMDBT) at 613K under 4MPa as total pressure. In the HDS of DBT, NiW/Al₂O₃ exhibited better catalytic activity than NiW/HM and NiW/HM-M. On the contrary, NiW/HM-M exhibited the highest catalytic performance in the HDS of 4,6-DMDBT, attributed to different factors: an improved dispersion of the active phase, a better accessibility to the active sites owing to the creation of mesoporosity, and an additional reaction route (isomerization) owing to the acidic properties of the zeolitic support. These results show the potential interest of hierarchical mordenite as support for HDS catalysts, with the presence of mesoporosity beneficial to the HDS of large and refractory molecules. Hierarchy rules: In the hydrodesulfurization of dibenzothiophene (DBT) and 4,6-dimethyldibenzothiophene (4,6-DMDBT), NiW on mesoporous mordenite shows higher activity than mordenite-based catalysts, and, in the latter case, also higher activity than on alumina, attributed to an improved dispersion of the active phase, a better accessibility to the active sites, and the additional isomerization route owing to the acidic properties of the zeolitic support.

Full text of DOI:10.1002/cctc.201500686

DOI: 10.1002/cctc.201500686
Full Papers
Hierarchization of Mordenite as NiW Sulfide Catalysts
Support: Towards Efficient Hydrodesulfurization
Yi Wang,[a, b] Christine Lancelot,*[b] Carole Lamonier,[b] FrØdØric Richard,[c] Kunyue Leng,[a]
Yinyong Sun,*[a] and Alain Rives[b]
NiW-based catalysts were supported on Al2O3, commercial
mordenite (HM), and hierarchical mordenite prepared by acid–
base–acid treatment (HM–M). Their catalytic performance was
evaluated in the hydrodesulfurization (HDS) of dibenzothio-
phene (DBT) and 4,6-dimethyldibenzothiophene (4,6-DMDBT)
at 613 K under 4 MPa as total pressure. In the HDS of DBT,
NiW/Al2O3 exhibited better catalytic activity than NiW/HM and
NiW/HM–M. On the contrary, NiW/HM–M exhibited the highest
catalytic performance in the HDS of 4,6-DMDBT, attributed to
different factors: an improved dispersion of the active phase,
a better accessibility to the active sites owing to the creation
of mesoporosity, and an additional reaction route (isomeriza-
tion) owing to the acidic properties of the zeolitic support.
These results show the potential interest of hierarchical mor-
denite as support for HDS catalysts, with the presence of
mesoporosity beneficial to the HDS of large and refractory
molecules.
Introduction
As one important refinery process, hydrodesulfurization (HDS)
of transportation fuels received much attention in recent years.
Owing to the stringent demand of the environmental regula-
tions, the sulfur level in fuels must be reduced to 10 ppm in
many countries.[1–3] However, owing to the presence of refrac-
tory sulfur-containing molecules such as 4,6-dimethyldibenzo-
thiophene (4,6-DMDBT), deep desulfurization of diesel fuels is
difficult to achieve because of the effect of steric hindrance
from the alkyl groups adjacent to the sulfur atom.[4–9] There-
fore, the design and development of highly efficient HDS cata-
lysts that can eliminate the effect of steric hindrance have
been pursued.
enhancement of activity was also related to the electron-defi-
cient character of the sulfide particles, owing to electronic
transfer between the acidic zeolite and the active-phase parti-
cles.[13] Additionally, it was also proposed that the interaction
of active phases with acidic supports could be weakened and
thus the sulfidation degree of active phases may be im-
proved.[14] Based on these reasons, zeolites with strong acidity
have been selected as supports of HDS catalysts.[15] However,
owing to their relatively small pore size, several problems such
as rapid deactivation[7,14] and diffusion limitation[16] were gener-
ally observed.
To remedy the disadvantage of microporous zeolites, hier-
archical zeolites were developed and have attracted much at-
tention because of the introduction of additional meso-
pores.[17–20] Several studies present their use as supports for
HDS catalysts. For example, noble-metal catalysts were sup-
ported on mesoporous zeolites such as ZSM-5,[21] Beta,[22] and
Y[23] for the HDS of 4,6-DMDBT. The results indicated that these
catalysts exhibited higher efficiency in desulfurization than
those on alumina or mesoporous silica. CoMoS sulfide phases
were supported on several mesoporous materials as a zeolite
L,[16,24,25] ZSM-5,[26] beta,[27] and mordenite (MOR),[28] allowing to
improve their activity in the HDS of FCC gasoline, DBT, and 4,6-
DMDBT, compared to those obtained on catalysts supported
on conventional zeolites.
The commercial catalysts are usually composed of molybde-
num disulfide promoted by Co or Ni atoms and supported on
alumina. The utilization of acidic supports has been suggested
to improve catalytic performance. One explanation is that they
enable dealkylation and isomerization of the alkyl substituents
so that the effect of steric hindrance may be relieved.[9–12] The
[a] Y. Wang, K. Leng, Prof. Y. Sun
School of Chemical Engineering and Technology
Harbin Institute of Technology
Harbin 150001 (P.R. China)
E-mail: yysun@hit.edu.cn
[b] Y. Wang, Dr. C. Lancelot, Prof. C. Lamonier, Dr. A. Rives
UnitØ de Catalyse et de Chimie du Solide, UMR CNRS 8181
UniversitØ Lille
On the other hand, it is well known that nitrogen-containing
and aromatic compounds in fuels would inhibit the HDS activi-
ty of catalysts.[29–31] Therefore, catalyst with good ability of
HDS, hydrodenitrogenation (HDN), and hydrodearomatization
(HDA) is required. Among the studied HDS catalysts, NiW cata-
lysts have better hydrogenation properties than NiMo catalysts
and have displayed superior catalytic activity in the HDA and
HDN reactions.[32] For this reason the use of NiW supported on
59655 Villeneuve d’Ascq (France)
E-mail: christine.lancelot@univ-lille1.fr
[c] Dr. F. Richard
Institut de Catalyse des Milieux et MatØriaux de Poitiers, UMR CNRS 7285
UniversitØ de Poitiers
86073 Poitiers Cedex 9 (France)
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/cctc.201500686.
ChemCatChem 2015, 7, 3936 – 3944
3936
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
zeolite as HDS catalysts was also reported in the litera-
ture.[14,32–39] However, to our best knowledge, the study of NiW
catalysts supported on hierarchical zeolites has not been re-
ported.
In the present work, we reported the synthesis and charac-
terization of NiW catalysts supported either on alumina, on
a mordenite commercial zeolite, and on a hierarchical (meso-
porous) mordenite obtained by acid–base–acid treatment. The
catalytic performance of these catalysts was measured by the
HDS of two model sulfur compounds: dibenzothiophene (DBT)
and 4,6-dimethyldibenzothiophene (4,6-DMDBT). The activity
and selectivity of the NiW catalyst based on hierarchical zeolite
was compared to those obtained using the parent zeolite and
alumina.
Results and Discussion
Characterization
The powder XRD patterns (Figure 1) of all the samples exhibit-
ed well resolved diffraction peaks, characteristics of MOR
framework structure (JCPDS 43-0171), indicating the preserva-
tion of the zeolitic structure after the post-treatment leading
Figure 2. N2 adsorption–desorption isotherms of supports and catalysts.
firming the mesoporosity formation as reported by Leng
et al.[40] If we assume that the area is only attributed to the
support, we can express the specific surface area (SSA) in
square meters of support instead of in square meters per gram
of catalyst by removing the contribution of the weight gain
consecutive to the introduction of the active phase, thus ena-
bling a direct comparison with the properties of the support
before impregnation. The textural properties of both HM and
HM–M decreased after impregnation (except for the micropore
volume), owing to possible pore plugging or small collapse of
the texture of the supports during impregnation. Moreover,
the decrease in BET surface area and mesopore volume during
Figure 1. X-ray diffraction patterns of supports and catalysts.
to the creation of mesopores and also after the impregnation
of Ni and W. Moreover, diffraction peaks assigned to WO3 and
NiO were never observed in these patterns (JCPDS 43-1035
and 022-1189, respectively). This means that tungsten and
nickel oxidic species were either completely amorphous or
composed of crystallites smaller than 4 nm, in agree-
ment with a good dispersion of these species.[14,32,34].
Table 1. Textural properties of supports and catalysts.
The N2 sorption isotherms of all samples exhibited
a hysteresis loop at high relative pressure characteris-
tic of a type IV isotherm (Figure 2). This indicates the
existence of hierarchical pores, particularly on HM–M
based materials. The detailed sorption data are pre-
sented in Table 1. After acid–base–acid treatment, the
micropore volume remained unchanged whereas the
external surface area (SA) and the mesopore volume
of the zeolite were increased by a factor of 2.5, con-
Samples
SSA
External SA[a] Mesopore volume[b] Micropore volume[a]
[m2 gÀ1
]
[m2 gÀ1
]
[cmÀ3 gÀ1
]
[cmÀ3 gÀ1
]
HM
NiW/HM
HM-M
411
45
0.07
0.17
283/341[c]
471
29/35[c]
0.03/0.04[c]
0.17
0.14/0.17[c]
0.18
112
61/73[c]
NiW/HM–M
360/434[c]
0.12/0.14[c]
0.15/0.18[c]
[a] t-plot method. [b] BJH method. [c] Values after correction because of the contribu-
tion of the weight gain after the introduction of the active phase.
ChemCatChem 2015, 7, 3936 – 3944
www.chemcatchem.org
3937
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
impregnation appeared to be lower on HM–M than on HM
support, 8% instead of 17% for the BET surface area and 17%
instead of 43% for the mesopore volume, respectively. This in-
dicates a better preservation of the textural properties in the
case of hierarchical mordenite. In addition, the micropore
volume remained constant, suggesting that most of the NiW
species were distributed on external surface and in the meso-
pores of mordenite. Notably, the external surface area and
mesopore volume for NiW/HM–M catalyst were respectively
twice and four times higher than that of NiW/HM, suggesting
a possible better accessibility to large molecules for this solid.
The acidity of both HM and HM–M samples was determined
by pyridine adsorption (Table 2). The acid–base–acid treatment
Table 2. Chemical composition of the catalysts and acidity of the sup-
ports.
[a]
Samples
Si/Al[a]
WO3
W/Ni[a]
Lewis acid
sites[b]
Brønsted acid
[wt%]
sites[b]
[mmolgÀ1
]
[mmolgÀ1
]
Figure 3. Emission region of a) W4f and b) Ni2p XPS spectra of various cata-
lysts sulfided at 673 K with their fit decomposition curves: a) c experi-
mental, c simulation, c WS2, c WOxSy, c W6+; b) c experimen-
NiW/HM
NiW/HM–M
11
16
14.1
15.1
2.5
2.6
23
87
104
201
tal, c simulation, c NixSy, c NiWS, c Ni2+
.
[a] Determined by ICP–OES; [b] adsorption–desorption of pyridine fol-
lowed by IR.
of the support led to an increase of the number of Lewis and
Brønsted sites, from 23 and 104 mmolgÀ1 on HM to 87 and
201 mmolgÀ1 on HM–M, respectively, despite the increase of
the Si/Al ratio induced by the treatment. The higher number
of acid sites in HM–M can be related to its larger specific sur-
face area as well as to better accessibility of pyridine owing to
mesopore creation.
zeolite, which might be attributed to the weak interaction be-
tween active phase and zeolite support compared to alumina
as stated in the literature,[14,39] similar observations were report-
ed by Duan et al. on CoMo catalysts.[25]
The XPS spectra for Ni2p are presented in Figure 3b and
their decomposition shows three contributions: sulfidic NixSy
(Ni2p3/2: 852.6 eV, Ni2p1/2: 869.8 eV), NiWS (Ni2p3/2: 853.7 eV,
In Figure 3a the X-ray photoelectron spectroscopy (XPS)
spectra of W4f for the sulfided catalysts are presented. The de-
composition of the photopeaks showed three contributions at-
tributed to W(6+) in an oxidic environment (W4f5/2: 36.2 and
Ni2p1/2: 870.9 eV), and an oxidic phase Ni(2+) (Ni2p3/2:
856.4 eV, Ni2p1/2: 873.8 eV).[41–44] The proportion of NiO, NixSy
and NiWS (promotion degree) in the catalysts were calculated
and are listed in Table 3. A proportion of 44% NiWS was ob-
tained in NiW/Al2O3, which was similar to the value of 37% re-
ported on a sulfided NiW/ASA catalyst.[44] Interestingly, the use
of HM and HM–M as supports led to better promotion of the
sulfide phase, with the proportion of NiWS of 82% in NiW/HM
and 73% in NiW/HM–M. If the sulfidation degree of molybde-
num or tungsten has been evaluated in the literature, no stud-
ies reported the promotion degree of the sulfide phase sup-
ported on zeolites determined by decomposition of the XPS
Ni2p photopeak. However Fan et al. also reported that NiWS
phase exists in majority on NiW supported on Al2O3–HY sup-
port.[39]
W4f3/2
: 38.3 eV), W(5+) as oxysulfide species (33.2 and
35.4 eV), and W(4+) sulfided species (32.3 and 34.4 eV).[41–44]
The results obtained from these decompositions are listed in
Table 3. The sulfidation degree of tungsten on NiW/Al2O3 was
found at 30%. This value can be compared to that of a NiW
catalyst with a sulfidation degree of tungsten of 49% if sup-
ported on alumina (with 27% WO3),[37] and of 42% if supported
on amorphous silica alumina (with 17% WO3).[42] Similar sulfida-
tion degrees were obtained for NiW/HM and NiW/HM–M, with
much higher values than on alumina, 65 and 69% respectively.
This result indicates that tungsten could be well sulfided over
Representative TEM images of all the sulfided catalysts are
shown in Figure 4. All samples exhibited the typical structure
of the layered WS2 phase, with an interplanar spacing of
0.65 nm, corresponding to the (002) planes of the WS2 crys-
tal.[45] The WS2 slabs appeared to be well dispersed over Al2O3
and HM–M (Figure 4a and 4c), the sulfided NiW/Al2O3 sample
showing mainly monoslabs, whereas more stacked slabs were
present on the sulfided NiW/HM–M. On the contrary, large ag-
glomerations of slabs were observed on NiW/HM (Figure 4b),
Table 3. XPS data for the catalysts sulfided at 673 K with H2/H2S (90:10).
Samples
WS2
[%]
WOxSy
[%]
W6+
[%]
NiWS
[%]
NixSy
[%]
Ni2+
[%]
NiW/Al2O3
NiW/HM
NiW/HM–M
30
65
69
24
14
14
46
21
17
44
82
73
27
18
23
29
<5
<5
ChemCatChem 2015, 7, 3936 – 3944
www.chemcatchem.org
3938
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
4.1 nm and N=1.8 for 27% WO3[37]). Comparatively, the length
of WS2 slabs in NiW/HM–M was distributed in a relatively wide
range, the majority of the slabs showing a length between
1 and 13 nm. The calculated average length of 5.6 nm and the
calculated average stacking degree of 4.8 on sulfided NiW/
HM–M appeared to be much larger than those obtained on
alumina, in agreement with literature data obtained on NiW
supported on Y zeolites.[14,39] Average length between 9 and
16 nm and average stacking between 2.4 and 3.7 were report-
ed by Ding et al., for NiW catalysts supported on zeolites
beta.[32] This can be explained by a weaker interaction devel-
oped on the zeolite support by the sulfide phase compared to
alumina, well known for its dispersion ability.[39]
Figure 4. High-resolution TEM images of NiW catalysts supported on
a) Al2O3, b) HM, and c) HM–M after sulfidation at 673 K.
probably in relation with the low external surface of the sup-
port. As observed on the images, the size of active phase is
larger than the micropore size in mordenite, indicating that
the WS2 slabs are mainly located outside the micropores, in ac-
cordance with results obtained from N2 sorption. These obser-
vations clearly show that the presence of mesopores on HM–M
allowed a better dispersion of the active phase, by offering
larger surface for WS2 deposition.
Catalytic properties of sulfide catalysts in HDS of DBT and
4,6-DMDBT
Total and HDS activity
By statistical analysis of the images, the distribution in slab
length and stacking of the WS2 slabs were obtained, as shown
in Figure 5. Owing the heavy agglomeration observed on sul-
fided NiW/HM, the statistical data is not available for this
sample. On the sulfided NiW/Al2O3, the WS2 slab length was
between 1 and 10 nm with an average length L of 4.8 nm and
an average stacking N of 1.2, values similar to those reported
in the literature (L=3.3 nm and N=1.3 for 19% WO3,[46] L=
To further test the deep desulfurization ability of the synthe-
sized catalysts (NiW/Al2O3, NiW/HM, and NiW/HM–M), their cat-
alytic performances were evaluated in desulfurization of DBT
and 4,6-DMDBT at 613 K under 4 MPa as total pressure. The re-
action routes of DBT and 4,6-DMDBT HDS are presented in
Scheme SI1 and Scheme SI2 in the Supporting Information, re-
spectively.
Both NiW sulfide phase supported on mordenite zeolites
(NiW/HM and HM–M) appeared to be less efficient than NiW/
Al2O3 for the HDS of DBT, with the HDS activity higher on alu-
mina (2.19 mmolhÀ1 gÀ1) than on zeolite-based catalysts
(0.37 mmolhÀ1 gÀ1 for NiW/HM and 0.98 mmolhÀ1 gÀ1 for NiW/
HM–M) as illustrated in Figure 6b. Nevertheless XPS analysis
evidenced a higher NiWS percentage on catalysts supported
on HM (82%) and HM–M (73%) than on alumina (44%), indi-
cating that the promotion degree is not linked to catalytic ac-
tivity. This suggests that a great amount of sites that are active
on the alumina supported catalyst are (i) less active or (ii) inac-
cessible if the mordenite zeolite is used as support. The use of
zeolite support could lead to less active sites than alumina as
a result of support effects on the active phase. Another possi-
bility could be that these promoted sites would not be accessi-
ble on zeolite support. This hypothesis is supported by the dif-
ferences in activity between NiW/HM and NiW/HM–M as
shown below.
The use of the modified mordenite zeolite as catalytic sup-
port of the sulfide phase (NiW/HM–M) instead of the parent
zeolite (NiW/HM) allows an increase of the HDS activity by
a factor of 2.6 (Figure 6b). This result can be attributed to the
presence of mesopores allowing a better dispersion of the
active phase as evidenced by TEM as well as a better diffusion
of reagents, and hence a higher accessibility of sulfur mole-
cules to active sites.
On NiW/Al2O3, the total activity using DBT and 4,6-DMDBT
was largely dominated by HDS activity. As expected, 4,6-
DMDBT was found 4.4 times less reactive than DBT over NiW/
Al2O3 (Figure 6a) and 5 times less reactive in desulfurization
(Figure 6b). This difference in reactivity between both model
Figure 5. Distribution of the length (top) and stacking (bottom) of the WS2
slabs over Al2O3 and HM–M supports after sulfidation at 673 K.
ChemCatChem 2015, 7, 3936 – 3944
www.chemcatchem.org
3939
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
and/or on the intrinsic nature induced by the nature of the
support. Zeolites-based catalysts could thus contain mainly
sites active in both DBT and 4,6-DMDBT.
4,6-DMDBT was 2.3 times more reactive over NiW/HM–M
than over the NiW/HM catalyst, this ratio is the same order of
magnitude as that mentioned above in the HDS of DBT. As
previously proposed, this improvement can be related at least
in part to the presence of mesopores in HM–M support.
Selectivity results
Over sulfide catalysts supported on alumina, it is generally ac-
cepted that DBT and 4,6-DMDBT convert by two routes: the
direct desulfurization pathway (DDS) yielding biphenyls and
the hydrogenation pathway (HYD) yielding cyclohexylben-
zenes.[8] Over NiW/Al2O3 DBT is mainly desulfurized by the DDS
route (DDS/HYD selectivity close to 2) as shown in Table 4. The
use of sulfide catalysts containing acidic supports (HM and
HM–M) in DBT HDS led to the presence of other products com-
pared to those obtained over NiW/Al2O3. In addition to prod-
ucts already observed from DBT on the conventional catalyst
(tetrahydrodibenzothiophene, THDBT; and biphenyl, BPh),
three methyldibenzothiophene isomers (MDBT), four dimethyl-
dibenzothiophene isomers (DMDBT), two methylbiphenyl iso-
mers (MBPh), three dimethylbiphenyl isomers (DMBPh) and
benzene (Bz) were observed in the reactional mixture, as indi-
cated in Table 4. Under these experimental conditions, we veri-
fied that cyclohexylbenzene (the main HYD product obtained
over NiW/Al2O3) was totally converted into cyclohexane and
benzene likely by hydrocracking owing to the presence of
strong acid sites on zeolites. This observation is in good agree-
ment with results reported by several authors.[5,53] Moreover, it
was verified that biphenyl was unreactive in the presence of
NiW/HM. In addition, MDBT and DMDBT isomers and their de-
sulfurized products were also observed in very low amounts
over NiW/HM and NiW/HM–M (Table 4). The formation of these
compounds can be explained by alkylation reaction (methyla-
tion) because of the presence of DMDS in the model feed.
Such reaction was already observed for the transformation of
DBT by Bataille et al. using a CoMo/HY catalyst.[7]
Figure 6. Transformation of DBT and 4,6-DMDBT at 613 K under 4 MPa of
total pressure over NiW/Al2O3, NiW/HM, and NiW/HM–M. a) Total activity;
b) desulfurization (HDS) activity.
molecules is of the same order of magnitude than the values
measured over a NiMo/Al2O3 commercial catalyst using the
same experimental conditions.[8] It is well known that the pres-
ence of methyl groups in position of 4 and 6 leads to a signifi-
cant inhibition of the desulfurization rate of dibenzothiophenic
compounds owing to steric hindrance.[8,47–51]
Interestingly, the 4,6-DMDBT HDS activity of NiW/Al2O3 is rel-
atively close to the value measured over NiW/HM
(0.44 mmolhÀ1 gÀ1 for the former and 0.37 mmolhÀ1 gÀ1 for the
latter), indicating that the activity of the sulfide phase in the
HDS of 4,6-DMDBT does not follow the trend observed for the
HDS of DBT. This could correspond to the presence of two
types of sites, active either only in DBT HDS or in both DBT
and 4,6-DMDBT HDS, hypothesis already suggested in the liter-
ature,[52] depending probably on their localization on the slabs
In Figure 7a the activity of the different routes of DBT trans-
formation described above over the three catalysts is shown.
In the presence of NiW/HM catalyst, both main desulfurization
routes (HYD and DDS) were affected compared to the NiW/
Al2O3 catalyst. Moreover the comparison of the DDS/HYD ratios
Table 4. Conversion and distribution of products obtained from the transformation of dibenzothiophene (DBT) over sulfided catalysts at 613 K under
4 MPa as total pressure.
Catalyst
LHSV
Conv.
[%]
Distribution of products [mol%]
[hÀ1
]
HYD route
CHB
DDS route
BPh
DDS(ALK) route
ALK route
DDS/HYD
selectivity
Bz
THDBT
iMBPh+iDMBPh
iMDBT+iDMDBT
NiW/Al2O3
NiW/HM
NiW/HM–M
79
26
65
66
46
48
0
34
32
31
0
0
2
5
6
67
48
47
0
8
10
0
5
5
1.98
1.25
1.27
ChemCatChem 2015, 7, 3936 – 3944
www.chemcatchem.org
3940
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
The use of the hierarchical zeolite as support led to the
same DDS/HYD ratio than on NiW/HM, indicating that the crea-
tion of mesoporosity has no influence on the selectivity and in
the nature of the active sites. Similar promotion degrees were
obtained by XPS over both NiW/HM and NiW/HM–M (Table 3).
Concerning the reactivity of 4,6-DMDBT over NiW/Al2O3, the
conversion was mainly through the HYD route leading to 3-(3’-
methylcyclohexyl)toluene (MCHT) and three dimethyltetrahy-
drodibenzothiophene isomers (iDMTHDBT), whereas 3,3’-dime-
thylbiphenyl was the only product formed by the DDS route
(Table 5). Contrary to the reactivity of DBT, the main reaction
route of 4,6-DMDBT was the HYD way (DDS/HYD selectivity
near 0.3). These results are in agreement with those reported
in the literature over Mo based catalysts,[8] confirming that the
presence of methyl groups in position 4 and 6 strongly inhibits
the DDS way whereas the HYD route is less affected.
In the presence of both NiW/HM and Ni/HM-M catalysts,
other sulfur and desulfurized products were detected (Table 5).
Two isomers of 4,6-dimethyldibenzothiophene (iDMDBT) were
observed in large quantities, 3,6-dimethyldibenzothiophene
being the main isomer formed. By the DDS route, 3,6-dimethyl-
dibenzothiophene yielded 3,4’-dimethylbiphenyl, the main
product of the DDS(ISOM) pathway. As in the case of DBT, the
main hydrogenation product of 4,6-DMDBT (MCHT) was totally
converted into toluene and methylcyclohexane. Similarly to re-
sults observed with DBT, alkylation products were also ob-
served in trace amounts.
Figure 7. Activity of the different desulfurization pathways for the transfor-
mation of a) DBT and b) DMDBT at 613 K under 4 MPa of total pressure over
NiW/Al2O3, NiW/HM and NiW/HM–M.
On the zeolite supported catalysts, the DDS/HYD ratios ob-
tained from 4,6-DMDBT transformation were lower than the
ratio measured over NiW/Al2O3 (Table 5), in accordance with
the results observed from DBT reaction already discussed.
Owing to the acidic properties of the zeolitic supports, 4,6-
DMDBT was mainly isomerized into 3,6-DMDBT (ISOM way),
which is known to be more reactive than 4,6-DMDBT.[5,7] The
use of a mesoporous support (NiW/HM–M) allowed the favor-
ing of the isomerization of 4,6-DMDBT, with the ISOM activity
doubled, and hence the production of dimethylbiphenyl iso-
mers (Figure 7b). This result explains why the desulfurization is
the most favored over NiW/HM–M, as shown in Figure 6b. The
higher reactivity of 4,6-DMDBT over the hierarchical zeolite
supports catalyst due to this new transformation route (ISOM)
can be related to the acidity properties of the supports:
indeed HM-M possessed a larger number of Lewis and Brønst-
ed acid sites (87 and 201mmolgÀ1 respectively) than HM (23
and 104 mmolgÀ1), as shown in Table 1. In addition, the rate of
on NiW/Al2O3 (1.98) and on NiW/HM (1.25) indicates that the
DDS route was more affected than the HYD route if zeolite
was used as support. As evidenced by TEM analysis (Figure 4),
the stacking of the WS2 phase appears to be much higher on
zeolite support than on alumina. In accordance with the rim–
edge model,[54] the DDS activity should be less reduced than
the HYD activity. As the opposite was observed, the selectivity
difference cannot be explained by the modification of the mor-
phology of the sulfide phase. The differences in DDS/HYD se-
lectivity could thus be related to the nature of the active sites
depending on the support used. Moreover XPS analysis indi-
cated that the promotion degree was much higher on zeolite-
supported catalyst (82%) compared to alumina catalyst (44%)
as illustrated in Table 3, which could support the fact that zeo-
lite used as support induces modifications in the active sites.
Table 5. Conversion and distribution of products obtained from the transformation of 4,6-dimethyldibenzothiophene (4,6-DMDBT) over sulfided catalysts
at 613 K under 4 MPa as total pressure.
Catalyst
LHSV Conv.
Distribution of products [mol%]
DDS(ISOM) route Other HC products ISOM route ALK route DDS/HYD
Tol MCHT iDMTHDBT 3,3’-DMBPh 3,4’-DMBPh+4,4’DMBPh iMBPh
[hÀ1
]
[%]
HYD route
DDS route
selectivity
iDMDBT
iTMDBT
NiW/Al2O3
NiW/HM
NiW/HM–M 65
26
44
54
44
58
0
36
32
63
0
0
9
7
6
23
8
9
0
10
12
0
3
7
5
33
33
0
3
1
0.32
0.18
0.24
ChemCatChem 2015, 7, 3936 – 3944
www.chemcatchem.org
3941
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
4,6-DMDBT isomerization is most likely higher than its direct
desulfurization because isomerized products were observed in
large quantities compared to the 3,3’-dimethylbiphenyl formed
by the DDS way (Table 5). Moreover, 3,6-DMDBT, the main
isomer of 4,6-DMDBT, can be converted into 3,4’-dimethylbi-
phenyl by the DDS(ISOM) route and into toluene and methyl-
cyclohexane by HYD(HCK). The highest value of the HYD activi-
ty measured over NiW/HM–M (0.54 mmolhÀ1 gÀ1) compared to
the one determined over NiW/Al2O3 (0.37 mmolhÀ1 gÀ1) could
be explained through the participation of the HYD route of
the isomerized products (3,6-DMDBT mainly), which led to the
same products (toluene and methylcyclohexane) as those
formed from the HYD route of 4,6-DMDBT, as indicated in
Scheme SI2.
Preparation of catalyst
Hierarchical mordenite was prepared by an acid–base–acid post-
treatment method following the protocol described in the litera-
ture.[40] The commercial and hierarchical mordenites were labeled
as HM and HM–M, respectively.
The NiW-supported catalysts were prepared by co-impregnation
using the incipient wetness technique. Typically, the impregnation
solution was obtained by dissolution of nickel nitrate hexahydrate
and ammonium metatungstate [(NH4)6H2W12O40] in water with a Ni/
W molar ratio of 0.36, which has been described as an appropriate
molar ratio.[41] The calculated content of WO3 was set at 15 wt% in
the calcined catalyst. After impregnation and maturation, the
solids were dried at 373 K overnight and calcined at 773 K for 4 h
in air, using a temperature rate of approximately 2 KminÀ1. The
chemical compositions of the catalysts, determined by inductively
coupled plasma optical emission spectroscopy are listed in Table 2.
The commercial and hierarchical mordenite supported NiW cata-
lysts were labeled as NiW/HM and NiW/HM–M, respectively. For
Conclusions
comparison, commercial alumina (BET surface area: 220 m2 gÀ1
,
pore volume obtained by water adsorption: 0.6 cm3 gÀ1) supported
NiW catalyst was prepared by the same method and named NiW/
Al2O3.
NiW catalysts supported on Al2O3, commercial mordenite and
hierarchical mordenite obtained by acid–base–acid treatment
were prepared by co-impregnation technique. NiW/Al2O3 cata-
lyst exhibited better HDS activity than NiW/HM and NiW/HM–
M in the transformation of DBT even if higher promotion
degree was determined by XPS on the zeolite supported cata-
lysts. This suggests that sites active on the alumina supported
catalyst are less active or inaccessible if mordenite, modified or
not, is used as support.
Catalyst characterization
X-ray powder diffraction (XRD) patterns were recorded by using
a Siemens d-5000 diffractometer equipped with the CuKa radiation
(wavelength of l=1.5418 Š) and at BM01B–SNBL beam line, ESRF,
France (k=0.5 A, Si(111) channel cut monochromatic) using the
two cycle diffractometer equipped by six detectors having Si(111)
analyzer crystals and Na–I scintillation counters.
Comparatively, NiW/HM and NiW/HM–M displayed superior
catalytic performance over NiW/Al2O3 in the transformation of
4,6-DMDBT as a result of the introduction of an additional reac-
tion route (isomerization) resulting from the acidity of the zeo-
litic supports. Furthermore, NiW/HM–M exhibited higher cata-
lytic activity than NiW/HM in the transformation of both DBT
and 4,6-DMDBT. This can be explained by (i) better dispersion
of the sulfide NiW phase observed by TEM when deposited on
the hierarchical zeolite; (ii) enhanced accessibility of the sulfur
model compounds to the active sites owing to the creation of
mesoporosity by acid–base–acid treatment; (iii) in the case of
4,6-DMDBT, the higher acidity of NiW/HM–M than that of NiW/
HM leading to an improvement of its reactivity in isomeriza-
tion.
Nitrogen sorption isotherms were obtained at 77 K on a Micromet-
rics TriStar II 3020 Gas Sorption and Porosimetry system. Prior to
the experiments, the samples were outgassed at 423 K under
vacuum for 3 h.
The IR studies of pyridine adsorption were performed to measure
the number of acid sites, recorded with a Nicolet Protege System
460 equipped with a DTGS detector. All samples were ground in
an agate mortar and were pressed into the form of self-supporting
wafers (5 mgcmÀ2), then heated at 723 K under high vacuum
(10À6 mbar) for 2 h before probe molecule adsorption at RT and de-
sorption at 423 K. All recorded spectra were recalculated to a nor-
malized wafer of 10 mg.
This study clearly demonstrates that the use of hierarchical
mordenite as support of NiW based HDS catalyst is beneficial
in the desulfurization of refractory sulfur molecules, combining
the acidic properties of the zeolite and the mesoporosity al-
lowing better accessibility to the active sites.
XPS analysis were performed on an Axis ultra DLD (Kratos analyti-
cal) using a monochromatic AlKa X-ray source (hn=1486.6 eV). The
emission voltage and the current of this source were set to 15 kV
and 10 mA, respectively. The pressure in the analyzing chamber
was maintained at 10À9 Pa or lower during analysis, and the size of
the analyzed area was 300”700 mm2. After their sulfidation under
H2/H2S (90/10) for 2 h at 673 K, samples were transferred in the
spectrometer chamber using a glove box to avoid any reoxidation.
Surveys (0–1300 eV) were recorded at pass energies of 160 eV with
a step of 1 eV and high-resolution (W4f, Ni2p, S2p, and C1s zone)
spectra were recorded at pass energies of 40 eV with a step of
0.1 eV. Data treatment and peak-fitting procedures were performed
using Casa XPS software. Obtained spectra were calibrated in re-
spect of C1s (CÀC bond) at 285 eV. The peaks were decomposed
using Gaussian–Lorentzian peak shapes.
Experimental Section
Materials
Commercial mordenite zeolite was purchased from the Catalyst
Plant of Nankai University. HNO3, NaOH, NH4NO3 were purchased
from Sinopharm Chemical Reagent Co. and used without further
purification. n-heptane (solvent) was purchased from Carlo–Erba,
n-decane from Sigma–Aldrich, dimethyl disulfide (DMDS) from
Fluka, DBT from Lancaster, and 4,6-DMDBT from Eburon.
High-resolution TEM was performed on a TECNAI microscope oper-
ating at 200 kV with a LaB6 crystal. Freshly sulfided samples were
ChemCatChem 2015, 7, 3936 – 3944
www.chemcatchem.org
3942
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
[1] C. Song, X. Ma, Appl. Catal. B 2003, 41, 207–238.
[2] A. Duan, G. Wan, Z. Zhao, C. Xu, Y. Zheng, Y. Zhang, T. Dou, X. Bao, K.
Chung, Catal. Today 2007, 119, 13–18.
[3] D. Zhang, A. Duan, Z. Zhao, X. Wang, G. Jiang, J. Liu, C. Wang, M. Jin,
Catal. Today 2011, 175, 477–484.
[4] V. Meille, E. Schulz, M. Lemaire, M. Vrinat, J. Catal. 1997, 170, 29–36.
[5] P. Michaud, J. L. Lemberton, G. Perot, Appl. Catal. A 1998, 169, 343–353.
[6] D. D. Whitehurst, T. Isoda, I. Mochida, Adv. Catal. 1998, 42, 345–471.
[7] F. Bataille, J. L. Lemberton, G. Perot, P. Leyrit, T. Cseri, N. Marchal, S. Kasz-
telan, Appl. Catal. A 2001, 220, 191–205.
ground under an inert atmosphere and dispersed in ethanol. The
suspension was collected on carbon films supported on copper
grids. Statistical analysis was performed on 600 WS2 slabs to obtain
a reliable determination of the distribution of size and stacking
degree of the slabs for each sample.
Catalytic tests
The HDS of DBT and 4,6-DMDBT were performed in a high-pressure
fixed-bed microreactor (length: 40 cm, inner diameter: 1.25 cm) at
613 K under 4 MPa of total pressure. Before catalytic test, the
solids were sulfided by a mixture containing DMDS (5.8 wt%) dilut-
ed in n-heptane introduced at 8 mLhÀ1 as volume flow in the pres-
ence of hydrogen (volume flow equal to 4.7 NLhÀ1). This sulfiding
solution was injected at 423 K. After 1 h, the temperature was in-
creased to 623 K by 5 KminÀ1 and was maintained for 14 h to
ensure a good sulfidation of the catalyst. Then, the temperature
was decreased to reaction temperature (613 K). The composition of
the model feedstock was designed to simulate a straight run gas
oil containing 10000 ppm of sulfur. DBT or 4,6-DMDBT (500 ppm S)
was dissolved in n-heptane to which DMDS (1.4 wt%) was added
to generate H2S. Decane was used as a reference for quantitative
analysis using a GC apparatus. The liquid effluents were collected
at regular intervals after stabilization of the catalyst (observed after
1 hour on stream irrespective of the type of solid used) and ana-
lyzed by gas chromatography (Varian 430) equipped with a FID de-
tector and a DB1 capillary column (length: 30 m; inside diameter:
0.25 mm; film thickness: 0.25 mm). To obtain similar conversion
(close to 50%), experiments were performed at different liquid
hourly space velocity (LHSV) values depending on the reactivity of
the sulfur reactant and on the activity of the sulfide catalyst.
[8] F. Bataille, J. L. Lemberton, P. Michaud, G. Perot, M. Vrinat, M. Lemaire, E.
Schulz, M. Breysse, S. Kasztelan, J. Catal. 2000, 191, 409–422.
[9] D. Solís, A. L. Agudo, J. Ramirez, T. Klimova, Catal. Today 2006, 116,
469–477.
[10] D. Solís, T. Klimova, R. Cuevas, J. Ramirez, A. Lopez-Agudo, Catal. Today
2004, 98, 201–206.
[11] C. Marín, J. Escobar, E. Galvan, F. Murrieta, R. Zarate, H. Vaca, Fuel Pro-
cess. Technol. 2004, 86, 391–405.
[12] F. Richard, T. Boita, G. Perot, Appl. Catal. A 2007, 320, 69–79.
[13] C. E. HØdoire, C. Louis, A. Davidson, M. Breysse, F. Mauge, M. Vrinat, J.
Catal. 2003, 220, 433–441.
[14] L. Ding, Y. Zheng, Z. Zhang, Z. Ring, J. Chen, J. Catal. 2006, 241, 435–
445.
[15] Y. Wang, B. Wang, A. Rives, Y. Sun, Energy Environ. Focus 2014, 3, 45–52.
[16] Q. Huo, T. Dou, Z. Zhao, H. Pan, Appl. Catal. A 2010, 381, 101–108.
[17] H. Yang, Z. Liu, H. Gao, Z. Xie, Appl. Catal. A 2010, 379, 166–171.
[18] M. Guidotti, C. Canaff, J.-M. Coustard, P. Magnoux, M. Guisnet, J. Catal.
2005, 230, 375–383.
[19] W. Kim, J.-C. Kim, J. Kim, Y. Seo, R. Ryoo, ACS Catal. 2013, 3, 192–195.
[20] J.-C. Kim, K. Cho, R. Ryoo, Appl. Catal. A 2014, 470, 420–426.
[21] Y. Sun, H. Wang, R. Prins, Catal. Today 2010, 150, 213–217.
[22] T. Tang, C. Yin, L. Wang, Y. Ji, F.-S. Xiao, J. Catal. 2008, 257, 125–133.
[23] W. Fu, L. Zhang, T. Tang, Q. Ke, S. Wang, J. Hu, G. Fang, J. Li, F.-S. Xiao, J.
Am. Chem. Soc. 2011, 133, 15346–15349.
[24] Q. Huo, Y. Gong, T. Dou, Z. Zhao, H. Pan, F. Deng, Energy Fuels 2010, 24,
3764–3771.
[25] A. Duan, T. Li, Z. Zhao, B. Liu, X. Zhou, G. Jiang, J. Liu, Y. Wei, H. Pan,
Appl. Catal. B 2015, 165, 763–773.
Assuming a first-order reaction, the total activity of sulfided cata-
lyst was determined according the following equation:
[26] K. T. Højholt, P. N. R. Vennestrom, R. Tiruvalam, P. Beato, Chem. Commun.
2011, 47, 12864–12866.
A ¼ ÀFS=Wcat lnð1ÀXsÞ
[27] D. Zhang, A. Duan, Z. Zhao, C. Xu, J. Catal. 2010, 274, 273–286.
[28] T. Tang, L. Zhang, W. Fu, Y. Ma, J. Xu, J. Jiang, G. Fang, F.-S. Xiao, J. Am.
Chem. Soc. 2013, 135, 11437–11440.
[29] N. Azizi, S. A. Ali, K. Alhooshani, T. Kim, Y. Lee, J.-I. Park, J. Miyawaki,
S.-H. Yoon, I. Mochida, Fuel Process. Technol. 2013, 109, 172–178.
[30] G. C. Laredo, E. Altamirano, J. A. De Los Reyes, Appl. Catal. A 2003, 242,
311–320.
where A is the total activity (expressed in molhÀ1 gÀ1), FS the molar
flow of the sulfur reactant (DBT or 4,6-DMDBT), Wcat the weight of
the sulfided catalyst, and XS the conversion of the sulfur reactant
(DBT or 4,6-DMDBT). The HDS activity was determined using XHDS
instead of Xs, XHDS being the conversion into desulfurized products.
[31] G. C. Laredo, E. Altamirano, J. A. De Los Reyes, Appl. Catal. A 2003, 243,
207–214.
[32] L. Ding, Y. Zheng, H. Yang, R. Parviz, Appl. Catal. A 2009, 353, 17–23.
[33] S. Bendezffl, R. Cid, J. L. G. Fierro, A. Lopez Agudo, Appl. Catal. A 2000,
197, 47–60.
[34] R. Tian, B. Shen, F. Wang, C. Lu, C. Xu, Energy Fuels 2009, 23, 55–59.
[35] M. Ali, T. Tatsumi, T. Masuda, Appl. Catal. A 2002, 233, 77–90.
[36] A. Duan, Z. Guo, Q. Huo, C. Wang, D. Zhang, M. Jin, G. Jiang, Z. Zhao, H.
Pan, K. Chung, Energy Fuels 2010, 24, 796–803.
[37] G. Wan, A. Duan, Y. Zhang, Z. Zhao, G. Jiang, D. Zhang, Z. Gao, J. Liu,
K. H. Chung, Energy Fuels 2009, 23, 3846–3852.
[38] G. M. Kumaran, S. Garg, M, Kumar, N. Viswanatham, J. K. Gupta, L. D.
Sharma, G. M. Dhar, Energy Fuels 2006, 20, 2308–2313.
[39] Y. Fan, H. Xiao, G. Shi, H. Liu, Y. Qian, T. Wang, G. Gong, X. Bao, J. Catal.
2011, 279, 27–35.
[40] K. Leng, Y. Wang, C. Hou, C. Lancelot, C. Lamonier, A. Rives, Y. Sun, J.
Catal. 2013, 306, 100–108.
[41] K. Ben Tayeb, C. Lamonier, C. Lancelot, M. Fournier, E. Payen, A. Bon-
duelle, F. Bertoncini, Catal. Today 2010, 150, 207–212.
[42] K. Ben Tayeb, C. Lamonier, C. Lancelot, M. Fournier, A. Bonduelle, F. Ber-
toncini, Appl. Catal. B 2012, 126, 55–63.
[43] J. N. Díaz de León, M. Picquart, L. Massin, M. Vrinat, J. A. de Los Reyes, J.
Mol. Catal. A 2012, 363–364, 311 –321.
Acknowledgements
Martine Trentesaux and Jean-Charles Morin from UCCS are ac-
knowledged for the catalysts characterization (XPS and pyridine
adsorption followed by IR, respectively). The authors acknowledge
for financial support the Fundamental Research Funds for the
Central Universities (Grant No. HIT. NSRIF. 2015046), the Scientific
Research Starting Funding from Harbin Institute of Technology,
the Scientific Research Foundation for the Returned Overseas Chi-
nese Scholars, State Education Ministry, and the Program
Caiyuanpei (project 30323NF) between China (CSC)-France
(Campus-France). The TEM facility in Lille (France) is supported
by the Conseil Regional du Nord-Pas de Calais, and the European
Regional Development Fund (ERDF).
Keywords: desulfurization · mesoporous materials · nickel ·
tungsten · zeolites
ChemCatChem 2015, 7, 3936 – 3944
www.chemcatchem.org
3943
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
[44] H. R. Reinhoudt, E. Crezee, A. D. van Langeveld, P. J. Kooyman, J. A. R.
van Veen, J. A. Mouljin, J. Catal. 2000, 196, 315–329.
[45] B. S. Zhang, Y. J. Yi, W. Zhang, C. H. Liang, D. S. SU, Mater. Charact. 2011,
62, 684–690.
[46] E. J. M. Hensen, Y. van der Meer, J. A. R. van Veen, J. W. Niemantsverdriet,
Appl. Catal. A 2007, 322, 16–32.
[51] Y. Saih, S. Kohichi, Catal. Surv. Asia 2003, 7, 235–249.
[52] M. Philippe, F. Richard, D. Hudebine, S. Brunet, Appl. Catal. B 2013, 132–
133, 493–498.
[53] M. V. Landau, D. Berger, M. Herskowitz, J. Catal. 1996, 159, 236–245.
[54] M. Daage, R. R. Chianelli, J. Catal. 1994, 149, 414–427.
[47] M. Vrinat, D. Laurenti, C. Geantet, Appl. Catal. B 2012, 128, 3–9.
[48] M. Egorova, R. Prins, J. Catal. 2004, 225, 417–427.
[49] M. Egorova, R. Prins, J. Catal. 2004, 224, 278–287.
[50] P. Da Costa, C. Potvin, J.-M. Manoli, J.-L. Lemberton, G. PØrot, G. DjØga-
Mariadassou, J. Mol. Catal. A 2002, 184, 323–333.
Received: June 15, 2015
Revised: July 15, 2015
Published online on September 17, 2015
ChemCatChem 2015, 7, 3936 – 3944
www.chemcatchem.org
3944
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim