NiMo catalysts supported on graphene-modified mesoporous TiO2 toward highly efficient hydrodesulfurization of dibenzothiophene

Article abstract of DOI:10.1016/j.apcata.2015.05.028

The graphene-covered mesoporous TiO₂ (S_BET > 90 m² g^-1) were successfully prepared by using graphene oxide (GO) as graphene source via one-step hydrothermal reduction, and then were prepared supported NiMo catalysts for hydrodesulfurization (HDS) of dibenzothiophene (DBT). The active species were incorporated by incipient impregnation. Compared with the unmodified NiMo/TiO₂ catalyst, all modified catalysts exhibited higher HDS activity under mild conditions of 280 °C, 2.0 MPa, liquid hourly space velocity of 4 h^-1 and hydrogen/feed ratio of 400 (V/V). Among of them, the catalyst with the appropriate mass ratio of rGO/TiO₂(0.5) showed the highest DBT conversion of 99.9%. The prepared catalysts were characterized by various techniques (N₂ adsorption-desorption, XRD, Raman, Pyridine FT-IR, NH₃-TPD, H₂-TPR, CO uptake, SEM and TEM). The enhanced HDS activity can be ascribed to (i) The incorporation of rGO (reduced graphene oxide) induces proper enhanced Lewis acidity of the modified catalysts, which promotes the adsorption of DBT molecules and facilitates the hydrogenation route. (ii) The introduction of rGO changes the HDS product (H₂S) desorption behaviors, accelerating H₂S desorption rate from catalyst surface and promoting the efficiency of catalyst immediately.

Full text of DOI:10.1016/j.apcata.2015.05.028

Accepted Manuscript
Title: NiMo catalysts supported on graphene-modified
mesoporous TiO toward highly efficient hydrodesulfurization
2
of dibenzothiophene
Author: Huaiyuan Wang Bo Xiao Xiaoshuang Cheng Chijia
Wang Li Zhao Yanji Zhu Jiahua Zhu Xiaohua Lu
PII:
S0926-860X(15)30006-5
DOI:
Reference:
http://dx.doi.org/doi:10.1016/j.apcata.2015.05.028
APCATA 15402
To appear in:
Applied Catalysis A: General
Received date:
Revised date:
Accepted date:
13-2-2015
7-5-2015
27-5-2015
Please cite this article as: Huaiyuan Wang, Bo Xiao, Xiaoshuang Cheng,
Chijia Wang, Li Zhao, Yanji Zhu, Jiahua Zhu, Xiaohua Lu, NiMo catalysts
supported on graphene-modified mesoporous TiO2 toward highly efficient
hydrodesulfurization of dibenzothiophene, Applied Catalysis A, General
http://dx.doi.org/10.1016/j.apcata.2015.05.028
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.

NiMo catalysts supported on graphene-modified
mesoporous TiO towards highly efficient
2
hydrodesulfurization of dibenzothiophene
a,*
a
a
a
a
a
Huaiyuan Wang , Bo Xiao , Xiaoshuang Cheng , Chijia Wang , Li Zhao , Yanji Zhu , Jiahua
b
c
Zhu , Xiaohua Lu
College of Chemistry and Chemical Engineering, Northeast Petroleum University, Daqing
63318, PR China. Fax: 0086-4596503498; Tel: 0086-4596503083; E-mail:
a
1
wanghyjiji@163.com.
b
Department of Chemical and Biomolecular Engineering, The University of Akron,
Akron, OH 44221 USA.
c
State Key Lab of Material-Oriented Chemical Engineering, Nanjing University of
Highlights
►
Technology, Nanjing 210009, China.
Highlights
ꢀ
ꢀ
ꢀ
The introduction of rGO accelerates HDS product (H S) desorption rate from
2
catalyst surface and promotes the efficiency of catalyst.
The incorporation of rGO induces proper enhanced Lewis acidity of the
modified catalysts.
NiMo/TGC-0.5 showed great HDS performance under a mild condition of 280
o
C, 2.0MPa.
Graphical abstract
Abstract

2
-1
2
The graphene-covered mesoporous TiO (SBET>90 m g ) were successfully prepared by using
graphene oxide (GO) as graphene source via one-step hydrothermal reduction, and then were
prepared supported NiMo catalysts for hydrodesulfurization (HDS) of dibenzothiophene (DBT).
The active species were incorporated by incipient impregnation. Compared with the unmodified
NiMo/TiO
2
catalyst, all modified catalysts exhibited higher HDS activity under mild conditions
o
−1
of 280 C, 2.0 MPa, liquid hourly space velocity of 4 h and hydrogen/feed ratio of 400 (V/V).
Among of them, the catalyst with the appropriate mass ratio of rGO/TiO
highest DBT conversion of 99.9%. The prepared catalysts were characterized by various
techniques (N adsorption-desorption, XRD, Raman, Pyridine FT-IR, NH -TPD, H -TPR, CO
2
(0.5) showed the
2
3
2
uptake, SEM and TEM). The enhanced HDS activity can be ascribed to (i) The incorporation of
rGO (reduced graphene oxide) induces proper enhanced Lewis acidity of the modified catalysts,
which promotes the adsorption of DBT molecules and facilitates the hydrogenation route. (ii)
2
The introduction of rGO changes the HDS product (H S) desorption behaviours, accelerating
H
2
S desorption rate from catalyst surface and promoting the efficiency of catalyst immediately.
1
. Introduction
Environmental concerns are forcing dramatic challenge in transportation fuel specifications
as sulfur-containing compounds in motor fuel are closely related to the emission of SO , which is
x
the main reason of air pollution. To meet the increasing stringent environmental regulations,
deep/ultra-deep desulfurization are in urgent demanding [1-4]. Hydrodesulfurization (HDS) is
one of the most widely used techniques for removing sulfur-containing compounds from
petroleum products in the refining industry, where HDS catalyst plays a vital role in the
determining process. Therefore, developing highly efficient catalyst exerts particularly important
[
5-7].
The support of catalyst is one of the key factors that impact its catalytic performance [8-10].
Recently, TiO has been recognized as a new generation of active support because of the
2
moderate acid properties and the relatively weak interaction with the active metals, which is
capable of improving the dispersion and reducibility of the supported active species [11,12].
2
However, the low surface area and poor thermal stability of traditional TiO limit its application.
To overcome the above drawbacks, a new type of mesoporous TiO
2
(anatase) whisker with high
o
-1
thermal stability (convert to rutile at 900 C), good mechanical property (>16.4 N mm ) and
2
-1
large surface area (>90 m g ) was developed in our previous work [13]. The HDS performance
of corresponding catalyst NiMo/TiO is investigated, but the result is not satisfactory. Generally,
2
the catalytic performance is mainly determined by two factors: active sites and mass transfer of
reaction on catalyst surface [14,15]. However, most current studies are focused on enhancing the
activity of reactive sites and catalytic reaction mechanisms. It should be pointed out that mass
transfer is equally important for the catalyst efficiency. According to the our previous molecular
2
dynamics simulation [16], the key HDS resultant H S (polar molecule) was prone to be absorbed

on catalyst surface, which decreased the transfer rate of molecules off interfacial layer drastically
and influenced the efficiency of NiMo/TiO directly.
2
In recent years, many studies have reported the carbon-modified supports to prepare
supported active metals catalysts using different materials as carbon sources, including phenyl-
phosphonic acid (PPA), citric acid (CA) and monocarboxylic acids [17-19]. It shows that carbon
modification on the surface of supports is also an alternative and effective way besides the
conventional methods [20]. To the best of our knowledge, there has been no research reporting
2
on graphene-doped TiO by surface modification in HDS reaction. Graphene, as one of the
advanced carbon nano-materials, is a 2-dimensional single sheet of carbon atoms arranged in a
hexagonal network. It has attracted a great deal of scientific interest and possess potentially
applications in various fields, such as composite materials, batteries, solar cells and catalysts, due
to its distinct properties [21-23]. Inspired from the above studies, we intend to use graphene to
2
modify the surface of TiO support. But the solubility of graphene is poor, which is unfavorable
for our work. Unlike graphene, graphene oxide (GO) is soluble due to the hydrophilic functional
groups on the sheet surface, which render GO a good substitute for synthesizing new composite
materials in liquid phase [24].
Herein, we first fabricated a reduced graphene oxide (rGO) surface modification on TiO
support via one-step hydrothermal reduction method. Through adjusting TiO surface adsorption
towards H S molecule, an optimization was realized between diffusion and HDS process.
Besides, the effects of rGO on the acidity of catalysts were investigated. Compared with the
unmodified catalyst of NiMo/TiO , the conversion of dibenzothiophene (DBT) increases from
8.6% to 99.9% under mild conditions.
. Experimental
.1. Materials
Graphite powder (180 mesh) was purchased from Qing-dao Black Dragon Reagent Co. Ltd.
Qing dao, China). (NH Mo O, Ni(NO •6H O and C OH (A.R.) were supplied by
Tianjin Yaohua Chemical Reagent Co. Ltd. (Tian jin, China). Decalin and CS (A.R. standard)
were purchased from Sinopharm Chemical Reagent Co. Ltd. Dibenzothiophene (C10 S, 99.9%)
2
2
2
2
7
2
2
(
4
)
6
7
O24•4H
2
3
)
2
2
2 5
H
2
H
8
was purchased from Acros Organics. Rapid hydrogen sulfide gas detector tube (200-5000 ppm)
was purchased from Beijing Environmental Applied Technology Research Centre. All the
reagents were used without further purification.
2
.2. Preparation of GO, supports and catalysts
GO was synthesized by the modified Hummers’ method [25]. TiO
2
was synthesized by ion-
was firstly hydrolyzed to be
) at the presence of hydrochloride (0.1 M). Then, H Ti was calcined in
-rGO composite (TGC) was
synthesized via hydrothermal reduction based on modified Rajamathi’s method [26,27]. Briefly,
exchange method described in the literature [13]. In short, K
titanate (H Ti
muffle oven at 500 C to obtain mesoporous TiO
2 2 5
Ti O
2
2
O
5
2
2 5
O
o
2
(anatase). TiO
2

a desired mount of GO was placed in a mixed solvent of distilled water (20 mL) and ethanol (10
mL). After 1 h ultrasonic, 2.0 g TiO was added to the above GO solution and stirred for 2 h to
2
get a homogeneous suspension. Then the suspension was transferred to a 40 mL Teflon-Sealed
o
autoclave and maintained at 120 C in electric oven for 3 h. Finally, the resulting composite was
o
obtained by sedimentation, rinsed by deionized water for several times and dried at 60 C for 5 h.
2
The supports obtained were named as TGC-x (x represents the mass ratio of rGO/TiO in
percentage, x=0.25, 0.5, 1.0). The active components were introduced by incipient impregnation
with nickel nitrate hexahydrate (0.12 g) and ammonium heptamolybdate (0.12 g). Finally, the as-
o
prepared powder was calcined at 500 C for 2 h to obtain the NiMo/TGC catalyst loading with
NiO(1.4wt.%)-MoO
3
(4.7wt.%), which was the precursor of the “Ni–Mo–S” active phase for
HDS [28].
2
.3. Characterization methods
o
2
Structural properties of samples were determined by N adsorption-desorption (at -196 C)
in a Micromeritics Tristar II 3020M apparatus. And the average pore size was calculated by
D≈4VBJH/SBET. Fourier transform infrared (FT-IR) spectra were acquired using the Tensor 27
-1
infrared spectrometer with dry KBr. The spectra were recorded from 400 to 4000 cm with a
-1
resolution of 2 cm .
Acidity of catalysts were analyzed by FT-IR of adsorbed pyridine using the Tensor 27. 10
o
mg of a sample is pressed into a self-supporting wafer followed by evacuation at 350 C for 0.5
o
h. After cooled down to 50 C, the background spectrum was recorded, the sample was exposed
o
to 1 µL pyridine for 0.5 h until equilibration at the temperature of 50 C, after that the catalyst
o
o
was heated in a linear rate of 10 C/min to 280 C. Subsequently, the samples were outgassed for
o
0
.5 h at 280 C.
X-ray diffraction (XRD) patterns of the samples were obtained on a D/max-2200PC using
o
Cu Kα radiation operated at 40 kV, 30 mA, and scan speed of 10 /min.
Raman spectra were obtained using a Horiba HR 800 spectrometer, equipped with a charge
coupled device (CCD) camera detector.
Transmission electron microscope (TEM) was performed using a JEM-1010 instrument
supplied by JEOL. The surface morphological details were studied by scanning electron
microscope (SEM, JSM-5600LV).
3 3
NH -temperature programmed desorption (NH -TPD) tests were carried out on a BELCAT-
A apparatus (Ankersmid, Netherlands). The synthesized catalyst (0.33 g, 20-40 mesh) was
o
o
pretreated with a pure He flow from room temperature to 300 C at a rate of 10 C/min and then
o
cooled down to room temperature. Then, NH
mixture gas of 5 vol.% NH
weakly adsorbed NH . Finally, the catalyst was heated to 600 C at a rate of 10 C/min in pure
He and desorption pattern was recorded.
3
adsorption was carried out at 35 C for 2 h using a
3
/He, and subsequently purged by pure He to remove excessive and
o
o
3

Temperature-programmed reduction (TPR) characterization was conducted on TP-5000
apparatus (Tianjin Xianquan Co., Ltd., China). 20 mg sample was taken each time and the
o
o
temperature was increased up to 900 C at a rate of 10 C/min, meanwhile, the gas mixture of 10
vol.% H /Ar was introduced at 20 mL/min into the quartz reactor.
2
Irreversible CO uptake measurements were used to estimate the active sites on the catalysts.
o
Usually, 0.3 g of sample was loaded into a quartz reactor and purged in He at 200 C for 3h.
3
-1
After cooling, pulses of CO in He carrier at 40 cm (NTP) min were injected at room
temperature through a sampling valve. CO uptake was calculated by measuring the decrease in
the peak areas caused by adsorption in comparison with the area of a calibrated volume.
2
.4. Catalytic activities
The DBT HDS activity tests were conducted in a continuous fixed-bed reactor with 0.7 g
catalyst. Prior to the reaction, all fresh catalysts were sulfided in a mixture of 2 wt.% CS
2
-
o
cyclohexane at 350 C for 4 h under 3.0 MPa H
2
pressure. After sulfidation, the temperature and
o
pressure were decreased to reaction conditions of 280 C and 2.0 MPa, the DBT feedstock with a
concentration of 1300 ppm (decalin was used as solvent) was fed into reactor under conditions
−
1
liquid hourly space velocity (LHSV) of 4 h and hydrogen/feed ratio 400 (V/V). 5 h later, when
the steady reaction conditions had been achieved, H S detector tube was installed at the position
16th) in Fig. 1 to detect the density of H S desorption in the next 5 h, meanwhile, liquid samples
2
(
2
were collected hourly and analyzed by gas chromatograph (GC-14C) installed with a flame
ionization detector and a 30 m KB-1 capillary column.
Fig. 1. Experimental flow chart (1) dual piston micropump, (2) check valve, (3) Mass flowmeter,
(
4, 5, 14, 15) valve, (6) reducing valve, (7) open tubular resistance furnace, (8) catalyst layer, (9)
quartz sand, 10 thermocouple, (11) temperature control, (12) liquid trap, (13) backpressure valve,
16) H S detector tube.
(
2

The DBT molecules undergo HDS through two parallel reaction routes: direct
desulfurization (DDS) and hydrogenation desulfurization (HYD) [29,30]. Biphenyl (BP) and
cyclohexylbenzene (CHB) correspond to the main product of the above processes, respectively.
In the experiment, the detected main products are BP and CHB, the HDS conversion and the
selectivity for HYD path could be defined as shown in Eq.(1) and Eq.(2):
(
1)
2)
(
3
3
. Results and discussion
.1. Textural properties
The nitrogen adsorption isotherms of the catalysts and their corresponding Barrett–Joyner–
Halenda (BJH) pore size distributions determined from the adsorption branch are exhibited in
Fig. 2. The isotherm pattern of all catalysts present the typical adsorption characteristics of
IUPAC V, the NiMo/GO shows obvious H4 type hysteresis loop due to layer structure.
However, the other catalysts exhibit typical H3 type hysteresis loop at relative high pressure
(
0.8<P/P
mesoporous structure [31,32]. Besides, compared with NiMo/TiO
almost no vertical shifts between P/P =0.8 and 0.99, which can be attributed to the pore structure
0
<0.99), suggesting that the as-prepared material is sheet/virgate and possesses
2
, the modified catalysts show
0
2
changes slightly indicating that there is no blockage of the mesoporous TiO [33]and this is also
verified in BET.
Fig. 2. Nitrogen adsorption–desorption isotherms and BJH pore size distributions of different
catalysts.
BET surface area (SBET), pore volume (VBJH), and average pore diameter of different
2
modified supports and catalyst are shown in Table 1. The original TiO shows the largest SBET of

2
-1
1
02.5 m g . With the incorporation of rGO, the SBET of modified supports decrease slightly
indicating that rGO has a good dispersion on TiO . Moreover, after the impregnation of active
2
component, the SBET of corresponding catalysts decrease in a limited extent. Compared with
supports, the VBJH and pore size of the corresponding catalysts do not change obviously. These
results show that the incorporation of active components do not affect the pore accessibility of all
supports and block the pores [33].
3
.2. FT-IR
FTIR spectra of GO and supports are shown in Fig. 3. The characterization method is
usually used to detect the changes of functional groups after materials modification. The broad
Fig. 3. FTIR spectra of GO and supports.
-
1
-
absorptions below 1000 cm are attributed to the vibration of Ti-O-Ti, the absorptions at 974 cm
1
for the modified supports are weaker than that for the pure TiO
2
, which is mainly due to the
partially coverage of the surface of TiO
2
by the rGO sheets. The absorption bonds (curve of
-1
modified supports and GO) appearing at 1365 and 1620 cm are ascribed to C=C stretching
-1
vibrations. Moreover, the absorption at 1690 cm (curve GO) is assigned to the C=O stretching
vibration [34] and it is absent in all the other supports, which indicates that GO is entirely
reduced to rGO during the hydrothermal process.[26] The above results confirm the coexistence
2
of TiO and rGO in the composite.
3
.3. XRD

Fig. 4. XRD patterns of catalysts-fresh.
Fig. 4 shows the XRD patterns of catalysts-fresh. As shown in Fig.4, the above four
samples show diffraction peaks at 25.3°, 37.8°, 48°, 53.9°, 55.1°, 62.6° and 75.2°, which are
2
ascribed to the diffraction of anatase TiO (JCPDS 21-1272). The bottom curve is NiMo/GO, the
peak around 13° is assigned as diffraction of GO, and peak around 28° is attributed to the
diffraction of graphite. No peaks of NiMo phases are observed in Fig. 4, which is due to the high
dispersion of NiMo species on the supports and this is consistent with the Nitrogen adsorption–
desorption results. Moreover, no diffraction peaks for graphene are exhibited in the curves of the
modified catalysts, which is attributed to the low amount and relatively low diffraction intensity
of graphene [25]. Furthermore, it can be concluded that the incorporation of rGO does not change
the crystalline phase of TiO
.4. Raman
2
.
3

2
Fig. 5. Raman spectra of (a) NiMo/TiO , (b) NiMo/TGC-0.25, (c) NiMo/TGC-0.5, (d)
NiMo/TGC-1.0.
Raman spectroscopy is widely used to determine the state of oxide (Mo and Ni) loaded on
-1
the supports. As presented in Fig. 5, typical anatase bonds at 395, 515, 640 cm are observed
-
1
[
35,36]. In the lower frequency region, the bands at 244 cm are assigned to the deformation
-1
vibration of the Mo–O–Mo bridges in polymolybdate surface while the bands around 300 cm
6
-
4-
are due to the bending vibrations of Mo=O in either Mo
7 8
O24 or Mo O26 [37,38].
The NiMo oxide species emerge mainly at the higher frequency region, for all catalysts, the
-
1
-1
peaks arising at 750-1050 cm are attributed to Mo species [37]. The peaks at 800 cm are
2
-
-1
originated from the species of MoO
4
[39]. The peaks at 948 and 958 cm are attributed to the
6
-
4-
symmetric and antisymmetric Mo=O terminal stretches of Mo
7 8
O24 and Mo O26 , respectively
6
-
-1
[
37,39,40]. However, the Mo
7
O
24 peak at 948 cm is weaker, which implies that the amount of
6
-
4-
Mo
7
O
24 is smaller than that of Mo
8
O
26 . Moreover, the existence of some NiMoO
4
species
-
1
(
Raman bands at 948 and 958 cm ) cannot be excluded [41,42]. Compared with the unmodified
-
1
catalysts, the new peaks at 875 cm can be ascribed to the β-NiMoO
easier to be reduced than α-NiMoO and shows better catalytic properties [41]. Meanwhile, the
well-defined bonds for crystalline MoO
4
phase, which is considered
4
-1
3
at 667, 818, 996 cm are not appeared, which imply
that highly dispersion of Mo oxide species in the as-prepared NiMo/TGC catalysts [43-45]. And
the results are consistent with the XRD characterization results. In addition, curve b, c and d
show the presence of D and G bond of graphene indicating that the successful incorporation of
rGO to the modified catalyst [46], and this is consistent with the FT-IR results.
3
3
.5. NH -TPD and Pyridine FT-IR
3
Fig. 6. (a) NH -TPD and (b) Pyridine FT-IR profiles of catalysts.

NH
base and can interact with OH species on catalyst, which represent the acidity of samples. The
absorption of NH can be divided into two models as the physical absorption and the chemical
3 3
-TPD is widely used to measure the acidic properties of catalysts, as NH is a strong
3
absorption. The effect of rGO on the acidity of catalysts is exhibited in Fig. 6a, the desorbed NH
3
o
below 105 C is attributed to the physically absorbed NH
3
while the chemically absorbed NH
3
is
o
desorbed between 105 and 610 C. Based on the temperature at which adsorbed NH
3
o
desorbs, the
o
o
acid sites are classified as weak (<230 C), medium (230–430 C) and strong (>430 C) acid sites
47]. Obviously, an intensive desorption peak (marked in red circle) is observed in demonstrating
[
the increase of medium acid sites and the type of acid will be discussed later.
In order to investigate the acid types of catalysts, Pyridine FT-IR analysis is adopted.
Generally, the adsorption of pyridine molecules are classified into three types: (i) the protonation
of pyridine characterizes the Brönsted acid sites, showing adsorption bonds around 1540 and
-
1
1
630 cm ; (ii) the coordination between pyridine molecules and the Lewis acid sites giving
-1
adsorption bonds at 1440-1460, 1575-1580, 1600-1620 cm ; (iii) the weak hydrogen bond
between pyridine and hydroxy groups, which exhibits adsorption bonds around 1447 and 1590
-1
cm [48]. As shown in Fig. 6b, NiMo/TGC-x show adsorption bands at 1447, 1490, 1558 and
-1
-1
1
607 cm . The bands at 1447 and 1607 cm are attributed to Lewis acid sites [49], with the
increasing of x, the Lewis acid sites increase at first and reach a maximum at x=0.5 then
-1
decrease. Duan et. al. [50] assigned the bond at 1490 cm to pyridine adsorption of both Lewis
-1
acid and Brönsted acid sites. The absorption bands at 1558 cm are due to Brönsted acid sites,
which exhibit no significant changes with the increasing of x.
In summary, NH
different degree. Among of them, the acidity of NiMo/TGC-0.5 increases at a maximum, but
NH -TPD cannot give the types of the increased acid sites. Pyridine FT-IR exhibits that the acid
3
-TPD analysis shows the overall acidity of modified catalysts increase in
3
sites increase mainly in the form of Lewis acid (Table 2, column 3), which is likely to facilitate
the DBT molecules adsorption on catalysts and promote the HYD route.
3
2
.6. H -TPR

Fig. 7. TPR profiles of different catalyst-fresh.
H
2
-TPR is adopted to inquire more characteristics into Ni and Mo oxide on the surface of
catalyst and their changes with the composition of support. TPR characterization results are
o
shown in Fig. 7, all of catalysts exhibit reduction peak at 340-360 C, the temperature peak is
6
+
4+
usually assigned to the first step of reduction(Mo to Mo ) of polymeric octahedral Mo species
o
bonded to TiO
2
surface weakly[51]. The higher temperature with a maxima 600-800 C is
4
+
0
ascribed to the second step of reduction(Mo to Mo ) of polymeric octahedral Mo and isolated
tetrahedral Mo species interacted with the support strongly[52,53]. It should be noted that with
the catalyst modified with rGO, the reduction temperature in high temperature region declines at
o
least 150 C. This indicates the incorporation of rGO can weaken the interaction between active
component and support, thereby reduce the reduction temperature for active species.
3
.7. TEM and SEM

Fig. 8. (a) TEM image of GO, (b) SEM image of TiO
TEM image of NiMo/TiO , (e) TEM image of NiMo/TGC-0.5, (f) TEM image of NiMo/TGC-
.5-spent, (g, h) HRTEM images of NiMo/TGC-0.5-spent.
2
, (c) SEM image of NiMo/TGC-0.5, (d)
2
0
The microphotographs of GO, TiO support, NiMo/TiO and NiMo/TGC-0.5 are shown in
2
2
Fig. 8. The synthesized GO nanosheet is investigated by TEM in Fig. 8a, which shows a
semitransparent micro-surface with a few thin ripples. As reported previously, the ripples are
intrinsic properties of thin GO sheets [54], which are due to the extra thermodynamic stability of
2
the 2D nanosheets arising from microscopic crumpling. The as-prepared TiO (Fig. 8b) possesses
a length/diameter ratio of 20, which is also called whisker. This is accordance with the results of
nitrogen adsorption–desorption isotherms. After the impregnation of Mo and Ni species,
NiMo/TGC-0.5 maintains the whisker morphology (Fig. 8c), but the surface becomes rough. The
TEM of fresh NiMo/TiO
2
and NiMo/TGC-0.5 are shown in Fig. 8d and 8e, through careful
observation, it can be obtained that the active species deposited on the modified catalyst shows

2
better dispersion than NiMo/TiO . Low-resolution TEM image of NiMo/TGC-0.5-spent is
exhibited in Fig. 8f, which shows the diameter of matrix is around 50 nm. Fig. 8g and 8h show
the high-resolution TEM images of NiMo/TGC-0.5-spent. It can be observed (Fig. 8g) that, the
active species arranges closely and has a good dispersion on the support, which is due to the
weak interaction between active metals and support. In the high magnification TEM micrographs
(Fig. 8h), the well-known Ni–Mo–S slabs (marked by arrows) are observed, exhibiting
multilayer stacking structure, which are the true reaction active centers and considered to have a
good deep HDS performance [55].
3
.8. XPS
Fig. 9. XPS envelope of the Ni 2p and Mo 3d of sulfided NiMo/TiO
Fig. 9a shows the XPS profiles of the Ni 2p region of sulfided NiMo/TiO
.5. The Ni 2p envelope is consist of a mixture with three substances (NiS, NiO and
2
and NiMo/TGC-0.5
2
and NiMo/TGC-
0
NiMoS)[56]. The binding energy of NiS for the two catalyst are around 875 eV, the peak
emerging at 861 eV is attributed to NiO and the peak of NiMoS appears at 855 eV[57,58]. It
should be mentioned that the modified catalyst shows a higher peak of NiMoS. The Mo 3d XPS
2
of sulfided NiMo/TiO and NiMo/TGC-0.5 are presented in Fig. 9b. Generally, the signals of Mo
4
+
5+
6+
3
d include Mo (MoS
2
), Mo (oxysulfide) and Mo (oxide)[59], each of Mo species appears in
4
+
5+
pairs such as, Mo 3d5/2 and 3d3/2. The deconvolution peaks of Mo 3d5/2 for Mo , Mo and
6
+
Mo are around 229 eV, 230 eV, 232.5 eV, respectively. The binding energies values of Mo
d3/2 are fixed 3.1 eV more than those of Mo 3d5/2. Moreover, the peak appearing about 226
3
4
+
eV is ascribed to the S 2s line. It is noted that the Mo ratio of NiMo/TGC-0.5 catalyst is higher
than that of NiMo/TiO . In conclusion, an appropriate of rGO can promote the formation of
active sites and enhance the efficiency of desulfurization.
.9. CO uptake
2
3

The kHDS, CO uptake and TOF of different catalysts are presented in Table 3. Among of
them ,CO uptake is undertaken over the spent catalyst at room temperature to evaluate the active
sites on catalyst. The kHDS and TOF are calculated by KHDS
and
,
-1
respectively, where F is the rate of DBT fed into the reactor (µmol s ), w is mass of catalyst (g),
y is the conversion of DBT and M is the active sites deposited on catalyst decided by CO
uptake[49,60]. As can be seen in Table 3, the higher conversion of DBT corresponding to a
higher value of KHDS, and the NiMo/TGC-0.5 catalyst shows the maximum of 0.76. The CO
2
uptake amount of NiMo/TGC-0.5 is lower than that of NiMo/TiO , however, and it possesses
relatively higher TOF value, which indicates rGO sheets play an indispensable role in HDS
reaction.
3
.10. Hydrodesulfurization
Fig. 10. (a) HDS performance and selectivity of catalysts, (b) The long term stability of
2
NiMo/TiO and NiMo/TGC-0.5
The activities of different catalysts were investigated under relative mild conditions (280
o
C, 2.0 MPa). As shown in Fig. 10a, the conversion of DBT increases with x and reaches a
maximum conversion of 99.9% at x=0.5. With the further increase of x, the conversion of DBT
decreases slightly, which may arise from the loss of synergistic effect between the support and
active species due to the excess incorporation of rGO. Moreover, the selectivity of catalysts show
a similar trend and this is attributed to the Lewis acid sites of NiMo/TGC change in different
degree, as the Lewis acid site can serve as the DBT adsorption sites and improve the HYD route.
1
00 h stability of NiMo/TiO
samples are collected at the 5th hour (starting time when pump is turned on). As shown in Fig.
0b, the HDS performance of NiMo/TGC-0.5 is stable at a value of 99.9% in the reaction 100 h.
2
and NiMo/TGC-0.5 catalyst are presented in Fig. 10b,
1

As a comparison sample, NiMo/TiO
exhibits a stable value of 78.8%.
2
is also considered in the long term stability experiment and
2
Fig. 11. Proposed mechanism for the HDS of NiMo/TGC-0.5 and H S desorption rate (inset)
detected by rapid gas detector tube.
On the basis of the above experimental results and theoretical analyses, a possible catalytic
mechanism of DBT on NiMo/TGC-0.5 is proposed and explain the increase in activity in two
aspects. (i) During the HDS process, hydrogen molecules dissociate into hydrogen atoms on the
active sites of the catalyst [61], and then transfer to adjacent surface of support by spillover and
diffusion. Meanwhile, DBT molecules can be absorbed on active metal sites and Lewis acid
a
sites. The former are assigned as DBT and they undergo HDS via direct desulfurization (DDS).
b
The latter are named as DBT and they undergo HDS via hydrogenation (HYD). As the Lewis
b
acid sites of NiMo/TGC-0.5 increase the most (Table 2, column 3), more DBT are absorbed on
the Lewis acid sites (Fig. 11b) undergoing HDS via HYD. Finally, when the dissociative
hydrogen atoms flow over the active sites, they participate in hydrogenation reaction and form
the resultant of H
dynamics simulation showed that polar molecule tended to be absorbed on mesoporous TiO
which would influence the activity of catalysts definitely [16]. Li and his co-workers [17] found
that the surface of carbon-modified TiO possessed excellent hydrophobic property, which can
2
S, realizing deep desulfurization. (ii) As mentioned previously, our molecular
2
,
2
weaken the adsorption of polar molecules and accelerate their transfer. This is beneficial to the
reaction whose product is polar molecule, such as the HDS. Consequently, a graphene
modification on TiO
HDS reaction moving forward and enhance the HDS performance. The rapid gas (H
tube is used to measure the H S desorption rate from catalysts, as exhibited the inset in Fig. 11
under 5 h steady-state of HDS reaction), the tube gives a value of 2000 ppm for the unmodified
catalyst, whereas the NiMo/TGC-0.5 is 4200 ppm. The result indicates that the H S desorption
2
support is proposed to accelerate the desorption of product H
2
S, promote
2
S) detector
2
(
2

rate from catalyst surface is faster in the presence of rGO sheets. And this is also an important
reason for improving the activity of the catalyst.
4
. Conclusions
In summary, we examined the influence of graphene heterogeneous surface modification on
the HDS performance of TiO -supported catalyst and found the improving effect of graphene
addition for the first time. Through adjusting TiO surface adsorption towards H S molecule,
realizing an optimization between diffusion and HDS process. The presence of rGO changes the
desorption behaviours of product H S molecules-accelerating their transfer rate on catalysts, and
2
2
2
2
promotes reaction moving forward. In addition, the incorporation of graphene induces an
enhanced Lewis acid site, which facilitates the adsorption of DBT molecules and promotes the
2
HYD route. When the mass ratio of rGO/TiO is 0.5 in percentage, the corresponding catalyst
shows the best HDS activity. Compared with the unmodified catalyst, the DBT conversion of
o
NiMo/TGC-0.5 increases from 78.6% to 99.9% under relative mild conditions of 280 C, 2.0
−
1
MPa, LHSV 4 h , and hydrogen/feed ratio of 400 (V/V). Our present work shows that the
heterogeneous modification on the surface of materials is an effective and alternative way to
2
open a new possibility in the investigation of TiO -rGO composites and promote their practical
application in addressing various environmental issues.
Acknowledgements
The research is financially supported by the National Young Top Talents Plan of China
(
(
2013042), National Science Foundation of China (Grant No.51175066), FANEDD
201164), New Century Excellent Talents in University (NCET-12-0704) and the Science
Foundation for Distinguished Young Scholars of Heilongjiang Province (JC201403).
References
[1]
[2]
[3]
H.M. Li, L.N. He, J.D. Lu, W.S. Zhu, X. Jiang, Y. Wang, Y.S. Yan, Energy Fuels 23 (2009)
1
354.
S. Nassreddine, S. Casu, J.L. Zotin, C. Geantet, L. Piccolo, Catal. Sci. Technol. 1 (2011)
08.
4
M. Breysse, G. Djega-Mariadassou, S. Pessayre, C. Geantet, M. Vrinat, G. Pérot, Catal.
Today 84 (2003) 129.
[4]
[5]
[6]
[7]
[8]
T.C. Ho, J.M. McConnachie, J. Catal. 277 (2011) 117.
C. Song, X. Ma, Appl. Catal., B 41 (2003) 207.
I. Fechete, Y. Wang, J.C. Védrine, Catal. Today 189 (2012) 2.
Y. Fan, H. Xiao, G. Shi, H.Y. Liu, Y. Qian, T.H. Wang, J. Catal. 279 (2011) 27.
M.J. Luo, Q.F. Wang, G.Z. Li, X.W. Zhang, L. Wang, T. Jiang, Catal. Sci. Technol. 4, (2014)
2
081.
[9]
F. Trejo, M.S. Rana, J. Ancheyta, A. Rueda, Fuel, 100 (2012) 163.
[
[
10]
J.C. Amezcua, L. Lizama, C. Salcedo, I. Puente, J.M. Domínguez, T. Klimova, Catal. Today
1
07 (2005) 578.
11]
T.K.T. Ninh, L. Massin, D. Laurenti, M. Vrinat, Appl. Catal., A 407 (2011) 29.

[
[
[
[
12]
13]
14]
15]
P. Castillo-Villalón, J. Ramírez, J. Catal. 268 (2009) 39.
M. He, X.H. Lu, X. Feng, L. Yu, Z.H. Yang, Chem. Commun. 19 (2004) 2202.
H.M. Wang, E. Iglesia, J. Catal. 273 (2010) 245.
R. Klaewkla, M. Arend, W.F. Hoelderich, de Mass Transfer-Advanced Aspects, Germany,
InTech (2011) 668.
[
[
16]
17]
M.J. Wei, J. Zhou, X.H. Lu, Y.D. Zhu, W.J. Liu, L.H. Lu, L.Z. Zhang, Fluid Phase Equilib.
3
02 (2011) 316.
L.C. Li, Y.D. Zhu, X.H. Lu, M.J. Wei, W. Zhuang, Z.H. Yang, Chem. Comun. 48 (2012)
1525.
1
[
[
[
[
[
[
[
[
[
[
[
[
[
[
18]
19]
20]
21]
22]
23]
24]
25]
26]
27]
28]
29]
30]
31]
H. Li, M. Li, Y. Chu, F. Liu, H. Nie, Appl. Catal., A 403 (2011) 75.
F. Cui, G.C. Li, X.B. Li, M.H. Lu, M.S. Li, Catal. Sci. Technol. DOI: 10.1039/c4cy00814f
L. Li, Y. Wang, K. Shi, S. Chen, Z. Yang, X. Lu, Catal. Lett. 142 (2012) 480.
V. Chabot, D. Higgins, A. Yu, Energ. Environmen. Sci. 7 (2014) 1564.
N. Li, G. Liu, C. Zhen, Adv. Funct. Mater. 21 (2011) 1717.
P.V. Kamat, J. Phys. Chem. Lett. 2 (2011) 242.
Y.Y. Gao, X.P. Pu, D.F. Zhang, G.Q. Ding, X. Shao, J. Ma, Carbon 50 (2012) 4093.
Y. Wang, Y.M. Li, L.H. Tang, J. Lu, J.H. Li, Electrochem Commun. 11 (2009) 889.
H. Zhang, X.J. Lv, Y.M. Li, Y. Wang, J.H. Li, ACS Nano 4 (2009) 380.
C. Nethravathi, M. Rajamathi, Carbon 46 (2008) 1994.
S.J. SardharBasha, N.R. Sasirekha, R. Maheswari, K. Shanthi, Appl. Catal., A 308 (2006) 91.
S.T. Oyama, Y.K. Lee, J. Catal. 258 (2008) 393.
X. Li, Z. Sun, A. Wang, X. Yang, Y. Wang, Appl. Catal., A 417 (2012) 19.
K.S.W. Sing, D.H. Everett, R.A.W. Haul, L. Moscou, R.A. Pierotti, J. Rouquerol, T.
Siemieniewska, Pure & Appl. Chem. 57 (1985) 603.
[
[
[
32]
33]
34]
J.L. Blin, C. Otjacques, G. Herrier, B.L. Su, Langmuir 16 (2000) 4229.
D. Ferdous, N.N. Bakhshi, A.K. Dalai, J. Adjaye, Appl. Catal., B 72 (2007) 118.
Y. Matsumoto, M. Koinuma, S. Ida, S. Hayami, T. Taniguchi, K. Hatakeyama, S. Amano, J.
Phys. Chem. C 115 (2011) 19280.
[
[
[
[
[
[
[
35]
36]
37]
38]
39]
40]
41]
Z. Cao, A. Duan, Z. Zhao, J. Li, Y. Wei, G. Jiang, J. Liu, J. Mater. Chem. A 2 (2014) 19738.
J. Zhang, M. Li, Z. Feng, J. Chen, C. Li, J. Phys. Chem. B 110 (2006) 927.
G. MESTL, T.K.K Srinivasan, Catal. Rev. 40 (1998) 451
A.N. Desikan, L. Huang, S.T. Oyama, J. Phys. Chem. 95 (1991) 10050.
V. La Parola, G. Deganello, C.R Tewell, A.M. Venezia, Appl. Catal., A 235 (2002) 171.
S. Badoga, K.C. Mouli, K.K. Soni, A.K. Dalai, J. Adjaye, Appl. Catal., B 125 (2012) 67.
Y.G. Wang, G. Xiong, X. Liu, X.C. Yu, L.P. Liu, J.Y. Wang, Z.C. Feng, C. Li, J. Phys. Chem.
B 112 (2008) 17265.
[
[
[
42]
43]
44]
Z. Liu, Y. Chen, J.Catal. 177 (1998) 314.
S. Sigurdson, V. Sundaramurthy, A.K. Dalai, J. Adjaye, J. Mol. Catal. A 291 (2008) 30.
R. Huirache-Acuña, B. Pawelec, E. Rivera-Muñoz, R. Nava, J. Espino, J.L.G. Fierro, Appl.
Catal., A 92 (2009) 168.
[
[
45]
46]
J.A. Mendoza-Nieto, O. Vera-Vallejo, L. Escobar-Alarcón, D. Solís-Casados, T. Klimova,
Fuel 110 (2013) 268.
P.N. Zhu, N. A.Sreekumaran, S.J. Peng, S.Y. Yang, R. Seeram, ACS Appl. Mater. Interfaces
4
(2012) 581.
[
[
47]
48]
M. Lewandowski, Z. Sarbak, Fuel 79 (2000) 487.
G.A. Eimer, S.G. Casuscelli, C.A. Chanquia, V. Elias, M.E. Crivello, E.R. Herrero, Catal.

Today 133 (2008) 639.
[
[
[
[
[
49]
50]
51]
52]
53]
H. Song, J. Wang, Z.D. Wang, H.L. Song, F. Li, Z.S. Jin, J. Catal. 311 (2014) 257.
X. Duan, Y. Teng, A. Wang, V. M. Kogan, X. Li, Y. Wang, J. Catal. 261 (2009) 232.
R. López Cordero, A. López Agudo, Appl. Catal., A 175 (2000) 23.
M.Á. Calderón-Magdaleno, J.A. Mendoza-Nieto, T.E. Klimova, Catal. Today 220 (2014) 78.
T.E. Klimova, D. Valencia, J.A. Mendoza-Nieto, P. Hernández-Hipólito, J. Catal. 304 (2013)
2
9.
H.A. Becerril, J. Mao, Z. Liu, R.M. Stoltenberg, Z. Bao and Y.S. Chen, ACS Nano 2 (2008)
63.
[
54]
4
[
[
55]
56]
N. Y. Topsøe, H. Topsøe, J. Catal. 84 (1983) 386.
W.K. Lai, W.J. Song, L.Q. Pang, Z.F. Wu, N. Zheng, J.J. Li, J.B. Zheng, X.D. Yi and W.P.
Fang, J. Catal. 303 (2013) 80.
[
[
57]
58]
J. Mérida-Robles, E. Rodrıguez-Castellón, A. Jiménez-López, J. Mol. Catal. A 145 (1999)
1
69.
B. Guichard, M. Roy-Auberger, E. Devers, C. Legens and P. Raybaud, Catal. Today 130
2008) 97.
(
[
[
59]
60]
P. Beccat, P.D. Silva, Y. Huiban and S. Kasztelan, Oil Gas Sci. Technol. 54 (1999) 487.
V.T. da Silva, L.A. Sousa, R.M Amorim, L. Andrini, S.J.A. Figueroa, F.G Requejo, F.C.
Vicentini, J. Catal. 279 (2011) 88.
[
61]
R. Prins, Chem. Rev. 112 (2012) 2714.
Table 1 BET surface area, pore volume and average pore diameter of different supports and
catalysts

2
-1
3
-1
Sample
Support
S
BET (m g )
V
BJH (cm g )
Average pore diameter (nm)
TiO
2
102.5
99.8
97.8
97.3
0.27
0.28
0.27
0.28
10.7
10.8
10.8
11.2
TGC-0.25
TGC-0.5
TGC-1.0
Catalyst
NiMo/GO
39.7
99.5
97.9
95.6
95.0
0.10
0.29
0.31
0.29
0.32
8.9
NiMo/TiO
2
10.4
10.5
11.0
11.2
NiMo/TGC-0.25
NiMo/TGC-0.5
NiMo/TGC-1.0
Table 2 Distribution of Brönsted and Lewis acidity of catalysts.
Sample
Brönsted acidity Lewis acidity Total acidity
-1
-1
-1
(
mmol g )
(mmol g )
(mmol g )
Catalyst
NiMo/TiO
2
0
0
0
0.008
0.030
0.050
0.034
0.008
0.030
0.050
0.043
NiMo/TGC-0.25
NiMo/TGC-0.5
NiMo/TGC-1.0
0.009
Table 3 HDS reaction rate constant (kHDS), CO uptake and TOF.
-1
-1
-1
-3 -1
Sample
KHDS (mmol h g )
CO uptake (µmol g )
TOF (10 s )
NiMo/GO
0.20
0.26
0.39
0.76
0.43
-
61.0
-
16.3
-
-
36.8
-
172.8
-
NiMo/TiO
2
NiMo/TGC-0.25
NiMo/TGC-0.5
NiMo/TGC-1.0