Catalytic oxidation of thiophene and its derivatives via dual activation for ultra-deep desulfurization of fuels

Article abstract of DOI:10.1016/j.jcat.2011.11.003

A catalyst system composed of tungstate and Bronsted acidic ionic liquids (BAILs) was found to be highly active for the oxidative desulfurization (ODS) removal of thiophene, benzothiophene (BT), and their derivatives from model oil using 30 wt.% H₂O₂ as the oxidant. Five BAILs, [Hnmp]BF₄, [Hmim]BF₄, [Hnmp]HSO₄, [Hmim]HSO₄, and [Hnmp][CH₃SO₃] (nmp = N-methyl-pyrrolidonium, mim = N-methylimidazolium), and various commercial tungstate compounds were investigated. High activity was obtained for the combination of ammonium tungstate and BAILs [Hnmp]BF₄. In this catalytic reaction, sulfur content of model oil containing BT could be decreased from 700 ng μL^-1 to less than 1 ng μL^-1. Turnover frequency (TOF) for BT oxidation is higher than 194 × 10^-3 s^-1 and that of in non-BAILs [Bmim]BF₄ is less than 11 × 10^-3 s^-1. Noteworthily, the ODS of thiophene, which has been regarded as difficult task, can be also achieved up to 99% conversion with a TOF of 7 × 10^-3 s^-1. The FT-IR, ¹H NMR and electrochemical measurements evidences indicate that the strong hydrogen bonding between the sulfur-containing compounds and [Hnmp]BF₄ and the oxidative function of the tungstate synergistically activate the reactants and result in the excellent catalytic performance.

Full text of DOI:10.1016/j.jcat.2011.11.003

Journal of Catalysis 287 (2012) 5–12
Contents lists available at SciVerse ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
Catalytic oxidation of thiophene and its derivatives via dual activation
for ultra-deep desulfurization of fuels
a,
Boyu Zhang a,b, Zongxuan Jiang a, Jun Li a, Yongna Zhang a,b, Feng Lin a,b, Yan Liu a, , Can Li
⇑
⇑
a State Key Laboratory of Catalysis, Dalian Institute of Chemical Physics, Chinese Academy of Sciences, 457 Zhongshan Road, Dalian 116023, China
b Graduate University of Chinese Academy of Sciences, Beijing 100049, China
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 10 July 2011
Revised 4 November 2011
Accepted 5 November 2011
Available online 26 January 2012
A catalyst system composed of tungstate and Brønsted acidic ionic liquids (BAILs) was found to be highly
active for the oxidative desulfurization (ODS) removal of thiophene, benzothiophene (BT), and their
derivatives from model oil using 30 wt.% H2O2 as the oxidant. Five BAILs, [Hnmp]BF4, [Hmim]BF4,
[Hnmp]HSO4, [Hmim]HSO4, and [Hnmp][CH3SO3] (nmp = N-methyl-pyrrolidonium, mim = N-methylimi-
dazolium), and various commercial tungstate compounds were investigated. High activity was obtained
for the combination of ammonium tungstate and BAILs [Hnmp]BF4. In this catalytic reaction, sulfur con-
Keywords:
Thiophene
Dual activation
Hydrogen bonding
Oxidative desulfurization (ODS)
tent of model oil containing BT could be decreased from 700 ng l l
LÀ1 to less than 1 ng LÀ1. Turnover fre-
quency (TOF) for BT oxidation is higher than 194 Â 10À3 sÀ1 and that of in non-BAILs [Bmim]BF4 is less
than 11 Â 10À3 sÀ1. Noteworthily, the ODS of thiophene, which has been regarded as difficult task, can
be also achieved up to 99% conversion with a TOF of 7 Â 10À3 sÀ1. The FT-IR, 1H NMR and electrochemical
measurements evidences indicate that the strong hydrogen bonding between the sulfur-containing com-
pounds and [Hnmp]BF4 and the oxidative function of the tungstate synergistically activate the reactants
and result in the excellent catalytic performance.
Ó 2011 Elsevier Inc. All rights reserved.
1. Introduction
compounds to sulfone in the first step followed by extraction sul-
fone from the oxidized fuels in the second step [2]. In addition to
The deep removal of sulfur-containing compounds from fuel
oils has become an increasing challenge as growing environmental
concern. The industrial desulfurization is traditionally carried out
by catalytic hydrodesulfurization (HDS) technology. However, the
HDS process requires both high temperatures and high pressures
of H2, which results in high operating-cost and octane-number loss
particularly for FCC gasoline [1]. Therefore, the development of
alternative ultra-deep desulfurization processes is highly desired.
A variety of alternative routes such as oxidative desulfurization
(ODS), biodesulfurization, adsorptive desulfurization, extractive
desulfurization of sulfur compounds using solvents and ionic liq-
uids, have been investigated to remove sulfur from diesel streams
or FCC gasoline to ultra low levels. Among them, selective catalytic
oxidation combined with extraction has been extensively studied
for ultra-deep desulfurization. Until 2010, five processes involving
the ODS route for deep desulfurization of diesel were in commer-
cialization stage. These include (i) Sulphco process, (ii) Lyondell
chemicals process, (iii) Enichem/UOP process, (iv) Unipure process,
and (v) PetroStar process, which utilize H2O2, t-butyl hydroperox-
ide (TBHP), and peroxyacid (peroxo acetic acid) as oxidizing agent.
These processes require two steps, oxidation of sulfur-containing
ODS method, there are abundant studies about extraction with io-
nic liquids (ILs) for ultra-desulfurization. However, most of the effi-
ciency of sulfur removal extracted merely with ILs is relatively low
due to the similar polarity between the sulfur-containing mole-
cules and remaining diesel fuels. More recently, a new and effec-
tive approach that is chemical oxidation in conjunction with ILs
extraction (ECODS) has been explored in order to improve the effi-
ciency of desulfurization [3]. However, to date, the efficiency of
ECODS system for FCC gasoline is still low [4–14], because the ma-
jor sulfur-containing compounds of FCC gasoline, thiophene, and
its derivatives are difficult to be oxidized [15].
The inertness of the thiophene in the oxidation is mainly due to
its aromaticity and low electron density on its sulfur atom. So
highly reactive oxidants, such as dimethyldioxirane [16] and per-
acids [17,18], were used to oxidize thiophene in early years. As a
clean oxidant, hydrogen peroxide has always been attractive. Oxi-
dation of thiophene was once achieved with titanium silicates-1
(TS-1) for ODS using aqueous H2O2 as oxidant [19,20]. However,
suffering from the small size of the micro pores, TS-1 is not suitable
for the oxidation of bulky sulfur compounds, like BT and its deriv-
atives. Therefore, the prerequisite pursuit in ultra-low sulfur gaso-
line is to explore an efficient catalytic system for broad reactant
scopes in the oxidation of sulfur-containing compounds, particu-
larly thiophene-like ones with electron-deficiency.
⇑
Corresponding authors. Fax: +86 411 84694447 (C. Li).
E-mail addresses: canli@dicp.ac.cn (C. Li), yanliu503@dicp.ac.cn (Y. Liu).
0021-9517/$ - see front matter Ó 2011 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2011.11.003

6
B. Zhang et al. / Journal of Catalysis 287 (2012) 5–12
In this work, we found that a catalyst consisted of commercial
2.2. General procedure for ODS of sulfur-containing compounds and
recycling of catalyst
ammonium tungstate and commonly used BAILs could achieve
high activity in the oxidation of thiophene, BT, and their deriva-
tives under mild conditions using aqueous hydrogen peroxide as
oxidant. We also discussed the mechanism based on a concept of
metal oxide-Brønsted acid dual activation where the interaction
between sulfur-containing compound and BAILs was believed to
be critical for the high oxidative activity.
In a typical experiment, a mixture including 20 mL pre-prepared
model oil with sulfur content 700 ng
LÀ1, 0.5 mL 30 wt.% H2O2, 4 g
l
of [Hnmp]BF4, and 4 mol% catalyst in 50 mL vessel was stirred vigor-
ously at 60 °C. After the reaction, the sulfur content of the model oil
phase was analyzed directly by GC-SCD or Antek 9000S total sulfur
analyzer. The recycling of catalyst was conducted as follows: after
the reaction, ionic liquid phase was separated and extracted with
diethyl ether (4 mL Â 3). The water in ionic liquid phase was re-
moved under reduced pressure by rotary evaporation equipment
at 90 °C for 2 h. Then, fresh aqueous H2O2 (30 wt.%) and model oil
were added to ionic liquid phase for the next reaction cycle.
2. Experimental
2.1. General
Thiophene, 2-methylthiophene (2-MT), 3-metylthiophene (3-
MT), 2,5-dimethylthiophene (2,5-DMT), 2,3,5-trimethylthiophene
(2,3,5-TMT), and BT were purchased from Alfa Aesar. Six model oils
were prepared by dissolving thiophene, 2-MT, 3-MT, 2,5-DMT,
2,3,5-TMT, and BT in n-octane to give solutions with S-contents
2.3. Product identification and analysis
Afterthe reaction, the ILs phasewas separated and extractedwith
diethyl ether (4 mL Â 3). The organic phase was collected, dried with
anhydrous Na2SO4, and concentrated for product analysis.
of 700 ng l
LÀ1. The BAILs [Hnmp]BF4, [Hmim]BF4, [Hnmp]HSO4,
[Hmim]HSO4, and [Hnmp]CH3SO3 used in this work were synthe-
sized according to the procedures described in the literatures
[13,14,21–23]. The ionic liquid [Bmim]BF4 was purchased from
Shanghai Cheng Jie Chemical Co. Ltd. (Scheme 1). Other reagents
were obtained from commercial sources and used without purifi-
cation. Aqueous H2O2 was stored at 4 °C.
3. Results and discussion
3.1. Effect of different ILs on ODS of thiophene
Infrared spectra (IR) were recorded on a Nicolet NEXUS 470 FT-IR
Spectrometer in ambient air at room temperature. The IR spectra of
the IL/thiophene mixture were taken by a KBr liquid cell from 400 to
4000 cmÀ1 with 4 cmÀ1 resolution and averaged 10 times. Other IR
spectra were collected in KBr pellets. GC–MS was conducted on Agi-
lent 6890N GC/5973 MS detector. 1H NMR spectra were recorded on
a Bruker DRX 400 MHz type spectrometer at room temperature
(400 MHz) and internally referenced to tetramethylsilane signal.
Coaxial NMR experiments were conducted as follows: the samples
were loaded in a stem coaxial NMR capillary tube which inserted
into 5 mm NMR tube with D2O; the residual DOH in D2O was used
as the 1H NMR external reference at 4.790 ppm. Electrochemical
measurements were carried out in the three-electrode type cell.
The volume of the electrolyte was 10 mL. A glassy carbon electrode
and a platinum disk with a surface area of 0.068 cm2 were used as
workingand counter electrodes, respectively. Theworking electrode
was polished with a 10 mm alumina, rinsed with distilled water, and
then electrochemically activated in 1 M H2SO4 solution. All poten-
tials were estimated using a saturated calomel electrode (SCE) as
the reference electrode and the scan rate of the potential was
4 mV sÀ1. The remained sulfur-containing compounds in model oil
after oxidation were analyzed by GC-SCD (sulfur chemilumines-
cence detection) (GC: Agilent 7890A equipped with a capillary
In our initial study, we selected liquid–liquid extraction and
catalytic oxidative desulfurization to investigate the oxidative re-
moval of thiophene from model oil (thiophene in n-octane) using
30 wt.% H2O2 as oxidant. ILs can partially extract sulfur-containing
compounds from n-octane [6]. H2O2 is miscible with selected ILs.
So thiophene is extracted by ILs firstly and then oxidized in ILs.
The data in Table 1 show desulfurization systems of extraction,
the extraction combined with a chemical oxidation, and the extrac-
tion combined with the catalytic oxidation reaction in six different
ILs. In the extraction coupled with catalytic oxidation process,
commercial ammonium tungstate was selected as catalyst. When
[Bmim]BF4 was used as an extractant to remove thiophene from
model oil, the desulfurization efficiency is poor (the removal of thi-
ophene is only 15.3%). Either extraction combined with chemical
oxidation or extraction combined with catalytic oxidation does
not show obvious removal of thiophene. However, when
[Bmim]BF4 was replaced with [Hnmp]BF4, the conversion of thio-
phene is increased sharply and can reach as high as 99.4% in the
case of the extraction coupled with catalytic oxidation. It indicates
that [Hnmp]BF4 not only serves as extractant and reaction media
but also acts as co-catalyst.
To investigate the effect of different BAILs on the reaction,
[Hmim]BF4, [Hnmp]HSO4, [Hmim]HSO4 and [Hnmp]CH3SO3 were
column [PONA, 50 m  0.2 mm i.d.  0.5
lm]; SCD: Agilent G6600A-
S355). The total sulfur contents of the samples were determined
by Antek 9000S total sulfur analyzer. (Precision: 5 ppb or 2% RSD).
Table 1
Desulfurization performances via extraction with different ILs, extraction coupled
with chemical oxidation and ECODS with catalysts containing ammonium tungstate
and different ILs.
Entry Type of ILs
Thiophene removal of different desulfurization
systems (%)
Extraction with IL IL + H2O2 IL + catalyst + H2O2
1
2
3
4
5
6
[Bmim]BF4
[Hmim]HSO4
[Hnmp]SO3CH3 9.4
[Hmim]BF4
[Hnmp]HSO4
[Hnmp]BF4
15.3
4.3
18.5
13.3
15.9
20.7
30.2
19.6
19
23.8
44.8
47.1
63.5
99.4
4.2
10.0
10.4
Experimental conditions: T = 60 °C, t = 5 h, 20 mL model oil of thiophene in n-octane
with S-contents 700 ng
LÀ1, ILs = 4 g, H2O2/thiophene = 10 (molar ratio), S/C = n
(S)/n (catalyst (based on W)) = 25, catalyst: ammonium tungstate.
l
Scheme 1. Structures of ILs used in this work.

B. Zhang et al. / Journal of Catalysis 287 (2012) 5–12
7
used for ODS. The removal of thiophene is in the range of
4.2–10.4% for these four ionic liquids by a single extraction. With
adding of H2O2, the removal of thiophene is still insufficient. The
results are 13.3%, 15.9%, 20.7%, and 30.2% while using [Hmim]H-
SO4, [Hnmp]SO3CH3, [Hmim]BF4, and [Hnmp]HSO4, respectively.
When catalytic amount of ammonium tungstates were added to
the reaction mixtures, the conversion of thiophene increases to
23.8% for [Hmim]HSO4, 44.8% for [Hnmp]SO3CH3, 47.1% for
[Hmim]BF4, and 63.5% for [Hnmp]HSO4. This indicates that these
BAILs play an important role in catalytic oxidation of thiophene.
3.2. Effect of different catalysts on ODS of thiophene
The data in Table 2 show that the removal of thiophene was
only 19.6% without catalyst. Among various commercial catalysts,
tungstate-based catalysts exhibit the relatively high activity, espe-
cially ammonium tungstate, H2WO4 and Na2WO4Á2H2O, which cat-
alyzed the oxidation of thiophene with over 97% conversion. The
high activities of tungstate-based catalysts might attribute to per-
Fig. 1. Conversion of sulfur-containing compounds versus the reaction time at (a) S/
C = 100 and (b) S/C = 25. Experimental conditions: T = 60 °C, 20 mL model oil of
sulfur-containing compounds in n-octane with S-contents 700 ng l ,
LÀ1
oxo-species [{W(@O)(O2)2(H2O)}2(l
-O)]2À which are formed in the
[Hnmp]BF4 = 4g, H2O2/S = 10 (molar ratio), S/C = n(S)/n (catalyst(based on W)),
catalyst: ammonium tungstate. TOF value was based on the <30% conversion, and
given in mole of conversion of reactant per mole of W in the catalyst per hour.
mixture of tungsten-based precursors with H2O2 [24]. Under the
same reaction conditions, WO3 and H3PW12O40Á6H2O (HPA) show
very low activities which might be due to the poor solubility of
WO3 and HPA in BAILs.
apparent activation energy of the oxidation in BAILs [Hnmp]BF4
and in non-BAILs [Bmim]BF4, the oxidation of BT in the presence
of [Hnmp]BF4 or [Bmim]BF4 was carried out under different tem-
peratures. Fig. 2a displayed conversion of BT vs. reaction time un-
der various temperatures using [Hnmp]BF4 as extractant and
reaction media with 1 mol% ammonium tungstate (based on tung-
sten content). The oxidation reaction was run efficiently under
60 °C, and the activity decreased with decreasing temperature. At
the same reaction time, the conversion of BT increased from 53%
at 20 °C to 96% at 60 °C in 1 h. When BAIL [Hnmp]BF4 was replaced
by non-BAIL [Bmim]BF4, the conversion of BT is only 49% at 60 °C
and less than 5% at 20 °C (Fig. 2b).
From the reaction rates determined at various temperatures,
the apparent activation energies for the oxidation of BT were calcu-
lated derived from the Arrhenius equation (Fig. 3). In the ODS sys-
tem in the presence of non-BAIL [Bmim]BF4, 76.8 kJ molÀ1
apparent activation energy was obtained. However, less than
40 kJ molÀ1 apparent activation energy was calculated when the
oxidation was carried out in the presence of BAIL [Hnmp]BF4.
3.3. Sulfur removal of different sulfur-containing compounds
The ODS of six model sulfur-containing compounds from model
oil was carried out under the same conditions. As shown in Fig. 1a,
when the ratio of substrate/catalyst (S/C) is equal to 100, the TOF in
the oxidation of different sulfur-containing compounds decreases
for ODS in the follow order: BT > 2,3,5-3MT > 2,5-2MT > 3-
MT > 2-MT > thiophene. This order agrees well with the electron
density on the sulfur atom [15]. When the ratio of S/C is 25
(Fig. 1b), all of sulfur-containing compounds we selected are re-
moved completely in 5 h. Significantly, for ODS of BT, excellent cat-
alytic activity is achieved with high TOF (194 Â 10À3 sÀ1).
Therefore, the catalytic system we developed has a broad substrate
scope in aromatic sulfur-containing compounds.
3.4. Influence of temperature on ODS of BT and difference of activation
energies for oxidation BT in presence of BAILs and non-BAILs
In the foregoing experiments, the catalyst consisted of ammo-
nium tungstate and [Hnmp]BF4 showed high catalytic activity for
ODS of thiophene, BT, and its derivates. In order to obtain the
Table 2
Conversion of thiophene via ECODS with catalyst containing various commercially
tungsten, molybdenum or vanadium containing compounds and [Hnmp]BF4.
Entry Catalyst
Thiophene removal of different
desulfurization systems (%)
1 h
3 h
5 h
1
2
3
4
5
6
7
8
9
(NH4)10H2(W2O7)6ÁH2O 64
96.4
86.6
82.5
13.5
15.3
25.4
21.5
17.9
10.2
16.3
99.4
98.3
97.7
18.3
15.6
30.8
31.2
21.4
10.9
17.3
H2WO4
45.5
Na2WO4ÁH2O
33.9
7.6
WO3
H3PW12O40ÁxH2O
(NH4)6Mo7O24Á4H2O
H2MoO4
12.5
18.3
17.9
12.8
9.5
Na2MoO4Á2H2O
MoO3
Fig. 2. Conversion of BT versus the reaction time on the reaction temperature via
ECODS with catalysts containing ammonium tungstate and (a) [Hnmp]BF4 (b)
[Bmim]BF4. Experimental conditions: 20 mL model oil of BT in n-octane with S-
contents 700 ng
ammonium tungstate.
10
V2O5
15.2
Experimental conditions: T = 60 °C, t = 5 h, 20 mL model oil of thiophene in n-octane
with S-contents 700 ng
LÀ1, [Hnmp]BF4 = 4 g, H2O2/thiophene = 10, thiophene/W,
Mo or V = 25 (molar ratio).
l , ILs = 4g, H2O2/BT = 2.5(molar ratio), S/C = 100, catalyst:
LÀ1
l

8
B. Zhang et al. / Journal of Catalysis 287 (2012) 5–12
Fig. 3. Arrhenius activation energies for BT oxidation in BAIL [Hnmp]BF4 and
[Bmim]BF4.
Therefore, the reaction rates in all the sulfur-containing
compounds are enhanced in the presence of the BAILs. The func-
tion of BAILs will be discussed in section below.
3.5. Oxidation products
Qualitative measurements of oxidative products were done as
follows. The sulfur species could not be detected in oil phase and
all of products exist in the ILs phase due to the different chemical
polarity between products and model oil. After the reaction, the ILs
phase was separated and extracted with diethyl ether (4 mL Â 3).
The organic phase was collected, dried with anhydrous Na2SO4,
and concentrated for GC–MS analysis. GC–MS analysis of the prod-
ucts from thiophene oxidation gives two main compounds for
qualitative analysis (Fig. S1 in Supporting information). The spectra
of products A and B match that of benzaldehyde (C7H6O, MW=106)
and benzoic acid (C7H6O2, MW=122), respectively. After a complete
reaction, Ba(NO3)2 solution was added to the ILs phase. A white
precipitate appeared immediately. After washed with an excess
of H2O and dried, the XRD pattern of the precipitate is attributed
to BaSO4 (Fig. 4a). It is clear that the oxidation reaction of thio-
phene takes place and yields the SO24À and oxygenates.
The mass spectra of products B and C from oxidation of 2-MT
match that of 2-methylbenzoic acid (C8H8O2, MW = 136.1) and 3-
methylbenzoic acid (C8H8O2, MW = 136.0), respectively (Fig. S2 in
Supporting information). The mass spectrum of product A shows
the molecular ion peak at m/z 130.0 and the strongest peak at m/
z 82.1 due to the furan radical cation (Fig. 4b). Based on the similar
results from characterizing thiophene 1,1-dioxide using EI-MS in
the literature [16], it implies that 2-MT 1,1-dioxide is the major
product after the ODS.
Fig. 4. (a) The powder-XRD patterns of BaSO4 precipitation; (b) the mass
spectrogram of one oxidized product of 2-MT.
acid. As the number of substitute-group on thiophene increasing,
the corresponding 1,1-dioxide becomes more stable. As a result,
1,1-dioxides of 2-MT, 2,5-DMT, and 2,3,5-TMT are detected by
the GC–MS.
As proposed formation of benzoic acid from thiophene in
Scheme 2, more H2O2 should be required in the oxidation of thio-
phene and 2-MT. As shown in Table 3, when 2.5 equivalents of
H2O2 were used in the oxidation of 2-MT, only 45.3% conversion
was obtained (Table 3, entry 3). As the equivalent of H2O2 was in-
creased to 10, the oxidation of thiophene and 2-MT could get com-
plete conversion (Table 3, entry 1 and 2). However, when 2.5
equivalents of H2O2 were used in the oxidation of BT, 2,3,5-3MT
and 2,5-2MT, over 90% conversions were achieved. Corresponding
efficiency of hydrogen peroxide utilization was around 60%.
One major product and some minor products were obtained
from the oxidation of 2,5-DMT and 2,3,5-TMT (Figs. S3 and S4 in
Supporting information). The MS of main product shows the
molecular ion peak at m/z 144.0 for 2,5-DMT and at m/z 158.0
for 2,3,5-TMT, which are assigned to corresponding 1,1-dioxides.
Thus, 2,5-DMT 1,1-dioxides and 2,3,5-TMT 1,1-dioxides are the
main oxidative products from 2,5-DMT and 2,3,5-TMT. The oxida-
tive product of BT is BT-sulfone which can be isolated by flash
chromatography on silica gel and characterized by 1H NMR and
IR (Fig. S5 in Supporting information).
3.6. Obtaining of ILs-free desulfurized oil and recycling of the catalyst
containing [Hnmp]BF4 and ammonium tungstate
During the course of desulfurization using BAILs [Hnmp]BF4, the
unwanted ‘‘contamination’’ of the desulfurized oil with [Hnmp]BF4
was investigated. After the oxidation, ionic liquid phase was sepa-
rated and the nitrogen content in oil phase is less than 50 ppm
through elementary analysis. We could obtain the ILs-free model
oil by washing once with H2O.
From above results, a summary of oxidative products of thio-
phene and its derivates is given in Scheme 2. It is clear that thio-
phene 1,1-dioxide is unstable and then the consequent reactions
are followed, yielding benzaldehyde, benzoic acid, and sulfuric
The recycling test of the catalyst containing [Hnmp]BF4 and
ammonium tungstate was also investigated. When the reaction
was finished, ionic liquid phase was separated and washed with
ethyl ether. After removing H2O from ionic liquid phase under

B. Zhang et al. / Journal of Catalysis 287 (2012) 5–12
9
Scheme 2. The oxidized products of thiophene and its derivatives and proposed formation process of Benzoic acid.
reduced pressure, fresh H2O2 and model oil were added to ionic li-
3.8. Mechanistic studies
quid phase for the next cycle of the reaction. Table 4 shows no sig-
nificant deterioration of activity for the recovered catalyst even
after seven times of use.
3.8.1. IR studies on interaction between thiophene and BAILs
[Hnmp]BF4
Infrared spectroscopy has been used to determine the interac-
tion between ILs and organic compound [25]. Fig. 5a shows the
IR spectra in the range from 650 to 850 cmÀ1 of [Hnmp]BF4, thio-
phene and the mixture of them at molar ratio of 2. The IR results
indicate that the structure of thiophene is distorted according to
the appearance of a new peak at 735 cmÀ1 and the weakening
one at 713 cmÀ1 which is assigned to @CAH bending vibrations
of thiophene [26], when BAILs are mixed with thiophene. The rel-
ative intensity of the new peak appeared at 735 cmÀ1 for different
BAILs containing thiophene is found to be in accord with the activ-
ity in ODS of thiophene (Table 1, Fig. 5b). The thiophene shows the
highest reactivity in the oxidation catalyzed by ammonium tung-
state in [Hnmp]BF4; accordingly, the new peak intensity of
[Hnmp]BF4 is the strongest one among various BAILs. This means
that the structure of thiophene is distorted in [Hnmp]BF4 stronger
than in others. This might be attributed to the acidity of BAILs and
the polarization of thiophene by fluorine ion [27].
3.7. The selectivity in the oxidation of thiophene catalyzed by
ammonium tungstate in the presence of [Hnmp]BF4
We developed a multi-component system to oxidize electron-
poor aromatic compound including thiophene, BT, and its deri-
vates. As the important constituents of gasoline, olefins and
aromatics were introduced to the reaction mixture to clarify the
effects for oxidation of thiophene. If 30 wt.% toluene was added
to the reaction mixture, less influence to desulfurization was ob-
served and 98% conversion of thiophene was obtained. However,
when 30 wt.% cyclohexene was added to the reaction system, the
removal of thiophene was sharply decreased and only 25% conver-
sion was obtained. We thought that cyclohexene is easier to be oxi-
dized than thiophene. Therefore, the selectivity in the oxidation of
thiophene in the presence of olefin is more challenging and the ef-
fort on this aspect is under processing.

10
B. Zhang et al. / Journal of Catalysis 287 (2012) 5–12
Table 3
Catalytic oxidation of various sulfur-containing compounds with aqueous 30% H2O2 under dual activation conditions
a
.
Entry
1c
Substrate
Time (h)
3
Conv. (%)b
95.4
Main producte
Yield (%)
–
H2O2 eff. (%)f
–
COOH
S
S
S
S
2c
3d
4c
3
3
3
98.0
45.3
98.7
–
–
COOH
COOH
–
–
88.3
17.6
S
O
O
O
O
5d
6d
3
3
93.5
92.5
87.6
90.3
58.4
60.2
S
S
S
S
S
O
O
S
7d
0.5
98.2
96.5
64.3
O
O
a
Experimental conditions: T = 60 °C, 20 mL model oil of thiophene in n-octane with S-contents 700 ng l
LÀ1, [Hnmp]BF4 = 4 g, S/C = 25 (molar ratio,
based on W), catalyst: ammonium tungstates.
b
Conversion was determined by GC-SCD with an external standard technique.
10 equivalent H2O2 was used.
2.5 equivalent H2O2 was used.
Main product was determined by GC–MS.
c
d
e
f
H2O2 efficiency was calculated as mole of main product produced per mole of H2O2 consumed.
3.8.2. 1H NMR studies on the interaction between thiophene and ionic
liquids
Table 4
The 1H NMR spectra of ILs containing thiophene were recorded
to further understand the interaction nature (Fig. 6). The 1H NMR
of the mixtures were examined in a stem coaxial NMR capillary
tube inserted into 5 mm NMR tube loaded with D2O. When thio-
phene is added into [Bmim]BF4, the resonances for all protons of
ILs significantly shift upfield because of the aromatic ring current
effects [28,29]. However, in the BAILs [Hnmp]BF4, the adding of
thiophene makes resonance of the acidic proton shift a little bit
downfield, and the resonances of other protons still shift upfield.
Recycling test of the catalyst containing [Hnmp]BF4 and ammonium tungstate.
Cycle
1
2
3
4
5
6
7
Thiophene removal (%)
99.4
99.3
99.0
98.8
98.4
98.0
97.5
Experimental conditions: T = 60 °C, t = 5 h, 20 mL model oil of thiophene in n-octane
with S-contents 700 ng
l , [Hnmp]BF4 = 4 g, thiophene/W = 25, H2O2/thio-
LÀ1
phene = 10 (molar ratio).
Fig. 5. FT-IR spectra of (a) thiophene, [Hnmp]BF4 and the mixture of them; (b) thiophene mixed with [Hnmp]BF4, [Hnmp]HSO4 or [Hmim]BF4.

B. Zhang et al. / Journal of Catalysis 287 (2012) 5–12
11
Fig. 6. 1H NMR signals recorded for thiophene and the mixture of thiophene with [Bmim]BF4 or [Hnmp]BF4.
3.8.3. Cyclic voltammetry measurement on the interaction between
thiophene and BAILs [Hnmp]BF4
To obtain the experimental evidences for the existence of
hydrogen bonding that may activate thiophene and promote the
oxidation, the cyclic voltammograms (CVA) of [Bmim]BF4,
[Hnmp]BF4, thiophene in ILs [Bmim]BF4 and in BAILs [Hnmp]BF4
were performed according to the literature [33]. The electrochem-
ical windows of [Bmim]BF4 and [Hnmp]BF4 are found to be 3.8 V
(from À1.8 to 2.0 V vs SCE) and 2.7 V (from À1.0 to 1.7 V vs SCE),
respectively. The first electrochemical oxidation peak of thiophene
occurs within the electrochemical window for thiophene in both
ionic liquids. The first oxidation peak is assigned to an electron loss
from the two highest occupied p-orbitals of six p electrons in aro-
matic system of thiophene [34,35]. Thus, it is considered as a mea-
surement of the conjugation extent of thiophene. Fig. 7 shows the
CVA curves of thiophene in [Bmim]BF4 and [Hnmp]BF4. The first
oxidation peak of thiophene in [Hnmp]BF4 occurs at 1.15 V ahead
of that in [Bmim]BF4. This indicates that thiophene is activated
by forming hydrogen bonding with [Hnmp]BF4. The aromaticity
of thiophene is partially destroyed by the hydrogen bonding, and
thus the oxidation takes place easily in the presence of [Hnmp]BF4.
Based on the above mechanistic study, a plausible dual active
model [36,37] was proposed (Scheme 3). [Hnmp]BF4 activates thi-
ophene through hydrogen bonding and polarization to destroy the
aromaticity of thiophene. Both of activations work synergistically
to make the oxidation of thiophene possible.
Fig. 7. Cyclic voltammograms curves of thiophene in [Bmim]BF4 and [Hnmp]BF4
(Thiophene in ILs with S-contents 700 ng
LÀ1).
l
These results clearly show the existence of strong hydrogen bond-
ing [30–32] between [Hnmp]BF4 and thiophene.
Scheme 3. Dual-activation model for the oxidation of thiophene on catalyst containing [Hnmp]BF4 and ammonium tungstate.

12
B. Zhang et al. / Journal of Catalysis 287 (2012) 5–12
4. Summaries
References
[1] C.S. Song, Catal. Today 86 (2003) 211.
The main results of this work can be summarized as follows:
[2] A. Stanislaus, A. Marafi, M.S. Rana, Catal. Today 153 (2010) 1.
[3] P.S. Kulkarni, C.A.M. Afonso, Green Chem. 12 (2010) 1139.
[4] C. Li, Z. Jiang, J. Gao, Y. Yang, S. Wang, F. Tian, F. Sun, X. Sun, P. Ying, C. Han,
Chem. Eur. J. 10 (2004) 2277.
[5] H. Lv, J. Gao, Z. Jiang, F. Jing, Y. Yang, G. Wang, C. Li, J. Catal. 239 (2006) 369.
[6] W. Zhu, H. Li, X. Jiang, Y. Yan, J. Lu, J. Xia, Energy Fuels 21 (2007) 2514.
[7] W. Zhu, H. Li, X. Jiang, Y. Yan, J. Lu, L. He, J. Xia, Green Chem. 10 (2008) 641.
[8] L. He, H. Li, W. Zhu, J. Guo, X. Jiang, J. Lu, Y. Yan, Ind. Eng. Chem. Res. 47 (2008)
6890.
[9] H. Li, L. He, J. Lu, W. Zhu, X. Jiang, Y. Wang, Y. Yan, Energy Fuels 23 (2009) 1354.
[10] D. Xu, W. Zhu, H. Li, J. Zhang, F. Zou, H. Shi, Y. Yan, Energy Fuels 23 (2009)
5929.
[11] H. Li, W. Zhu, Y. Wang, J. Zhang, J. Lu, Y. Yan, Green Chem. 11 (2009) 810.
[12] J. Zhang, W. Zhu, H. Li, W. Jiang, Y. Jiang, W. Huang, Y. Yan, Green Chem. 11
(2009) 1801.
1. ODS catalyst composed of tungstate compounds and BAILs can
effectively catalyze the oxidation of BT, 2,3,5-TMT, 2,5-DMT, 3-
MT, 2-MT, and thiophene using H2O2 as the oxidant under mild
condition. The catalytic oxidation reactivity of the sulfur-con-
taining compounds is in the following order: thiophene < 2-
MT < 3-MT < 2,5-DMT < 2,3,5-TMT < BT.
2. The TOF more than 194 Â 10À3 sÀ1 was observed for ODS of BT
using catalyst containing tungstate compounds and BAIL
[Hnmp]BF4 in comparison with less than 11 Â 10À3 sÀ1 using
catalyst containing tungstate compounds and non-BAIL
[Bmim]BF4.
[13] D. Zhao, J. Wang, E. Zhou, Green Chem. 9 (2007) 1219.
[14] L. Lu, S. Cheng, J. Gao, G. Gao, M. He, Energy Fuels 21 (2007) 383.
[15] S. Otsuki, T. Nonaka, N. Takashima, W. Qian, A. Ishihara, T. Imai, T. Kabe, Energy
Fuels 14 (2000) 1232.
3. The sulfur atom of thiophene is oxidized to SO24À and the hydro-
carbon part of thiophene is to oxygenates.
[16] J. Nakayama, H. Nagasawa, Y. Sugihara, A. Ishii, J. Am. Chem. Soc. 119 (1997)
9077.
[17] S. Rozen, Y. Bareket, J. Org. Chem. 62 (1997) 1457.
[18] A. Treiber, J. Org. Chem. 67 (2002) 7261.
4. The ODS catalyst composed of [Hnmp]BF4 and ammonium
tungstate can be recycled seven times without significant loss
in activity.
5. Based on FT-IR, 1H NMR and electrochemical measurements,
the hydrogen bonding between thiophene and [Hnmp]BF4 and
the polaraztion induced by ILs distorte the planar structure of
thiophene, and decrease the aromaticity extent. This distortion
together with the catalytic oxidation function of tungstate syn-
ergistically activates the thiophene and makes the ODS of thio-
phene efficiently.
6. The oxidation of thiophene, usually difficult, can be efficiently
realized by using the catalyst containing tungstate compound
and BAILs which play a triple role in ODS as extractant, reaction
media and more importantly, as a catalyst. For [Hnmp]BF4 and
ammonium tungstate, up to 99% conversion can be achieved at
60 °C.
[19] L. Kong, G. Li, X. Wang, Catal. Today 93 (2004) 341.
[20] A. Chica, A. Corma, M.E. Dómine, J. Catal. 242 (2006) 299.
[21] A. Hajipour, A. Rajaei, A. Ruoho, Tetrahedron Lett. 50 (2009) 708.
[22] A. Hajipour, F. Rafiee, A. Ruoho, Synlett 7 (2007) 1118.
[23] L. Zhang, M. Xian, Y. He, L. Li, J. Yang, S. Yu, X. Xu, Bioresour. Technol. 100
(2009) 4368.
[24] K. Kamata, K. Yamaguchi, N. Mizuno, Chem. Eur. J. 10 (2004) 4728.
[25] J. Wu, N. Li, K.A. Li, F. Liu, J. Phys. Chem. B 112 (2008) 8134.
[26] M. Rico, J.M. Orza, J. Morcillo, Spectrochim. Acta 21 (1965) 689.
[27] A. Bongini, G. Barbarella, M. Zambianchi, C. Arbizzani, M. Mastragostino, Chem.
Commun. (2000) 439.
[28] B. Su, S. Zhang, Z. Zhang, J. Phys. Chem. B 108 (2004) 19510.
[29] A.D. Headley, N.M. Jackson, J. Phys. Org. Chem. 15 (2002) 52.
[30] M.H.F. Kox, K.F. Domke, J.P.R. Day, G. Rago, E. Stavitski, M. Bonn, B.M.
Weckhuysen, Angew. Chem. Int. Ed. 48 (2009) 8990.
[31] S.Y. Yu, J. Garcia-Martinez, W. Li, G.D. Meitzner, E. Iglesia, Phys. Chem. Chem.
Phys. 4 (2002) 1241.
[32] C.L. Garcia, J.A. Lercher, J. Phys. Chem. 96 (1992) 2669.
[33] M. Nefedieva, O. Lebedeva, D. Kultin, L. Kustov, S. Borisenkova, V. Krasovskiy,
Green Chem. 12 (2010) 346.
Acknowledgments
[34] A.D. Baker, D. Betteridge, N.R. Kemp, R.E. Kirby, Anal. Chem. 42 (1970)
1064.
[35] F. Pak, D. Ekinci, F. Tümer, Ü. Demir, Macromol. Chem. Phys. 208 (2007)
2367.
[36] H. Liu, T. Jiang, B. Han, S. Liang, Y. Zhou, Science 326 (2009) 1250.
[37] R.R. Deshmukh, J.W. Lee, U.S. Shin, J.Y. Lee, C.E. Song, Angew. Chem. Int. Ed. 47
(2008) 8615.
This work is financially supported by Natural Science Founda-
tion of China (NSFC Grant No. 20921092, NSFC Grant No.
20673114) and the State Key Project of China (No. 2006CB
202506). We thank Dr. Shengmei Lu and Miss Lina Norberg Samu-
elsson for helpful discussion.
Appendix A. Supplementary material
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.jcat.2011.11.003.