Peroxopolyoxometalate-based room temperature ionic liquid as a self-separation catalyst for epoxidation of olefins

Article abstract of DOI:10.1016/j.catcom.2009.11.025

A new peroxopolyoxometalate-based room temperature ionic liquid (POM-RTIL) has been synthesized and used as catalyst for efficient epoxidation of various olefins. The RTIL catalyst was found to be well dissolved in the solvent (ethyl acetate) during the reaction, while it can self-separate from the reaction media at room temperature after the reaction completed, which made the recovery and reuse of the IL catalyst convenient. The POM-RTIL catalyst can be recycled for five times without significant loss of activity.

Full text of DOI:10.1016/j.catcom.2009.11.025

Catalysis Communications 11 (2010) 470–475
Contents lists available at ScienceDirect
Catalysis Communications
journal homepage: www.elsevier.com/locate/catcom
Peroxopolyoxometalate-based room temperature ionic liquid as
a self-separation catalyst for epoxidation of olefins
*
Huan Li, Zhenshan Hou , Yunxiang Qiao, Bo Feng, Yu Hu, Xiangrui Wang, Xiuge Zhao
Key Laboratory for Advanced Materials, Research Institute of Industrial Catalysis, East China University of Science and Technology, Shanghai 200237, People’s Republic of China
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 29 July 2009
Received in revised form 16 November 2009
Accepted 30 November 2009
Available online 6 December 2009
A new peroxopolyoxometalate-based room temperature ionic liquid (POM-RTIL) has been synthesized
and used as catalyst for efficient epoxidation of various olefins. The RTIL catalyst was found to be well
dissolved in the solvent (ethyl acetate) during the reaction, while it can self-separate from the reaction
media at room temperature after the reaction completed, which made the recovery and reuse of the IL
catalyst convenient. The POM-RTIL catalyst can be recycled for five times without significant loss of
activity.
Keywords:
Ó 2009 Elsevier B.V. All rights reserved.
Polyoxometalate
Epoxidation
Olefin
Ionic liquid
Self-separation
1. Introduction
successful alternative solvents for homogeneous biphasic catalysis
have been designed for immobilizing catalysts, facilitating product
Epoxidation of olefins is an important reaction in laboratory as
well as in the chemical industry, because epoxides are widely used
for manufacturing epoxy resins, surfactants, and intermediates,
etc. in organic synthesis [1]. H2O2-based catalytic epoxidation
has received much attention from an economic and environmental
point of view [2]. It has been known that the tungsten-based
catalysts like polyoxometalate (POM) show high efficiency of
hydrogen peroxide utilization and high selectivity towards epox-
ides [3,4]. Among different polyoxometalates, the classic Ishii-Ven-
turello system [5,6] has attracted huge interests and has been
investigated extensively by different groups. The Venturello anion
({PO4[WO(O2)2]4}3À) was found to be the most important active
species in Ishii-Venturello system [7–11].
As a highly active oxidation species, Venturello anion was com-
bined with quaternary ammonium in conjunction with H2O2 as
primary oxidant for epoxidation of various olefins under mild con-
ditions [6,12,13]. Venturello anion has also been immobilized on
macroreticular Amberlite IRA-900 (exchange resin) [14], silicate
xerogel [15] and dendrimer [16–18] to facilitate the recycling of
catalyst. When coupled with 1-butyl-3-methylimidazolium cation,
the Venturello anion can be immobilized in ionic liquid for the oxi-
dation of secondary alcohols [19].
isolation and recycling the catalyst. The first POM-based RTIL was
reported by Bourlinos et al. [20]. In their communication they
showed the possibility to yield a liquid derivative of POM by partial
exchange of the protons of the POM with a PEG-containing quater-
nary ammonium. This new IL showed similar properties compared
with other conventional ILs, such as low melting point, comparable
viscosity and thermal stability. Then Rickert et al. reported that the
combination of a Keggin or Lindqvist anion with tetraalkylphospho-
nium cation can also yield POM-RTILs [21,22]. Though many
Venturello anion-based catalytic systems have been reported, as
far as we know, no Venturello anion-based RTIL is known and
utilized in catalysis yet.
In this work, we present two novel POM-RTILs prepared by pair-
ing alkylimidazolium cation with the well-defined Venturello an-
ion. The RTILs were highly active for the epoxidation of various
olefins and proved to be self-separation catalysts, which led to easy
separation of the products from the IL catalysts via simple
decantation.
2. Experimental
Room temperature ionic liquids (RTILs) are ionic liquids (ILs) that
are liquid at room temperatureor slightly higher (20–30 °C). RTILs as
2.1. General remarks
All manipulations involving air-sensitive materials were carried
out using standard Schlenk line techniques under an atmosphere of
nitrogen. All solvents (A.R. grade) were dried with the standard
* Corresponding author. Tel.: +86 21 64251686; fax: +86 21 64253703.
E-mail address: houzhenshan@ecust.edu.cn (Z. Hou).
1566-7367/$ - see front matter Ó 2009 Elsevier B.V. All rights reserved.
doi:10.1016/j.catcom.2009.11.025

H. Li et al. / Catalysis Communications 11 (2010) 470–475
471
methods. Triethylamine, N-butylimidazole and thionyl chloride
(Sinopharm Chemical Reagent Co., Ltd, Shanghai, A.R. grade) were
distilled and PEG-300 was dried at 60 °C under vacuum for 4 h
prior to use. Imidazole (Sinopharm Chemical Reagent Co., Ltd,
Shanghai, A.R. grade) was used as received without further purifi-
cation. All NMR spectra were recorded on a Bruker Avance 500
instrument (500 MHz 1H) using CDCl3 and TMS as solvent and ref-
erence, respectively. Chemical shifts (d) were given in parts per
million and coupling constants (J) in Hertz. The products were ana-
lyzed by GC and GC–MS equipped with a HP-5MS column (30 m
long, 0.25 mm i.d.).
6H, CH3). 13C NMR(125 MHz, CDCl3) 136.7, 123.3, 121.6, 70.2-
69.9, 68.8, 65.5, 49.6, 29.2, 14.9, 13.8.
Correspondingly, the ionic liquid catalyst (5b) was prepared
conveniently according to the following steps. After the solution
of H3PW12O40 (1.6 g, 0.6 mmol) in 12 ml of 30% H2O2 was stirred
for 30 min, 4b (0.7 g, 0.9 mmol) in 10 ml of dichloromethane was
added dropwise. The resulting biphasic mixture was thoroughly
stirred for another 30 min. Then the organic layer was separated
and dried under anhydrous air flow at 40 °C for 4 h to give an al-
most colorless transparent viscous ionic liquid catalyst (5b). IR
(cmÀ1): 3141, 3106, 2924 and 2854 (CH); 1736 and 1563 (CC and
CN); 1465 and 1350 (CH); 1162 and 1105 (CO); 844 (OO); 523
(WOO) (Fig. 2). The bands resulted from PAO and W@O were over-
lapped by the characteristic bands of PEG chain ranged from 900 to
1200 cmÀ1. As shown in Fig. 3, 31P NMR (CDCl3) of IL 5b gave
mainly five signals at different chemical shift ranged from À10 to
10 ppm, which were ascribed to the ‘‘aged” Venturello anion since
it was very liable to change. This phenomenon was found by Hill’s
group [11] in the in-depth study of the Venturello complex and
also observed by Kozhevnikov’s group [13] in the epoxidation of
oleic acid catalyzed by Q3{PO4[WO(O2)2]4} (Q = quaternary
ammonium).
2.2. Synthesis of ionic liquid catalyst (5a, 5b)
The synthesis of the ionic liquid catalyst was shown in Fig. 1. 5b
was taken as an example. Firstly, N-dodecylimidazole (3b, 1H-NMR
(500 MHz, CDCl3) d = 7.45(s, 1H, CH), 7.02(s, 1H, CH), 6.88(s, 1H,
CH), 3.90(t, 2H, CH2), 1.74(m, 2H, CH2), 1.27(m, 18H, CH2), 0.86(t,
3H, CH3)) and PEG dichlorides (IR cmÀ1: 2880, 1963, 1672, 1468,
1279, 1113, 946, 842, 745, 665(CACl), 531) were synthesized
according to the previously reported procedures, respectively
[23,24].
The PEG chain-functionalized N-dodecylimidazolium dichlo-
rides (4b) were synthesized by the following steps. 1.5 mmol chlo-
rinated PEG-300 and 4 mmol 3b were charged into a steel vessel,
and allowed to react at 90 °C for 48 h under 1.0 MPa nitrogen with
magnetically stirring. Then the product (4b) was washed with
diethyl ether for four times and treated under vacuum to give
the product. 1H-NMR (500 MHz, CDCl3) d = 10.68(s, 2H, CH),
7.64(s, 2H, CH), 7.18(s, 2H, CH), 4.60(t, 4H, CH2), 4.21(t, 4H, CH2),
3.30-3.80(m, PEG), 1.75(m, 4H, CH2), 1.28(m, 36H, CH2), 0.86(t,
2.3. Typical procedure for epoxidation of cyclooctene by using RTIL
catalysts
For the reaction in the absence of solvent, IL catalyst (50 mg)
was charged into a 50 ml steel vessel, followed by adding 3 mmol
H2O2 and 3 mmol cyclooctene. Then the vessel was heated to the
required temperature with magnetically stirring for a certain time.
After the reaction, the product was extracted with cyclohexane for
Fig. 1. Typical preparation procedure of the ionic liquid catalysts (5a, 5b). Reagents and conditions: (a) Triethylamine, CH2Cl2, 0 °C, 24 h. (b) DMF, NaH. (c) Benzene, pyridine,
reflux, 10 h. (d) 90 °C, 48 h. (e) CH2Cl2, H2O2.

472
H. Li et al. / Catalysis Communications 11 (2010) 470–475
Fig. 2. The FT-IR spectrum of RTIL 5b.
Fig. 3. The 31P spectrum of isolated 5b in CDC13.
3 times for the GC analysis. The catalyst remained in the steel ves-
sel was treated under vacuum at 60°C for 2 h, and then reused for
the next cycle.
For the reaction in the presence of organic solvent (5 ml), the
similar procedure was adopted except that 3 ml cyclohexane was
added to the reaction mixture to facilitate the catalyst separation
after the reaction. Then solvent phase was removed, and the cata-
lyst was washed with cyclohexane and further treated under vac-
uum for the next run.
(Fig. 1). When the PEG functionalized N-hexyl imidazolium dichlo-
rides (4a) were used instead of 4b, Venturello anion-based RTIL
(5a) can also be separated from the organic layer. But the PEG chain
functionalized N-butyl imidazolium dichlorides were proved to be
too hydrophilic to transfer the Venturello anion into the organic
phase. Accordingly, no Venturello anion-based RTIL was obtained
when organic solvent was removed.
The ILs 5a and 5b were first employed to epoxidize cyclooctene
in the absence of solvent. As shown in Table 1, IL catalyst 5a with a
shorter alkyl chain showed lower activity than that of 5b at 20 °C
(Entries 1 and 2, Table 1). But at 60 °C, a nearly full conversion of
cyclooctene was achieved by using both of catalysts after 4 h (En-
tries 3 and 4, Table 1). The reaction was obviously accelerated at
higher temperature (Entries 1 and 3, Table 1). Because the IL cata-
lysts (5a, 5b) were immiscible with cyclohexane, the extraction of
products and the separation of IL catalyst were achieved conve-
niently by adding cyclohexane. After extraction, the catalyst was
treated under vacuum for the next run. It can be reused for another
four times without obvious loss of activity (Entries 4–8, Table 1).
3. Results and discussion
Fig. 1 showed the synthetic route of the ionic liquid catalysts
(5a and 5b). The reaction of N-dodecylimidazole with PEG dichlo-
rides afforded dicationic dichlorides bridged by polyether linkage
chains (4b) in high yield. Venturello anion was obtained by the
degradation of H3PW12O40 in excess of H2O2 [11]. Then the IL 5b
was obtained by ion-exchange in aqueous-organic biphasic system

H. Li et al. / Catalysis Communications 11 (2010) 470–475
473
Table 1
from the acidic hydrolysis of the epoxide ring. Although chlorocar-
bons have been widely used as reaction media for the epoxidation,
their high toxicity largely defeated the environmental advantages
of using the green oxidant H2O2 Therefore, the application of envi-
ronmentally benign solvents in epoxidation would be highly
requirable. As a result, several different organic solvents have been
screened by choosing the epoxidation of cyclooctene as a model
reaction. As shown in Table 2, the cyclooctene was transformed
into the corresponding epoxide efficiently in most selected organic
solvents (Entries 1–3, Table 2). The lowest activity was obtained
with cyclohexane as reaction media (Entry 4, Table 2), which can
be explained by the negligible solubility of IL 5b in nonpolar cyclo-
hexane. Because ethyl acetate is a low toxic solvent, it was sequen-
tially chosen as reaction media for all the reactions.
Although the molar ratio of substrate to H2O2 had little influ-
ence on the selectivity of cyclooctene epoxidation (Entries 3 and
5, Table 2), it affected that of cyclohexene significantly. When the
molar ratio of substrate to H2O2 was kept at 3:1, the oxidation of
cyclohexene can afford epoxide in high selectivity (97%). However,
the selectivity to cyclohexene oxide decreased to 75% when the
molar ratio of substrate to H2O2 was lower (1:1) (Entries 6 and 7,
Table 2).
Epoxidation of olefins without solvent.a
Entry Substrates
IL
Temperature Concentration Selectivity
catalyst (°C)
(%)
(%)
1
2
3
4
5
6
7
8
9
10
Cyclooctene
Cyclooctene
Cyclooctene
cyclooctene
2nd Run
3rd Run
4th Run
5th Run
Styrene
5a
5b
5a
5b
5b
5b
5b
5b
5b
20
20
60
60
60
60
60
60
60
60
25
60
96
95
91
92
90
89
94
89
99
99
99
99
99
99
99
99
86b
23c
Cyclohexene 5b
a
Fifty milligram IL catalyst, 3 mmol substrate, 3 mmol H2O2 Reaction time: 4 h.
The main products were benzyl aldehyde and benzoic acid. The selectivity to
b
styrene oxide was less than 2%.
c
The 1,2-cyclohexanediol and 2-hydroxy cyclohexanone were detected as main
by-products.
When styrene or cyclohexene was used as substrate in the ab-
sence of solvent, although high conversions were achieved in both
cases, the selectivities to the epoxide were always poor due to the
acidic hydrolysis of the epoxide ring by direct contact with the acid
aqueous phase where acidity likely resulted from the degradation
of the anion [13] (Entries 9 and 10, Table 1).
In the next step, we examined the reaction kinetics over cata-
lysts 5a and 5b in ethyl acetate. The time profile of cyclooctene
conversion is given in Fig. 4, which showed that catalyst 5b with
a longer alkyl chain was more efficient than catalyst 5a. According
to the mechanism proposed by Ishii [25], in the epoxidation of ole-
In a conventional Ishii-Venturello system, an organic phase is
always introduced into the reaction system to protect the epoxides
Table 2
Epoxidation of olefins by using IL 5b in different organic solvents.a
Entry
Substrates
Substrate/H2O2 (molar ratio)
Solvents
Concentration (%)
Selectivity (%)
1
2
3
4
5
6
7
Cyclooctene
Cyclooctene
Cyclooctene
Cyclooctene
Cyclooctene
Cyclohexene
Cyclohexene
3:1
3:1
3:1
3:1
1:1
3:1
1:1
Dichloromethane
Benzene
98
96
97
45
91
90
86
99
99
99
99
99
97
75
Ethyl acetate
Cyclohexane
Ethyl acetate
Ethyl acetate
Ethyl acetate
a
Fifty milligram IL 5b, 3 mmol substrate, 5 ml solvent. Reaction conditions: 60 °C, 4 h. Conversion based on the H2O2 charged.
Fig. 4. Time profile of epoxidation of cyclooctene catalyzed by IL catalysts 5a and 5b. Reaction conditions: 50 mg IL, 3 mmol cyclooctene, 1 mmol H2O2, 5 ml ethyl acetate,
60°C. The selectivity to cyclooctene oxide was always over 99%.

474
H. Li et al. / Catalysis Communications 11 (2010) 470–475
Table 3
Epoxidation of various olefins by using catalyst 5a and 5b.a
b
Entry
Substrate
Catalyst
Temperature (°C)
Time (h)
Concentration (%)
Selectivity (%)
1
2
3
4
5
6
7
8
Cyclooctene
2nd Run
3rd Run
4th Run
5th Run
5b
5b
5b
5b
5b
5a
5b
5a
5b
5a
5b
5b
60
60
60
60
60
60
60
60
60
70
70
40
4
4
4
4
4
4
4
4
4
6
6
4
97
97
96
91
92
45
93
42
90
30
87
85
99
99
99
99
99
77c
92
85d
97
97
98
74e
Styrene
Styrene
Cyclohexene
Cyclohexene
1-Hexene
1-Hexene
Geraniol
9
10
11
12
a
b
c
Fifty milligram IL, 3 mmol substrate, 1 mmol H2O2, 5 ml ethyl acetate.
Conversion based on the H2O2 charged.
The benzaldehyde and 1-phenyl-1,2-ethanediol were detected as main byproducts.
The cyclohexanediol and 2-hydroxy cyclohexanone were detected as main byproducts.
Selectivity to 2,3-epoxy alcohol.
d
e
fins involving a phase transfer reagent and PW12OÀ403, the formed
peroxo species needed to be transferred into the organic phase
(ethyl acetate in our case) where the reaction took place. The more
hydrophobic 5b can transfer the active species to the organic phase
more efficiently and then gave higher catalytic activity.
The scope of the substrates was also examined. As shown in Ta-
ble 3, styrene, cyclohexene, 1-hexene and an allylic alcohol (gera-
niol) were all converted into the corresponding epoxides efficiently
by using IL 5b in the H2O2/ethyl acetate system (Entries 7, 9, 11 and
12, Table 3). 5a was found to be much less active than 5b (Entries 6
vs 7; 8 vs 9; 10 vs 11, Table 3). Furthermore, lower selectivity to
epoxide was always obtained by using the IL catalyst with shorter
alkyl chain IL (5a). This can be explained by the fact that more
hydrophilic 5a was more distributed in aqueous phase and then re-
sulted in the further hydrolysis of epoxide product in acidic aque-
ous phase during the reaction. As a result, the length of alkyl chains
attached to imidazolium cation affected not only the activity but
also the selectivity.
transformed to epoxides efficiently with high selectivity using a
relatively low toxic ethyl acetate as reaction media. The RTIL can
be regarded as a self-separation catalyst and reused for five times
without significant loss of activity. It is certain that many more cat-
alytically active POM-based RTILs need to be identified. The efforts
to explore them are underway in our laboratory.
Acknowledgement
The authors are grateful for the support from the National Nat-
ural Science Foundation of China (No. 20773037), ECUST (No.
YJ0142136), and the Commission of Science and Technology of
Shanghai Municipality (No. 07PJ14023), China. The authors would
also like to express their thanks to Prof. Walter Leitner’s helpful
suggestions and Dr. Nils Theyssen’s experimental assistance.
References
IL catalyst 5b was immiscible with cyclohexane and water, and
the solubility of IL 5b in ethyl acetate was found to be less than 1%
at room temperature. However, when 1 mmol of H2O2 was added,
5b (50 mg) can be dissolved well in ethyl acetate phase (5 ml) at
elevated temperature (40–60 °C). At the end of the reaction, when
the mixture continued on cooling to room temperature, most of IL
5b sank to the bottom of the vessel and leaching of the ionic liquid
into ethyl acetate was very low. Moreover, the addition of cyclo-
hexane to the reaction solution allowed IL 5b to separate further
from organic phase more rapidly. The above observation offered
an effective approach to separate the catalyst easily by adding
cyclohexane at room temperature after reaction. As shown in Table
3, IL 5b can be employed as self-separable catalyst for five consec-
utive catalytic cycles in the epoxidation of cyclooctene with a
slight decrease in conversion from 97% to 92% (Entries 1–5, Table
3), which might be indicative of the slight deactivation of the
recovered catalyst.
[1] K. Weissermel, H.J. Arpe, Industrial Organic Chemistry, third ed., VCH,
Weinheim, 1997.
[2] N. Mizuno, K. Yamaguchi, K. Kamata, Coord. Chem. Rev. 249 (2005) 1944–
1956.
[3] R. Neumman, in: J.J. Borrás-Almenar, E. Coronada, A. Müller, M.T. Pope (Eds.),
Polyoxometalate Molecular Science. NATO SCIENCE SERIES: II: Mathematics,
Physics and Chemistry, vol. 98, Kluwer Academic Publishers, Dordrecht, 2003,
pp. 327–349.
[4] I.V. Kozhevnikov, Catalysis by Polyoxometalate, John Wiley & Sons, Chichester,
UK., 2002.
[5] Y. Ishii, K. Yamawaki, T. Yoshida, T. Ura, M. Ogawa, J. Org. Chem. 52 (1987)
1868–1870.
[6] C. Venturello, E. Alneri, M. Ricci, J. Org. Chem. 48 (1983) 3831–3833.
[7] S. Sakaue, T. Tsubakino, Y. Nishiyama, Y. Ishii, J. Org. Chem. 58 (1993) 3633–
3638.
[8] C. Aubry, G. Chottard, N. Platzer, J.M. Brkgeault, R. Thouvenot, F. Chauveau, C.
Huet, H. Ledon, Inorg. Chem. 30 (1991) 4409–4415.
[9] A.C. Dengel, W.P. Griffith, B.C. Parkin, J. Chem. Soc. Dalton Trans. (1993) 2683–
2688.
[10] L. Salles, C. Aubry, F. Robert, G. Chottard, R. Thouvenot, H. Ledon, J.M.
Bregeault, New J. Chem. 17 (1993) 367–375.
[11] D.C. Duncan, R.C. Chambers, E. Hecht, C.L. Hill, J. Am. Chem. Soc. 117 (1995)
681–691.
[12] C. Venturello, R. D’Aloiso, J. Org. Chem. 53 (1988) 1553–1557.
[13] I.V. Kozhevnikov, G.P. Mulder, M.C. Steverink-de Zoete, M.G. Oostwal, J. Mol.
Catal. 134 (1998) 223–228.
[14] P.A.L. Villa, B.F. Sels, D.E. Vos, P.A. Jacobs, J. Org. Chem. 64 (1999) 7267–7270.
[15] R. Neumman, H. Miller, J. Chem. Soc. Chem. Commun. (1995) 2277–2278.
[16] L. Plault, A. Hauseler, S. Nlate, D. Astruc, J. Ruiz, S. Gatard, R. Neumann, Angew.
Chem., Int. Ed. 43 (2004) 2924–2928.
Although the precise self-separation mechanism of this new
peroxopolyoxometalate-based RTIL was not very clear at this stage,
the preliminary experimental results showed that the RTILs as no-
vel environmentally friendly and recyclable catalysts were not only
possible but also quite satisfactory.
[17] M.V. Vasylyev, D. Astruc, R. Neumann, Adv. Synth. Catal. 347 (2005) 39–44.
[18] S. Nlate, L. Plault, D. Astruc, Chem.-Eur. J. 12 (2006) 903–914.
[19] S.B. Chhikara, R. Chandra, V. Tandon, J. Catal. 230 (2005) 436–439.
[20] A.B. Bourlinos, K. Raman, H. Herrera, Q. Zhang, L.A. Acher, E.P. Giannelis, J. Am.
Chem. Soc. 126 (2004) 15358–15359.
4. Conclusion
In summary, the first example of epoxidation of olefins cata-
lyzed by a POM-based RTIL was reported here. Various olefins were

H. Li et al. / Catalysis Communications 11 (2010) 470–475
475
[21] P.G. Rickert, M.R. Antonio, M.A. Firestone, K. Kubatko, T. Szreder, J.F. Wishart,
[23] W.F. Hart, M.E. Mcgreal, J. Org. Chem. 22 (1957) 86–88.
M.L. Dietz, Dalton. Trans. (2007) 529–531.
[24] V.S. Pore, N.G. Aher, M. Kumar, P.K. Shukla, Tetrahedron (2006) 11178.
[22] P.G. Rickert, M.R. Antonio, M.A. Firestone, K. Kubatko, T. Szreder, J.F. Wishart,
M.L. Dietz, J. Phys. Chem. B 111 (2007) 4685–4692.
[25] Y. Ishii, M. Ogawa, in: A. Ohno, N. Furukawa (Eds.), Reviews on Heteroatom
Chemistry, vol. 3, MY, Tokyo, 1990.