Silica-gel-confined ionic liquids: A new attempt for the development of supported nanoliquid catalysis

Article abstract of DOI:10.1002/chem.200500107

A new concept of designing and synthesizing highly dispersed ionic-liquid catalysts was developed through physical confinement or encapsulation of ionic liquids (with or without metal complex) in a silica-gel matrix through a sol-gel process. We studied ionic liquids such as EMImBF₄, BuMImBF ₄, DMImBF₄, CMImBF₄, BuMImPF₆, either with or without [Pd(PPh₃)₂Cl₂] and [Rh(PPh₃)₃Cl], in a silica-gel matrix (E = ethyl, Bu = butyl M = methyl, D = decyl, C = cetyl and Im = imidazolium). The contents of ionic liquids and loadings of Pd or Rh were 853 wt% and 0.1 - 0.15 wt%, respectively. Analyses of FT-Raman spectra showed that abnormal Raman spectra of the confined ionic liquids were observed in comparison with the bulk and pure ionic liquids. EMImBF₄ and BuMImBF₄ ionic liquids could be completely washed out from the silica-gel matrix under vigorous reflux conditions, but ionic liquids with larger molecular size, for example, DMImBF₄ or CMImBF₄, could be confined into the silica-gel nanopores relatively firmly. These results suggested that the ionic liquids were physically confined or encapsulated into the silica gel. The N₂ adsorption measurements indicated that the silica-gel skeleton was mesoporous with 50-110 A pore size after the BuMImBF₄ ionic liquid was removed completely. Transmission electron microscopy (TEM) and X-ray diffraction (XRD) analysis showed that the silicagel matrix was amorphous and nonuniformly mesoporous. Carbonylation of aniline and nitrobenzene for synthesis of diphenyl urea, carbonylation of aniline for synthesis of carbamates, and oxime transformation between cyclo-hexanone oxime and acetone were used as test reactions for these catalysts. Catalytic activities were remarkably enhanced with much lower amounts of ionic liquids needed with respect to bulk ionic-liquid catalysts or silica-supported ionic-liquid catalysts prepared with simple impregnation, in which the ionic liquid may be deposited as a thin layer on the support. Such unusual enhancement in catalytic activities may be attributed to the formation of nanoscale and high-concentration ionic liquids due to the confinement of the ionic liquid in silica gel; this results in unusual changes in the symmetry and coordination geometry of the ionic liquids.

Full text of DOI:10.1002/chem.200500107

FULL PAPER
Silica-Gel-Confined Ionic Liquids: A New Attempt for the Development of
Supported Nanoliquid Catalysis
Feng Shi, Qinghua Zhang, Dongmei Li, and Youquan Deng*[a]
Abstract: A new concept of designing
and synthesizing highly dispersed ionic-
liquid catalysts was developed through
physical confinement or encapsulation
of ionic liquids (with or without metal
complex) in a silica-gel matrix through
a sol–gel process. We studied ionic liq-
uids such as EMImBF4, BuMImBF4,
DMImBF4, CMImBF4, BuMImPF6,
either with or without [Pd(PPh3)2Cl2]
pletely washed out from the silica-gel
matrixunder vigorous refluxcondi-
tions, but ionic liquids with larger mo-
lecular size, for example, DMImBF4 or
CMImBF4, could be confined into the
silica-gel nanopores relatively firmly.
These results suggested that the ionic
liquids were physically confined or en-
capsulated into the silica gel. The N2
adsorption measurements indicated
that the silica-gel skeleton was mesopo-
rous with 50–110 Š pore size after the
BuMImBF4 ionic liquid was removed
completely. Transmission electron mi-
croscopy (TEM) and X-ray diffraction
(XRD) analysis showed that the silica-
gel matrixwas amorphous and non-
uniformly mesoporous. Carbonylation
of aniline and nitrobenzene for synthe-
sis of diphenyl urea, carbonylation of
aniline for synthesis of carbamates, and
oxime transformation between cyclo-
hexanone oxime and acetone were
used as test reactions for these cata-
lysts. Catalytic activities were remarka-
bly enhanced with much lower
amounts of ionic liquids needed with
respect to bulk ionic-liquid catalysts or
silica-supported ionic-liquid catalysts
prepared with simple impregnation, in
which the ionic liquid may be deposit-
ed as a thin layer on the support. Such
unusual enhancement in catalytic activ-
ities may be attributed to the formation
of nanoscale and high-concentration
ionic liquids due to the confinement of
the ionic liquid in silica gel; this results
in unusual changes in the symmetry
and coordination geometry of the ionic
liquids.
and [Rh(PPh3)3Cl], in
a
silica-gel
matrix(E =ethyl, Bu=butyl M=
methyl, D=decyl, C=cetyl and Im=
imidazolium). The contents of ionic liq-
uids and loadings of Pd or Rh were 8–
53 wt% and 0.1~0.15 wt%, respective-
ly. Analyses of FT-Raman spectra
showed that abnormal Raman spectra
of the confined ionic liquids were ob-
served in comparison with the bulk and
pure ionic liquids. EMImBF4 and Bu-
MImBF4 ionic liquids could be com-
Keywords: carbonylation · cataly-
sis · ionic liquids · nanotechnology ·
synthesis design
Introduction
representation of supported solid nanocatalysts would be
metal oxides or polymer-supported Au nanocatalysts, which
have exhibited high catalytic activity for CO oxidation or
carbonylations.[3–5]
Supported solid nanocatalysis has been the subject of enor-
mous interest in catalysis science and technology, because of
its potential and wide-ranging applications in the chemical
industry and environmental protection.[1,2] In previous stud-
ies and applications, supported solid nanocatalysts were ba-
sically metals, oxides, or zeolites; that is, the catalytically
active nanomaterials were all macroscopic solids. A typical
On the other hand, homogeneous or liquid-phase catalysts
offer a number of important advantages; for example, all
catalytically active sites are accessible and uniform. Usually,
solvents are indispensable as reaction media and have even
played an important role in efficient homogeneous catalysis.
Different or unexpected catalytic performance may occur if
a bulk homogeneous catalyst system, that is, a reaction
medium plus metal complexcatalyst, is highly dispersed on
the nanoscale. Nanoliquid catalyst production has not yet
been reported in the literature, probably because of difficul-
ties in obtaining stable nanoliquid catalysts due to the vola-
tility of ordinary organic solvents as reaction media for the
[a] Dr. F. Shi, Q. Zhang, D. Li, Prof. Y. Deng
Center for Green Chemistry and Catalysis
Lanzhou Institute of Chemical Physics
Chinese Academy of Sciences, Lanzhou, 730000 (China)
Fax: ( +86)931-496-8116
E-mail: ydeng@lzb.ac.cn
Chem. Eur. J. 2005, 11, 5279 – 5288
DOI: 10.1002/chem.200500107
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
5279

metal complexcatalyst. However, some studies on confined
nanoliquids have been reported recently.[6–8]
silica surface or immobilized by dipping the porous support
in the ionic liquid containing the metal complex. Chemical
bonding of the dialkyl imidazolium cation to a solid surface,
however, may limit the degrees of freedom of the dialkyl
imidazolium cation and even change the physicochemical
properties of the ionic liquids. Hence the leaching of the
ionic liquids and metal-complexcatalyst was unavoidable
under rigorous reaction conditions if the supported ionic-
liquid catalyst was prepared by physical adsorption or by
having the ionic liquid as a thin layer on a support material.
Though not new, room-temperature ionic liquids have re-
ceived much attention as being potentially environmentally
benign or as suitable reaction media for organic syntheses,
nanosolid catalysts,[9–16] and as novel liquid catalysts.[17–26]
Their favorable properties, such as thermal stability, negligi-
ble vapor pressure, peculiar ion environment, and diversity,
may provide good opportunities to develop new kinds of
supported nanoliquid catalyst systems. Here, an attempt has
been made to confine ionic liquids either with or without
metal complex, into the nanopores of solid silica gel with a
“one-pot assembly” of ionic liquid (or ionic liquid plus
metal complex) and silicate ester through a sol–gel process.
Under these conditions, a new kind of mesoporous silica-
gel-supported ionic-liquid nanocatalyst could be established,
since ionic liquids could also be used as appropriate tem-
plates to prepare nanostructured materials.[27,28] The key
concept of designing and synthesizing such a catalyst system
involves the physical confinement or encapsulation of ionic
liquid (either with or without metal complex) through a tra-
ditional sol–gel process, based on the hydrolysis of silicate
esters to produce a solid matrixwith desired pore sizes or
cavities and channels. The ionic liquid acts as both catalyst
and reaction media, and the solid matrixacts as the nano-
scale reactor connected with inlets and outlets to contain
the ionic liquid and metal complexand to allow reactants
and products to be transported in or out. The pore size of
the nanoreactor should be large enough to contain ionic liq-
uids, while the channel size of the nanoreactor should be
small enough to prevent ionic liquids or metal complexfrom
leaching; however, the channel size must also be large
enough for free transportation of reactant and product,
(Figure 1). Therefore, they are intrinsically different from
previously reported supported or immobilized ionic-liquid
catalysts,[28–30] in which the ionic-liquid fragment, such as the
dialkyl imidazolium cation, was covalently bound to the
Results and Discussion
FT-Raman characterization of silica-gel-confined ionic-
liquid catalysts: Since the concentration of metal complexin
the silica-gel-confined ionic-liquid catalysts was too low to
be detected in the FT-Raman spectrometer, only the sam-
ples of silica-gel-confined ionic liquid without the metal
complexare shown here. BuMImBF 4/silica gel and
DMImBF4/silica gel (Scheme 1) with different ionic liquid
Scheme 1. General structure for the ionic liquids. X=BF4: n=1
EMImBF4, n=3 BuMImBF4, n=9 DMImBF4, n=15 CMImBF4; X=
PF6: n=3 BuMImPF6. E=Ethyl; M=Methyl; Im=imidazolium; Bu=
Butyl; D=Decyl; C=Cetyl.
loadings were firstly characterized with FT-Raman spectros-
copy, Figures 2 and 3. The results revealed that the Raman
spectra at 8–53 wt%Bu(or D)MImBF4/silica gel samples
were different from pure bulk ionic liquids, that is, abnormal
FT-Raman spectra were observed. In comparison with cor-
responding bulk ionic liquids, the bands at 2967, 1421, and
766 cmÀ1 for BuMImBF4 (or
2898, 1421, 1025, and 766 cmÀ1
for DMImBF4), which may be
attributable to several skeletal
modes of the side alkyl chains
of imidazolium,[31–34] were re-
markably inhibited, or even
disappeared and then reap-
peared, with increasing loading
of the ionic liquids. At the
same time, bands at 2941,
1449,
BuMImBF4 (or 2930, 1453, and
879 cmÀ1
for DMImBF4),
and
880 cmÀ1
for
which may be attributable to
several skeletal modes of the
imidazolium ring, appeared or
strengthened, and then disap-
peared or weakened again,
when the loadings of confined
Figure 1. Illustration of the synthesis of silica-gel-confined ionic liquids with and without the metal complex
(Mc).
5280
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2005, 11, 5279 – 5288

Ionic-Liquid Catalysts
FULL PAPER
ionic liquids were increased from 8 and 17 wt% and then to
35 and 53 wt%. It is worth noting that such unusual spectral
peaks were much stronger when the loadings of confined
ionic liquids ranged from 8 to 17 wt%, and such aberration
became small and then reverted back to the same or normal
spectral peaks as the bulk ionic liquids when the loadings of
confined ionic liquids were further increased to 53 wt%.
Unusual FT-Raman spectra similar to those for BuMImBF4/
silica gel were also observed for 17 wt%EMImBF4/silica
gel, 17 wt%CMImBF4/silica gel (Figure 4), and 17 wt%Bu-
MImPF6/silica gel samples (though not shown here; see
Figure 2. FT-Raman spectra of pure BuMImBF4 and BuMImBF4/silica
gel with different ionic liquid loadings: a) 8 wt%, b) 17 wt%, c) 35 wt%,
d) 53 wt%, and e) pure BuMImBF4.
Figure 4. FT-Raman spectra of a) 17 wt%EMImBF4/silica gel, b) pure
EMImBF4, c) 17 wt%CMImBF4/silica gel, and d) pure CMImBF4 .
Scheme 1 for structures of these ionic liquids). This suggests
that such aberration in the Raman spectra may be mainly
related to skeletal-mode modification of the imidazolium
ring, since variations in side-chain length of imidazolium or
in anions have less impact on such phenomena.
To investigate the stability of the ionic liquids confined in
silica gel, the prepared samples in which the ionic liquids
had different lengths of side chain attached to the imidazoli-
um group were washed in acetone (30 mL) under vigorous
refluxat 60 8C for 3 h. This was done because BuMImBF4,
DMImBF4, CMImBF4, and BuMImPF6 ionic liquids are
completely soluble in acetone at this temperature. Then FT-
Raman analysis was conducted (Figure 5). It can be seen
that ionic liquids DMImBF4 and CMImBF4, which have
longer side chains, could be firmly confined into the silica
gel, while BuMImBF4 with shorter side chains was, since the
Raman spectral peaks of DMImBF4 and CMImBF4 re-
Figure 3. FT-Raman spectra of pure DMImBF4 and DMImBF4/silica gel
with different ionic liquid loadings: a) 8 wt%, b) 17 wt%, c) 35 wt%,
d) 53 wt%, and e) pure DMImBF4.
Chem. Eur. J. 2005, 11, 5279 – 5288
www.chemeurj.org
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
5281

Y. Deng et al.
ionic liquid involvement indicated the formation of the typi-
cal structures of mesoporous materials (Figure 6). The BET
surface areas were decreased from 933 to 322 m2 gÀ1, but the
average pore volume and average pore diameters increased
from 0.298 to 1.35 cm3 gÀ1 and 29 to 109 Š, respectively,
when amounts of confined ionic liquids were increased from
0 to 53 wt% (Table 1). Furthermore, it can be seen that the
pore-size distributions were quite narrow and then became
wider when the amounts of confined BuMImBF4 ionic
liquid was below 35 wt% or near to 53 wt%, respectively;
this suggests that ionic liquids may be well confined in the
pores of silica gel when the loadings of ionic liquids were
lower than 35 wt%. In the case of silica-gel-confined Bu-
MImPF6, mesoporous silica gel was also obtained when the
ionic liquid was removed (Figure 7), but the corresponding
average pore diameters were much larger (168 and 343 Š)
and the pore-size distribution became quite wide relative to
that of BuMImBF4/silica gels (42 and 71 Š and a relatively
narrow pore-size distribution) when the same amount of
ionic liquids by weight were confined.
TEM and XRD analysis: We analyzed pure silica gel
(that is 0 wt% ionic liquid loading) and 17 wt%DMImBF4/
silica gel by high-resolution transmission electron microsco-
py (HRTEM), see Figure 8. Uniform but amorphous tex-
tures were observed for both the pure silica gel and silica-
gel-confined ionic liquids. Unfortunately, the “particles” of
ionic liquid over or in the silica gel were not visible, since
the ionic liquid could not be distinguished from the silica
gel substrate. Although not shown here, XRD results
of pure silica gel, 17 wt%BuMImBF4/silica gel, and
17 wt%DMImBF4/silica gel showed that the structures of
silica gel and the ionic liquids were amorphous.
Figure 5. FT-Raman spectra of 17 wt%B(or D/C)MImBF4/silica gel
before and after washing. a),b) 17 wt%BuMImBF4/silica gel before and
after washing; c),d) 17 wt%DMImBF4/silica gel before and after wash-
ing; e),f) 17 wt%CMImBF4/silica gel before and after washing.
mained, while those of BuMImBF4 disappeared almost com-
pletely after these samples were washed under the above-
mentioned conditions. Since CMImBF4 and DMImBF4 ionic
liquids are completely soluble in acetone at 608C, the fact
that they could not be washed out from the silica gel should
not be attributed to their lower solubility in acetone. This
may also imply that there was no chemical bonding between
the silica-gel surface and ionic liquid fragments, that is, the
ionic liquids were physically confined or encapsulated into
the silica gel, and porous materials of silica gel may be pro-
duced if the confined ionic liquids or liquids acting as a
“template” were removed completely.
Based on the results of FT-Raman spectroscopy, N2 ad-
sorption measurements, TEM and XRD analysis, some in-
sight into the silica-gel-confined ionic liquids could be
gained.
BET characterization of the silica-gel matrix: A porous
silica-gel matrixcould be formed if the confined ionic liq-
uids with a smaller relative molecular size, such as Bu-
MImBF4, could be removed without destroying the remain-
ing skeleton of silica gel. N2 adsorption measurements,
which have been a powerful tool for nano- or mesoporous
material characterization,[35–37] were performed to attain
more insight of the porous silica gels prepared by using
ionic liquids as templates. Since DMImBF4 or CMImBF4
were confined into the pores of silica gel relatively firmly
and could not be washed out in acetone under vigorous re-
fluxing, the accurate data of average pore volume and diam-
1) the size of particles of ionic liquids confined in the silica-
gel matrixmay range into the nanoscale; for example,
ionic-liquid particles 40–110 Š in diameter may be
formed over 7–53 wt%BuMImBF4/silica gel.
2) The unusual FT-Raman spectra could be attributed to
the nanoeffect; that is, changes in the ionic liquid sym-
metry and unusual coordination geometry occurred
when ionic liquids were confined or encapsulated into
the nanopores of silica gel.[38–41] Also, unusual FT-Raman
spectra showing strong peaks occurred when the amount
of confined ionic liquids was about 17 wt%.
3) The fact that BuMIm cations could move in the channels
easily but DMIm and CMIm cations could not suggested
that the silica-gel matrices containing BuMImBF4 and
DMImBF4 may possess similar sizes of channels, al-
though the pore-size distributions of silica gel matrices
containing DMImBF4 were unknown.
4) The particle sizes of DMImBF4 and CMImBF4 in
DMIm(or CMIm)BF4/silica gels, though not derivable
through N2 adsorption measurement, should also be in
eter of the silica-gel matrixcontaining DMImBF
and
4
CMImBF4 are not available. We only measured the remain-
ing skeleton of silica gel in which BuMImBF4 and Bu-
MImPF6 ionic liquids were confined at different loadings
ranging from 0–53 wt%. For the skeletons of silica gel
which had originally contained confined BuMImBF4 ionic
liquid, the isotherm plots showed distinct hysteresis loops,
whereas isotherm plots of silica gel prepared without any
5282
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2005, 11, 5279 – 5288

Ionic-Liquid Catalysts
FULL PAPER
Figure 6. Pore-size distribution and N2 adsorption–desorption isotherms of silica gel matrices after the confined BuMImBF4 was removed: a) silica gel
skeleton from 8 wt%BuMImBF4/silica gel, b) silica gel skeleton from 17 wt%BuMImBF4/silica gel, c) silica gel skeleton from 35 wt%BuMImBF4/silica
gel, and d) silica gel skeleton from 53 wt%BuMImBF4/silica gel.
Table 1. Results of N2 adsorption measurements of the remaining skeletons of silica gel after being washed to remove confined BuMImBF4 and
BuMImPF6 ionic liquids.
Original samples
BET surface
Single desorption
BJH desorption
average pore
diameter [Š]
area [m2 gÀ1
]
pore volume [cm3 gÀ1
]
pure silica gel
933
634
415
360
322
212
108
0.298
0.878
1.059
1.044
1.35
29
42
71
8%BuMImBF4/silica gel
17%BuMImBF4/silica gel
35%BuMImBF4/silica gel
53%BuMImBF4/silica gel
8%BuMImPF6/silica gel
17%BuMImPF6/silica gel
81
109
168
343
1.32
1.15
Chem. Eur. J. 2005, 11, 5279 – 5288
www.chemeurj.org
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
5283

Y. Deng et al.
Figure 8. TEM pictures of a) pure silica gel and b) 17 wt%DMImBF4/
silica gel.
first (Table 2, entries 1–3). Poor catalytic performances were
observed over these catalysts. However, the carbonylation
of aniline and nitrobenzene proceeded efficiently over Rh-
(or Pd)–D(or E)MImBF4/silica gel as catalysts with conver-
sions of 81–93% and selectivity >98% (Table 2, entries 4–
6). The best catalytic performance was achieved over Rh-
DMImBF4/silica gel, and the corresponding TOF exceeded
11000 molmolÀ1 hÀ1. In comparison with the results shown
in Table 2 (entries 1–3), it is worth noting that although the
amount of ionic liquid used was much less, the catalytic per-
formances were much better for the metal complex/ionic
liquid/silica gel catalysts. This means that more enhanced
catalytic activity was exhibited through the confinement or
high concentration of an ionic liquid containing a metal
complexin the pores or cavities of the silica-gel matrix;
however, such an effect would disappear when the bulk
ionic liquid and metal complexes were completely diluted
with large amounts of the reactant molecules, thus resulting
in poor catalytic performance. For the catalyst regeneration,
Figure 7. Pore-size distribution and N2 adsorption–desorption isotherms
of silica gel matrices after the confined BuMImPF6 was removed: Top:
silica gel skeleton from 8 wt%BuMImPF6/silica. Bottom: silica gel skele-
ton from 17 wt%BuMImPF6/silica gel.
the nanoscale range, since very similar unusual FT-
Raman spectra were exhibited.
5) Non-uniform mesoporous silica gels confined with nano-
scale ionic-liquid particles were formed.
Synthesis of diphenyl urea:
Table 2. Results of carbonylation of nitrobenzene and aniline to diphenyl urea.
Carbonylation of aniline and
nitrobenzene for the synthesis
of diphenyl is an important
process that generally proceeds
in the presence of a suitable
noble-metal complexas cata-
lyst. For the purpose of com-
parison, [Rh(PPh3)3Cl] only,
Catalysts
Ionic liquid/
Conversion [%]
Sel
TOF[a]
metal complex[mg] nitrobenzene aniline [%]
1
2
3
4
5
6
[Rh(PPh3)3Cl]
DMImBF4 +silica gel
[Rh(PPh3)3Cl]+DMImBF4 +pure silica gel 200/1
0/1
200/0
8
–
9
–
>95
–
>98
>98 11548
>98 7440
>98 10550
>98 10800
1586
–
7463
41
93
81
85
83
16
3
39
92
79
84
81
16
4
0.11 wt%Rh–35 wt%DMImBF4/silica gel
0.15 wt%Pd–35 wt%DMImBF4/silica gel
0.13 wt%Rh–35 wt%EMImBF4/silica gel
~35/~1
~35/~1
~35/~1
~35/~1
~35/~1
~35/~1
7[b] 0.11 wt%Rh–35 wt%DMImBF4/silica gel
8[b] 0.13 wt%Rh–35 wt%EMImBF4/silica gel
physical
mixtures
of
>98
>98
9987
2180
DMImBF4/silica gel, and [Rh-
(PPh3)3Cl]/DMImBF4/silica gel
as catalysts were examined
9
0.11 wt%Rh–35 wt%DMImBF4/SiO2
[a] TOF=mol substrates converted per mol metal complexes per hour. [b] The catalyst system had been re-
covered and reused for the second time.
5284
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2005, 11, 5279 – 5288

Ionic-Liquid Catalysts
FULL PAPER
methanol (20 mL) was added to the resultant solid mixture
(if the conversion and selectivity was high enough) to dis-
solve the unreacted substrates from the catalyst and desired
product. The mixture was filtered to remove methanol con-
taining unreacted substrates, and water (20 mL) was added
to the remaining solid mixture containing diphenyl urea and
the catalyst. The diphenyl urea floats on the water surface
and the catalyst is deposited on the base of the reaction
vessel. Hence, they could be separated easily with a simple
dredge and filtration, and the catalyst could be reused after
drying. An aniline conversion of 81% and >98% selectivity
were maintained when 0.11 wt%Rh-35 wt%DMImBF4/
silica gel was reused for the second time (Table 2, entry 7).
The catalytic activity was greatly reduced when the used
Rh-EMImBF4/silica gel was reused for the second time
(Table 2, entry 8). Analyses of Rh content in 0.11 wt%Rh-
35 wt%DMImBF4/silica gel and 0.13 wt%Rh-35 wt%
EMImBF4/silica gel after being used for twice gave
results of 0.08 and 0.09 wt%, respectively. FT-Raman analy-
ses of the ionic liquids contents in 0.11 wt%Rh-
35 wt%DMImBF4/silica gel and 0.13 wt%Rh-35 wt%-
EMImBF4/silica gel before and after reaction showed that
the characteristic peaks of EMImBF4 had almost disap-
peared, whilst the characteristic peaks of DMImBF4 could
be observed clearly but were slightly weakened. These re-
sults suggested that the deactivation of 0.13 wt%Rh-
35 wt%EMImBF4/silica gel was mainly due to EMImBF4
leaching from silica gel, because of its smaller molecular
Figure 9. FT-Raman
b) 17 wt%DMImBF4/SiO2, and c) pure DMImBF4 ionic liquid.
spectra
of
a) 17 wt%DMImBF4/silica
gel,
nates,[42–44] and transformation between cyclohexanone
oxime and acetone were also conducted over the Rh(or Pd)-
DMImBF4/silica
gel,
DMImBF4/silica
gel,
and
17 wt%DMImBF4/SiO2 catalysts. For oxidative carbonyla-
tion of aniline, lower catalyst activities were obtained with
the bulk DMImBF4 ionic liquid+Pd(PPh3)2Cl2 or Rh-
(PPh3)3Cl catalyst system (Table 3, entries 1 and 2). Greatly
enhanced catalytic activities were observed with the
Pd(Rh)-35 wt%DMImBF4/silica gel although the amount of
ionic liquid used was greatly reduced (Table 3, entries 3 and
4).
The transformation reaction of cyclohexanone oxime and
acetone[45] can be catalyzed with ionic liquids. Only 22.5%
of conversion was achieved when the bulk ionic liquid
DMImBF4 was used as catalyst directly and similar results
were obtained if a physical mixture of pure silica gel and the
ionic liquid DMImBF4 was used (Table 4, entries 1 and 2).
Surprisingly, more than 90% of conversions with nearly
100% of selectivity were achieved when using
17 wt%DMImBF4/silica gel as catalyst (Table 4, entry 3),
for which the TOFs were 16–27 times higher than that of
size rather than Rh complexleaching, while DMImBF
4
could be confined into the pores of silica gel relatively
firmly due to its larger molecular size, thus maintaining a
more stable activity. Though not shown here, carbonylation
of aniline and nitrobenzene were also examined over Rh-D-
(or E)MImPF6/silica gels. It was found that BuMImPF6 was
liable to decompose in the presence of water or acid during
reaction.
For the purpose of comparison, a silica-supported ionic-
liquid catalyst containing Rh(PPh3)3Cl, that is, 0.11%Rh-
35 wt%DMImBF4/SiO2, in which the ionic liquid was de-
posited only as a thin layer on the support according to the
FT-Raman characterization (Figure 9), was also tested with
the carbonylation of nitroben-
zene and aniline to diphenyl
Table 3. Results of oxidative carbonylation of aniline to phenyl carbamate in methanol.
carbonate; however, the con-
versions were only 3 and 4%
respectively (Table 2, entry 9).
Catalyst
Ionic liquid/
Conv [%]
Sel [%]
TOF
metal complex[mg]
1
2
3
4
[Pd(PPh3)2Cl2]–DMImBF4
[Rh(PPh3)3Cl]–DMImBF4
0.15 wt%Pd–35 wt%DMImBF4/silica gel
0.11 wt%Rh–35 wt%DMImBF4/silica gel
2500/1
2500/1
~35/~1
~35/~1
13
15
63
89
>99
96
>98
>98
494
752
2650
4960
Syntheses of phenyl carba-
mates and oxime transforma-
tion: To examine the possible
universality of such a silica-gel-
confined ionic-liquid catalyst
system, oxidative carbonyla-
tion of aniline to afford carba-
mate, a key intermediate for
the nonphosgene synthesis
of 4,4’-dipheylmethyldiisocya-
Table 4. Results of C=N and C=O bond exchange between cyclohexanone oxime and acetone.
Catalyst
Ionic liquid [mg]
Conv [%]
Sel [%]
TOF
1
2
3
4
DMImBF4
10
10
~1.7
~1.7
22.5
22
92.1
24.7
>99
>99
>99
>99
6
6
100
27
DMImBF4 +silica gel
17 wt%DMImBF4/silica gel
17 wt%DMImBF4/SiO2
Chem. Eur. J. 2005, 11, 5279 – 5288
www.chemeurj.org
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
5285

Y. Deng et al.
liquid was obtained even with ionic liquid loading of up to 53 wt%, indi-
cating that the ionic liquid added could be completely encapsulated into
sol–gel. Therefore, the loadings of confined ionic liquids could be calcu-
lated according to the weight of ionic liquids added originally. Solid sam-
ples exhibited different colors depending on the different ionic liquids,
their loadings, and on whether they contained metal complexor not
(Figure 10).
the bulk ionic-liquid catalyst. If 17 wt%DMImBF4/SiO2 was
used as the catalyst for the transformation reaction of cyclo-
hexanone oxime and acetone, only 24.7% of conversion was
obtained (Table 4, entry 4); this also showed the silica-gel-
confined ionic liquid was more effective than the bulk or
silica-supported ionic liquid as catalysts. Both the results of
the carbonylation reaction and oxime transformation also
confirmed that higher catalytic activities could be exhibited
with silica-gel-confined ionic-liquid nanocatalyst systems.
These results further confirmed that the enhancement in
catalytic activities may be derived from the formation of
nanoscale and high-concentration ionic liquids due to the
confinement in the nanopores of silica gel, in which the re-
strictive dimensions of the silica gel interior may force un-
usual compound symmetry, coordination geometry, and co-
ordinative unsaturation upon the entrapped ionic liquid and
metal complex, rather than being related to the use of a sup-
port to improve the possible transport restrictions occurring
in bulk ionic liquid. As a matter of fact, the problem of
mass transfer may indeed occur when the bulk ionic liquids
are immobilized into a porous support.
Conclusion
Figure 10. Two kinds of typical silica-gel-confined ionic liquid samples:
53 wt%BuMImBF4/silica gel (top) and 0.11 wt%Rh-35wt%DMImBF4/
silica gel (bottom).
In summary, physically confined ionic liquids on the nano-
scale, with or without a metal complex, in silica gel were de-
veloped through a “one-pot” sol–gel method as a new kind
of supported nanoliquid catalyst. The confined ionic liquids
on the nanoscale exhibited abnormal Raman spectra, and
particularly important was that they exhibited greatly en-
hanced catalytic performance with less ionic liquid required
relative to the bulk and pure ionic liquids. The preliminary
investigations showed that such silica-gel-confined ionic-
liquid catalysts may be effective for a wider range of reac-
tions. The preparation of these supported nanoliquid cata-
lysts could be further optimized and modified, and their ap-
plication expanded to other reactions and processes.
A series of samples of silica-gel-confined ionic liquids (either with or
without [Pd(PPh3)2Cl2] and [Rh(PPh3)3Cl]) and with various ionic liquid
or metal loadings were denoted as follows: BuMImBF4: 8 wt%Bu-
MImBF4/silica gel, 17 wt%BuMImBF4/silica gel, 35 wt%BuMImBF4/
silica gel, 53 wt%BuMImBF4/silica gel; DMImBF4: 8 wt%DMImBF4/
silica gel, 17 wt%DMImBF4/silica gel, 35 wt%DMImBF4/silica gel,
53 wt%DMImBF4/silica gel; EMImBF4: 17 wt%EMImBF4/silica gel;
CMImBF4: 17 wt%CMImBF4/silica gel; 0.15 wt%Pd-35 wt%DMImBF4/
silica gel, 0.11 wt%Rh-35 wt%DMImBF4/silica gel, 0.13 wt%Rh-
35 wt%EMImBF4/silica gel.
Synthesis of SiO2-supported ionic-liquid catalysts: For the purpose of
comparison, synthesis and catalytic performance of SiO2-supported ionic
liquid as catalysts, in which the ionic liquid was deposited as a thin layer
on the support, were also conducted. SiO2 pellets (4.9 g, 1 0.5–0.6 mm,
average pore diameter: 52 Š, pore volume: 0.75 mLgÀ1, surface area:
578 m2 gÀ1) calcined at 4008C for 6 h, were charged into a 50 mL three-
necked round-bottomed flask. Under the conditions of vacuum (approxi-
mately 5 mmHg) and room temperature, solutions of DMImBF4 ionic
liquid (1 g, or 2.7 g DMImBF4 ionic liquid containing 70 mg [Rh-
(PPh3)3Cl]) and ethanol (6 mL) were transferred into the flask, and the
adsorption/absorption was conducted for 1 h. The resultant catalyst pre-
cursors were then dried under vacuum (approximately 5 mmHg) at
1508C for another 4 h to remove the residual ethanol completely. Two
kinds of silica-supported ionic liquids (with and without [Rh(PPh)3Cl])
were denoted as follows: 17 wt%DMImBF4/SiO2; 0.11 wt%Rh-
35wt%DMImBF4/SiO2.
Experimental Section
Ionic liquids and metal complex synthesis: All chemicals used were ana-
lytical grade and used without further purification. The ionic liquids used
in this work were synthesized according to the literature.[46, 47] Two kinds
of typical ionic liquids, in which the cations and anions were dialkyl imi-
dazolium and BF4À and PF6À, respectively, were employed for the synthe-
sis of silica-gel-confined ionic-liquid catalysts:
The two metal complexes employed, [Pd(PPh3)2Cl2] and [Rh(PPh3)3Cl],
were synthesized according to the literature.[48,49]
Synthesis of silica-gel-confined ionic-liquid catalysts: A mixture of tetrae-
thoxyorthosilicate (TEOS, 10 mL) and EtOH (7 mL) was heated to 608C
and then ionic liquids (0.2–4 g) or ionic-liquids (2 g) containing [Pd-
(PPh3)2Cl2] or [Rh(PPh3)3Cl] (50 mg) were immediately transferred into
the TEOS. After the formation of a clear and homogeneous liquid mix-
ture, hydrochloric acid (5m, 5 mL) was added and the mixture gradually
coagulated. After aging at 608C for 12 h, the resultant solid material was
dried in vacuum at 1508C for 3 h and 3–7 g of solid sample was obtained.
It is worth noting that a dry solid sample of silica-gel-confined ionic
Catalyst sample characterization: FT-Raman analysis was conducted with
a Nicolet 910 Raman spectrometer. Only the samples of silica-gel-con-
fined ionic liquid without metal complexes were characterized, because
the loadings of metal complexes were too low to be observed in the FT-
Raman spectra.
BET surface areas (SBET) were obtained by physisorption of N2 at 77 K
by using a Micromeritics ASAP 2020. Prior to measurement, samples
were degassed to 0.1 Pa at 373 K.
5286
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2005, 11, 5279 – 5288

Ionic-Liquid Catalysts
FULL PAPER
TEM analysis was performed on a JEOL JEM-100CXII transmission
electron microscope. The specimens for electron microscopy were pre-
pared by gently grinding the powder samples suspended in ethanol in an
agate mortar, and then the resultant solution was dropped onto the
carbon film of the copper grid.
tant solution containing reactants and products was directly analyzed
without further treatment.
Analyses: Qualitative analyses were conducted with a HP 6890/5973 GC-
MS with a 30 m”0.25 mm”0.33 mm capillary column and a chemstation
containing a NIST Mass Spectral Database. Quantitative analyses were
conducted with a Agilent (Shanghai) 1790 GC equipped with FID detec-
tor and 30 m”0.25 mm”0.33 mm capillary column.
XRD was performed on a Philips XꢁPert Pro powder X-ray diffractome-
ter.
The metal Rh or Pd in the samples of silica-gel-confined ionic liquids
were analyzed with a 3520 ICT AES instrument (ARL Co. USA).
Synthesis of diphenyl urea:
Acknowledgement
This work was financially supported by the National Natural Science
Foundation of China (No. 20233040 and No. 20225309).
[1] Y. Wada, Surf. Sci. 1997, 386, 276.
[2] R. Bauer, G. Waldner, S. Fallmann, S. Hager, T. Krutzler, S. Malato,
P. Maletzky, Catal. Today 1999, 53, 131.
[3] C. B. Geoffrey, T. T. David, Catal. Rev. 1999, 319.
[4] S. Miao, Y. Deng, Appl. Catal. B. 2001, 31, L1.
[5] F. Shi, Y. Deng, T. SiMa, H. Yang, J. Catal. 2002, 211, 548.
[6] J. Jortner, Z. Phys. Chem. 1994, 184, 283.
The carbonylation of aniline and nitrobenzene for the synthesis of di-
phenyl urea was conducted as follows: Pd(or Rh)-EMImBF4/silica gel or
Pd(or Rh)-DMImBF4/silica gel (0.1 g, 50–60 mesh), or DMImBF4 ionic
liquid (2.5 g) containing Rh complex(1 mg), or 0.11 wt%Rh-
35 wt%DMImBF4/SiO2 (0.1 g, for the purpose of comparison), nitroben-
zene (10 mmol), aniline (10 mmol), and CO gas (5.0MPa, purity 99.99%)
were successively introduced into an autoclave (90 mL) without any addi-
tional organic solvent. The reaction proceeded at 1808C for 1.5 h and
then was cooled to room temperature. Since the desired diphenyl urea
was insoluble in methanol and water, methanol (20 mL) was added to the
resultant solid mixture (if the conversion and selectivity was enough
high) to dissolve the unreacted substrates for analysis and, hence, obtain
the conversion of nitrobenzene and aniline by means of GC with an out-
standard method. Then, filtration was conducted to remove methanol,
and water (20 mL) was added to the solid mixture containing diphenyl
urea and the catalyst. The diphenyl urea floated on the water and the cat-
alyst deposited on the bottom. They could be easily separated with a
simple dredge and filtration method, and the isolated yield of diphenyl
urea was obtained. The catalyst could be reused after drying. The selec-
tivity was calculated as follows: selectivity (%)=isolated yield/conver-
sion”100%.
[7] P. Alivisatos, Science 1996, 271, 933.
[8] I. Last, J. Jorter, Phys. Rev. Lett. 2001, 87, 3401.
[9] J. Dupont, G. S. Fonseca, A. P. Umpierre, P. F. P. Fichtner, S. R. Teix-
eira, J. Am. Chem. Soc. 2002, 124, 4228.
[10] V. Calo, A. Nacci, A. Monopoli, S. Laera, N. Cioffi, J. Org. Chem.
2003, 68, 2929.
[11] D. Zhao, Z. Fei, T. J. Geldbach, R. Scopelliti, P. J. Dyson, J. Am.
Chem. Soc. 2004, 126, 15876.
[12] K. He, Z. Zhou, L. Wang, K. Li, G. Zhao, Q. Zhou, C. Tang, Synlett
2004, 9, 1525.
[13] L. M. Rossi, G. Machado, P. F. P. Fichtner, S. R. Teixeira, J. Dupont,
Catal. Lett. 2004, 92, 149.
[14] K. Anderson, S. C. Fernµndez, C. Hardacre, P. C. Marr, Inorg.
Chem. Commun. 2004, 7, 73.
[15] J. Huang, T. Jiang, B. Han, H. Gao, Y. Chang, G. Zhao, W. Wu,
Chem. Commun. 2003, 1654.
Synthesis of phenyl carbamate:
[16] E. T. Silveira, A. P. Umpierre, L. M. Rossi, G. Machado, J. Morais,
G. V. Soares, I. J. R. Baumvol, S. R. Teixeira, P. F. P. Fichtner, J.
Dupont, Chem. Eur. J. 2004, 10, 3734.
[17] W. Liu, A. G. Fadeev, B. Qi, E. Smela, B. R. Mattes, J. Ding, G. M.
Spinks, J. Mazurkiewicz, D. Zhou, G. G. Wallace, D. R. MacFarlane,
S. A. Forsyth, M. Forsyth, Science 2002, 297, 983.
[18] J. Dupont, R. F. de Souza, P. A. Z. Suarez, Chem. Rev. 2002, 102,
3667.
[19] P. Wasserscheid, W. Keim, Angew. Chem. 2000, 112, 3926; Angew.
Chem. Int. Ed. 2000, 39, 3772.
[20] T. Welton, Chem. Rev. 1999, 99, 2071.
[21] W. Chen, L. Xu, C. Chatterton, J. Xiao, Chem. Commun. 1999, 1247.
[22] B. Ellis, W. Keim, P. Wasserscheid, Chem. Commun. 1999, 337.
[23] J. Peng, Y. Deng, New J. Chem. 2001, 25, 639.
[24] J. Peng, Y. Deng, Tetrahedron Lett. 2001, 42, 403.
[25] J. Peng, Y. Deng, Tetrahedron Lett. 2001, 42, 5917.
[26] M. Antonietti, D. Kuang, B. Smarsly, Y. Zhou, Angew. Chem. 2004,
116, 5096; Angew. Chem. Int. Ed. 2004, 43, 4988.
[27] J. Dupont, J. Braz. Chem. Soc. 2004, 15, 341.
[28] C. P. Mehnert, R. A. Cook, N. C. Dispenziere, M. Afeworki, J. Am.
Chem. Soc. 2002, 124, 12932.
The carbonylation of aniline for synthesis of carbamates was conducted
as follows: Pd(or Rh)-DMImBF4/silica gel (0.1 g, 50–60 mesh), or
DMImBF4 ionic liquid (2.5 g) containing Pd(or Rh) complex(1 mg, for
the purpose of comparison), aniline (0.5 mL), methanol (5 mL) and a
CO/O2 gas mixture (5.0MPa: CO purity 99.99% 4.5MPa and O2 99.99%
purity 0.5MPa) were successively introduced into an autoclave (90 mL)
without any additional organic solvent. The reaction proceeded at 1358C
for 1 h and the autoclave was then cooled to room temperature, and
methanol (20 mL) was added to dissolve the resultant mixture for analy-
ses.
Oxime transformation between cyclohexanone oxime and acetone:
[29] M. H. Valkenberg, C. deCastro, W. F. Hçlderich, Green Chem. 2002,
4, 88.
[30] H. Hagiwara, Y. Sugawara, K. Isobe, T. Hoshi, T, Suzuki, Org.
Lett. 2004, 6, 2325.
[31] A. Elaiwi, P. B. Hitchcock, K. R. Seddon, N. Srinivasan, Y. Tan, T.
Welton, J. A. Zora, J. Chem. Soc. Dalton Trans. 1995, 3467.
[32] N. Nanbu, Y. Sasaki, F. Kitamura, Electrochem. Commun. 2003, 5,
383.
The oxime transformation between cyclohexanone oxime and acetone
was carried out as follows: cyclohexanone oxime (0.2 g),
17 wt%DMImBF4/silica gel (0.01 g, 50–60 mesh) or pure DMImBF4
ionic liquid (0.01 g) or 17wt%DMImBF4/SiO2 (0.01 g, for the purpose of
comparison), acetone (2 mL) and water (6 mL) were successively added
into a 25 mL round-bottomed flask and the mixture was stirred magneti-
cally at room temperature (approximately 208C) for 2 h. Then, the resul-
[33] C. Perchard, A. Novak, Spectrochim. Acta A 1967, 23, 1953.
Chem. Eur. J. 2005, 11, 5279 – 5288
www.chemeurj.org
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
5287

Y. Deng et al.
[34] M. Majoube, M. Henry, L. Chinsky, P. Y. Turpin, Chem. Phys. 1993,
169, 231.
[35] D. Garfinkel, J. T. Edsall, J. Am. Chem. Soc. 1958, 80, 3807.
[36] A. Toyama, K. Ono, S. Hashimoto, H. Takeuchi, J. Phys. Chem. A.
2002, 106, 3403.
[37] S. L. Burkett, S. D. Sims, S. Mann, Chem. Commun. 1996, 1367.
[38] N. Herron, G. D. Stucky, C. A. Tolman, Inorg. Chim. Acta 1986, 25,
4717.
[42] V. L. K. Valli, H. Alper, J. Am. Chem. Soc. 1993, 115, 3778.
[43] I. Pri-Bar and J. Schwartz, J. Org. Chem. 1995, 60, 8124.
[44] F. Shi, J. Peng, Y. Deng, J. Catal. 2003, 219, 372.
[45] D. Li, F. Shi, S. Guo, Y. Deng, Tetrahedron Lett. 2004, 45, 265.
[46] P. Bonhote, A. P. Dias, N. Papageorgiou, K. Kalyanasundaram, M.
Grꢂtzel, Inorg. Chem. 1996, 35, 1168.
[47] P. A. Z. Suarez, J. E. L. Dullius, S. Einloft, R. F. de Souza, J. Dupont,
Polyhedron, 1996, 15, 1217.
[39] M. Ichikawa, T. Kimura, A. Fukuoka, Stud. Surf. Sci. Catal. 1991,
60, 335.
[48] H. Itatani, J. C. Bailar Jr., J. Am. Oil Chem. Soc. 1967, 44, 147.
[49] T. A. Stepheson, G. Wilkison, J. Inorg. Nucl. Chem. 1966, 28, 945.
[40] B. V. Romanovsky, A. G. Gabrielov, J. Mol. Catal. 1992, 74, 293.
[41] S. Kowalak, R. C. Weiss, K. J. Balkus Jr., J. Chem. Soc. Chem.
Commun. 1991, 57.
Received: January 29, 2005
Published online: July 4, 2005
5288
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2005, 11, 5279 – 5288