Novel cyclic sulfonium-based ionic liquids: Synthesis, characterization, and physicochemical properties

Article abstract of DOI:10.1002/chem.200800610

A new series of ionic liquids composed of three cyclic sulfonium cations and four anions has been synthesized and characterized. Their physicochemical properties, including their spectroscopic characteristics, ion cluster behavior, surface properties, phase transitions, thermal stability, density, viscosity, refractive index, tribological properties, ion conductivity, and electrochemical window have been comprehensively studied. Eight of these salts are liquids at room temperature, at which some salts based on [NO₃]^- and [NTf₂]^- ions exhibit organic plastic crystal behaviors, and all the saccharin-based salts display relatively high refractive indices (1.442-1.594). In addition, some ionic liquids with the [NTf₂] ^- ion exhibit peculiar spectroscopic characteristics in FTIR and UV/Vis regions, whilst those salts based on the [DCA]^- ion show lower viscosities (34.2-62.6 mPa s at 20°C) and much higher conductivities (7.6-17.6 mS cm^-1 at 20°C) than most traditional 1,3-dialkylimidazolium salts.

Full text of DOI:10.1002/chem.200800610

FULL PAPER
DOI: 10.1002/chem.200800610
Novel Cyclic Sulfonium-Based Ionic Liquids: Synthesis, Characterization, and
Physicochemical Properties
Qinghua Zhang,[a, b, c] Shimin Liu,[a] Zuopeng Li,[a] Jian Li,[a] Zhengjian Chen,[a]
Ruifeng Wang,[a] Liujin Lu,[a] and Youquan Deng*[a]
Abstract: A new series of ionic liquids
composed of three cyclic sulfonium cat-
ions and four anions has been synthe-
sized and characterized. Their physico-
chemical properties, including their
spectroscopic characteristics, ion cluster
behavior, surface properties, phase
transitions, thermal stability, density,
viscosity, refractive index, tribological
properties, ion conductivity, and elec-
trochemical window have been com-
prehensively studied. Eight of these
salts are liquids at room temperature,
at which some salts based on [NO3]ꢀ
and [NTf2]ꢀ ions exhibit organic plastic
crystal behaviors, and all the saccharin-
based salts display relatively high re-
fractive indices (1.442–1.594). In addi-
tion, some ionic liquids with the
[NTf2]ꢀ ion exhibit peculiar spectro-
scopic characteristics in FTIR and UV/
Vis regions, whilst those salts based on
the [DCA]ꢀ ion show lower viscosities
(34.2–62.6 mPas at 208C) and much
higher conductivities (7.6–17.6 mScmꢀ1
at 208C) than most traditional 1,3-di-
Keywords: electrochemistry · ionic
liquids
·
performance
·
physico-
chemical properties
ions
·
sulfonium
AHCTUNGERTGaNNUN lkylimidazolium salts.
Introduction
um,[10] pyrrolidinium,[11] tetraalkylammonium,[12] pyridini-
um,[13] and quaternary phosphonium[14] ions. Thus, one of the
hot topics in the field of IL research is the molecular design
and synthesis of novel ILs, in particular ILs with peculiar
physicochemical characteristics and special applications.
After a period of prosperity in which conventional ILs were
predominantly used as environmentally benign media in
synthesis and catalysis, interest in ILs by the scientific com-
munity has recently begun to revert to the field of electro-
chemistry,[15] for which these innovative liquids were initially
developed. One key area of current research is the develop-
ment of IL-based electrolytes for lithium ion batteries,[16]
fuel cells,[17] capacitors,[18] solar cells,[19] and other electro-
chemical devices.
Over the past decade, the interest in ionic liquids (ILs) has
been increasing with surprising speed, mainly stimulated by
the appearance of numerous new ILs with peculiar proper-
ties[1–3] and ever-increasing applications in both academia
and industry in fields as diverse as electrochemistry,[4] syn-
thesis,[5] catalysis,[6] materials,[7] separation,[8] and biotech-
nology.[9] Although there are, in theory, potentially millions
of ILs available through the combination of different cations
and anions, up to now only thousands of ILs have been de-
veloped and used, most of which are based on imidazoli-
The emergence of functionalized ILs has endowed them
with huge diversity both in quantity and properties, and has
left more scope for further development.[20] Unfortunately,
for ILs with NH2,[21] SH,[22] OH,[23] COOH,[24] CN,[25] or
[a] Dr. Q. Zhang, S. Liu, Z. Li, J. Li, Z. Chen, R. Wang, L. Lu,
Prof. Y. Deng
Centre for Green Chemistry and Catalysis
Lanzhou Institute of Chemical Physics
Chinese Academy of Sciences
Lanzhou, 730000 (China)
[26]
NHCONH2 as functional groups attached, a salient prob-
lem blocking their further applications is the relatively
higher viscosity than organic solvents and the lower conduc-
tivity resulting from the high viscosity. From an electro-
chemical point of view, an ideal ionic liquid should possess
the following desirable properties: 1) low melting point,
<258C; 2) high ion conductivity; 3) good electrochemical
stability; and 4) relatively low viscosity. To achieve the
above targets, more effort should be dedicated to the devel-
Fax : (+86)931-496-8116
E-mail: ydeng@lzb.ac.cn
[b] Dr. Q. Zhang
Graduate School of the Chinese Academy of Sciences
Beijing, 100039, PRC (China)
[c] Dr. Q. Zhang
School of Information Science and Engineering and
The National Laboratory of Applied Organic Chemistry
Lanzhou University, Lanzhou, 73000 (China)
Chem. Eur. J. 2009, 15, 765 – 778
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
765

opment of new ILs with the characteristics of ideal electro-
lytes.[27,28] Recently, a new family of ILs based on symmetri-
cal and asymmetric trialkylsulfonium cations began to re-
ceive the interest of IL researchers, owing to the low viscosi-
ty and high conductivity of these systems.[29–32] For example,
Wasserscheid et al. developed a series of ILs based on trial-
kylsulfonium dicyanamides, which were found to have low
viscosity at extremely low temperature, for example, h=
94.9 mPas at ꢀ208C for triethylsulfonium dicyanamide.[33]
Hydrophobic trialkylsulfonium ILs based on perfluorinated
anions were also reported, and their potential applications
as promising electrolytes in lithium batteries and dye-sensi-
tized solar cells were preliminarily investigated.[34] However,
information on the synthesis of ILs based on a cyclic sulfoni-
um cation is relatively limited in previous literature.[35]
Therefore, there is still some room to develop new sulfoni-
um-based ILs that may be liquids at lower temperatures,
and it should be expected that these novel ILs would possess
the characteristics of low viscosity and high conductivity.
In this work, a new family of cyclic sulfonium-based ILs
was systematically synthesized, and the detailed study of
their physicochemical characteristics, including their spectro-
scopic properties, ion cluster behavior, surface properties,
phase transitions, heat capacity, thermal stability, density,
viscosity, refractive index, tribological properties, ion con-
ductivity, and electrochemical stability, is particularly em-
phasized.
ed tetrahydrothiophene iodide ([S2]I) and butylated tetrahy-
drothiophene iodide ([S4]I) were prepared by using a simi-
lar reaction of tetrahydrothiophene with iodoethane or n-
butyl iodide in acetone or anhydrous diethyl ether. The
silver dicyanamide (AgDCA) was precipitated by mixing
aqueous solutions of silver nitrate and sodium dicyanamide
(1:1 molar ratio). The saccharin silver (AgSac) was also pre-
pared by a similar method. The desired sulfonium salts with
the target anion could be obtained easily and with high
quality through an anion-exchange reaction between the
above-mentioned three iodides (i.e., [S1]I, [S2]I, and [S4]I)
with the corresponding inorganic salts (i.e., AgDCA, AgSac,
and LiNTf2) with a molar ratio of 1:1.1. All new cyclic sulfo-
nium-based ILs were obtained with yields of 89–94%, and
they are all stable to moisture. Among these ILs, [S1]DCA,
[S2]DCA, [S4]DCA, [S4]NTf2, [S1]Sac, [S2]Sac, [S4]Sac, and
[S2]NO3 are colorless liquids at room temperature and are
miscible with water except for [S4]NTf2; the others are
white solids at room temperature. The structures and com-
1
positions of all ILs were confirmed by H NMR and FTIR
spectroscopy, and ESI mass spectrometry. The water content
in the ILs was detected by a coulometric Karl-Fischer titra-
tion. The thermal properties of all the prepared salts were
characterized by differential scanning calorimetry (DSC)
and thermogravimetric analysis (TGA). For those salts that
are liquid at room temperature, their densities (1), viscosi-
ties (h), refractive indices (n), friction coefficients (FC),
ionic conductivities (k), and electrochemical windows (EW)
were measured. All the measured data are summarized in
Tables 1–4. The effect of structural variations in the cation
and anion on the above properties of these salts will be dis-
cussed below based on the results presented in these tables.
Results and Discussion
Synthesis and characterization: The structures and abbrevia-
tions of the cations and anions of twelve new ILs are shown
in Scheme 1. All the cyclic sulfonium salts were prepared ac-
cording to a similar approach reported previously.[33] Methy-
lated tetrahydrothiophene iodide ([S1]I) was prepared with
a high yield from the reaction of tetrahydrothiophene and
the appropriate amount of iodomethane in acetone. Ethylat-
Product purity: Data on the purity of any new class of ILs
are critical to the assessment of their physicochemical prop-
erties. Thus, the purity of the ILs was determined before
each test of physicochemical properties. To ensure that the
amount of water and other volatile solvents in the ILs was
reduced as much as possible, each IL was kept in a vacuum
(pressure 10ꢀ2–10ꢀ3 mbar) at 90–1008C for 2–4 h before
every test. The purity of each IL was first verified by NMR
spectroscopy to check for residues of unreacted reactants or
residual solvents, and the analysis results indicated that
these compounds had been completely eliminated. The
water content in the ILs was determined by means of a
Karl–Fischer titration (Mitsubishi CA-06 Moisturemeter).
The water content in [S4]NTf2 was 32 ppm (after drying at
1008C for 2 h) owing to its hydrophobicity, whereas the hy-
drophilic [DCA]ꢀ, [NO3]ꢀ, and [Sac]ꢀ salts still contained a
water content of 200–400 ppm, even after rigorous drying at
908C and 10ꢀ2–10ꢀ3 mbar for 4 h (see Table 2). It is worth
noting that during the measurement of some properties
(e.g., viscosity, conductivity, and refractive index) the ILs
need to be exposed to the open ambient air, and the water
in the air will inevitably pollute the ILs, especially those hy-
drophilic ILs. To find out how quickly these ILs absorb
water from the ambient air, the hydrophilic ILs [S2]DCA
Scheme 1. Structures and abbreviations of the cations and anions em-
ployed in this work.
766
www.chemeurj.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 765 – 778

Cyclic Sulfonium-Based Ionic Liquids
FULL PAPER
and [EMIm]BF4 as well as the
hydrophobic [S4]NTf2 were ex-
posed to the open ambient for
1, 5, and 30 min. The results
showed that the rate of water
absorption was relatively faster
at the initial stages, and the
rates for the hydrophilic
[S2]DCA and [EMIm]BF4 ILs
were remarkably higher than
that of the hydrophobic
[S4]NTf2 (Figure 1). For exam-
ple, when the hydrophilic ILs
[S2]DCA and [EMIm]BF4
were exposed to the ambient
for 5 min, approximately 70–
Figure 2. 1H NMR spectrum of [S4]DCA.
and the information obtained showed that they are all the
expected compounds (see Experimental Section). As shown
1
by the H NMR spectra of [S4]DCA in Figure 2, although
ꢀ
ꢀ
ꢀ
ꢀ
the hydrogen atoms of C1 H and C4 H (or C2 H and C3
H) are chemically equivalent in the tetrahydrothiophene
molecule, a chemical shift difference was clearly observed
when an alkyl group was attached to the sulfur atom to
form the sulfonium cation. The splitting of the signal related
ꢀ
ꢀ
to these two kinds of CH2 group (1,4 CH2 and 2,3 CH2)
demonstrated that the introduction of the alkyl group had
some influence on the steric location of the hydrogen atoms
of the four CH2 groups, resulting in an emerging difference
in the chemical shifts. Moreover, the extent of peak splitting
can be assigned to the electronic environment provided by
the [S]+ cation, that is, the splitting of the 1- and 4-position
hydrogen atoms is more obvious than that of the 2- and 3-
position hydrogen atoms because the former are attached
directly to the [S]+ ion.
Figure 1. Water contents of four selected ILs at different exposure times.
For the FTIR spectra of all new ILs, the absorption fre-
ꢀ
80 ppm water was absorbed into the ILs, whereas under the
same conditions the water content increased from 50 ppm to
80 ppm for hydrophobic [S4]NTf2. Further prolonging the
exposed time to 30 min, around 250–300 ppm water was ab-
sorbed into [EMIm]BF4 and [S2]DCA, but just 120 ppm for
hydrophobic [S4]NTf2. This means that changes in water
content in the ILs (particularly for the hydrophilic ILs) oc-
curred during those measurements conducted under open
ambient conditions (viscosity, conductivity, and refractive
index). However, these measurements could be performed
within 1–3 min, which means that although the slight in-
crease in water content would certainly cause some mea-
surement errors, such errors should be acceptable. The
cyclic voltammetry analysis showed that no reducible or oxi-
dizable species such as Iꢀ and Ag+ were detectable in the
ILs, indicating that the metathesis reaction had been con-
ducted completely and the purities of the ILs were all of
electrochemical grade.
quencies of the aliphatic C H vibrations in the cation ring
or the alkyl group were mainly observed in the range of
2875–3021 cmꢀ1. Figure 3 gives the molecular vibration spec-
1
NMR and FTIR spectra: H NMR and FTIR spectroscopy
were used to characterize the structures of the ILs prepared,
Figure 3. The FTIR spectra of a) [S4]DCA and b) [S4]NTf2.
Chem. Eur. J. 2009, 15, 765 – 778
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
767

Y. Deng et al.
tra of [S4]DCA and [S4]NTf2. From Figure 3b, we can see a
their isotope patterns can also be detected. These peaks, cor-
responding to the aggregates of cations and anions, are very
consistent with the theoretical values, thus confirming their
identities. The relative abundances of positively charged ion
clusters in the ESI-MS spectrum for [S1]DCA (Figure 3a)
indicated that the positively charged dimer cluster (m/z
272.1158) is more easily formed than the trimer cluster (m/z
441.1627) and the tetramer cluster (m/z=610.1998), and is
even more abundant than the dissociated cation [S1]+ in
[S1]DCA.
ꢀ
good example of a C H stretch of an aliphatic heterocyclic
ꢀ
compound, where we observe the asymmetric C H stretch
of the methyl and methylene groups (2967 and 2940 cmꢀ1,
respectively) occurring at slightly higher frequency than the
symmetric vibration of the methyl group at 2879 cmꢀ1. Sur-
prisingly, no strong absorptions in the range 1352–2880 cmꢀ1
were observed in the FTIR spectrum of [S4]NTf2 (Fig-
ure 3b). Thus, this salt might be a potential IR-transparent
medium. As for the [DCA]ꢀ ILs, the main feature in the
FTIR spectrum was the strong characteristic absorptions of
The ion cluster behavior of [S4]NTf2 was also studied, and
its ESIMS spectrum is shown in Figure 4b. Apart from the
detection of charged [S4]+ (m/z 145.1047), the charged ion
clusters (m/z 570.0869, 995.0367, 1419.9098, etc.) were also
observed, and the peak values are very consistent with the
theoretical values. As for the relative abundances of differ-
ent ions and clusters in [S4]NTf2, although a distribution
trend similar to [S1]DCA was obtained, the charged dimer
cluster [(S4)2NTf2]+ obviously dominates the spectrum (Fig-
ꢀ
ꢀ
the CN group in the [DCA] ion ranging from 2134–
2236 cmꢀ1. The spectrum of [S4]DCA shows (Figure 3a)
three strong absorptions of the [DCA]ꢀ ion at 2135, 2195,
and 2235 cmꢀ1. Moreover, the absorptions in the fingerprint
region (523 and 663 cmꢀ1 in Figure 3a) were assigned to the
bending vibration resulting from the [DCA]ꢀ ion.
ESI-MS: Although ILs are always denoted as liquid salts
composed entirely of ions, there is clear experimental evi-
dence showing that a lot of cation–anion pairs and larger
charged and neutral ion clusters exist in most ILs.[28,36] Un-
doubtedly, knowledge of the electrostatic interactions and
supramolecular assembly behavior of ILs by atmospheric-
pressure chemical ionization mass spectrometry (APCI-
MS)[37] atmospheric-pressure thermal desorption ion mass
spectrometry (APTDI-MS),[38] and electrospray ionization
mass spectrometry (ESI-MS),[39] will be of great value for
understanding intrinsically the ionic nature of ILs, and is
therefore fundamental to understanding their unique prop-
erties. Herein, experimental evidence resulting from ESI-
MS studies is provided, giving important information about
the structural organization of these cyclic sulfonium salts.
Dilute methanol solutions of the ILs were used for ESI-MS
and two representative positive-ion spectra are provided in
Figure 4. For [S1]DCA, in addition to the observation of the
parent cation [S1]+ (m/z 103.0579), positively charged ion
clusters (m/z 272.1158 for [(S1)2DCA]+, m/z 441.1627 for
ure 4b). In addition, the larger cluster [(S4)5ACTHNUTRGNEUNG
(NTf2)4]+ (m/z
1844.7584) was also detected, although its signal was very
weak in comparison with those of other smaller ion clusters.
Combining the ESI-MS results with the fact that these cyclic
sulfonium [DCA]ꢀ and [NTf2]ꢀ salts have lower viscosities
and high conductivities (detailed discussions will be provid-
ed in the following sections), we think that they may be
forming supramolecular clusters of lower nuclearity (in both
positive and negative modes), as observed for the imidazoli-
um salts in solution[39] or even in the gas phase.[37a,38]
UV/Vis absorptions and fluorescence behavior: The UV/Vis
absorptions of the new ILs were also studied, and represen-
tative spectra of four ILs (0.1 wt% IL in ethanol solution)
were selected for comparison. Although it is not shown in
Figure 5, for a given anion the ILs exhibited similar absorp-
tion characteristics, a feature attributed to the fact that the
sulfonium cations contributed only extremely weak absorp-
tions in the UV regions due to their lack of chromophore.
The UV/Vis absorptions of the precursors [S1]I, [S2]I, and
[S4]I also proved this point.
[(S1)3ACHTUNGTRENNUNG 4ACHTUNGTRENNGUN
(DCA)2]+, and m/z 610.1998 for [(S1) (DCA)3]+) and
Moreover, according to the
absorbance spectra, the absorp-
tion intensity decreased in the
anionic order: [Sac]ꢀ > [NO3]ꢀ
>
[DCA]ꢀ [NTf2]ꢀ. Since
>
the absorption in the UV
region is related to the excita-
tion of p electrons, this means
that the [Sac]ꢀ ion should be
more easily excited than other
anions under the same condi-
tions. In Figure 5, although only
0.1 wt% [S2]Sac was dissolved
in ethanol, the strong absorp-
tion of the solution can be ob-
served below 300 nm, and the
absorption tail extends beyond
Figure 4. The positive-ion ESI mass spectra of [S1]DCA and [S4]NTf2.
768
www.chemeurj.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 765 – 778

Cyclic Sulfonium-Based Ionic Liquids
FULL PAPER
curred at 279 nm, much lower than those of other studied
ILs. As for the three ILs with the [DCA]ꢀ ion, relatively
strong emissions could be observed with excitation band
ranging from 370–385 nm, and the maximum excited wave-
lengths of [S1]DCA, [S2]DCA, and [S4]DCA were 384 nm,
371 nm, and 373 nm, respectively. In addition, the fluores-
cence intensities of ILs with the [DCA]ꢀ ion were weaker
than those of ILs with the [NTf2]ꢀ or [Sac]ꢀ ions, which
could be caused by the stronger p-conjugated system of
[Sac]ꢀ or [NTf2]ꢀ than [DCA]ꢀ.
Moreover, the fluorescence behavior of pure ILs is strong-
ly dependent on the excitation wavelength, as seen in
Figure 7. For [S2]DCA, the fluorescence became very weak
Figure 5. UV/Vis spectra of the ethanol solutions containing 0.1 wt% ILs.
a)[S2]Sac, b) [S4]NO3, c) [S2]DCA, d) [S4]NTf2.
330 nm. However, almost no absorptions in UV/Vis regions
occurred for the ethanol solution containing 0.1 wt%
[S4]NTf2, and the corresponding absorbance was less than
4.0 even when neat [S4]NTf2 was employed, indicating that
the absorption of this IL was comparatively weak. In com-
parison with the absorptions of [S2]Sac and [S4]NTf2, the
absorption intensity of [DCA]ꢀ-based ILs was moderate,
that is, weaker than that of [S2]Sac and stronger than that of
[S4]NTf2.
Figure 6 gives the emission spectra of six selected ILs
(neat) at their maximum excited wavelengths. The results
show that the fluorescence behavior of these sulfonium-
Figure 7. Excitation wavelength-dependent emission behavior of neat
[S2]DCA. lexc (nm)=340 (a), 360 (b), 380 (c), 400 (d), 420 (e), 440 (f),
460 (g).
or could not be observed when the excited wavelengths
were below 320 nm or over 480 nm. The fluorescence ap-
peared at longer wavelength and became stronger in intensi-
ty with its emission band centered at 370 nm, and then the
intensity decreased quickly with the maximum emission
shifted to 491 nm from 406 nm as the excited wavelengths
increased to 460 nm from 340 nm. Also, it is worth noting
that no fluorescence was observed for any of the IL precur-
sors (i.e., [S1]I, [S2]I, and [S4]I), whereas a strong intensity
of >6000 for NaSac, NaDCA, and LiNTf2 was observed and
only weak fluorescence was exhibited for NaNO3. This indi-
cates that the fluorescence behaviors of all sulfonium-based
ILs are attributable mainly to the contribution of the anion
and are independent of the sulfonium cations.
Figure 6. Emission behavior of six ILs at their maximum excited wave-
lengths, lexc (max, nm)= a) 279, [S4]NTf2; b) 350, [S4]Sac; c) 373,
[S4]DCA; d) 371, [S2]DCA; e) 384, [S1]DCA; f) 373, [S2]NO3.
Surface property: Surface property studies of ILs are partic-
ularly important for understanding many mechanisms in-
volving interfacial chemistry of ILs, for example, biphasic
homogeneous catalysis or supported IL phase catalysis. In
comparison with the physicochemical “bulk” properties of
ILs, XPS studies on ILs are relatively limited, but it has
been proven that this technique is a powerful tool in the sur-
face science of ILs.[40–42] Therefore, surface studies of these
cyclic sulfonium-based ILs utilizing XPS were also per-
based ILs is strongly dependent on the anion and there is
only a little change resulting from the increase in the alkyl
chain length of the cations. For four ILs with the [S4]+ ion,
the fluorescence intensity decreased in the order [S4]Sac >
[S4]NTf2 > [S4]DCA, and the strongest fluorescence of
[S4]Sac was observed at 426 nm when the excited wave-
length was 350 nm. The maximum excitation of [S4]NTf2 oc-
Chem. Eur. J. 2009, 15, 765 – 778
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
769

Y. Deng et al.
formed, and representative spectral information from two
ILs ([S4]NTf2 and [S2]DCA) is presented. Under our analy-
sis conditions, both sulfonium-based ILs showed well-de-
fined characteristic emissions, and no obvious degradation
of the ILs was observed during or after measuring, which in-
dicated that these sulfonium-based ILs are very stable
during exposure to an X-ray source and ultra-high vacuum
conditions.
energies: one peak at 292.02 eV originating from CF3 in the
[NTf2]ꢀ ion, and another peak at 285.00 eV, which can be as-
signed to eight carbon atoms in the [S4]+ ion. Although
there are at least three different chemical shifts for the eight
carbon atoms in the cation, considering their similar binding
energies, we are reluctant to make any further interpretation
of the C 1s signal at 285.00 eV. According to the XPS inten-
sities of various elements, we deduced the atomic ratio of
five elements as follows: N:S:O:F:C = 1:3.1:4.7:6.4:11.3
(the theoretical ratio is 1:3:4:6:10). Clearly, the intensities of
C and O are higher than the theoretical values, which may
be caused by trace organic contaminations introduced
during synthesis. In addition, the presence of Si-containing
surface contaminants might also contribute to the excessive
oxygen concentration,[43] although no Si signal was detected
(possibly <0.3 wt%). The enrichment of F in the IL surface
can be explained by the facts that fluorine has a high rela-
tive sensitivity factor (RSF=1), and that the atoms of fluo-
rine-containing anions involved in emissive processes are
not affected by shake-up/off events.[44]
Figure 8 shows the overview scan of a [S4]NTf2 film and
the high-resolution spectra for S 2p and C 1s. From the
survey spectrum in Figure 8a, the expected elements F, N,
Figure 9a gives the XPS spectra of [S2]DCA, which are
free from O and F. Besides the IL-related signals (C, S, and
N), we observed trace Si-containing contaminants, which
Figure 8. XPS spectra of [S4]NTf2: survey spectrum (a) and high-resolu-
tion spectra (b, c) dealing with the S 2p, O 1s, and C 1s photoemission.
O, S, and C were all detected, and no evidence of impurities
of other elements was found, indicating that the surface con-
centration of other impurities was below the detectable
level for XPS. The electron emissions from N 1s, O 1s, F 1s,
S 2p, and C 1s emissions all displayed well-resolved and
narrow spectral features. For F 1s, N 1s, and O 1s, their 1s
core level spectra were fitted with only one peak at the
binding energies (B.E.) of 688.0 (F), 398.58 (N), and
531.8 eV (O), respectively. For sulfur, the S 2p signals origi-
nate from the two chemically different types of sulfur atoms
in [S4]NTf2, one from the cation and another from the
anion. This is clearly reflected in the S 2p spectrum (Fig-
ure 8b). Furthermore, each S 2p signal can be fitted with
Figure 9. XPS spectra of [S2]DCA: survey spectrum (a) and high-resolu-
tion spectra (b, c) dealing with the S 2p and N 1s photoemission.
might originate from the sealing of the storage container or
the silicone grease introduced during synthesis.[43,44] The S 2p
spectrum of [S2]DCA is shown in Figure 9b. The B.E. of
S 2p3/2 is 165.38 eV, which is assigned to the sulfur atom in
the [S2]+ ion. The high-resolution profile of the N 1s peak is
shown in Figure 9c. It is clear that this photoemission con-
sists of two resolved peaks, one peak at 398.97 eV for the ni-
trogen atom with the negative charge, and one peak at
two components due to the contribution of 2p1/2 and 2p3/2
,
respectively. The B.E. for the two types of S 2p peaks are
165.40 eV for S 2p in the [S4]+ ion and 168.14 eV for S 2p in
the [NTf2]ꢀ ion. The atomic ratio of S and N is 3.1:1 (the
theoretical ratio is 3:1), suggesting that the actual distribu-
tion of the [S4]+ ion and the [NTf2]ꢀ ion at the IL surface is
close to 1:1. As shown in the C 1s spectrum (Figure 8c), at
least two peaks are clearly distinguishable by their binding
ꢀ
397.69 eV for the nitrogen atom of the CN group in
[DCA]ꢀ. In addition, the B.E. of O 1s in [S2]DCA is
770
www.chemeurj.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 765 – 778

Cyclic Sulfonium-Based Ionic Liquids
FULL PAPER
531.68 eV, suggesting that it is very close to the chemical
state of the O species in [S4]NTf2 (B.E. is 531.8 eV). Al-
though the nature of the impurities is not clear yet, it is
likely that the excessive O and C signals in the XPS of
[S2]DCA also originate in part from organic contamination
introduced during synthesis. Similarly, the XPS spectra for
other cyclic sulfonium-based ILs give well-defined character-
istic photoemissions, indicating that XPS can indeed give
large amounts of surface information for these novel liquids.
ate, are collected in Table 1. In general, four types of phase
transition behavior were observed for these cyclic sulfoni-
um-based salts. Typical DSC traces for these four types of
behavior are shown in Figure 10.
The first type of behavior is characterized by [S1]DCA,
which has a single melting transition (Figure 10a, entry 1 in
Table 1). The second type of behavior is represented by four
salts that exhibited one or two solid–solid transitions (Ts–s)
before melting (Figure 10b and c, entries 4, 5, 10, and 12 in
Table 1). For example, two solid–solid transitions were ob-
served for [S1]NTf2 at ꢀ148C and 168C, respectively, before
final melting at 988C, indicating a progressive transforma-
tion from a fully ordered state through a series of increas-
ingly disordered phases. These solid–solid transitions, which
are characteristic of plastic crystal, result from dynamically
disordered orientations of IL constituents about the molecu-
lar axis.[8,45] Similarly, the salt [S2]NTf2 also displayed the
solid–solid transition, but only a single solid–solid transition
could be observed before melting. The third type of behav-
ior is represented by the salts that exhibited only a glass
transition in a heating and cooling cycle without melting
(entries 2, 3, and 6–9 in Table 1). From the DSC traces of
this type, an obvious enthalpy relaxation during the glass
transition could usually be observed (Figure 10d), indicating
that these salts were not under thermodynamic equilibrium
below Tg and therefore showed an endothermal phenomen-
on during glass transition to relax to equilibrium. The fourth
type of behavior was observed only in [S2]NO3 (Figure 10e).
Here, only a glass transition occurred at ꢀ808C during cool-
ing from 1008C to ꢀ1008C. However, upon heating, the salt
first passed from the glass state to a subcooled phase at Tg,
and then a cold crystallization occurred at ꢀ358C, followed
by a preliminary melting at ꢀ58C and finally a full melting
transition at 208C. In general, this phase behavior is depen-
dent upon many factors and therefore is less predictable.
Phase behavior, melting point, and glass transition tempera-
ture: The solid–solid or solid–liquid phase transitions for
these sulfonium salts were investigated by differential scan-
ning calorimetry (DSC), and the data for melting points
(Tm), freezing points (Tf), glass phase transition (Tg), solid–
solid transition (Ts–s), and entropy change (DSm), if appropri-
Table 1. Physical and thermal properties of cyclic sulfonium-based ILs.
[a]
[c]
[d]
[e]
[f]
Entry Salts
Tg
Tf[b]
Ts–s
Tm
Cp
Td
[8C] [8C]
[8C]
[8C]
[Jgꢀ1 Kꢀ1
]
[8C]
1
2
3
4
5
6
7
8
9
[S1]DCA
ꢀ19
13
2.00
1.83
1.87
2.08[h]
1.73
1.60
2.01
1.89
1.86
2.01
1.59
1.83[h]
179
189
187
308
298
283
161
159
168
179
172
171
[S2]DCA ꢀ88
[S4]DCA ꢀ80
[S1]NTf2
85
51
ꢀ14, 16
ꢀ19
98
69
[S2]NTf2
[S4]NTf2
[S1]Sac
[S2]Sac
[S4]Sac
[S1]NO3
[S2]NO3
[S4]NO3
ꢀ82
ꢀ43
ꢀ53
ꢀ33
10
11
12
108
ꢀ29, ꢀ6 116
ꢀ80 ꢀ15[g]
18
54
23
ꢀ35, 20
[a] Glass-transition temperature. [b] Freezing point. [c] Solid–solid transi-
tion. [d] Melting point. [e] Heat capacity at 258C. [f] Thermal decomposi-
tion temperature. [g] Cold-crystal temperature. [h] The sample was mea-
sured at 308C.
Figure 10. Representative DSC traces recorded at a rate of 108Cminꢀ1 in the heating process. a) [S1]DCA: single melting point (Tm); b) [S2]NTf2: single
solid-solid transition (Ts–s) before melting (Tm); c) [S1]NTf2: two solid-solid transition (Ts–s) before melting (Tm); d) [S2]Sac: only glass transition (Tg); e)
[S2]NO2: glass transition (Tg) followed by crystallization (Tc), and melting (Tm).
Chem. Eur. J. 2009, 15, 765 – 778
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
771

Y. Deng et al.
Generally, the melting point (Tm) of an IL is determined
mainly by the following three factors: molecular symmetry,
intermolecular forces, and conformational degrees of free-
dom of the molecule.[46] For the twelve salts presented in
this paper, except for [S1]NO3 (Tm =1168C) all the salts
have a melting point of <1008C (some salts only displayed
a glass transition and had no melting point). Obviously, the
structure variation in the cation and anion has some impact
on the melting point of these sulfonium salts. For a given
anion, the Tm values generally increase in the order [S2] <
[S4] < [S1], indicating that a medium cationic size might be
beneficial to reducing ion symmetry and increase the ionꢁs
conformational degrees of freedom. In the family of salts
containing the same cation, the Tm values are in the follow-
ing anion order: [DCA]ꢀ < [NTf2]ꢀ < [NO3]ꢀ. Since all
three salts with the [Sac]ꢀ ion had no melting point during
the temperature increase, the effect of the [Sac]ꢀ ion on Tm
could not be compared with the salts based on other anions.
From the above discussions, it can also be concluded that a
medium cationic size combined with low symmetry and ef-
fective charge distribution in the anion favor destabilization
of the packing efficiency in the crystal lattice of the ILs,
thus decreasing their Tm values. However, no clear correla-
tions are observed between the structures of these salts and
their melting points.
capacities (^2.0 Jgꢀ1 Kꢀ1) than those with other anions. For
example, the heat capacities of [S1]Sac and [S1]NO3 are
both 2.01 Jgꢀ1 Kꢀ1, which is even higher than thermal oil
(1.907 Jgꢀ1 Kꢀ1).
Moreover, the thermal stability of the salts was investigat-
ed by thermogravimetric analysis (TGA). As shown in
Table 1, the temperatures of thermal decomposition (Td, i.e.,
the temperatures at which the weight loss of a compound in
an aluminum pan under an N2 atmosphere reached 5 wt%)
of the twelve salts employed in this work are in the range of
159–3088C, which are lower than for 1,3-dialkylimidazoli-
um-based ILs (in general, Td >3008C), and even lower than
for the structurally related pyrrolidinium salts such as
[P1,2]DCA (Td =2508C).[47] As shown in Figure 11, the salts
The glass transitions of the new sulfonium-based ILs were
also investigated. As seen in Table 1, seven sulfonium salts
exhibited a glass transition, and the Tg values ranged from
ꢀ88 to ꢀ338C. The structure of the anion clearly affected
the Tg values of ILs. Among the developed ILs, the Tg
values of the salts with the [DCA]ꢀ or [NTf2]ꢀ ion are all ꢁ
ꢀ808C, much lower than for those with the large [Sac]ꢀ ion
(Tg =ꢀ53 to 338C). This can be explained by the results de-
rived from the bulky, non-fluorinated, rigid, and polarizable
[Sac]ꢀ ion, that is, the relatively large size, high polarizabili-
ty, and rigid structure as well as the non-fluorinated nature
of [Sac]ꢀ ion enhances possible hydrogen bonding interac-
tions between cation and anion, which in turn increases the
Tg of the ILs. On the other hand, for ILs with a [Sac]ꢀ ion,
the Tg values generally increase in the cationic order [S2] <
[S1] < [S4], which is different from the trend observed for
the effect of cationic structure on Tm. This also indicates
that the influence of cationic structure on Tm comes from
many factors, such as the strength of ion interactions, the
ion size, as well as the flexibility and polarizability of the
anion, and a medium cationic size might be helpful to de-
crease the Tg of ILs by means of weakening electrostatic in-
teractions between cation and anion. For a given anion, no
clear correlations are observed between the structure of the
cation and the glass transition temperature.
Figure 11. TGA traces of six cyclic sulfonium-based salts.
containing [NTf2]ꢀ exhibited much higher thermal stability
than those with the other three anions (i.e., [DCA]ꢀ, [Sac]ꢀ,
and [NO3]ꢀ). Irrespective of the cations, the thermal stabili-
ties of these salts generally increase in the order [Sac]ꢀ
<
[NO3]ꢀ < [DCA]ꢀ < [NTf2]ꢀ. This trend indicates that the
anion has a significant impact on the thermal stability of the
sulfonium salts.
Density, viscosity, and refractive index: For the salts that are
liquids at 258C, the fundamental properties including densi-
ties (1), viscosities (h), and refractive indices (n) were also
measured, and the corresponding characterization data are
presented in Table 2. The 1 values of eight RTILs lie in
range of 1.098–1.472 gcmꢀ3. For a given anion, the densities
of these cyclic sulfonium-based ILs decreased gradually with
increasing formula weight (Mw) of the cation, which is at-
tributed to the relatively smaller van der Waals volume of
the cation along with the increase of the alkyl chain length.
For the salts with the same cation, the 1 values decrease in
the following anion order: [DCA]ꢀ < [Sac]ꢀ < [NO3]ꢀ.
The ILs based on [DCA]ꢀ displayed relatively lower densi-
ties than ILs with other anions, and [S4]DCA gave the
smallest 1 value (1.098 gcmꢀ3 at 258C), which is still slightly
higher than the structurally related pyrrolidinium salts (e.g.,
1=0.92, 0.95, and 0.92 gcmꢀ3 for [P13]DCA, [P14]DCA, and
Heat capacity and thermal stability: The heat capacities (Cp)
of ten ILs at 258C were measured (Table 1), and the values
range from 1.59–2.08 Jgꢀ1 Kꢀ1. Since a solid–solid transition
at about 258C occurred for [S1]NTf2 and [S4]NO3, their
heat capacities were not measured. The data in Table 1 indi-
cate that the ILs based on the [S1] cation have higher heat
772
www.chemeurj.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 765 – 778

Cyclic Sulfonium-Based Ionic Liquids
FULL PAPER
Table 2. Density (1), viscosity (h), and refractive index (n) of eight sulfo-
nium-based ILs.
The temperature dependence of viscosity was investigated
for five ILs over the temperature range ꢀ20 to 808C, and
the Arrhenius plots of viscosity according to Equation (1)
are shown in Figure 12.
ILs
Water 1[a]
content
[gcmꢀ3
[ppm]
Viscosity (h):ln h=ln A + Eh/RT n[c]
U
]
ACHTUNGTRENNUNG
G
H[b]
Eh
R2
A
] ACHTUNGRTENNUGN
lnh ¼ ln A þ Eh=RT
ð1Þ
[S1]DCA 248.1
[S2]DCA 297.2
[S4]DCA 292.2
[S4]NTf2 50.3
1.169
1.135
1.098
1.456
1.423
49.8
34.2
62.6
104.8
>
25.20
25.23
31.66
36.44
–
16.3
14.5
1.68
0.39
–
0.995 1.5502
0.991 1.5426
0.988 1.5315
0.991 1.4420
[S1]Sac
[S2]Sac
[S4]Sac
317.2
259.4
287.3
–
–
–
1.5940
1.5795
1.5785
2000
>
1.350
1.265
1.472
–
–
–
2000
>
–
2000
[S2]NO3 384.7
156.8[d] 31.27
6.12
0.997 1.5401
[a] Density at 258C. [b] Viscosity at 208C. [c] Refractive index at 258C.
[d] The sample was measured at 308C.
[P16]DCA at 258C, respectively).[47] Because the effect of
temperature on the ILꢁs density is not as obvious as for the
viscosity and conductivity, the temperature dependence of
density was not investigated in this work.
Figure 12. Arrhenius plot of viscosity for five sulfonium salts.
The viscosities of the eight RTILs were measured. Except
for the ILs with a [Sac]ꢀ ion, most cyclic sulfonium-based
ILs showed relatively low viscosities (34.2–104.8 mPas at
208C, Table 2), much lower than those of traditional 1,3-dia-
lkylimidazolium [BF4]ꢀ and [PF6]ꢀ salts (e.g. h=112 mPas
for [BMIm]BF4 and h=201 mPas for [BMIm]PF6 at
208C).[48] [S2]DCA had the lowest viscosity of all salts under
investigation (h=34.2 mPas at 208C), even lower than for
the pyrrolidinium [DCA]ꢀ salts (h=45, 50, and 45 mPas for
[P13]DCA, [P14]DCA, and [P16]DCA at 208C, respective-
ly),[47] whereas the ILs based on the [Sac]ꢀ ion displayed
high viscosities of >2000 mPas at the same temperature.
Moreover, the relationships between ion structure and vis-
cosity were also noticeable. For a given anion, the viscosity
decreases in the order [S4]+ > [S1]+ > [S2]+, for example,
the viscosities of [S4]DCA, [S1]DCA, and [S2]DCA are
62.6, 49.8, and 34.2 mPas, respectively. In comparison with
the inconspicuous impact induced by cationic variation, the
effect of the anion on the viscosity is remarkable. Obviously,
the [DCA]ꢀ ILs had much lower viscosities than the ILs
with other anions. For example, the h value of [S4]DCA is
only 62.6 mPas at 208C, but the h value of [S4]NTf2 reaches
104.8 mPas at the same temperature. The low viscosities of
the [DCA]ꢀ ILs may be explained by the increased ion mo-
bility owing to smaller anion size and better charge delocali-
zation in [DCA]ꢀ than other anions. Although the [Sac]ꢀ
ion also possesses good charge distribution, extremely high
viscosities (>2000 mPas at 208C) were observed for all
[Sac]ꢀ salts, which could be attributed to high rigidity (com-
prising two conjugated rings) and strong ion interactions
(mainly van der Waals interactions induced by the C=O
group and two S=O groups), and the larger size of the
[Sac]ꢀ ion.
Eh is the activation energy for viscous flow, which is always
regarded as the energy barrier to be overcome by ions to
move past each other, and thus its value is correlated with
the structure of both cation and anion. The values of Eh, A,
and the linear fitting parameter (R2) for all the ILs are cal-
culated and listed in Table 2. According to the values of R2
in Table 2, the five salts were approximately fit by the Ar-
rhenius model over the temperature range ꢀ20 to 808C. In
comparison with the [S4]+ salts, [S1]DCA, and [S2]DCA
gave relatively smaller Eh values (25.20 and 25.23 kJmolꢀ1,
respectively), indicating that both ILs have a lower energy
barrier and thus exhibit lower viscosity.
The refractive indices (n) of eight sulfonium-based salts at
258C ranged from 1.4420–1.5940 (Table 2), higher than
those of the traditional dialkylimidazolium [BF4] and [PF6]
salts (e.g., n=1.4188 for [BMIm]BF4 and n=1.4083 for
[BMIm]PF6 at 258C).[49] Similarly to the viscosity, the struc-
ture variation in the cation and anion has some impact on
the refractive indices of ILs. In general, for the same cation,
the high refractive indices of the ILs reflect particularly high
electron mobility in the anion. Among the studied ILs, the
refractive indices decreased in the order [Sac]ꢀ > [DCA]ꢀ
> [NO3]ꢀ > [NTf2]ꢀ, and [S1]Sac gave the highest refrac-
tive index (1.5940). The cationic structure also influences
the refractive indices of the salts. For the same anion, the re-
fractive indices increase with increasing chain length of the
cations. The effect of temperature on the refractive index
was also studied (Figure 13), and the n values decreased lin-
early with increasing temperature. As the curve shows, the
attenuation slope of the eight lines is approximate. This
means that although the cationic or anionic structures of the
eight salts were different, the effect of the temperature on
the refractive index was almost the same.
Chem. Eur. J. 2009, 15, 765 – 778
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
773

Y. Deng et al.
Conductivity and electrochemical stability: The conductivity
of an IL is of vital importance if it is to be considered as a
supporting electrolyte in electrochemical devices. According
to the conclusions drawn by Grꢂtzel et al.,[6] the ionic con-
ductivity (k) of an IL should be related to its viscosity (h),
formula weight (FM), and density (1), and the radii of its
cation and anion (rc and ra), as described by Equation (2), in
which 0<y<1 is the degree of dissociation, F is the Faraday
constant, za and zc are the anion and cation microviscosity
factors, respectively.
k ¼ yF21=ð6pNAFWhÞ½ðzaraÞꢀ1 þ ðzcrcÞꢀ1Þꢂ
ð2Þ
Figure 13. Refractive indices of eight ILs as a function of temperature.
Therefore, besides the obvious influence of the viscosity,
the effect of ion size and weight must be stressed. In the
search for highly conductive ILs, one must not focus on vis-
cosity only, but keep in mind the importance of ion size.
Figure 14 shows the temperature dependence of the specific
Tribological properties: The ILs were found to possess ex-
cellent tribological properties, and thus could be used as val-
uable lubricants. However, very few studies on the tribologi-
cal performance of sulfonium-based ILs have been reported
so far. Many S-containing compounds, such as MoS2, have
been proven to be good solid lubricants. Therefore, the tri-
bological performance of novel cyclic sulfonium-based ILs
as lubricants for various sliding contacts was also evaluated
in this work.
For the purpose of comparison, a typical IL [BMIm]NTf2
was also employed. From the test results of friction and
wear of various ball-on-disc systems lubricated with four
studied ILs as lubricants (Table 3), it is found that the struc-
Table 3. Friction coefficients for different frictional pairs lubricated with
four ILs.[a]
Frictional
pair
Friction coefficient
(ball/disk)
[S2]Sac
[S4]NTf2
G
[S4]DCA
Figure 14. Change of ion conductivity with increasing temperature for
new ILs.
steel/steel
steel/Al
steel/Cu
Si3N4/Cu
0.15
0.16
0.13
0.20
0.13
0.14
0.15
0.13
0.15
0.13
0.12
0.11
conductivities (k) for four representative ILs based on
[DCA]ꢀ and [NTf2]ꢀ in the temperature range from ꢀ30 to
808C. At 20 8C, the ionic conductivities of four examined
ILs range from 3.2 to 17.6 mScmꢀ1 (Table 4). [S2]DCA dis-
[a] SRV tester, load 200 N, frequency 25 Hz, amplitude 1 mm.
ture of the IL has some impact on its tribological perfor-
mance. In comparison with [BMIm]NTf2, [S4]NTf2 showed
much lower friction coefficients; that is, [S4]NTf2 has better
tribological properties than [BMIm]NTf2. For [S4]DCA,
much lower friction coefficients than [BMIm]NTf2 were ob-
tained in steel/Cu and Si3N4/Cu tests, but only similar tribo-
logical performance to [BMIm]NTf2 was observed in steel/
steel and steel/Al tests. As for [S2]Sac, poor wearing perfor-
mance was observed in steel/steel, steel/Al, and Si3N4/Cu
tests. However, the friction coefficient of [S2]Sac in the
steel/Cu test was 0.13, much lower than the 0.22 for
[BMIm]NTf2, that is, [S2]Sac showed better tribological
properties than [BMIm]NTf2 when tested for the steel/Cu
system, although the detailed mechanism is still not clear at
this stage.
Table 4. Electrochemical properties of four selected ILs with low viscosi-
ties
ILs
Reductive
voltage [V] voltage [V] window [V]
Oxidative
Electrochemical Conductivity[a]
[mScmꢀ1
N
]
[S1]DCA ꢀ1.8
[S2]DCA ꢀ1.8
[S4]DCA ꢀ1.8
1.4
1.4
1.4
2.0
3.2
3.2
3.2
3.8
13.6
17.6
7.6
[S4]NTf2
ꢀ1.8
3.2
[a] Conductivities were measured at 208C.
played the highest conductivity (17.6 mScmꢀ1 at 208C), even
higher than that of [EMIm]BF4 (14 mScmꢀ1 at 208C).[49]
From Figure 14, it is clear that the conductivities were more
774
www.chemeurj.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 765 – 778

Cyclic Sulfonium-Based Ionic Liquids
FULL PAPER
affected at high temperature. The reason might be that at
low temperatures the k value is strongly affected by viscosi-
ty, and at high temperature the intrinsic structure and char-
acteristics of the ILs may determine the k value. Similarly to
the viscosity, the structure of cation and anion has some
effect on the conductivities. For the same anion, the conduc-
Conclusion
A novel family of cyclic sulfonium-based ILs has been pre-
pared and characterized. Their physicochemical properties,
including molecular vibrations, UV/Vis absorptions, fluores-
cence behavior, ion aggregation behavior, surface properties,
thermal properties, densities, viscosities, refractive index, tri-
bological property, and electrochemical properties, have
been studied in detail. Among twelve developed ILs, eight
of them are liquids at room temperature. The [DCA]ꢀ-based
ILs exhibit much lower viscosities (34–62 mPas at 208C)
than most 1,3-dialkylimidazolium [BF4]ꢀ and [PF6]ꢀ salts (>
100 mPas at 208C). Four of the ILs exhibit organic plastic
crystal behavior, and to our knowledge, no example of IL
materials based on sulfonium salts has been previously re-
ported to possess this characteristic. [S4]NTf2 has almost no
absorptions in the 1352–2880 cmꢀ1 IR region and only weak
absorption in the whole UV region. ILs based on the [Sac]ꢀ
ion display high refractive indices (1.5785–1.5940 at 208C)
owing to the high electron mobility of [Sac]ꢀ. In addition,
some ILs with the [NTf2]ꢀ and [Sac]ꢀ ion show better tribo-
logical performance than [BMIm]NTf2 in the steel/Cu wear-
ing test, and those [DCA]ꢀ-based ILs show much higher
conductivities (7.6–17.6 mScmꢀ1 at 208C) than most tradi-
tional 1,3-dialkylimidazolium salts. Hence, these new sulfo-
nium salts could be promising electrolytes for electrochemi-
cal devices and good potential lubricants.
tivity values of anions are in the following order: [S4]+
<
[S1]+ < [S2]+. This is contrary to the viscosity order of cat-
ions; for a given cation, the [DCA]ꢀ ILs have much higher
conductivities than the ILs with other anions, for example,
at 208C [S4]DCA exhibits a conductivity of 7.6 mScmꢀ1, but
the k value of [S4]NTf2 is only 3.2 mScmꢀ1. This can be ex-
plained by the lower viscosity and smaller anion size of the
former, which can improve the rate of ion mobility.
The electrochemical stability of most of the ILs was eval-
uated by cyclic voltammetry. Figure 15 shows a typical cyclic
voltammetry curve for [S4]NTf2. Compared to the conven-
Experimental Section
General: Commercially available regents were purchased from Aldrich,
Tianjin Chemical Reagent Corporation Ltd., or Alfa Aesar Inc., were of
analytical grade and were used in this work as received.
Figure 15. Cyclic voltammogram of [S4]NTf2; GC working electrode.
NMR, FTIR, and ESI-MS: Before NMR, FTIR and ESI-MS analyses,
each sample was obtained by filtration through a silica column, and dried
at 90–1008C and 10ꢀ2–10ꢀ3 mbar for 4 h. 1H NMR spectra were recorded
on a Bruker AMX FT 400 MHz NMR spectrometer. Chemical shifts
were reported downfield in parts per million (ppm, d) from a tetrame-
thylsilane reference. IR spectra were recorded on a Thermo Nicolet 5700
FTIR spectrophotometer. Electrospray ionization mass spectra were re-
corded on a Bruker Daltonics APEX II 47e FTMS. The samples were
dissolved in methanol.
tional 1,3-dialkylimidazolium-based ILs (e.g., EW=3.5 V
for [EMIm]DCA and 4.1 V for [EMIm]NTf2),[10b,27a] these
cyclic sulfonium-based salts exhibited lower electrochemical
stabilities, and their electrochemical windows (EWs) ranged
only from 3.2–3.8 V (Table 4). For [S1]DCA, [S2]DCA, and
[S4]DCA, an approximate reductive potential was observed
at ꢀ1.8 V, which meant that the electrochemical stabilities
of the cations were influenced very little by the alkyl substi-
tution pattern in the ILꢁs cation. Moreover, by comparing
the oxidative potentials of [S4]NTf2 and [S4]DCA, it was
found that the oxidative voltage of 2.0 V for [NTf2]ꢀ was ob-
viously higher than 1.4 V for [DCA]ꢀ, indicating that the
UV/Vis, fluorescence, and XPS: UV/Vis spectra were conducted on an
Agilent 8453 UV/Vis spectrophotometer, and the fluorescence spectra
were recorded at room temperature on a Hitachi model F-7000 FL spec-
trophotometer at a scan speed of 1200 nmminꢀ1. Before both UV/Vis
and fluorescence analysis each sample was dried at 90–1008C and 10ꢀ2
–
10ꢀ3 mbar for 4 h. The sample were contained in 1 cm quartz cuvettes
during measurements, and were measured in open ambient conditions
within 1–3 min. X-Ray Photoelectron Spectroscopy (XPS) analyses were
performed on
a VG ESCALAB 210 instrument with MgKa source
ꢀ
[DCA]ꢀ ion was more easily oxidized than the NTf2 ion.
(1253.6 eV) and calibrated versus the C 1s peak at 285.0 eV. A thin layer
of IL was deposited on a polycrystalline gold substrate, and was kept
under moderate vacuum for at least 12 h before introduction into the an-
alytical chamber of the XPS instrument. Spectrometer pass energies of
100 eV for the survey spectra and 30 eV for high resolution spectra were
used for all elemental spectral regions. The pressure in the analyser
chamber was 10ꢀ9 Torr.
Moreover, one remarkable characteristic for [S4]NTf2 is that
no obvious oxidation peak appeared at 2.0 V (i.e. the oxida-
tive potential of the anion, Figure 15), which can be ex-
plained by the passivation of the anodic electrode induced
by the trace S-containing contaminations during cyclic vol-
tammetry. Similar phenomena could also be observed for
the electrochemical behaviors of three [DCA]ꢀ ILs.
Water content: The water content in ILs was detected by means of a cou-
lometric Karl–Fischer titration using a Mitsubishi CA-06 Moisturemeter,
and each test was conducted in an anhydrous sealed chamber and was
Chem. Eur. J. 2009, 15, 765 – 778
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
775

Y. Deng et al.
finished within 1 min. Duplicate measurements were performed on each
sample with results agreeing to within 5%. Before the water content test,
each sample were kept at 908C and 10ꢀ2–10ꢀ3 mbar for 4 h, except for
the ILs based on the [NTf2]ꢀ ion, which were dried at 1008C.
General procedure for cyclic sulfonium iodides: The desired cyclic sulfo-
nium iodide precursors were synthesized according to an approach simi-
lar to that reported previously.[32,33,35] A solution of tetrahydrothiophene
(44.0 g, 0.50 mol) in dry acetone (200 mL) was prepared. A small excess
of alkyl iodide (0.55 mol) was added slowly and the resulting mixture was
stirred vigorously in darkness at room temperature for several days. The
white precipitate produced was removed by filtration and the filter cake
was washed with anhydrous diethyl ether. The resulting solid was dried
in clean and dry reagent bottles under reduced pressure for 2 h. The puri-
fied product was wrapped in aluminum foil and stored in a desiccator
under dry nitrogen.
[S1]I: White solid, yield: 91%; MW, 229.96. 1H NMR (DMSO,
500 MHz): d=3.44 (m, 2H), 3.26 (m, 2H), 2.75 (s, 3H), 2.21 (m, 2H),
2.10 ppm (m, 2H); IR: n˜ =2984, 2938, 2875, 1460, 1424, 1401 cmꢀ1 (nC–H
aliphatic); MS (ESI+) calcd: m/z: 103.06 ([C5H11S]+); found: m/z:
103.0579.
[S2]I: White crystal, yield: 85%; MW, 243.98. 1H NMR (DMSO,
500 MHz): d=3.45 (m, 2H), 3.37 (m, 2H), 3.21 (m, 2H), 2.18 (m, 2H),
2.11 (m, 2H), 1.29 ppm (t, 3H); IR: n˜ =2948, 2875, 1450, 1416, 1385 cmꢀ1
(nC–H aliphatic); MS (ESI+) calcd: m/z: 117.07 ([C6H13S]+); found: m/z:
117.0734.
[S4]I: White crystal, yield: 64%; MW, 272.01. 1H NMR (DMSO,
500 MHz): d=3.47 (m, 2H), 3.37 (m, 2H), 3.18 (m, 2H), 2.22 (m, 2H),
2.11 (m, 2H), 1.64 (m, 2H), 1.36 (m, 2H), 0.88 ppm (t, 3H); IR: n˜ =2954,
2931, 2869, 1463, 1409 cmꢀ1 (nC–H aliphatic); MS (ESI+) calcd: m/z: 145.1
([C8H17S]+); found: m/z: 145.1047.
General procedure for cyclic sulfonium salts with [NTf2]ꢀ ions: The pro-
duced cyclic sulfonium iodide precursors and lithium bis(trifluoromethyl-
sulfonyl)imide (LiNTf2, ꢄ99.95%, Aldrich), were dissolved in deionized
water and mixed for 2 h at ambient temperature (a 1:1.1 molar ratio of
iodide/LiNTf2 was employed to ensure the metathesis reaction ran as
completely as possible). The crude products (solid precipitates or liquid)
were separated from water, followed by thorough washing with deionized
water until no residual iodide anions in the deionized water were detect-
ed by use of AgNO3. The final products were dried under high vacuum
for more than 4 h at 1008C, and then were wrapped in aluminum foil and
stored in a desiccator under dry nitrogen.
Phase transitions and thermal stability: Measurements of glass-transition
temperatures, melting and freezing points, and heat capacities were car-
ried out on a Mettler–Toledo differential scanning calorimeter (DSC),
model DSC822e, and the data were evaluated by using the Mettler–
Toledo STARe software version 7.01. The instrument was calibrated for
temperature and heat flow with zinc and indium reference samples pro-
vided by Mettler–Toledo. Samples were placed in a 40 mL hermetically
sealed aluminum pan with a pinhole at the top of the pan. An empty alu-
minum pan was used as the reference. The samples inside the differential
scanning calorimeter furnace were exposed to a flowing N2 atmosphere.
Before the DSC test, each sample was dried at 90–1008C and 10ꢀ2
–
10ꢀ3 mbar for 4 h, and was further dried in situ on the differential scan-
ning calorimeter by holding the sample at 1208C for 15 min. This is im-
portant because the presence of volatiles, especially water, can affect the
glass-transition and melting temperatures. Melting, crystallization, and
glass-transition temperatures were determined by cooling the samples
from 1508C to ꢀ1008C, followed by heating from ꢀ1008C to 1508C,
both at a rate of 108Cminꢀ1. The glass-transition temperature was deter-
mined as the midpoint of a heat capacity change, whereas the melting
and crystallization temperatures were determined as the onset of the
transition. The decomposition temperature (Td) was recorded with 10%
of mass loss by using a Pyris Diamond Perkin-Elmer TG/DTA with scan
rate of 208Cminꢀ1 under a N2 atmosphere, and the samples for TG/DTA
measurements were sealed tightly in Al2O3 pans. Each sample before the
TGA test was also dried at 90–1008C and 10ꢀ2–10ꢀ3 mbar for 4 h.
Measurement of the viscosity, density, and refractive index: The viscosity
of each IL was measured on a Stabinger Viscosimeter SVM 3000/GR.
The density was examined by the weight method at 258C. Measurements
of refractive indices were conducted with a WAY-2s Abbe refractometer
(Shanghai Precision & Scientific Instrument Co.), calibrated by the re-
fractive indices of deionized water. Each sample, before viscosity, density,
and refractive index analyses, underwent removal of water as before the
test of water content. For these three analyses, all the samples were mea-
sured in an open ambient atmosphere and each test at a specified tem-
perature was conducted within 1–3 min.
1
[S1]NTf2: White crystal, yield: 94%; MW, 382.96. H NMR (CD3COCD3,
Conductivity and electrochemical stability: The ion conductivity was
measured by using a Mettler–Toledo Seven Muliti meter, and the sample
underwent removal of water as before the test of water content. Similarly
to the viscosity and refractive index, the IL conductivities were measured
under open ambient conditions and each measurement at a specified
temperature was completed within 1–3 min. Cyclic voltammetry was con-
ducted by using a CHI 660A Electrochemical Work Station. The working
electrode was a glassy carbon electrode (3 mm diameter) and the auxili-
ary electrode was a platinum wire. The reference electrode was Ag/AgCl,
and porous glass frits were used to separate the internal solutions from
the working solutions and 3m KCl solution. Before testing, each sample
was kept at 908C and in a vacuum at <5 mm Hg for 1 h. Electrolysis was
500 MHz): d=3.79 (m, 2H), 3.61 (m, 2H), 3.05 (s, 3H), 2.51 (m, 2H),
2.42 ppm (m, 2H); IR: n˜ =3033, 2964, 2892, 1432 (nC–H aliphatic), 1354,
1194, 1137, 1056, 794, 741 cmꢀ1; MS (ESI+) calcd: m/z: 103.06 ([C5H11S]+);
found: m/z: 103.0579.
[S2]NTf2: White solid, yield: 93%; MW, 396.98. 1H NMR (CD3COCD3,
500 MHz): d=3.78 (m, 2H), 3.66 (m, 2H), 3.47 (m, 2H), 2.49 (m, 2H),
2.39 (m, 2H), 1.54 ppm (t, 3H); IR: n˜ =2985, 2959, 2888, 1457 (nC–H ali-
phatic), 1352, 1194, 1140, 1059, 790, 740 cmꢀ1; MS (ESI+) calcd: m/z:
117.07 ([C6H13S]+); found: m/z: 117.0734.
[S4]NTf2: Colorless liquid, yield: 89%; MW, 425.02. 1H NMR (DMSO,
500 MHz): d=3.79 (m, 2H), 3.66 (m, 2H), 3.44 (t, 2H), 2.50 (m, 2H),
2.44 (m, 2H), 1.91 (m, 2H), 1.56 (m, 2H), 0.96 ppm (t, 3H); IR: n˜ =2967,
carried out in
a conventional three-electrode electrochemical cell
2940, 2879, 1468 (nC–H aliphatic), 1352, 1194, 1140, 1059, 790, 740 cmꢀ1
;
equipped with a sealed chamber with which temperatures could be moni-
tored and adjusted from room temperature to about 908C through a cir-
cular water bath system, and vacuum treating (ꢃ1 mm Hg) and N2 filling
could also be implemented so as to minimize the influence of O2 and
H2O during experiments.
MS (ESI+) calcd: m/z: 145.1 ([C8H17S]+); found: m/z: 145.1048.
General procedure for cyclic sulfonium salts with [DCA]ꢀ, [NO3]ꢀ, and
[Sac]ꢀ ions: An aqueous solution of the cyclic sulfonium iodides was
added to aqueous slurry of the excess silver salts containing the target
anion (1:1.1 molar ratio for AgDCA/iodide and AgSac/iodide, and 1:1
molar ratio for AgNO3/iodide) and the solution was heated to 408C with
stirring for 4 h. Silver iodide was removed from the solutions by filtra-
tion. After complete removal of water, the resulting viscous liquids or
solids were washed with diethyl ether under vigorous stirring three times
and the volatile component removed. To ensure complete removal of
silver salts from the product, the dried products were dissolved in dry di-
chloromethane or acetone and cooled in a freezer overnight before fur-
ther filtration. The final products were dried under high vacuum for
more than 4 h at 1008C, and then were wrapped in aluminum foil and
stored in a desiccator under dry nitrogen.
Tribological coefficient: Before the tribological test, each sample under-
went removal of water as before the test of water content. The tribologi-
cal coefficients of the selected ILs were evaluated by using an Optimol
SRV (SRV is the abridged name for German Schwingung, Beibung, Vers-
chleiss) oscillating friction and wear tester. The friction and wear tests
were performed at room temperature in a ball-on-disc configuration, by
oscillating an AISI52100 alloy steel ball (or Si3N4 ball) over a steel, Al,
or Cu block at a frequency of 25 Hz, with a sliding amplitude of 1 mm.
Prior to the friction and wear test, two drops of the lubricant were let
into to the ball–disc contact area. The friction coefficient curve was re-
corded automatically with a chart attached to the SRV test rig.
776
www.chemeurj.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 765 – 778

Cyclic Sulfonium-Based Ionic Liquids
FULL PAPER
[S1]DCA: Colorless liquid, yield: 93%; MW, 169.07. 1H NMR
(CD3COCD3, 500 MHz): d=3.78 (m, 2H), 3.60 (m, 2H), 2.92 (m, 3H),
2.55 (m, 2H), 2.42 ppm (m, 2H); IR: n˜ =2953, 2884 (nC–H aliphatic),
[4] a) H. Ohno, Electrochemical Aspects of Ionic Liquids, Wiley, New
Jersey, 2005; b) B. OꢁRegan, M. Grꢂtzel, Nature 1991, 353, 737–740.
[5] P. Wasserscheid, T. Welton, Ionic Liquids in Synthesis, VCH, Wein-
heim, 2003.
2236, 2196, 2135 cmꢀ1 (nC N); MS (ESI+) calcd: m/z: 103.06 ([C5H11S]+);
=
found: m/z: 103.0579.
[6] a) P. Wasserscheid, W. Keim, Angew. Chem. 2000, 112, 3926–3945;
Angew. Chem. Int. Ed. 2000, 39, 3772–3789; b) T. J. Geldbach, P. J.
Dyson, J. Am. Chem. Soc. 2004, 126, 8114–8115.
[7] a) C. Ye, W. Liu, Y. Chen, L. Yu, Chem. Commun. 2001, 2244–2245;
b) Y. Wang, H. Yang, J. Am. Chem. Soc. 2005, 127, 5316–5317; c) K.
Ding, Z. Miao, Z. Liu, Z. Zhang, B. Han, G. An, S. Miao, Y. Xie, J.
Am. Chem. Soc. 2007, 129, 6362–6363.
[8] a) J. G. Huddlestou, H. D. Willauer, R. P. Swatloski, A. E. Visser,
R. D. Rogers, Chem. Commun. 1998, 1765–1766; b) A. Bçsmann, L.
Datsevich, A. Jess, A. Lauter, C. Schmitz, P. Wasserscheid, Chem.
Commun. 2001, 2494–2495.
[9] a) R. M. Lau, F. Rantwijk, K. R. Seddon, R. A. Sheldon, Org. Lett.
2000, 2, 4189–4191; b) J. L. Kaar, A. M. Jesionowski, J. A. Berber-
ich, R. Moulton, A. J. Russell, J. Am. Chem. Soc. 2003, 125, 4125–
4131.
[S2]DCA: Colorless liquid, yield: 92%; MW, 183.08. 1H NMR
(CD3COCD3, 500 MHz): d=3.75 (m, 2H), 3.65 (m, 2H), 3.46 (m, 2H),
2.48 (m, 2H), 2.41 (m, 2H), 1.53 ppm (t, 3H); IR: n˜ =2977, 2950, 2884
(nC–H aliphatic), 2234, 2195, 2134 cmꢀ1 (nC N); MS (ESI+) calcd: m/z:
=
117.07 ([C6H13S]+); found: m/z: 117.0734.
[S4]DCA: Colorless liquid, yield: 90%; MW, 211.11. 1H NMR
(CD3COCD3, 500 MHz): d=3.76 (m, 2H), 3.64 (m, 2H), 3.43 (t, 2H),
2.49 (m, 2H), 2.40 (m, 2H), 1.88 (m, 2H), 1.53 (m, 2H), 0.97 ppm (t,
3H); IR: n˜ =2962, 2937, 2875 (nC–H aliphatic), 2235, 2195, 2135 cmꢀ1
(nC N); MS (ESI+) calcd: m/z: 145.1 ([C8H17S]+); found: m/z: 145.1047.
=
[S1]NO3: White solid, yield: 92%; MW, 165.05. 1H NMR (DMSO,
500 MHz): d=3.43 (m, 2H), 3.24 (m, 2H), 2.73 (s, 3H), 2.20 (m, 2H),
2.09 ppm (m, 2H); IR: n˜ =2983, 2948 (nC–H aliphatic), 2395, 1762,
1382 cmꢀ1 (nNO ); MS (ESI+) calcd: m/z: 103.06 ([C5H11S]+); found: m/z:
3
[10] a) J. S. Wilkes, M. J. Zaworotko, J. Chem. Soc. Chem. Commun.
1992, 965–967; b) P. Bonhꢃte, A. P. Dias, N. Papageorgiou, K. Ka-
lyanasundaram, M. Grꢂtzel, Inorg. Chem. 1996, 35, 1168–1178.
[11] a) J. Sun, D. R. MacFarlane, M. Forsyth, Electrochim. Acta 2003, 48,
1707–1711; b) J. Golding, N. Hamid, D. R. MacFarlane, M. Forsyth,
C. Forsyth, C. Collins, J. Huang, Chem. Mater. 2001, 13, 558–564.
[12] a) Z. B. Zhou, H. Matsumoto, K. Tatsumi, Chem. Eur. J. 2005, 11,
752–766; b) J. Pernak, A. Syguda, I. Mirska, A. Pernak, J. Nawrot,
A. PradzyÇska, S. T. Griffin, R. D. Rogers, Chem. Eur. J. 2007, 13,
6817–6827.
[13] a) R. P. Singh, J. M. Shreeve, Inorg. Chem. 2003, 42, 7416–7421; b) J.
Zhang, G. R. Martin, D. D. DesMarteau, Chem. Commun. 2003,
2334–2335.
[14] a) N. M. M. Mateus, L. C. Branco, N. M. T. LourenÅo, C. A. M.
Afonso, Green Chem. 2003, 5, 347–352; b) L. Yu, D. Garcia, R.
Ren, X. Zeng, Chem. Commun. 2005, 2277–2279.
[15] See: Proceedings of the Second International Congress on Ionic Liq-
uids, Yokohama, Japan, August, 5–7, 2007.
[16] a) B. Garcia, S. Lavallꢄe, G. Perron, C. Michot, M. Armand, Electro-
chim. Acta 2004, 49, 4583–4588; b) H. Sakaebe, H. Matsumoto,
Electrochem. Commun. 2003, 5, 594–598.
[17] a) M. A. B. H. Susan, A. Noda, S. Mitsushima, M. Watanabe, Chem.
Commun. 2003, 938–939; b) R. F. Souza, J. C. Padilha, R. S. Gon-
calves, J. Dupont, Electrochem. Commun. 2003, 5, 728–731.
[18] a) M. Ue, M. Takeda, A. Toriumi, A. Kominato, R. Hagiwara, Y.
Ito, J. Electrochem. Soc. 2003, 150, 499–502; b) Y. J. Kim, Y. Matsu-
zawa, S. Ozaki, K. C. Park, C. Kim, M. Endo, H. Yoshida, G.
Masuda, T. Sato, M. S. Dresselhausc, J. Electrochem. Soc. 2005, 152,
710–715.
103.0579.
[S2]NO3: Colorless liquid, yield: 94%; MW, 179.06. 1H NMR (DMSO,
500 MHz): d=3.44 (m, 2H), 3.35 (m, 2H), 3.16 (m, 2H), 2.16 (m, 2H),
2.10 (m, 2H), 1.29 ppm (t, 3H); IR: n˜ =2977, 2948, 2881 (nC–H aliphatic),
2394, 1762, 1349 cmꢀ1 (nNO ); MS (ESI+) calcd: m/z: 117.07 ([C6H13S]+);
3
found: m/z: 117.0734.
[S4]NO3: White solid, yield: 89%; MW, 207.09. 1H NMR (DMSO,
500 MHz): d=3.46 (m, 2H), 3.34 (m, 2H), 3.15 (m, 2H), 2.18 (m, 2H),
2.12 (m, 2H), 1.64 (m, 2H), 1.37 (m, 2H), 0.89 ppm (t, 3H); IR: n˜ =2960,
2940, 2874 (nC–H aliphatic), 2396, 1762, 1377 cmꢀ1 (nNO ); MS (ESI+)
3
calcd: m/z: 145.1 ([C8H17S]+); found: m/z: 145.1047.
[S1]Sac: Colorless viscous liquid, yield: 91%; MW, 285.05. 1H NMR
(CD3COCD3, 500 MHz): d=7.64 (m, 4H), 3.82 (m, 2H), 3.60 (m, 2H),
3.07 (s, 3H), 1.93 (m, 2H), 1.73 ppm (m, 2H); IR: n˜ =2942, 2870, 1460
(nC–H aliphatic), 1333 (nCNS), 1258, 1145 (nSO ), 950 cmꢀ1 (nCNS); MS (ESI+)
2
calcd: m/z: 103.06 ([C5H11S]+).
[S2]Sac: Colorless viscous liquid, yield: 92%; MW, 299.06. 1H NMR
(CD3COCD3, 500 MHz): d=7.63 (m, 4H), 3.81 (m, 2H), 3.68 (m, 2H),
3.52 (m, 2H), 2.42 (m, 2H), 2.35 (m, 2H), 1.47 ppm (t, 3H); IR: n˜ =2944,
2875, 1455, 1417, 1386 (nC–H aliphatic), 1329 (nCNS), 1257, 1145 (nSO ),
2
949 cmꢀ1 (nCNS); MS (ESI+) calcd: m/z: 117.07 ([C6H13S]+); found: m/z:
117.0734.
[S4]Sac: Colorless viscous liquid, yield: 87%; MW, 327.10. 1H NMR
(DMSO, 500 MHz): d=7.63 (m, 4H), 3.54 (m, 2H), 2.58 (m, 2H), 3.32
(m, 2H), 1.84 (m, 2H), 1.80 (m, 2H), 1.55 (m, 2H), 1.50 (m, 2H),
1.38 ppm (t, 3H); IR: n˜ =2959, 2934, 2873, 1455 (nC–H aliphatic), 1329
(nCNS), 1256, 1145 (nSO ), 949 cmꢀ1 (nCNS); MS (ESI+) calcd: m/z: 145.1
2
([C8H17S]+); found: m/z: 145.1047.
[19] a) N. Papageorgiou, Y. Athanassov, M. Armand, P. Bonh ote, H.
Pettersson, A. Azam, M. Gr¨atzel, J. Electrochem. Soc. 1996, 143,
3099–3108; b) P. Wang, S. M. Zakeeruddin, J.-E. Moser, R. Humph-
ry-Baker, M. Grꢂtzel, J. Am. Chem. Soc. 2004, 126, 7164–7165.
[20] Z. Fei, T. J. Geldbach, D. Zhao, P. J. Dyson, Chem. Eur. J. 2006, 12,
2122–2130.
Acknowledgements
This work was supported by the National Natural Science Foundation of
China (No. 20533080, 20225309, and 50421502). We are grateful to Ms.
Ling Gao and Mr. Qixiu Zhu for their assistance with this work.
[21] E. D. Bates, R. D. Mayton, I. Ntai, J. H. Davis, Jr., J. Am. Chem.
Soc. 2002, 124, 926–927.
[22] H. Itoh, K. Naka, Y. Chujo, J. Am. Chem. Soc. 2004, 126, 3026–
3027.
[23] W. Miao, T. H. Chan, J. Org. Chem. 2005, 70, 3251–3255.
[24] Z. Fei, D. Zhao, T. J. Geldbach, R. Scopelliti, P. J. Dyson, Chem.
Eur. J. 2004, 10, 4886–4893.
[25] D. Zhao, Z. Fei, T. J. Geldbach, R. Scopelliti, P. J. Dyson, J. Am.
Chem. Soc. 2004, 126, 15876–15882.
[1] a) T. Welton, Chem. Rev. 1999, 99, 2071–2083; b) J. C. B. Richard,
J. D. Paul, J. E. David, W. Thomas, Chem. Commun. 2001, 1862–
1863; c) E. Kuhlmann, S. Himmler, H. Giebelhaus, P. Wasserscheid,
Green Chem. 2007, 9, 233–242.
[26] A. E. Visser, R. P. Swatloski, W. M. Reichert, R. Mayton, S. Sheff,
A. Wierzbicki, J. H. Davis, Jr., R. D. Rogers, Chem. Commun. 2001,
135–136.
[2] a) C. Ye, J. M. Shreeve, J. Org. Chem. 2004, 69, 6511–6513; b) Y.
Fukaya, Y. Iizuka, K. Sekikawa, H. Ohno, Green Chem. 2007, 9,
1155–1157.
[27] a) D. R. MacFarlane, J. Golding, S. Forsyth, M. Forsyth, G. B.
Deacon, Chem. Commun. 2001, 1430–1431; b) M. Yoshizawa, H.
Ohno, Chem. Commun. 2004, 1828–1829.
[3] a) A. J. Boydston, C. S. Pecinovsky, S. T. Chao, C. W. Bielawski, J.
Am. Chem. Soc. 2007, 129, 14550–14551; b) K. Fukumoto, M. Yosh-
izawa, H. Ohno, J. Am. Chem. Soc. 2005, 127, 2398–2399.
Chem. Eur. J. 2009, 15, 765 – 778
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
777

Y. Deng et al.
[28] M. Yoshizawa, W. Xu, C. A. Angell, J. Am. Chem. Soc. 2003, 125,
15411–15419.
[29] J. N. Barisci, G. G. Wallace, D. R. MacFarlane, R. H. Baughman,
Electrochem. Commun. 2004, 6, 22–27.
[30] H. Matsumoto, H. Sakaebe, K. Tatsumi, J. Power Sources 2005, 146,
45–50.
[31] O. O. Okoturo, T. J. VanderNoot, J. Electroanal. Chem. 2004, 568,
167–181.
[32] H. Matsumoto, T. Matsuda, Y. Miyazaki, Chem. Lett. 2000, 1430–
1431.
[33] D. Gerhard, S. C. Alpaslan, H. J. Gores, M. Uerdingen, P. Wassersc-
heid, Chem. Commun. 2005, 5080–5081.
[34] a) L. Yang, Z. X. Zhang, X. H. Gao, H. Q. Zhang, K. Mashita, J.
Power Sources 2006, 162, 614–619; b) S. Fang, L. Yang, C. Wei, C.
Peng, K. Tachibana, K. Kamijima, Electrochem. Commun. 2007, 9,
2696–2702; c) H. Paulsson, A. Hagfeldt, L. Kloo, J. Phys. Chem. B
2003, 107, 13665–13670.
[35] a) P. Wang, F. Gao, Y. Cao, J. Zhang, Chinese Patent, CN
200710055896.3; b) C. Xi, Y. Cao, Y. Cheng, M. Wang, X. Jing, S. M.
Zakeeruddin, M. Grꢂtzel, P. Wang, J. Phys. Chem. C 2008, 112,
11063–11067.
[36] a) H. Shirota, E. W. Castner, Jr., J. Phys. Chem. B 2005, 109, 21576–
21585; b) D. Zhao, Aust. J. Chem. 2004, 57, 509; c) J. Wang, H.
Wang, S. Zhang, H. Zhang, Y. Zhao, J. Phys. Chem. B 2007, 111,
6181–6188.
[37] a) B. A. DaSilveira, Neto, L. S. Santos, F. M. Nachtigall, M. N. Eber-
lin, J. Dupont, Angew. Chem. 2006, 118, 7409–7412; Angew. Chem.
Int. Ed. 2006, 45, 7251–7254; b) B. A. DaSilveira, Neto, L. S. Santos,
F. M. Nachtigall, M. N. Eberlin, J. Dupont, Angew. Chem. 2006, 118,
1–1; Angew. Chem. Int. Ed. 2006, 45, 1–5.
[38] H. Chen, Z. Ouyang, R. G. Cooks, Angew. Chem. 2006, 118, 3738–
3742; Angew. Chem. Int. Ed. 2006, 45, 3656–3660.
[39] a) F. C. Gozzo, L. S. Santos, R. Augusti, C. S. Consorti, J. Dupont,
M. N. Eberlin, Chem. Eur. J. 2004, 10, 6187–6193; b) B. K. Ku, J. F.
de. la. Mora, J. Phys. Chem. B 2004, 108, 14915–14923; c) R. Bini,
O. Bortolini, C. Chiappe, D. Pieraccini, T. Siciliano, J. Phys. Chem.
B 2007, 111, 598–604; d) J. W. Remsburg, R. J. Soukup-Hein, J. A.
Crank, Z. S. Breitbach, T. Payagala, D. W. Armstrong, J. Am. Soc.
Mass Spectrom. 2008, 19, 261–269.
[40] a) E. F. Smith, I. J. V. Garcia, D. Briggs, P. Licence, Chem. Commun.
2005, 5633–5635; b) D. S. Silvester, T. L. Broder, L. Aldous, C. Har-
dacre, A. Crossley, R. G. Compton, Analyst 2007, 132, 196–198;
c) F. Maier, J. M. Gottfried, J. Rossa, D. Gerhard, P. S. Schulz, W.
Schwieger, P. Wasserscheid, H.-P. Steinrꢅck, Angew. Chem. 2006,
118, 7778–7780; Angew. Chem. Int. Ed. 2006, 45, 7942–7944.
[41] S. Caporali, U. Bardi, A. Lavacchi, J. Electron Spectrosc. Relat.
Phenom. 2006, 151, 4–8.
[42] a) O. Hçfft, S. Bahr, M. Himmerlich, S. Krischok, J. A. Schaefer, V.
Kempter, Langmuir 2006, 22, 7120–7123; b) V. Lockett, R. Sedev,
C. Bassell, J. Ralston, Phys. Chem. Chem. Phys. 2008, 10, 1330–
1335.
[43] M. Gottfried, F. Maier, J. Rossa, D. Gerhard, P. S. Schulz, P. Was-
serscheid, H.-P. Steinrꢅck, Z. Phys. Chem. 2006, 220, 1439–1453.
[44] E. F. Smith, F. J. M. Rutten, I. J. Villar-Garcia, D. Briggs, P. Licence,
Langmuir, 2006, 22, 9386–9392.
[45] a) D. R. MacFarlane, J. Huang, M. Forsyth, Nature, 1999, 402, 792–
794; b) C. M. Forsyth, D. R. MacFarlane, J. J. Golding, J. Huang, J.
Sun, M. Forsyth, Chem. Mater. 2002, 14, 2103–2108.
[46] a) J. C. Dearden, Sci. Total Environ. 1991, 109–110, 59–68;
b) R. J. C. Brown, R. F. C. Brown, J. Chem. Educ. 2000, 77, 724
ꢀ731.
[47] D. R. MacFarlane, S. A. Forsyth, J. Golding, G. B. Deacon, Green
Chem. 2002, 4, 444–448.
[48] O. O. Okoturo, T. J. VanderNoot, J. Electroanal. Chem. 2004, 568,
167–181.
[49] Q. Zhang, Z. Li, J. Zhang, S. Zhang, L. Zhu, J. Yang, X. Zhang, Y.
Deng, J. Phys. Chem. B 2007, 111, 2864–2872.
[50] J. Fuller, R. Carlin, R. Osteryoung, J. Electrochem. Soc. 1997, 144,
3881–3885.
Received: April 2, 2008
Revised: August 14, 2008
Published online: November 28, 2008
778
www.chemeurj.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 765 – 778