Silica ionogels for proton transport

Article abstract of DOI:10.1039/c2jm00037g

A number of ionogels - silica-ionic liquid (IL) hybrid materials - were synthesized and studied for their ionic conductivity. The materials are based on a sulfonated IL, 1-methyl-3-(3-sulfopropyl-)-imidazolium p-toluenesulfonate, [PmimSO₃H][PTS], which contains a sulfonic acid/sulfonate group both in the IL anion and in the side chain of the IL cation. By way of the sulfonate-sulfonic acid proton transfer, the IL imparts the ionogel with a high ionic conductivity of ca. 10^-2 S cm^-1 in the as-synthesized state at 120 °C and 10^-3 S cm^-1 in the dry state at 120 °C. The ionogels are stable up to ca. 150 °C in dynamic thermogravimetric analysis. This suggests that these materials, which are relatively cheap and easily fabricated, could find application in fuel cells in intermediate temperature ranges where many other membrane materials are not suitable. The Royal Society of Chemistry 2012.

Full text of DOI:10.1039/c2jm00037g

View Online / Journal Homepage / Table of Contents for this issue
Dynamic Article Links <
C
Journal of
Materials Chemistry
Cite this: J. Mater. Chem., 2012, 22, 17140
www.rsc.org/materials
PAPER
Silica ionogels for proton transport†
a
b
b
c
a
Emilie Delahaye, Ronald Gobel, Ruben Lobbicke, Regis Guillot, Christoph Sieber and Andreas Taubert
bd
*
€
€
ꢀ
Received 3rd January 2012, Accepted 27th June 2012
DOI: 10.1039/c2jm00037g
A number of ionogels – silica–ionic liquid (IL) hybrid materials – were synthesized and studied for their
ionic conductivity. The materials are based on a sulfonated IL, 1-methyl-3-(3-sulfopropyl-)-
imidazolium p-toluenesulfonate, [PmimSO3H][PTS], which contains a sulfonic acid/sulfonate group
both in the IL anion and in the side chain of the IL cation. By way of the sulfonate–sulfonic acid proton
transfer, the IL imparts the ionogel with a high ionic conductivity of ca. 10ꢀ2 S cmꢀ1 in the as-
synthesized state at 120 ꢁC and 10ꢀ3 S cmꢀ1 in the dry state at 120 ꢁC. The ionogels are stable up to ca.
150 ꢁC in dynamic thermogravimetric analysis. This suggests that these materials, which are relatively
cheap and easily fabricated, could find application in fuel cells in intermediate temperature ranges
where many other membrane materials are not suitable.
Computer simulations suggest that negative surface charges lead
to bilayer formation in 1-ethyl-3-methylimidazolium hexa-
fluorophosphate [Emim][PF6].20 The common finding of all
studies is that there is a distinct and highly specific interplay
between the IL and the matrix, which strongly depends on the
chemistry of the individual components. As a result, although a
number of studies have been published on the topic, the
community lacks a sufficient experimental and theoretical
foundation of ionogel composition–property relationships. For
the development of quantitative and predictive models, a much
larger dataset is necessary. This particularly applies to transport
properties in silica-based ionogels. There are, so far, only a few
studies available in the literature and, unfortunately, these few
studies are on different materials.3,9–11,21
Introduction
Ionic liquids (ILs) have in the recent past been studied in virtually
all areas of chemistry.1,2 An emerging field is the general area of
IL–inorganic hybrid materials, or ionogels;3,4 for a detailed
treatise of ionogels see ref. 5–7. As of now, the matrix in inor-
ganic ionogel research is (mesoporous) silica. This is, among
others, due to the fact that silica chemistry, including pore size
adjustment and functionalization along with pore organization,
is well established.8 Depending on the properties of the silica
matrix and the IL, functional materials, such as materials with
high ionic conductivity,3,9–11 can be synthesized. These materials
could, for example, be useful for lithium ion batteries,5,6,10,12 in
particular, because ionogels are stable over a large temperature
window and ILs often have liquid ranges that are much larger
than those of molecular solvents.3,9,13–15
The current study therefore focuses on proton transport in a set
of specifically designed ionogels. We have made a sulfonate-based
IL, 1-methyl-3-(3-sulfopropyl)-imidazolium para-toluenesulfonate,
[PmimSO3H][PTS], which carries a sulfonic acid/sulfonate group
both in the IL anion and in a side chain attached to the imidazo-
lium cation, Scheme 1. Formally, there is one acidic proton on
either the side chain sulfonate of the imidazolium ring or the
tosylate counterion; the other sulfonate site is not protonated and
can take up the proton. As a result, the IL should be capable of
rapid proton transport via a hopping mechanism to free sites. As
the matrix, we have used three silica materials, (i) pure mesoporous
silica15 or (ii) mesoporous silica bearing different amounts of
We have recently shown that ILs confined in a silica matrix
exhibit a complex phase behavior.14,15 Both the chemical
composition of the IL and the chemical modification of the pores
of the mesoporous silica matrix affect, for example, the glass
transition or the crystallization process of the IL in the ionogel.
Similar observations have been made by other research
groups.9,16–18 The surface of SiOx nanoparticles lowers the
melting point of imidazolium ILs compared to the bulk IL.19
aMax Planck Institute for Polymer Research, D-15528 Mainz, Germany
bInstitute of Chemistry, University of Potsdam, Karl-Liebknecht-Str. 24-
25, Building 26, Golm, D-14476 Potsdam, Germany. E-mail: ataubert@
uni-potsdam.de; Tel: +49 (0)331 977 5773
c
0
ꢀ
Institut de Chimie Moleculaire et des Materiaux d Orsay, UMR 8182,
Universiteꢀ Paris-Sud, F-91405 Orsay, France
ꢀ
dMax Planck Institute of Colloids and Interfaces, D-14476 Potsdam,
Germany
† Electronic supplementary information (ESI) available: CCDC 870864.
For ESI and crystallographic data in CIF or other electronic format see
DOI: 10.1039/c2jm00037g
Scheme 1 The IL 1-methyl-3-(3-sulfopropyl)-imidazolium para-tolue-
nesulfonate, [PmimSO3H][PTS].
17140 | J. Mater. Chem., 2012, 22, 17140–17146
This journal is ª The Royal Society of Chemistry 2012

View Online
pyridyl dangling groups on the walls of the mesopores.14 The latter
are interesting, because the free nitrogen atoms in the pyridine rings
can in principle also act as proton acceptors. This could, possibly,
enhance the proton transfer and hence the overall proton
conductivity beyond what is possible with the IL and the silanol
groups of the silica matrix alone.
transparent liquid. After a few months, this viscous liquid turned
into a solid. This phenomenon has been already described for ILs
bearing –SO3H groups.24–27 Yield ¼ 80%. Elemental analysis:
C14H20N2S2O6 (M ¼ 376.45 g molꢀ1): found (calc.%): C 44.41
1
(44.67); H 5.37 (5.35); N 7.13 (7.44); S 16.93 (17.04). H-NMR
(D2O): 8.60 (s, 1), 7.55 (d, 2), 7.35 (s, 1), 7.30 (s, 1), 7.20 (d, 2),
4.15 (t, 2), 3.70 (s, 3), 2.70 (t, 2), 2.25 (s, 3), 2.15 (m, 2). 13C-NMR
(D2O + 5 drops of DMSO-d6): 144, 140, 136, 130, 126, 124, 122,
48, 37, 27, 21. IR (reflectance, cmꢀ1): 3458 w, 3153 m, 3111 m,
2959 w, 2923 w, 2861 w, 1694 m, 16040 w, 1569 m, 1494 w, 1453
m, 1231 s, 1168 s, 1122 s, 1031 s, 1004 s, 904 w, 818 m, 742 m, 677
m, 620 w.
Experimental
Chemicals
Pyridine, 3-fluoropyridine, tetramethylorthosilicate (TMOS),
methanol, p-toluenesulfonic acid (ABCR), 1-methylimidazole,
and 1,3-propanesultone (VWR) were used as received.
Silica monolith synthesis
Ionic liquid synthesis22,23
Silica monoliths were synthesized according to our previous
protocol.14,15 Fluoropyridine was used as the catalytic base. In the
following, Fpy refers to silica monoliths obtained from tetrame-
thoxyorthosilicate (TMOS) and fluoropyridine only. PyX refers to
silica monoliths obtained from TMOS and 2-(4-pyridylethyl)trie-
thoxysilane (PEOS).14 X ¼ 6, 8 refers to the TMOS : PEOS molar
ratio used in the sol–gel synthesis of the monoliths. Py6 indicates a
molar ratio of 6 : 1 and Py8 a molar ratio of 8 : 1. These monoliths
contain dangling pyridyl units on the pore walls, while the mate-
rials Fpy (synthesized from TMOS only) exclusively exhibit silanol
surface groups.
PmimSO3 and [PmimSO3H][PTS] were prepared via a two-step
procedure reported previously.7 In the first step, the zwitterionic
compound PmimSO3 is formed where the positive charge is
located on the imidazolium ring and the negative charge on the
sulfonate group in the side chain. In the second step, the addition
of the strong acid p-toluenesulfonic acid leads to the formation
of the highly viscous, transparent, IL [PmimSO3H][PTS]
(Scheme 2).
Preparation of 3-(1-methylimidazolium-3-yl)propane-sulfonate
PmimSO3
Loading of [PmimSO3H][PTS] into the silica monoliths
A solution of 1-methylimidazole (1.642 g, 20 mmol) in acetone
(20 mL) was prepared in a three-necked round bottom flask.
Then a solution of 1,3-propanesultone (2.443 g, 20 mmol) in
acetone (20 mL) was slowly added to this solution. The mixture
was stirred at room temperature under N2 for 4 days and a white
precipitate formed progressively. The mixture was filtered and
the white powder was washed with acetone, followed by drying
under vacuum. Yield ¼ 85%. Elemental analysis: C7H12N2SO3
(M ¼ 204.25 g molꢀ1): found (calc.%): C 41.17 (41.16); H 5.91
(5.92); N 13.68 (13.72); S 15.75 (15.70). 1H-NMR (D2O): 8.65 (s,
1), 7.45 (d, 2), 7.35 (d, 1), 4.70 (d, 1), 3.80 (s, 3), 2.8 (d, 2), 2.25 (t,
2). IR (reflectance, cmꢀ1): 3453 w, 3133 s, 3075 m, 3035 s, 2951 w,
2928 w, 2871 w, 2840 w, 1575 m, 1557 m, 1467 m, 1334 s, 1267 m,
1203 s, 1171 s, 1149 s, 1119 m, 1079 s, 1031 s, 877 m, 805 s, 755 s,
723 s, 651 s, 626 s, 589 s, 530 s, 520 m.
Methanol-filled monoliths obtained after Soxhlet extraction14,15
were placed in a vial filled with IL, placed on a laboratory shaker
ꢁ
and held at 90 C. After 24 hours, the monoliths were removed
from the IL and dried in vacuum overnight to remove excess
solvent. These samples will be denoted Fpy/IL or PyX/IL in the
remainder of the text.
Characterization
C, H, N, S, Si analysis was carried out at the Analytische Labo-
ratorien Lindlar. Infrared spectra were recorded in ATR mode on
a Nicolet FT-IR 730 spectrometer (32 scans, 2 cmꢀ1 resolution).
Raman spectra were obtained on a Bruker RFS100/S spectrometer
with an excitation wavelength of 632 nm and a resolution of
1 cmꢀ1. Solution NMR spectroscopy was done on a Bruker Avance
250 in D2O. Spectra were calibrated using the solvent peak (HDO/
D2O). Thermogravimetric analysis was done on a TA Instruments
Preparation of 1-methyl-3-(3-sulfopropyl)-imidazolium para-
toluenesulfonate [PmimSO3H][PTS]
SDT Q600 in air from 25 to 900 C at 5ꢁ C minꢀ1. Differential
ꢁ
scanning calorimetry was done on a Mettler Toledo DSC 822.
PmimSO3 (2.041 g, 10 mmol) was added to solid para-toluene-
sulfonic acid (1.901 g, 10 mmol) at room temperature and the
mixture was stirred at 110 ꢁC overnight yielding a viscous
Samples of about 5 mg were placed in aluminum pans with pierced
ꢁ
ꢁ
lids. DSC traces were recorded from ꢀ80 C to 200 C at 10 K
minꢀ1. To remove water traces the samples were heated to 200 ꢁC
before the first cooling cycle. Heating and cooling cycles were
repeated twice for reproducibility. Isothermal times were 5
minutes. Impedance spectroscopy data were recorded as a function
of temperature on a Solartron SI 1260 impedance/gain phase
analyzer (Schlumberger) with a broadband dielectric converter
(BDC; Novocontrol) from 10ꢀ1 to 106 Hz. The measurements
were performed using platinum electrodes with a temperature-
controlled cryostat under nitrogen (Novocontrol). The dc-
conductivity was estimated from Cole–Cole plots (Nyquist plot) at
Scheme 2 Synthesis of [PmimSO3H][PTS]; (a) RT, 4 days, under N2 and
(b) 110 ꢁC overnight.
This journal is ª The Royal Society of Chemistry 2012
J. Mater. Chem., 2012, 22, 17140–17146 | 17141

View Online
the point where the imaginary part was minimal or from Bode
plots taking the plateau values of the ac-conductivity. Single
crystal X-ray diffraction was done on a Bruker Kappa X8 APEX II
diffractometer (graphite-monochromatized Mo Ka radiation).
The temperature of the crystal was maintained at 200 ꢂ 2 K with a
700 series Cryostream cooling device. Intensity data were cor-
rected for Lorentz-polarization and absorption factors. The
structure was solved by direct methods using SHELXS-97 (ref. 28)
and refined against F2 by full-matrix least-squares methods using
SHELXL-97 (ref. 29) with anisotropic displacement parameters
for all non-hydrogen atoms. All calculations were performed with
the software package WINGX.30 The structure was drawn using
ORTEP3.31 All hydrogen atoms were localized in difference
Fourier maps and introduced into the calculations as a riding
model with isotropic thermal parameters. Crystallographic data
can be found in the ESI, Tables S1 and S2.†
Fig. 2 shows representative Raman and infrared (IR) spectra
of the IL, the neat silica monoliths, and the resulting ionogels. In
the spectra of the neat IL [PmimSO3H][PTS], the bands can be
assigned to the C–N and C–H stretching vibration modes of the
imidazolium ring (3160, 3115, 2960, 2915 cmꢀ1), CH3–N and
CH2–N stretching modes of the imidazolium ring, C–C stretch-
ing vibrations (1600, 1568, 1457, 1231 cmꢀ1), and out-of-plane
C–H bending vibrations (817, 742, 684, 522 cmꢀ1). Moreover,
two bands at 1122 and 1032 cmꢀ1 can be ascribed to the asym-
metric and symmetric stretching vibration of the sulfonate group.
The infrared spectra of the neat silica monoliths show a broad
band centered around 3300–3400 cmꢀ1, which is attributed to Si–
OH and H2O vibrations. Bands at 1045, 950, 800, and 560 cmꢀ1
can be attributed to the asymmetric Si–O–Si stretching vibration,
the Si–OH silanol groups on the silica surface, and the symmetric
Si–O–Si stretching vibration and rocking modes, respectively.
The IR spectra of the ionogels (that is, after filling the IL into
the silica host) are dominated by the vibration bands from the IL
indicating that the IL is the major component of the ionogels.
This is confirmed by elemental analysis, see below. The spectra
exhibit a broad band around 3300 cmꢀ1 which can mainly be
attributed to water molecules that are still present in the mate-
rials and to silanol groups. Between 1700 cmꢀ1 and 600 cmꢀ1, the
spectra of the ionogels are virtually identical to the spectra of the
neat IL (which are different from the IR spectra of the silica
host).14,15 In particular, the asymmetric nasymSO3 and symmetric
nsymSO3 vibration bands of the sulfonate groups can clearly be
identified (Table 1). Table 1 also shows that the difference Dn
between nasymSO3 and nsymSO3 is smaller in the ionogels (40 to
72 cmꢀ1) than in the neat IL (84 cmꢀ1). This suggests that the
sulfonate groups are not directly linked to the monolith but
involved in hydrogen bonds.33–35 This would be consistent with
the data from the single crystal structure (Fig. 1), where extensive
hydrogen bonding between sulfonate/sulfonic acid moieties is
observed. It must be noted at this point, however, that the
characteristic bands of the silica monoliths at around 1100 cmꢀ1
(attributed to the Si–O–Si asymmetric stretching vibration) may
overlap with the sulfonate bands described above. The details of
the hydrogen bonding pattern in the liquid state are therefore
difficult to assign.
Results
Materials
The as-prepared pure IL [PmimSO3H][PTS] is a transparent,
highly viscous fluid. Over a few months, however, the liquid
becomes turbid and a solid progressively forms. Single crystal X-
ray diffraction shows that this solid is made of colorless crystals
of the IL. This demonstrates that the slow formation of a solid is
in fact a very slow crystallization, which, presumably, is strongly
kinetically hindered due to the high viscosity of the initial IL.
Crystallization could therefore also be the reason for the slowly
appearing turbidity in other examples of sulfonate ILs described
in the literature.24–26
The asymmetric unit of the crystal structure contains two IL
molecules (Fig. 1a). The SO3 function in the side chain of the
imidazolium cation shares a proton with an SO3 group attached
on the side chain of a neighboring imidazolium cation. This
behavior is confirmed by the S–O distances (Table S2 in the
ESI†) since the bond length of S–O bearing the proton is slightly
larger than the other two S–O distances. In contrast, the S–O
distances in the betaine PmimSO3 (where there is no S–O–H
bond) are essentially identical and consistent with the literature.32
For this reason, the H has a formal occupation of ½ at each
sulfonate. The same behavior is observed on the SO3 group of the
IL anion, p-toluenesulfonate. However, it is interesting to note
that the residual density peak is the highest on the SO3 function
Raman spectroscopy complements IR spectroscopy in the
sense that also in the former, the presence of the IL in the
ionogels is clearly visible. Bands around 3067, 2965, and
2931 cmꢀ1 can be attributed to the C–N and C–H stretching
vibration modes of the imidazolium ring. The bands at 1593,
1416, 1400, and 1120 cmꢀ1 can be attributed to CH3–N and
CH2–N stretching vibrations of the imidazolium ring, and C–C
stretching vibrations in both the ring and the side chains. The
bands located at 1032 cmꢀ1 can be ascribed to the symmetric
vibrations of the sulfonate groups. Raman spectroscopy there-
fore allows for a better identification of nsymSO3 and confirms the
IR data above. Overall, IR and Raman spectroscopies thus show
that the IL is incorporated in significant amounts in the ionogels.
Fig. 3 shows representative thermogravimetric analysis–
differential thermal analysis (TGA–DTA) data of the IL and the
ionogels. TGA shows that the IL and all ionogels decompose via
the same three steps. The first step between 25 and 150 ꢁC is
attributed to the loss of water. This first weight loss up to 150 ꢁC
is below 10% in the as-synthesized ionogels. The large weight loss
ꢁꢀ3
attached on the imidazolium cation (Q1 ¼ 0.32 e A ), indicating
a slightly higher overall presence of H at these groups.
A view along the crystallographic b-axis (Fig. 1b) reveals the
existence of alternating layers of the IL betaine with an overall
positive charge, [(PmimSO3H)2]+, and tosylate anion with an
overall negative charge, [(PTSH)2]ꢀ. It must be noted, however,
that the crystal structure and in particular the position of H on
the SO3 functions does not necessarily reflect the liquid state of
the IL since the infrared spectra of the neat IL and the crystal-
lized IL show some differences in the position of the vibration
bands of the SO3 and CH functions (Fig. S1 in the ESI†). In spite
of this, X-ray crystallography clearly shows that the protons do
in principle have the possibility to move from one sulfonate to
another one, which is a prerequisite for successful proton
transport.
17142 | J. Mater. Chem., 2012, 22, 17140–17146
This journal is ª The Royal Society of Chemistry 2012

View Online
Fig. 1 Crystal structure of [PmimSO3H][PTS] obtained from the white precipitate isolated after few months. (a) Asymmetric unit (since the H on the
SO3 functions are shared between two SO3 functions, they are formally noted as ½ for their occupation rate). Displacement ellipsoids are drawn at the
30% probability level. (b) View along the b axis showing the packing of the cationic and anionic parts of the IL.
and the organic pore wall functionalization in Py6 and Py8)36 and
the condensation of silanol groups, which is associated with a
further water elimination. In the pure monoliths (that is, without
IL), the last weight loss occurs at ca. 550 ꢁC. The residual weight
in the TGA curves at 900 ꢁC can be assigned to silica and carbon
residues. Overall these data show that, consistent with earlier
data,14,15 the overall fraction of IL in the ionogels is ca. 80%.
DTA confirms the above assignments as we observe an
endothermic signal until ca. 150 ꢁC (water evaporation) followed
by another endothermic signal between 300 and 400 ꢁC (ther-
mally induced decomposition of the IL and silanol crosslinking
reactions) and finally a strongly exothermic signal between ca.
400 and 630 ꢁC. This last signal is assigned to the thermal
decomposition of residual organic and carbonaceous material,
which is consistent with earlier data.36
Fig. 2 IR spectra (a) and Raman spectra (b) of the IL [PmimSO3H]
[PTS] (black), Fpy/IL (red), Py6/IL (blue) and Py8/IL (green). Dotted
lines are IR spectra of pure silica monoliths, solid lines are spectra of
ionogels.
between 250 and ca. 390 ꢁC can be attributed to the main
decomposition of the IL. The last step between ca. 400 and
ꢁ
ꢁ
650 C (400 and 500 C for the neat IL) can be ascribed to the
complete oxidative elimination of organic residues (from the IL
Table 1 Wavenumbers of the asymmetric and symmetric sulfonate
vibration and the respective differences in the neat IL and the ionogels
(cmꢀ1
)
Compound
nasym(SO3)
nsym(SO3)
Dn(SO3)
Fpy/IL
Py6/IL
Py8/IL
[PmimSO3H][PTS]
1070
1079
1104
1116
1030
1030
1032
1032
40
49
72
84
Fig. 3 TGA (solid lines) and DTA (dashed lines) data of IL (black),
Fpy/IL (red), Py6/IL (blue) and Py8/IL (green).
This journal is ª The Royal Society of Chemistry 2012
J. Mater. Chem., 2012, 22, 17140–17146 | 17143

View Online
Table 2 Elemental analysis data (%C, H, N, S, and Si) of the ionogels, calculated stoichiometry, and water content from EA and TGA
Samples
Fpya
%Si
%C
%H
%N
%S
Formula molecular weight
%Water
Calc.
Found
Calc.
Found
Calc.
Found
Calc.
Found
Calc.
—
—
4.28
—
—
—
—
13.13
12.71
—
42.30
7.67
7.67
36.69
36.70
6.76
6.76
39.04
39.00
5.83
5.83
1.76
4.82
5.28
1.58
2.26
5.02
5.38
1.25
2.05
5.42
5.47
0.12
5.51
5.95
2.56
2.43
6.07
5.98
2.14
2.02
5.96
5.71
Fpy/IL
Py6
34.41
33.52
15.36
15.75
36.40
36.55
12.85
14.09
35.75
35.58
SiO2(C14H20N2S2O6)0.75(H2O)1.33
M ¼ 366.11 g molꢀ1
6.54 (EA)
7.00 (TGA)
SiO2(C7H8N)0.14(H2O)0.09
M ¼ 76.51 g molꢀ1
—
Py6/IL
Py8
12.81
12.86
—
—
12.91
12.92
SiO2(C7H8N)0.14(C14H20N2S2O6)0.83(H2O)1.57
M ¼ 415.36 g molꢀ1
6.80 (EA)
8.20 (TGA)
SiO2(C7H8N)0.11(H2O)0.01
M ¼ 71.92 g molꢀ1
Found
Calc.
Found
Py8/IL
SiO2(C7H8N)0.11(C14H20N2S2O6)0.97(H2O)2.92
M ¼ 481.61 g molꢀ1
10.91 (EA)
10.10 (TGA)
a
The Fpy monolith does not contain organic functionalities; fractions of C, H, N, S are therefore due to the residual trace of Fpy.
Table 2 shows the results of elemental analysis (EA). The
amount of pyridyl dangling groups in the neat silica monoliths
Py6 or Py8 can directly be determined on the basis of the fraction
of C and N. The amount of adsorbed water can then be calcu-
lated via the difference between the total H content and the
content of H from the pyridyl dangling groups.
that were not cooled prior to the first heating cycle in the DSC
(Fig. 4b).
Upon incorporation into the silica hosts the DSC traces of the
IL do not change significantly. All ionogels exhibit a glass
transition, which only slightly varies with the silica matrix,
Tg(Fpy/IL) ¼ ꢀ12.8 ꢁC, Tg(Py6/IL) ¼ ꢀ9.7 ꢁC, and Tg(Py8/
IL) ¼ ꢀ10.3 ꢁC. A closer inspection of the DSC traces of the
ionogels, however, reveals that there is a weak additional signal at
around 50 to 65 ꢁC. As the signal is very weak, a definite assign-
ment is not possible, but the fact that it is endothermic suggests
that it could also originate from the melting process described
before. The inverse, exothermic, signal is not observed in the
cooling curves, again supporting the notion of significant
supercooling.
A similar calculation can be done for the ionogels. The amount
of IL in the ionogels Py6/IL and Py8/IL can be determined from
the total amount of S. This value can further be confirmed with
the values obtained for C and N. In the case of Fpy/IL, the
amount of IL in the ionogels can be determined directly based on
the amounts of C, N, and S since there is no dangling group in the
silica matrix (only Si, O, and H) and the amount of water is
calculated by the difference between the total amount of H and
the amount of H from the IL. The calculated values are in good
agreement with TGA. In particular, the quantities of water and
IL determined by EA are fairly close to those determined by
TGA. Table 2 clearly shows that the water fraction is highest in
the Py8-based and Py6-based ionogels, indicating that the addi-
tional functionalization with pendant pyridyl units of the pore
wall aids water uptake.
Ionic conductivity
Fig. 5 shows the ionic conductivity of the neat IL [PmimSO3H]
[PTS] as a function of temperature. Not surprisingly, the
conductivity increases with increasing temperature and reaches
ca. 10ꢀ3 S cmꢀ1 at 120 ꢁC. This is in agreement with the fact that
Fig. 4 shows representative DSC traces of the IL [PmimSO3H]
[PTS] before and after drying and of the ionogels. The as-prepared
neat IL only exhibits a glass transition at Tg ¼ ꢀ14.5 ꢁC. No
crystallization or melting peaks are observed. The cooling curves
suggest that the dynamics of the IL are rather slow, because Tg is
only poorly visible upon cooling. This suggests that [PmimSO3H]
[PTS], similar to other sulfonated ILs,24–26 shows significant
supercooling. It is also consistent with the findings on the very
slow crystallization of [PmimSO3H][PTS] described above.
DSC analysis of a thoroughly dried IL further confirms these
findings: even drying over 3 days at 60 ꢁC and subsequent heating
at 160 ꢁC for 10 minutes does not produce additional signals or
shift the previously observed signals. This confirms that the
signals assigned to Tg are indeed due to a glass transition and not
plasticization or similar effects of the IL by water incorporation.
However, as stated before, the IL does crystallize after a few
months and indeed DSC traces of these samples do show a weak
and broad but clearly visible endothermic peak at ca. 47 ꢁC.
Upon cooling, the corresponding exothermic peak is not
observed. We currently assign this signal to a poorly defined
melting transition because a similar but more pronounced and
slightly shifted endothermic signal has been observed in samples
Fig. 4 DSC curves (a, b) of as-synthesized IL (black), of dried IL
(violet), of crystallized IL (after several months of storage, orange) and
(c) of ionogels Fpy/IL (red), Py6/IL (blue), and Py8/IL (green). Panel (b)
shows measurements of samples that were directly heated from room
temperature (no prior cooling) to 150 ꢁC. These samples show a broad
endothermic signal that can be assigned to a melting transition in the
crystalline IL. The signal is most pronounced in the samples isolated after
several months of storage (orange curve). The notation of a strong
supercooling in the IL (see text for details) is supported by the fact that no
crystallization peak is observed on cooling.
17144 | J. Mater. Chem., 2012, 22, 17140–17146
This journal is ª The Royal Society of Chemistry 2012

View Online
cooling curves reaching a conductivity of 10ꢀ3 S cmꢀ1 at 120 ꢁC,
Fig. 6 and Fig. S3 in the ESI.†
Discussion
We have investigated ionogels based on three different silica
matrix materials and an acidic IL, [PmimSO3H][PTS]. Both the
synthesis of the IL and the silica matrix are straightforward, as is
the preparation of the final ionogels. The neat IL shows signifi-
cant supercooling, but does crystallize after a few months at
ambient conditions. Interestingly, the IL crystallizes without
taking up significant amounts of water in the crystal structure.
The single crystal structure also shows significant hydrogen
bonding (Fig. 1) between sulfonate moieties, where two anionic
groups share one proton. Moreover, the crystalline IL exhibits a
layered structure of sulfonate/proton-rich layers separated by
more hydrophobic and cationic layers, which, possibly favors
anisotropic proton transport.
Similar to other ionogels,3,4,9,14–16,39,40 IR and Raman spec-
troscopies (Fig. 2 and Table 1) along with TGA and EA (Fig. 3
and Table 2) show that the major component of the ionogels
made by incorporating the IL into the silica hosts is the IL; the
main role of the matrix is the fixation of the IL in a macroscopic
object. IR and Raman spectroscopies further suggest that the IL
essentially hydrogen bonds to itself (Table 1), consistent with the
IL crystal structure. The interaction between the IL and the silica
host appears much weaker.
Fig. 5 Conductivity of the neat IL [PmimSO3H][PTS] vs. temperature
during heating (open circles) and cooling (open squares).
the IL viscosity decreases with increasing temperature leading to
a higher mobility and hence higher conductivity (since viscosity
and conductivity are inversely related).
Fig. 6 shows the ionic conductivity of the ionogels vs. temper-
ature. The conductivity is on the same order of magnitude as that
of the neat IL (Fig. 5). As in the case of the neat IL, the conduc-
tivity increases with increasing temperature and reaches ca. 10ꢀ2 S
cmꢀ1 at 120 ꢁC. It must be noted that the ionic conductivity,
similar to Nafion,37,38 depends on the hydration. Indeed, contrary
to the neat IL, the as-synthesized ionogels show a clear difference
between the heating and cooling curves. This difference can be
explained by the presence of water in the ionogels (see Table 2),
which evaporates during the course of the experiment and directly
affects the conductivity.
DSC (Fig. 4) and ionic conductivity measurements (Fig. 5 and
6) clearly show that the ionogels exhibit useful, temperature-
dependent properties. DSC suggests that the IL exhibits strong
supercooling both in the neat IL and in the ionogels. This is
comparable to other sulfonated ILs24–26 (and it would be inter-
esting to also evaluate the crystal structures of these ILs) and is
also further supported by macroscopic observation of the IL,
which only crystallizes after a few months. The ionic conductivity
reaches ca. 10ꢀ2 S cmꢀ1 at 120 ꢁC in the hydrated and ca. 10ꢀ3 S
cmꢀ1 at 120 ꢁC in the dry state. In combination with the observed
supercooling, the ionogels could be interesting as proton trans-
port membranes, similar to a recent report by Ye et al.41 who
showed that composites of ordered mesoporous SBA-15 silica,
1-butyl-3-methylimidazolium bis(trifluoromethyl)sulfonyl imide
[Bmim][NTf2] and poly(methyl methacrylate) also have
conductivities on the order of 10ꢀ2 S cmꢀ1. The advantage of the
current system lies in the fact that the synthesis is much simpler
and there is no need for a polymeric component, which makes the
material attractive for higher operation temperatures.
To analyze the effect of water on the ionic conductivity, the
ionogels were therefore dried under vacuum at 55 ꢁC overnight,
followed by another measurement of the conductivity. For
comparison, the ionogels were also dried at 120 ꢁC for two hours
under N2 (which is a fairly harsh process for ILs). All treatments
leave the ionogels virtually unaffected, that is, there is no color
change and the IR spectra are identical to those of the as-
prepared samples (Fig. S2 in the ESI†). However, the drying
process does have aꢁn influence on the conductivity; the higher
temperature of 120 C leads to a somewhat lower conductivity
than the drying at 55 ꢁC. This suggests that the lower drying
temperature is not sufficient to completely dry the ionogel; this
assumption is further supported by the observation that only the
ꢁ
ionogels dried at 120 C show superposition of the heating and
The main advantage over the currently used Nafion
membranes is the fact that the ionogels are stable up to at least
130 ꢁC, without a significant loss of their conductivity (as the
ꢁ
ionogels as such are stable up to ca. 200 C in dynamic TGA,
even higher operation temperatures may be possible; these were,
however not accessible with the existing impedance spectroscopy
setup). In contrast, the operating temperature of conventional
Nafion membranes is limited to ca. 80 ꢁC because the water
necessary for proper operation will evaporate at higher temper-
atures.42 As [PmimSO3H][PTS] does not have a measurable
vapor pressure, the temperature window for application is
essentially limited by the thermal stability of the ionogel (upper
limit) and possibly by the supercooling/crystallization (lower
Fig. 6 (a) Conductivity of Fpy/IL (red), Py6/IL (blue) and Py8/IL
(green) vs. temperature during heating (open symbols) and cooling
(closed symbols). (b) Conductivity of Fpy/IL (red), Py6/IL (blue) and
Py8/IL (green) vs. temperature (after drying for 2 hours at 120 ꢁC under
N2 stream) during heating (open symbols) and cooling (closed symbols).
This journal is ª The Royal Society of Chemistry 2012
J. Mater. Chem., 2012, 22, 17140–17146 | 17145

View Online
7 E. Delahaye, Z. L. Xie, A. Schaefer, L. Douce, G. Rogez, P. Rabu,
€
C. Gunter, J. S. Gutmann and A. Taubert, Dalton Trans., 2011, 40,
9977.
8 H. E. Bergna and W. O. Roberts, Colloidal Silica – Fundamentals and
Applications, Taylor & Francis, Boca Raton - London - New York,
2006.
9 M. A. Neouze, J. Le Bideau and A. Vioux, Prog. Solid State Chem.,
2005, 33, 217.
10 T. Echelmeyer, H. W. Meyer and L. van Wullen, Chem. Mater., 2009,
21, 2280.
11 D. Petit, J. P. Korb, P. Levitz, J. LeBideau and D. Brevet, Comptes
Rend. Chim., 2010, 13, 409.
limit). In contrast, the chemical functionalization of the pore wall
(Py6 and Py8) does not appear to have a measurable effect on the
ionic conductivity. This is consistent with the IR data (Table 1),
which suggest a preference for intra-IL hydrogen bonding and
much less interaction with the wall of the silica host. This may
however strongly depend on the specific nature of the chemical
groups immobilized on the silica, as there are several reports
showing strong influence of surfaces on the IL
behavior.11,14–16,19,20,40,43
At this point it must be noted that there are also examples of
polymer-based IL membranes. Some of these membranes reach
conductivities up to 10ꢀ2 S cmꢀ1 and exhibit a very high thermal
stability as well. They are based on Nafion-115 and [Bmim][Cl] or
[Bmim][OH].44 An example reported by Hoarfrost and Segalman
is based on poly(styrene-block-vinylpyridine) block copoly-
mers.45 The advantage of the current material over these systems
is the low cost and the ease of preparation.
12 E. Quartarone and P. Mustarelli, Chem. Soc. Rev., 2011, 40, 2525.
13 F. Gayet, L. Viau, F. Leroux, F. Mabille, S. Monge, J. J. Robin and
A. Vioux, Chem. Mater., 2009, 21, 5575.
€
14 R. Gobel, A. Friedrich and A. Taubert, Dalton Trans., 2010, 39, 603.
€
€
15 R. Gobel, P. Hesemann, J. Weber, E. Moller, A. Friedrich,
S. Beuermann and A. Taubert, Phys. Chem. Chem. Phys., 2009, 11,
3653.
16 J. Le Bideau, P. Gaveau, S. Bellayer, M. A. Neouze and A. Vioux,
Phys. Chem. Chem. Phys., 2007, 9, 5419.
17 S. Bellayer, L. Viau, Z. Tebby, T. Toupance, J. Le Bideau and
A. Vioux, Dalton Trans., 2009, 1307.
18 M. A. Neouze, J. LeBideau, F. Leroux and A. Vioux, Chem.
Commun., 2005, 1082.
19 Y. S. Liu, G. Z. Wu, H. Y. Fu, Z. Jiang, S. M. Chen and M. L. Sha,
Dalton Trans., 2010, 39, 3190.
It is interesting to note that, to date, all studies have been per-
formed with commercially available and fairly standard ILs based
ꢀ
on halide or NTf2 anions. The resulting materials have conduc-
tivities that are comparable to more traditional membranes,46 but
have the potential to operate at higher temperatures. The use of
sulfonated ILs, such as those studied here, should in principle
enable a much more efficient proton transport and also avoid
problems associated with the redox chemistry associated with, e.g.,
halide anions. The major issue in this context is the very high
viscosity of sulfonated ILs24–26 which apparently impedes a better
performance of these ILs in proton transport. This will therefore
be a key issue to address in the future.
20 M. Sha, G. Wu, Q. Dou, T. Tang and H. Fang, Langmuir, 2010, 26,
12667.
21 Z. L. Xie, F. Wang, A. Jelicic, A. Friedrich, P. Rabu, S. Beuermann
and A. Taubert, J. Mater. Chem., 2010, 20, 9543.
22 Z. Du, Z. Li and Y. Deng, Synth. Commun., 2005, 35, 1343.
23 D.-Q. Xu, J. Wu, S.-P. Luo, J.-X. Zhang, J.-Y. Wu, X.-H. Du and
Z.-Y. Xu, Green Chem., 2009, 11, 1239.
24 C. P. Fredlake, J. M. Crosthwaite, D. G. Hert, S. N. V. K. Aki and
J. F. Brennecke, J. Chem. Eng. Data, 2004, 49, 954.
25 H. Xing, T. Wang, Z. Zhou and Y. Dai, Ind. Eng. Chem. Res., 2005,
44, 4147.
26 Q. Yang, Z. Wei, H. Xing and Q. Ren, Catal. Commun., 2008, 9, 1307.
27 X. Liu, J. Zhou, X. Guo, M. Liu, X. Ma, C. Song and C. Wang, Ind.
Eng. Chem. Res., 2008, 47, 5298.
Conclusion
28 G. M. Sheldrick, SHELXS-97 Program for Crystal Structure Solution,
€
Ionogels based on sulfonic acid/sulfonate ILs are attractive
candidates for proton transport membranes. Of particular
interest is the fact that the ionogels are stable at temperatures
that are interesting for fuel cells. The current approach provides
a strategy for further membrane development. In particular the
use of thermally stable sulfonated ILs with a low viscosity could
become a key approach towards advanced membranes.
€
University of Gottingen, Gottingen, 1997.
29 G. M. Sheldrick, SHELXL-97 Program for the Refinement of Crystal
€
Structures from Diffraction Data, University of Gottingen, Gottingen,
€
1997.
30 L. J. Farrugia, J. Appl. Crystallogr., 1999, 32, 837.
31 L. J. Farrugia, J. Appl. Crystallogr., 1997, 30, 565.
32 W. M. Reichert, P. C. Trulove and H. C. De Long, Acta Crystallogr.,
2010, E66, o591.
33 C. Bauer, P. Jacques and A. Kalt, Chem. Phys. Lett., 1999, 307, 397.
34 E. Delahaye, S. Eyele-Mezui, J. F. Bardeau, C. Leuvrey, L. Mager,
P. Rabu and G. Rogez, J. Mater. Chem., 2009, 19, 6106.
35 E. Delahaye, S. Eyele-Mezui, M. Diop, C. Leuvrey, P. Rabu and
G. Rogez, Dalton Trans., 2010, 39, 10577.
36 Z. Li, A. Friedrich and A. Taubert, J. Mater. Chem., 2008, 24, 1008.
37 T. A. Zawodzinski, M. Neeman, L. O. Sillerud and S. Cottesfeld,
J. Phys. Chem., 1991, 95, 6040.
Acknowledgements
E.D. acknowledges an MPG-CNRS postdoctoral fellowship.
Financial support by the Deutsche Forschungsgemeinschaft
(TA571/2-1 and 7-1), the University of Potsdam, the MPI of
Colloids and Interfaces (Colloid Chemistry Department), and
the MPI for Polymer Research is gratefully acknowledged.
38 T. A. Zawodzinski, C. Derouin, S. Radzinski, R. J. Sherman,
V. T. Smith, T. E. Springer and S. Gottesfeld, J. Electrochem. Soc.,
1993, 140, 1041.
39 M. A. Neouze, J. Mater. Chem., 2010, 20, 9593.
40 M. A. Neouze and M. Litschauer, J. Phys. Chem. B, 2008, 112, 16721.
41 Y.-S. Ye, G.-W. Liang, B.-H. Chen, W.-C. Shen, C.-Y. Tseng,
M.-Y. Cheng, J. Rick, Y.-J. Huang, F.-C. Chang and B.-J. Hwang,
J. Power Sources, 2011, 196, 5408.
References
1 P. Wasserscheid and T. Welton, Ionic Liquids in Synthesis, Wiley-
VCH, Weinheim, 2007.
42 G. Lakshminarayana and M. Nogami, Solid State Ionics, 2010, 18,
760.
2 B. Kirchner, Ionic Liquids, Springer, Heidelberg - Dordrecht -
London - New York, 2009.
43 B. Coasne, L. Viau and A. Vioux, J. Phys. Chem. Lett., 2011, 2, 1150.
44 J. Yang, Q. Che, L. Zhou, R. He and R. F. Savinell, Electrochim.
Acta, 2011, 56, 5940.
45 M. L. Hoarfrost and R. A. Segalman, Macromolecules, 2011, 44,
5281.
3 M. A. Neouze, J. Le Bideau, P. Gaveau, S. Bellayer and A. Vioux,
Chem. Mater., 2006, 18, 3931.
4 M. A. Neouze, J. Le Bideau, F. Leroux and A. Vioux, Chem.
Commun., 2004, 1082.
5 A. Vioux, L. Viau, S. Volland and J. Le Bideau, Comptes Rend. Chim.,
2010, 13, 242.
6 J. Le Bideau, L. Viau and A. Vioux, Chem. Soc. Rev., 2011, 40, 907.
46 N. W. Deluca and Y. A. Elabd, J. Polym. Sci., Part B: Polym. Phys.,
2006, 44, 2201.
17146 | J. Mater. Chem., 2012, 22, 17140–17146
This journal is ª The Royal Society of Chemistry 2012