Introduction

Nanostructured polymeric materials have attracted much attention as platforms for developing new series of ionic liquids (ILs), because there is much potential to demonstrate advanced functions originating not only from the component molecules but also from their structural and morphological characteristics.1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 The increase in molecular weight and the presence of charges often cause an increase in the melting temperature. We have recently reported a polyhedral oligomeric silsesquioxane (POSS)-based room temperature IL that displays a melting temperature below 25 °C.16 The bonding of ion pairs to POSS significantly improved the thermal stability and decreased the melting temperatures, even with the significant increase in molecular weight. However, the mechanism of these changes caused by bonding to POSS remains unclear. In particular, which characteristics of POSS, such as the cubic structure or the rigidity of the silica cage, predominantly affect the thermal properties of ion pairs needs to be investigated. An understanding of the contribution of POSS to the thermodynamics of ILs can provide guidelines for future molecular designs for polymer- and macromolecule-based ILs.

The typical POSS molecule possesses a cubic structure represented by the formula R8Si8O12, in which the central hydrophobic core (Si8O12) is functionalized with organic moieties (R) at the eight vertices.17, 18, 19, 20, 21, 22, 23, 24 In polar solvents, POSS can form a globular conformation and create a distinct hydrophobic space around the POSS core.25, 26, 27, 28 Such structural features can contribute to reduce the packing density and isolating the distal ion pairs. Moreover, the molecular motion of substituents of the POSS core can be efficiently restricted because of the rigidity of the silica framework.29 Indeed, POSS has been used as a filler for improving the thermal and mechanical properties of conventional plastics by suppressing the molecular rotation of polymer chains.30

Herein, we prepared a series of modified POSS with variable numbers of imidazolium–carboxylate pairs (n=2, 4, 6 and 8) and investigated the relationships between structures and thermal properties (Figure 1). We hypothesized that the larger number of ion pairs connected to POSS induced the star-shaped structure31 because of electron repulsion. Therefore, the modified POSS molecules should show low packing density. In contrast, the modified POSS molecules were able to aggregate and condense into smaller numbers of ion pairs because of the strong hydrophobicity of POSS. On the basis of these ideas about the conformation, we considered the thermodynamic behaviors of POSS-tethered ion pairs in this article.

Figure 1
figure 1

Chemical structures and proposed conformations of the polyhedral oligomeric silsesquioxane-based ionic compounds. A full color version of this figure is available at Polymer Journal online.

Full size image

Experimental Procedure

General

1H nucleic magnetic resonance (NMR) and 13C NMR spectra were measured with a JEOL EX–400 (400 MHz for 1H and 100 MHz for 13C) spectrometer (JEOL Ltd., Tokyo, Japan). 29Si NMR spectra were measured with a JEOL JNM–A400 (80 MHz) spectrometer. The coupling constants (J value) are reported in Hertz. The chemical shifts are expressed in p.p.m. downfield from tetramethylsilane, using residual chloroform (δ=7.24 in 1H NMR, δ=77.0 in 13C NMR) as an internal standard. The mass spectra were obtained on a JEOL JMS–SX102A. Water abundances were evaluated with a Karl–Fischer Moisture Titrator MKC–501, Kyoto Electronics Manufacturing (Kyoto, Japan).

Bromide salt of imidazolium cation 1, [Bmim]Br

Freshly distilled 1-bromobutane (42.82 ml, 0.400 mol) was added to distilled 1-methylimidazole (16.42 g, 0.200 mol) with vigorous stirring under nitrogen at 0 °C. The mixture was stirred under nitrogen for 1 week at ambient temperature. After the reaction was completed, the excessive phase-separated 1-bromobutane was decanted. Then, the molten salt was washed with ethyl acetate and dried in vacuo at 50 °C for 24 h to obtain compound 1 as a white solid (42.8 g, 0.196 mmol, 98%). The characterization was executed according to Huddleston et al.32

Octaammonium POSS 3

(3-Aminopropyl)triethoxysilane (100 ml, 0.427 mol)25, 33, 34, 35 and concentrated hydrochloric acid (35–37%, 135 ml) in methanol (800 ml) produced compound 5 as a white precipitate after 5 days at room temperature. The product was obtained after filtration, washing with cold methanol and drying. Compound 3 was spectroscopically pure at a 30% yield (18.8 g). 1H NMR (dimethyl sulfoxide (DMSO)-d6) δ 8.23 (s, 24H), 2.76 (t, 16H), 1.71 (m, 16H), 0.72 (t, 16H). 13C NMR (DMSO-d6) δ 40.53, 20.13 and 7.96. 29SiNMR (DMSO-d6) δ −66.4 (s).

Compound 4, POSS-(COOH)8

To a solution of compound 3 (20.0 g, 17.0 mmol) and triethylamine (20 ml, 14.4 mmol) in methanol (1 l), succinic anhydride (80 g, 0.80 mol) was added, and the reaction mixture was stirred at room temperature for 2 h and then condensed by evaporation.25 Chloroform was poured into the reaction solution, and the white precipitation was collected via filtration and washed with chloroform and tetrahydrofuran. Subsequently, the product was resolved in 60 ml of formic acid and reprecipitated by the addition of 1 l of water. The white precipitation was filtered and washed with water until the filtrate indicated pH 7. The desired product 4 was obtained as a white solid after being dried in vacuo (19.2 g, 11.4 mmol, 67%). 1H NMR (DMSO-d6) δ 0.59 (br, 2H), 1.44 (br, 2H), 2.30 (br, 2H), 2.41 (br, 2H), 3.02 (br, 2H). 13C NMR (DMSO-d6) δ 8.74, 22.48, 29.15, 29.98, 41.00, 170.92 and 173.89. 29Si NMR (DMSO-d6) δ −65.3. Low-resolution mass spectrometry (LRMS; 3-nitrobenzyl alcohol (NBA)) [(M+H)+] calcd. 1680, found 1680. HRMS (NBA) [M+H+] calcd. 1680.4081, found 1680.4041.

Compound 6, Arm-COOH

Succinic anhydride (60.0 g, 0.6 mol) was dissolved into 200 ml of dioxane, and propylamine (35.4 g, 0.6 mol) in 200 ml of dioxane was slowly added.16 The solution was warmed to 80 °C and stirred for 40 min. The succinamic acid was crystallized by cooling after the reaction. The white crystal 6 was filtered, dried and recrystallized from dioxane (76.2 g, 0.48 mol, 80%). 1H NMR (DMSO-d6) δ 0.79 (t, 3H, J=7.43 Hz), 1.37 (m, 2H), 2.28 (t, 2H, J=6.76 Hz), 2.40 (t, 2H, J=6.76 Hz), 2.97 (m, 2H). 13C NMR (DMSO-d6) δ 11.37, 22.37, 29.20, 30.00, 40.29, 170.74, 173.84: LRMS (NBA) [(M+H)+] calcd. 160, found 160. HRMS (NBA) [M+H+] calcd. 160.0974, found 160.0978. Anal. calcd. for C7H13NO3: C, 52.82; H, 8.23; N, 8.80; O, 30.15. Found: C, 52.71; H, 8.08; N, 8.80; O, 29.99.

General procedure for the preparation of the ILs (5a–d, 7)

The desired equivalent mole of bromide anion 1 for the carboxyl groups was converted into compound 2 by anion exchange resin (Amberlite–IRA400) in water and neutralized with each carboxyl compound suspended in methanol (2 l).16 The aqueous solution was concentrated by a rotary evaporator, and the residual liquid was freeze dehydrated to give a white solid. The solid was dried in vacuo and stored in a glove box. Arm-Im, 7: 1H NMR (DMSO-d6) δ 0.82 (t, 3H, J=7.43 Hz), 0.90 (t, 3H, J=7.31 Hz) 1.25 (m, 2H) 1.35 (m, 2H) 1.74 (m, 2H) 2.01 (t, 2H, J=6.76 Hz), 2.13 (t, 2H, J=6.76 Hz), 2.94 (m, 2H) 3.85 (s, 3H) 4.16 (t, 2H, J=7.19 Hz) 7.71 (m, 1H) 7.78 (m, 1H) 9.04 (s, 1H) 9.31 (s, 1H). 13C NMR (DMSO-d6): δ 11.41, 13.25, 18.76, 22.47, 31.38, 33.80, 34.90, 35.59, 40.12, 48.36, 122.21, 123.55, 137.24, 173.38, 174.70. LRMS (NBA) [(M+[Bmim+])+] calcd. 436, found 436. HRMS (NBA) [(M+[Bmim+])+] calcd. 436.3282, found 436.3297. Anal. calcd. for C15H27N3O3: C, 60.58; H, 9.15; N, 14.13; Br, 0. Found: C, 57.26; H, 9.40; N, 13.43; Br, 0. POSS-Im2, 5a: 1H NMR (DMSO-d6) δ 0.59 (t, 16H, J=7.92 Hz) 0.90 (t, 6H, J=7.31 Hz) 1.26 (m, 4H) 1.44 (br, 16H) 1.76 (m, 4H) 2.27 (br, 16H), 2.35 (br, 16H), 3.00 (m, 16H) 3.85 (s, 6H) 4.16 (t, 4H, J=7.07 Hz) 7.74 (s, 8H) 7.81 (s, 8H) 8.98 (s, 8H) 9.59 (s, 8H). 13C NMR (DMSO-d6) δ 8.75, 13.27, 18.77, 22.51, 30.61, 30.97, 31.34, 35.70, 40.93, 48.45, 122.13, 123.48, 136.42, 171.28, 174.28. 29Si NMR (DMSO-d6) δ −66.1. LRMS (NBA) [(M+H)+] calcd. 1959, found 1959. Anal. calcd. for C72H124N12O36Si8: C, 44.15; H, 6.38; N, 8.58; Br, 0. Found: C, 40.11; H, 6.17; N, 8.41; Br, 0. POSS-Im4, 5b: 1H NMR (DMSO-d6) δ 0.57 (t, 16H, J=7.92 Hz) 0.90 (t, 12H, J=7.31 Hz) 1.25 (m, 8H) 1.42 (br, 16H) 1.76 (m, 8H) 2.25 (br, 32H), 2.94 (m, 16H) 3.87 (s, 24H) 4.16 (t, 8H, J=7.19 Hz) 7.74 (s, 8H) 7.81 (s, 8H) 8.98 (s, 8H) 9.59 (s, 8H). 13C NMR (DMSO-d6) δ 8.77, 13.27, 18.77, 22.58, 31.35, 31.93, 32.07, 35.67, 40.83, 48.43, 122.11, 123.46, 136.57, 171.87, 174.78. 29Si NMR (DMSO-d6) δ −66.1. LRMS (NBA) [(M+H)+] calcd. 2236, found 2236. Anal. calcd. for C88H152N16O36Si8: C, 47.29; H, 6.86; N, 10.03; Br, 0. Found: C, 42.54; H, 6.92; N, 9.70; Br, 0. POSS-Im6, 5c: 1H NMR (DMSO-d6) δ 0.53 (t, 16H, J=8.16 Hz) 0.89 (t, 18H, J=7.31 Hz) 1.25 (m, 12H) 1.41 (br, 16H) 1.76 (m, 12H) 2.15 (br, 16H), 2.19 (br, 16H), 2.95 (m, 16H) 3.87 (s, 18H) 4.17 (t, 12H, J=7.19 Hz) 7.72 (s, 8H) 7.79 (s, 8H) 8.71 (s, 8H) 9.38 (s, 8H). 13C NMR (DMSO-d6) δ 8.82, 13.26, 18.77, 22.65, 31.37, 32.74, 33.44, 35.59, 40.84, 48.36, 122.09, 123.43, 136.95, 172.51, 174.86. 29Si NMR (DMSO-d6) δ −66.2. LRMS (NBA) [(M+H)+] calcd.2511, found 2511. Anal. calcd. for C104H180N20O36Si8: C, 49.74; H, 7.22; N, 11.15; Br, 0. Found: C, 44.76; H, 7.68; N, 10.71; Br, 0. POSS-Im8, 5d: 1H NMR (DMSO-d6) δ 0.53 (t, 16H, J=8.16 Hz) 0.88 (t, 24H, J=7.31 Hz) 1.23 (m, 16H) 1.38 (br, 16H) 1.75 (m, 16H) 2.05 (t, 16H, J=7.09 Hz), 2.18 (t, 16H, J=7.09 Hz), 2.94 (m, 16H) 3.87 (s, 24H) 4.17 (t, 16H, J=7.19 Hz) 7.74 (s, 8H) 7.81 (s, 8H) 8.98 (s, 8H) 9.59 (s, 8H). 13C NMR (DMSO-d6) δ 8.88, 13.26, 18.77, 22.68, 31.39, 33.51, 34.74, 35.54, 40.83, 48.30, 122.09, 123.41, 137.28, 173.12, 174.76. 29Si NMR (DMSO-d6) δ −66.2. LRMS (NBA) [(M+H)+] calcd. 2787, found 2787. Anal. calcd. for C120H208N24O36Si8: C, 51.70; H, 7.52; N, 12.06; Br, 0. Found: C, 45.21; H, 7.70; N, 11.30; Br, 0.

Differential scanning calorimetry (DSC)

DSC thermograms were carried out on an SII DSC 6220 instrument using approximately 10 mg of exactly weighed samples. The sample on the aluminum open pan was cooled to −130 °C at a rate of 10 °C min−1 under flowing nitrogen (30 ml min−1) and then heated from −130 to 80 °C at the same rate. The glass transition (Tg) and melting temperatures (Tm) were determined as the onset of the second curves to eliminate heat history. The fusion enthalpy (ΔHfus) was calculated from the areas of the endothermic peaks at the first cycle with the completely crystallized samples soaked in the liquid nitrogen before the measurements.

Thermogravimetric analysis (TGA)

TGA was performed on an EXSTAR TG/DTA6220, Seiko Instrument (Seiko Instruments Inc., Chiba, Japan), with a heating rate of 10 °C min−1 up to 900 °C under flowing nitrogen (200 ml min−1). Residual water was removed by keeping the sample on the platinum pan at 110 °C for 1 h before the curve profiling. The decomposition temperatures (Td) were determined from the onset of the weight loss.

Results and discussion

The series of POSS samples with variable numbers of ion pairs were prepared according to Scheme 1.16, 25, 32, 33, 34, 35 The formation of the ion pairs, [R-COO−][Bmim+], was performed via acid–base neutralization using POSS-(COOH)8 and [Bmim]OH.16 The stoichiometry between POSS and ion pairs was adjusted by changing the feeding ratios. To avoid the degradation of the POSS cage in the presence of excess [Bmim]OH, the reaction was performed by titration of the diluted [Bmim]OH with a suspension of POSS-(COOH)8. After being dried in vacuo, each product was obtained as a colorless and transparent product.

All samples containing the POSS moiety produced single peaks at −66 p.p.m. in the 29Si NMR spectra assigned to the T8 POSS structure. Integration of the peaks in the 1H NMR spectrum indicated the formation of ion pairs with desired ratios and an imidazolium cation. The samples were stored in a glove box under an argon atmosphere, and the water abundance was kept below 1.5 wt%, as determined by the Karl Fischer method. The concentration of residual bromide ion was lower than the detectable level in the elemental analysis. Therefore, we concluded that all products were sufficiently pure for the following analyses. We prepared the ion pair, Arm-Im, for comparison purposes to evaluate the effects of the connection to POSS.

The thermal stability of the synthetic compounds against pyrolysis was investigated by TGA. Figure 2a shows the TGA profiles of the synthetic compounds, and Table 1 summarizes the Td values evaluated from the onsets of the TGA profiles. We observed two significant weight losses in all samples. Because similar molecular weights were obtained after the first weight loss, this step can be assigned as the degradation of the carboxyl groups (Figure 2b). Therefore, the Td value represents the influence on the thermal stability of the carboxyl groups by introducing [Bmim+] to POSS. The Td values of POSS-Im6 and POSS-Im8 were higher than those of Arm-Im and POSS-(COOH)8. On the contrary, POSS-Im2 and POSS-Im4 showed lower Td values than those of Arm-Im and POSS-(COOH)8. These results indicate that the larger number of ion pairs accumulated on POSS may be necessary to express the characteristic changes of thermal stability observed from POSS-Im6 and POSS-Im8. This is likely because the probability of blocking by imidazolium should increase in these compounds from pyrolysis of free carboxyl groups. In addition, the existence of POSS can hardly contribute to improve thermal stability because less significant enhancements were observed from POSS-Im2 and POSS-Im4. These results imply that the thermal properties of POSS-based ILs might be influenced not by the rigidity but by the structural feature of POSS. It should be mentioned that there should be distinct differences in properties between POSS-Im4 and POSS-Im6.

Figure 2
figure 2

Thermogravimetric analysis thermograms of the ionic compounds under flowing nitrogen. POSS, polyhedral oligomeric silsesquioxane. A full color version of this figure is available at Polymer Journal online.

Full size image
Table 1 Thermodynamic parameters of the ionic compounds determined from the DSC and TGA curves
Full size table

The DSC analysis was performed at a heating rate of 10 °C min−1 (Figure 3). The Tm values of the obtained compounds are listed in Table 1. Endothermic peaks assigned to the melting process were observed below 100 °C from each of the POSS-containing compounds except POSS-Im2. Therefore, POSS-Im4, POSS-Im6 and POSS-Im8 can be categorized as ILS. Remarkably, POSS-Im6 and POSS-Im8 showed melting below 25 °C for the second cycle. Thus, they can be classified as a room temperature ILs. The Tm values of POSS-Im6 and POSS-Im8 were 19 and 23 °C lower than that of Arm-Im. In contrast, POSS-Im2 and POSS-Im4 showed less significant Tm values below 100 °C. These data clearly indicate that a large number of ion pairs may be required for decreasing the Tms of POSS-tethered ion compounds. From these results involving the TGA data, it can be summarized that the accumulation of a larger number of ion pairs on POSS is essential to obtain the characteristic thermal property of POSS-based ILs. In addition, the rigidity of POSS seems to have less of an effect on the thermal properties. Corresponding to the results of TGA, there should be significant differences in the properties of POSS-Im4 and POSS-Im6.

Figure 3
figure 3

The differential scanning calorimetry curves of the ionic compounds for the second scan. POSS, polyhedral oligomeric silsesquioxane. A full color version of this figure is available at Polymer Journal online.

Full size image

To understand the thermodynamics of the melting process of ion pairs, the fusion enthalpies and entropies (ΔHfus and ΔSfus) per single imidazolium molecule were evaluated from the areas of the endothermic peaks observed in the DSC profiles (Table 1). Lower ΔHfus values than that of Arm-Im, which is a characteristic of a POSS-based IL, were observed for POSS-Im6 and POSS-Im8. Conversely, POSS-Im2 and POSS-Im4 showed larger values of ΔHfus. These results can be explained by the structural features of POSS. The star-shaped structures of the modified POSS molecules could be induced by electrostatic repulsion, resulting in the isolation of the distal ion pairs and large exclusive volumes.31 These structures should disrupt the formation of the thermally stable conformation and reduce the packing density. Thus, the interaction between the POSS molecules could be weakened in POSS-Im6 and POSS-Im8. Moreover, smaller values of ΔSfus for POSS-Im6 and POSS-Im8 can also support the contribution of the structural features of POSS to the thermal properties. The star-shaped structure should reduce the conformational variety of the ion pairs because of the higher symmetry.36 In addition to the intrinsic symmetry of the cubic structure of POSS, POSS-Im6 and POSS-Im8 should show much smaller ΔSfus values. These data can be summarized as follows. When a larger number of ion pairs are accumulated on POSS, the star-shaped structure should be induced. Larger exclusive volumes and higher symmetry of the POSS molecules could reduce the intermolecular interaction. Hence, the melting temperatures could decrease.

To support the conformational speculation, X-ray diffraction analysis analysis was performed (Figure 4). From the samples of POSS-Im6 and POSS-Im8, several peaks in the region from 2θ=5° (d=17 Å) to 15° (d=6 Å) were observed. In contrast, dispersed patterns were obtained around 2θ=23° (d=4 Å) from POSS-Im2 and POSS-Im4. These results indicate that POSS-Im6 and POSS-Im8 can involve large regular structures. It is implied that star-shaped structures of POSS-Im6 and POSS-Im8 could create the exclusive areas around the POSS core. In the case of POSS-Im2 and POSS-Im4, it was presumed that the aggregation at the POSS moiety because of the hydrophobic interaction and at the ionic groups because of the electrostatic interaction could induce the miscellaneous conformation. These assumptions can be supported by the conformation.

Figure 4
figure 4

X-ray diffraction analysis patterns of the ionic compounds. POSS, polyhedral oligomeric silsesquioxane. A full color version of this figure is available at Polymer Journal online.

Full size image

The chemical shifts of the peak assigned as the two-position of imidazolium were compared in the 1H NMR spectra (Figure 5). It has been reported that the strength of the solubilizing ability of cellulose by imidazolium-based ILs can be evaluated from the width of the downfield shift of this peak.37 Correspondingly, by increasing the number of ion pairs, the peak position showed a larger downfield shift. These data suggest that the strength of hydrogen bonds was significantly enhanced in POSS-Im6 and POSS-Im8. Exclusion volumes originating from the star-shaped structure could generate hydrophobic spaces inside molecules. It was implied that the strength of the hydrogen bonding might be enhanced by locating it at relatively hydrophobic spaces.

Figure 5
figure 5

1H nucleic magnetic resonance spectra of the ionic compounds in dimethyl sulfoxide at 25 °C. POSS, polyhedral oligomeric silsesquioxane. A full color version of this figure is available at Polymer Journal online.

Full size image

Conclusion

In conclusion, we present findings indicating that POSS-tethered imidazolium salts can work as room temperature ILs, despite the inclusion of modified POSS, which has multiple ion pairs in a single molecule. We demonstrate the significance of the cubic structure of POSS to show the characteristic thermal properties in POSS-based molten salts. It was suggested that the increase of thermal stability and the decrease of melting temperatures of ion pairs tethered to POSS should originate from the star-shaped structure created by the cubic framework. These results imply that by replacing the carboxyl group with other thermally stable anions, POSS-based ILs could be improved. In addition, it was demonstrated that the strength of the hydrogen bonding of carboxyl groups can be significantly enhanced by connecting them to POSS. POSS might create distinct hydrophobic spaces. These results suggest that POSS-ILs might be applicable as a superior medium for solubilizing polysaccharides or other insoluble materials. Our findings described here are important not only for modulating the functions of ILs but also for designing new advanced ILs based on POSS.

scheme 1

Synthesis of the ionic compounds. Reagents and conditions: (a) butyl bromide, 0 °C to room temperature, 1 week, 98%; (b) amberlite–IRA400, water, room temperature; (c) methanol, concentrated hydrochloric acid, room temperature, 5 days, 30%; (d) succinic anhydride, triethylamine, methanol, room temperature, 2 h, 67%; (e) compound 2, methanol, room temperature; (f) succinic anhydride, dioxane, 80 °C, 40 min, 80%; (g) compound 2, methanol, room temperature. POSS, polyhedral oligomeric silsesquioxane.