High-pressure glass formation of a series of 1-alkyl-3-methylimidazolium bis(trifluoromethanesulfonyl)imide homologues

Abstract

We investigated the stability of the liquid phase of a series of 1-alkyl-3-methylimidazolium bis(trifluoromethanesulfonyl)imide ([C_nmim][TFSI]) homologues with different alkyl chain lengths for 3 ≤ n ≤ 10 at room temperature. We found that all [C_nmim][TFSI] samples (n = 3–10) formed a glassy state when pressure was applied. Intriguingly, the glass transition pressure (p_g) slightly increases up to n = 5, reaches a plateau at n ≧ 8, and increases again at n = 10. This is completely different from the high-pressure glass formation of [C_nmim][BF₄], where the p_g decreases as n increases. We discussed the local structural changes occurring in [C_nmim][TFSI] in view of the conformational changes of the cation and anion, and small-angle X-ray scattering data. It seems that [C_nmim][TFSI] is resistant to external pressure and retains its local liquid structure by conformational adjustments of the cation and anion.

---

Introduction

Recently, there has been growing interest in the phase transition behavior of room-temperature ionic liquids (ILs) at high pressure,^1–8 especially with respect to glass formation, in which the main governing interactions are the interplay between Coulomb forces and van der Waals interactions among constituent ions.^9,10

In a previous study,¹¹ we reported the phase transition behavior of typical ILs, a series of 1-alkyl-3-methylimidazolium tetrafluoroborate (hereafter abbreviated as [C_nmim][BF₄]) homologues with different alkyl chain lengths (2 ≤ n ≤ 8) at 298 K. We showed that [C_nmim][BF₄] exhibited superpressurization to form a glassy state at around 2–3 GPa upon compression. As n increases, the glass transition pressure (p_g) decreases as might be expected. The results indicate that the p–T range of the liquid phase is regulated by the alkyl chain length of [C_nmim][BF₄] homologues. Subsequently, we proposed that the flexibility of the alkyl chain of the cations (conformational adjustment) plays an important role in avoiding crystallization, leading to high-pressure glass formation.¹²

The mechanism by which pressure affects the phase behavior of ILs is thus worthy of investigation. Alternatively, studies of ILs at high pressure would contribute to a basic understanding of these compounds that would allow their use under stressed conditions, e.g., tribological applications.^13–15 Moreover, this information might be useful with respect to recycling and purification using high-pressure crystallization in future applications.^16,17

Bis(trifluoromethanesulfonyl)imide, usually abbreviated as TFSI, is also one of the most popular anions in ILs because of its attractive properties, which result in low melting point salts, relatively low viscosity, and high electrical conductivity^18,19 when combined with common organic cations. Apart from the BF₄ anion, it is noted that the TFSI anion has C₁ (cisoid) and C₂ (transoid) conformers.^20,21 A schematic representation of the C_nmim cation and TFSI anion is shown in Fig. 1. Therefore, there might be competition among the different types of interactions in [C_nmim][TFSI], which probably determines its phase transitions and affects its behavior. Indeed, the molecular configurations of anions and cations are key to understanding the dependence of the macroscopic properties of the ILs on their ionic composition and therefore to tailoring them to a particular application. The structure–function link in ILs is, however, sometimes counterintuitive, and predicting the phase behavior is often difficult.

Fig. 1 A schematic representation of the structure of 1-alkyl-3-methylimidazolium bis(trifluoromethanesulfonyl)imide ([C_nmim][TFSI]).

Motivated by this background, and especially as a counterpart to the [C_nmim][BF₄] homologues, in this study we investigated the phase behavior of [C_nmim][TFSI] as a function of alkyl chain length (n = 3–10) under high pressure at room temperature (298 K). We then compared the results for two typical ILs, i.e., [C_nmim][TFSI] and [C_nmim][BF₄], where the TFSI anion has a conformational equilibrium.

Experimental

The [C_nmim][TFSI] used in this study had n values ranging from 3 to 10 (Kanto Chemical Co. for n = 4; Ionic Liquids Technologies Inc. for n = 3, 5–10).

Raman spectra were typically measured using a JASCO NRS-5100 Raman spectrophotometer equipped with a single monochromator and a Peltier cooled device. The spectra were excited at 532 nm with a green laser having a power of 5.8 mW. The Gaussian–Lorentzian mixed function was used to analyze the observed Raman spectra using GRAMS/386 software (Galactic Industries Corp.). The pressure-loading device was a screw-type diamond anvil cell (DAC, SR-DAC-KYO3-3d, Kyowa) with 0.6 mm culet anvils. We used a T301 stainless steel gasket with a 0.3 mm hole. The sample was normally compressed to ∼3 GPa in steps of approximately 1 GPa h⁻¹, including the duration time.

For [C_nmim][TFSI] (n = 8–10), synchrotron radiation (SR) small-angle X-ray scattering (SAXS) measurements were performed. A sample was encapsulated into a hole of 0.3 mm diameter drilled in the T301 gasket and loaded on a modified Mao-Bell-type DAC with 0.6 mm culet anvils. The samples were compressed at a rate of ∼0.5 GPa h⁻¹. The measurements were carried out at beamline BL18C in Photon Factory, KEK, Tsukuba. Incident X-rays at wavelengths of 0.08014 nm for n = 8 and 10 and 0.08293 nm for n = 9 were focused and collimated by the shaping and additional collimators with a hole of 35 and 130 μm in diameter. An imaging plate, used as a detector, was placed at a distance of 1285 mm from the sample in a vacuum chamber with a polyimide film window. The acquisition time for each scattering pattern was typically 1800 s. The intensity of the incident X-rays was continuously monitored with a photodiode and used to normalize measured scattering intensities. Background scattering and the diffraction around 4 nm⁻¹ from a polyimide film for an X-ray window were then subtracted from measured scattering patterns.

The pressure was determined from the spectral shift of the R₁ fluorescence line of several ruby balls (0.015 mm diameter) placed in the sample chamber of the DAC.²² The error of the estimated pressure was ±0.05 GPa. The sample was prepared in a glovebox filled with argon gas to avoid contamination by water in air. All the Raman and SR-SAXS measurements were performed at room temperature (298 K).

Results and discussion

Firstly, as a typical example of Raman spectral changes, Raman spectra of the C–H stretching vibrational modes (ν_CH) of the imidazolium cation in [C_nmim][TFSI] (n = 3, 5, 8) are shown as a function of pressure in Fig. 2. These modes are commonly used as an indicator of crystallization. We found that as the pressure increases, the spectral resolution of each peak at 0.1 MPa in the ν_CH band becomes less clear and the spectra show a different feature from that at 0.1 MPa accompanied by a higher frequency shift. Unfortunately, the spectra present difficulties with respect to quantitative interpretation due to the broad spectral features at high pressure. Visual observations at elevated pressures showed that the sample remained transparent within the studied range of pressures.

Fig. 2 Typical example of pressure-induced Raman CH stretching spectral changes in [C_nmim][TFSI] (n = 3, 5, and 8) as a function of pressure.

As with previous studies,^3,11 we used the line broadening method with the sharp ruby R₁ fluorescence line to distinguish high-pressure vitrification from crystallization. As an example of the results, we show the spectral change in the ruby R₁ line of [C₃mim][TFSI] in Fig. 3(A). Fig. 3(B) shows the variations in the full width at half-maximum (FWHM) of the ruby R₁ line relative to the 0.1 MPa line for [C₃mim][TFSI] during the compression process. The pressure at which the line broadening begins is defined as the glass transition pressure p_g.²³ We can see that the ΔFWHM drastically increases with pressure above the initiation point, p_g. We found that the [C_nmim][TFSI] homologues for n = 3–10 were superpressed into a metastable (glassy) state.

Fig. 3 (A) An example of variations in the spectra ([C₃mim][TFSI]) of the ruby R₁ line relative to the width of the 0.1 MPa line. (B) Pressure broadening of the sharp ruby R₁ fluorescence line (full width at half-maximum, FWHM) relative to the 0.1 MPa line width. The lines are guides for the eye.

The observed p_g values are plotted in Fig. 4, including their experimental uncertainties based on several runs. Data on p_gs for [C₄mim][TFSI] and [C₆mim][TFSI] were presented by Ribeiro et al.²⁴ (p_g = 1.6 GPa for n = 4 and p_g = 1.7 GPa for n = 6) and Wu et al.²⁵ (p_g = 1.8 GPa for n = 4). We suspect that the slight differences between the values given by these authors and those found in the present studies are due to differences in experimental conditions, such as the compression rate and the measurements of the apparatus used (such as gasket and diamond culet sizes). Notably, as the alkyl chain length increases, p_g shifts to slightly higher pressures up to n = ∼5 and then exhibits almost a plateau region up to n = 8. However, we found that the p_gs for n = 9 and 10 increase again.

Fig. 4 Changes in the glass transition pressure p_g (●) for [C_nmim][TFSI] and p_g (○) for [C_nmim][BF₄]¹¹ as a function of alkyl chain length (n).

It should now be interesting to compare the present results with those for [C_nmim][BF₄]. These comparisons are made in Fig. 4. Intriguingly, the results are completely different from those for [C_nmim][BF₄], where p_g shifts to lower pressures as the alkyl chain length increases. We wonder whether the different phase diagrams of the two typical ILs are the result of competition among the anion–cation and intermolecular interactions. As mentioned in the Introduction, the TFSI anion is a flexible molecule that can adopt two energetically different conformations, the C₁ and C₂ forms, whose concentrations in the liquid and/or solid phases affect the physical properties of the ILs.²⁶ At normal pressure, the C₂ conformer is slightly more stable than the C₁ conformer with an energy difference of only 3.5 kJ mol⁻¹,²⁰ and both conformers are present in the liquid state.^27,28 The alkyl chain plays an important role in determining the morphology of ILs, but it seems that the TFSI anion also contributes significantly to the structure of [C_nmim][TFSI]. Thus, there must be complex interactions in [C_nmim][TFSI] originating from the cationic and anionic conformations.

We then consider local structural changes at high pressure, especially conformational changes in the C_nmim cation and TFSI anion. As mentioned in the Introduction, the C_nmim cation exhibits conformational equilibria between the trans and gauche conformers of the main long alkyl chain. As shown in Fig. 5(A), the ∼625 and ∼600 cm⁻¹ peaks for the CH₂ rocking modes of the trans and gauche conformers are usually used for the analysis.^29,30 Unfortunately, at n = 9 and 10, strong fluorescence emerging in the observed spectra at high pressures prevents precise local conformational analysis. The intensity fractions (f) of gauche and trans conformers were determined by curve fitting spectra of the CH₂ rocking mode. The results for [C₃mim][TFSI] and [C₈mim][TFSI] are shown in Fig. 6(A). Fig. 6(A) shows that both cations favor the trans conformer in the liquid state under normal pressure, but the population of gauche conformers increases upon compression. It should be noted that the change in [C₃mim][TFSI] against pressure is small and the populations for both conformers tend to equalize at higher pressures compared to [C₈mim][TFSI]. These results imply greater dynamic heterogeneity for main chains with longer alkyl chain segments.

Fig. 5 (A) Effect of pressure on changes in Raman spectra of [C₃mim][TFSI] for the CH₂ rocking mode. The dotted line indicates the gauche conformer and the dashed line the trans conformer. (B) Pressure-induced Raman spectral changes in the region from 200 to 500 cm⁻¹ of the TFSI anion in [C₃mim][TFSI]. The dashed line indicates the C₁ conformer and the dotted line the C₂ conformer.

Fig. 6 (A) Intensity fractions (f) of the conformers of the cations in [C₃mim][TFSI] and [C₈mim][TFSI] as a function of pressure. The open and closed circles represent the trans and gauche conformers of C₃mim and C₈mim cations, respectively. (B) Intensity fractions (f) of the conformers of the anions in [C₃mim][TFSI] and [C₈mim][TFSI] as a function of pressure. The closed and open circles represent the C₁ and C₂ conformers of the TFSI anion, respectively. Here, the f values of the cation and anion are determined by the equation f_A = [I_A/(I_A + I_B)] (where I_A and I_B are the integrated Raman intensities of each conformer).

On the other hand, Fig. 5(B) shows pressure-induced Raman spectral changes in the region from 200 to 500 cm⁻¹ of the TFSI anion in [C₃mim][TFSI]. As with previous studies,^21,31 we used this band for conformational equilibrium analysis of the TFSI anion in an effort to gain further insight into the local structures. The fractions (C₂ and C₁) for [C₃mim][TFSI] and [C₈mim][TFSI] are plotted in Fig. 6(B). Under normal pressure, the population of C₂ conformers is more favorable, but with increasing pressure, the fraction of C₂ decreases, whereas that of C₁ increases. Interestingly, in contrast to the conformational changes in the cation, the dependence on pressure for [C₈mim][TFSI] is less and the fractions remain constant upon compression. That is, the conformational changes in the anion against pressure are relatively smaller, but those of the cation are larger when the alkyl chain is long, and vice versa. It appears that the local structures of the C_nmim cation and TFSI anion change cooperatively against pressure depending on the alkyl chain length n.

From combined vibrational and NMR spectroscopic measurements, Garage et al.³² assumed that TFSI locates above/below the imidazolium ring, with the –S=O groups pointing to the ring plane and the –CF₃ groups pointing away from the ring and forming fluorine rich inter-ring layers. Because of this, the π–π stacking that can be established in imidazolium ILs³³ is less likely for intermediate chain lengths (2 ≦ n ≦ 6). An increase in the C₁/C₂ was reported for TFSI with longer chains, where the TFSI is slightly displaced from a position above the ring to a location closer to the C² and C⁶ sites without significantly changing the cation and anion interactions.

Simulations³⁴ and experimental³⁵ studies have shown that imidazolium-based ILs with alkyl chain lengths greater than that of the C₆mim cation form nanostructural domains permeated by a charged (polar) network. Moreover, Moschovi et al.³⁶ studied the structure of [C_nmim][TFSI] using vibrational spectroscopy (FT-IR/ATR and FT-Raman). They proposed a sponge-like structure for the system. Finally, we show the pressure dependence of SAXS normalized intensities for [C₈mim][TFSI] at room temperature in Fig. 7(A). The lowest Q peak (Q ∼ 3 nm⁻¹) is known as the pre-peak. The molecular origins of the pre-peak are still under debate, but recent reports state that it is indicative of intermediate range order associated with the size of the apolar domain in ILs, which is the average separation of segments in a charge network spaced by apolar domains.^37,38 On the other hand, Fujii et al.³⁹ concluded that for the n = 8 and 10 systems the low Q peak intensity corresponds to correlations between anions with no contribution from alkyl-group aggregates. As seen in the figure, with the initial increase in pressure to ∼0.5 GPa, the pre-peak position slightly shifted to higher Q (Fig. 7(D)) and the peak intensity decreased. However, intriguingly, the peak intensity remains almost unchanged with a further increase in pressure, even above the p_g. It seems that [C₈mim][TFSI] resists external pressure and almost retains the local liquid structure, as if a sponge absorbs a stimulus. In view of the results shown in Fig. 6, the adjustments are achieved by a combination of cation and anion conformational changes. The results for the [C₉mim][TFSI] and [C₁₀mim][TFSI] systems follow a similar line to those for [C₈mim][TFSI] as shown in Fig. 7(B) and (C). Overall, the results indicate that even pressures as high as ∼3 GPa change but do not diminish the polar–apolar alternation and barely distort the liquid structure. On the contrary, in a previous study,¹¹ we reported that the pre-peak almost disappeared above ∼2 GPa, at which [C₈mim][BF₄] falls into a superpressed state. Thus, the results are completely different for the two typical ILs.

Fig. 7 Pressure-induced changes in the SAXS patterns for (A) [C₈mim][TFSI], (B) [C₉mim][TFSI], and (C) [C₁₀mim][TFSI] and (D) changes in the respective pre-peak positions of [C_nmim][TFSI] (n = 8–10) at room temperature as a function of pressure. The closed and open circles and closed triangles represent [C₈mim][TFSI], [C₉mim][TFSI], and [C₁₀mim][TFSI], respectively. (B) A small shoulder around 4 nm⁻¹ in the SAXS pattern is due to the residual diffraction peak from a polyimide film.

We should note that in a recent report Dhungana et al.⁴⁰ predicted the structural response to pressure of [C₁₀mim][TFSI] by computer simulation. Overall, their results are consistent with the present study, but the intensity of the pre-peak for [C₁₀mim][TFSI] diminished at higher pressures up to ∼1.2 GPa in their prediction. They concluded that polar–apolar alternation is more affected by pressure than charge alternation and that curlier tail conformations of the cation dominate. At higher pressure, the population of gauche states is higher, while the tails deform to lie flat at the interface with the charge network.⁴⁰ This leads to smaller apolar pockets without breaking the charge networks. They ascribed this to the larger Q shift for changes in polarity alternation. They also predicted changes in the anionic conformation with pressure, i.e., between two conformers with C₁ and C₂ configurations. As the pressure is increased, the C₂ configuration loses population to other dihedral conformations. All these conformational changes are consistent with the present results.

In addition, we should mention about the hydrogen bonds’ role in the behavior of the pressure-induced glass transition of the [C_nmim][TFSI] series. Hydrogen bonds play an important role in the dynamics of this series of ILs, other than the Coulomb and/or dispersion forces.⁴¹ Recently, Roth et al.⁴¹ investigated three [C_nmim][TFSI] (n = 1, 2, 8) with infrared, linear Raman and multiplex coherent anti-Stokes Raman scattering (CARS) spectroscopy at normal pressure. They showed that the C⁴–H and C⁵–H groups form weak hydrogen bonds with the TFSI anion, whereas the hydrogen bonds of the C²–H group are relatively stronger. Whether hydrogen bonds in ILs strengthen or weaken with increasing pressure would be intriguing issue. We suspect that the differences observed in conformational changes between cations with short and long alkyl chains together with the respective behavior of anions may reflect changes in the hydrogen bonds by applied pressure. In particular, for the p_g behavior with the longest side alkyl chains (i.e., C₉ and C₁₀) in the TFSI series studied, change in the hydrogen bond strength may be responsible for the jump in p_gs. Another possibility is that some plastic or liquid crystalline-type ordering in the sample with shorter side chains than that is observed at normal pressure⁴² may be induced by application of pressure, which is beyond the present study and needs further experimental studies. As shown in Fig. 2, the CH stretching band of the imidazolium ring (C², C⁴, and C⁵) at ∼3175 cm⁻¹ shifts to a higher wavenumber with increasing pressure. However, as mentioned earlier, the broad spectral features at high pressure could not allow us to perform further full-detailed quantitative analysis. A plausible explanation for the results shown in Fig. 4 is that these features are probably key to the relatively higher p_g values of longer alkyl chains in [C_nmim][TFSI], though the overall picture that emerges from the present results is of complex layers of structure and dynamics arising from the spatial heterogeneity.

Conclusions

In summary, based on the present study, we have provided stability limits for [C_nmim][TFSI] homologues with different alkyl chain lengths (3 ≤ n ≤ 10) under stress at room temperature. We found that [C_nmim][TFSI] (n = 3–10) formed a glassy state with applied pressure. The p_g increased with the length of the alkyl chain. An interesting difference was found between the p_g behavior against pressure of two ILs sharing the same C_nmim cation but having different anions, i.e., [C_nmim][TFSI] and [C_nmim][BF₄]. Clearly, the more complex molecular structure of the TFSI anion, which existed as C₁ and C₂ conformations, was responsible for the differing dependence of the p_g behavior on the alkyl chain lengths. A more interesting point is that the liquid structure of [C_nmim][TFSI] tends to resist shear stress through cooperative conformational changes in the cation and anion. Each contribution shifts depending on the alkyl chain length n, where the anion contribution is relatively larger than that of the cation for shorter alkyl chains. In other words, [C_nmim][TFSI] can adapt to external stimuli such as applied pressure. We believe that the information deduced from the present study would be valuable with respect to optimizing the properties of this IL for particular applications. The use of pressure is a useful tool for fine-tuning the strength of intermolecular interactions without the unavoidable perturbations and side effects experienced when exploiting changes in temperature and chemical composition.³ Thus, we need to further explore an extended p–T region of ILs, which might allow us to draw general conclusions on the phase behavior of the ILs.⁴³ Of course, more detailed simulations based on information from scattering experiments at high pressure will be especially helpful for analyzing actual local structures.

Author contributions

Y. Y., H. A., K. M. and N. H. designed the experiments. T. T., Y. K., M. T., M. Y., N. K., D. W., N. F., H. A., and N. H. performed experiments. Y. Y., T. T. and N. H. wrote the manuscript.

Conflicts of interest

There are no conflicts of interest to declare.

Acknowledgements

The authors would like to thank Crimson Interactive Japan for the English language review. Part of this work was performed under the approval of the Photon Factory Program Advisory Committee (Proposal Number: 2015G124).

References

1. L. Su, M. Li, X. Zhu, Z. Wang, Z. Chen, F. Li, Q. Zhou and S. Hong, J. Phys. Chem. B, 2010, 114, 5061–5065.
2. O. Russina, B. Fazio, C. Schmidt and A. Triolo, Phys. Chem. Chem. Phys., 2011, 13, 12067–12074.
3. Y. Yoshimura, H. Abe, Y. Imai, T. Takekiyo and N. Hamaya, J. Phys. Chem. B, 2013, 117, 3264–3269.
4. M. Shigemi, T. Takekiyo, H. Abe, N. Hamaya and Y. Yoshimura, J. Solution Chem., 2014, 43, 1614–1624.
5. F. Capitani, C. Fasolato, S. Mangialardo, S. Signorelli, L. Gontrani and P. Postorino, J. Phys. Chem. Solids, 2015, 84, 13–16.
6. H.-C. Chang, J.-C. Jiang, M.-H. Kuo, D.-T. Hsu and S. H. Lin, Phys. Chem. Chem. Phys., 2015, 17, 21143–21148.
7. K. Hirosawa, K. Fujii, T. Ueki, Y. Kitazawa, M. Watanabe and M. Shibayama, Macromolecules, 2016, 49, 8249–8253.
8. V. H. Paschoal, L. F. O. Faria and M. C. C. Ribeiro, Chem. Rev., 2017, 117, 7053–7112.
9. T. Welton, Chem. Rev., 1999, 99, 2071–2084.
10. M. J. Earle and R. Seddon, Pure Appl. Chem., 2000, 72, 1391–1398.
11. Y. Yoshimura, M. Shigemi, M. Takaku, M. Yamamura, T. Takekiyo, H. Abe, N. Hamaya, D. Wakabayashi, K. Nishida, N. Funamori, T. Sato and T. Kikegawa, J. Phys. Chem. B, 2015, 119, 8146–8153.
12. T. Takekiyo, Y. Koyama, M. Shigemi, K. Matsuishi, H. Abe, N. Hamaya and Y. Yoshimura, Phys. Chem. Chem. Phys., 2017, 19, 863–870.
13. C. Ye, W. Liu, Y. Chen and L. Yu, Chem. Commun., 2001, 2244–2245.
14. P. K. Cooper, H. Li, M. W. Rutland, G. B. Webbera and R. Atkin, Phys. Chem. Chem. Phys., 2016, 18, 23657–23662.
15. L. F. O. de Faria, M. M. Nobrega, M. L. A. Temperini and M. C. C. Ribeiro, J. Raman Spectrosc., 2013, 44, 481–484.
16. L. Su, L. Li, Y. Hu, C. Yuan, C. Shao and S. Hong, J. Chem. Phys., 2009, 130, 184503.
17. H. Li, L. Su, X. Zhu, X. Cheng, K. Yang and G. Yang, J. Phys. Chem. B, 2014, 118, 8684–8690.
18. P. Bonhote, A. Dias, N. Papageorgiou, K. Kalyanasundaram and M. Gratzel, Inorg. Chem., 1996, 35, 1168–1178.
19. V. R. Koch, C. Nanjundah, G. B. Appetecchi and B. Scrosati, J. Electrochem. Soc., 1995, 142, L116–L118.
20. K. Fujii, T. Fujimori, T. Takamuku, R. Kanzaki, Y. Umebayasi and S. Ishiguro, J. Phys. Chem. B, 2006, 110, 8179–8183.
21. J. C. Lassègues, J. Grondin, R. Holomb and P. Johansson, J. Raman Spectrosc., 2007, 38, 551–558.
22. H. K. Mao and P. M. Bell, J. Appl. Phys., 1978, 49, 3276–3283.
23. G. J. Piermarini, S. Block and J. D. Barnett, J. Appl. Phys., 1973, 44, 5377–5382.
24. M. C. C. Ribeiro, A. A. H. Pádua and M. F. Costa Gomes, J. Chem. Phys., 2014, 140, 244514.
25. J. Wu, X. Zhu, L. Su, K. Yang, X. Cheng, G. Yang and J. Liu, J. Solution Chem., 2015, 44, 2106–2116.
26. F. Capitani, F. Trequattrini, O. Palumbo, A. Paolone and P. Postorino, J. Phys. Chem. B, 2016, 120, 2921–2928.
27. J. C. Lassègues, J. Grondin, C. Aupetit and P. Johansson, J. Phys. Chem. A, 2009, 113, 305–314.
28. A. Martinelli, A. Matic, P. Johansson, P. Jacobsson, L. Börjesson, A. Fernicola, S. Panero, B. Scrosati and H. Ohno, J. Raman Spectrosc., 2011, 42, 522–528.
29. H. Hamaguchi and R. Ozawa, Adv. Chem. Phys., 2005, 131, 85–104.
30. R. Ozawa, S. Hayashi, S. Saha, A. Kobayashi and H. Hamaguchi, Chem. Lett., 2003, 32, 948–949.
31. Y. Yoshimura, T. Takekiyo, Y. Imai and H. Abe, J. Phys. Chem. C, 2012, 116, 2097–2101.
32. M. N. Garage, M. Nayeri and A. Martinelli, J. Mol. Liq., 2015, 210, 169–177.
33. R. P. Matthews, T. Welton and P. A. Hunt, Phys. Chem. Chem. Phys., 2014, 16, 3238–3253.
34. J. N. A. C. Lopes and A. A. H. Padua, J. Phys. Chem. B, 2006, 110, 3330–3335.
35. A. Triolo, O. Russina, H. J. Bleif and E. Di Cola, J. Phys. Chem. B, 2007, 111, 4641–4644.
36. A. M. Moschovi and V. Dracopulos, J. Mol. Liq., 2015, 210, 189–199.
37. H. K. Kashyap, J. J. Hettige, H. V. R. Annapureddy and C. J. Margulis, Chem. Commun., 2012, 48, 5103–5105.
38. H. V. R. Annapureddy, H. K. Kashyap, P. M. De Biase and C. J. Margulis, J. Phys. Chem. B, 2010, 114, 16838–16846.
39. K. Fujii, S. Kohara and Y. Umebayashi, Phys. Chem. Chem. Phys., 2015, 17, 17838–17843.
40. K. B. Dhungana and C. J. Margulis, J. Phys. Chem. B, 2017, 121, 6890–6897.
41. C. Roth, S. Chatzipapadopoulos, D. Kerlé, F. Friedriszik, M. Lütgens, S. Lochbrunner, O. Kühn and R. Ludwig, New J. Phys., 2012, 14, 105026.
42. Y. Nozaki, K. Yamaguchi, K. Tomida, N. Taniguchi, H. Hara, Y. Takisawa, K. Sadakane, K. Nakamura, T. Konishi and K. Fukao, J. Phys. Chem. B, 2016, 120, 5291–5300.
43. S. Mangialardo, L. Baldassarre, E. Bodo and P. Postorino, in The Structure of Ionic Liquids, ed., R. Caminiti and L. Gontrani, Springer International Publishing, Switzerland, 2014.

---