Speciation and dipole reorientation dynamics of glass-forming liquid electrolytes: Li[N(SO₂CF₃)₂] mixtures of 1,3-propane sultone or tetrahydrothiophene-1,1-dioxide†

Yasuhiro Umebayashi*^a, Erika Otani^a, Hikari Watanabe^b and Jihae Han^a
aGraduate School of Science and Technology, Niigata University, 8050 Ikarashi, 2-no-cho, Nishi-ku, Niigata, 950-2181, Japan. E-mail: yumescc@chem.sc.niigata-u.ac.jp
^bDepartment of Pure and Applied Chemistry, Faculty of Science and Technology, Tokyo University of Science, 2641 Yamazaki, Noda, Chiba 278-8510, Japan

First published on 10th June 2024

---

---

Introduction

To achieve a sustainable society, it is crucial to realize next-generation electrochemical devices and/or processes, such as secondary batteries of higher energy density and electrodeposition of higher efficiency. Room-temperature ionic liquids have attracted much attention as a new class of electrolytes during the last few decades. Further research on new electrolytes is currently focused on super-concentrated solutions,^1,2 solvate ionic liquids^3–6 and deep eutectic solvents⁷ via room-temperature ionic liquids. These new electrolytes are frequently composed of similar components: a solvent and a large/excess amount of salt mixtures. As reported by Angell, the first examples of this class of electrolyte were hydrate-melt electrolytes like CaNO₃·4H₂O.^8–10 It is easy to speculate that the salt lattice energy could be reduced by incorporating small amounts of solvent molecules to lower the melting point. However, physicochemical or thermodynamic definition is still unclear for these electrolytes. Recently, we attempted to clarify the definition and proposed that the relative solvent activity was less than 1% that of the neat liquid.¹¹

From both the scientific and industrial viewpoints, some of the most important properties of liquid electrolytes are their ionic transport properties, such as conductivity and viscosity. As proposed by Angell, the Walden rule or product is useful because this empirical rule generally holds well for a wide range of liquid electrolytes.¹² In this sense, rotational motion of ions and solvents in solution is significant, as has been well established by the Stokes–Einstein–Debye (SED) equation.¹³ Fig. 1 shows a viscosity vs. relaxation time plot for 43 polar organic solvents that demonstrated dipole reorientation by means of dielectric relaxation spectroscopy (DRS). According to the SED relationship, viscosity should be proportional to the relaxation time. The slope is unity for the solid line shown in the figure, which clearly shows that the SED relationship generally holds for a large number of polar solvents. On the other hand, we found super-ionic proton conduction in pseudo-protic ionic liquids and demonstrated that the acetate/acetic acid rotational motion plays a crucial role in this super-ionic proton conduction with the DRS technique.¹⁴ Similarly, we recently found that quite-slow relaxations exist in glyme-type solvate ionic liquids, and pointed out that this quite-slow relaxation could be closely related with specific lithium ion conduction due to a combination of the solvent/anion exchange of lithium ions and their rotational motion.¹⁵

In this contribution, we mainly discuss two subjects: (1) glass-forming liquid electrolytes as a new class of liquids, and (2) their speciation and dipole reorientation dynamics. 1,3-Propane sultone (propane sultone: PS) is a cyclic sulfonate ester comprising a 5-membered ring and it is an analogue of tetrahydrothiophene-1,1-dioxide (sulfolane: SL). As expected from their molecular structures and electrostatic potentials shown in Fig. 2 and Table 1, the solvating ability of the lithium ion of PS should be weaker than that of the SL. Alvarado et al.,¹⁶ Ueno, Dokko and Shinoda et al.^17–19 found that Li⁺ ions diffuse faster than the anions and the SL in super-concentrated lithium salt SL solutions. This unique behavior is similar for the lithium salt PS solutions, as shown in Fig. S1 in the ESI.†

---

Results and discussion

Glass-forming liquid electrolytes

Before a detailed discussion, it is worth mentioning the concept of recent new ionic fluids such as super-concentrated solutions, solvate ionic liquids and deep eutectic solvents. As aforementioned, the concepts of these ionic fluids are currently similar; i.e., they consist of a solvent and a large or excess amount of salt. Although there are several expressions regarding ionic liquids,^20–23 the definition is currently clear. Angell suggested that ionic liquids be defined as liquids composed solely of ions, in principle, with melting points below 100 °C, and also classified ionic liquids into four categories.²⁴ As aforementioned, we recently proposed that salt and solvent mixtures that have less than 1% component solvent activity relative to that of the neat solvent could be practically identical to the ionic liquids.¹¹ In addition, the definition of super-concentrated electrolyte solutions such as water-in-salt electrolytes is not clear. It is unclear whether the super-concentrated electrolyte solutions are the same as, or different from, concentrated solutions as an extension of conventional dilute solutions.

On the other hand, eutectic mixtures should be absolutely different solutions compared to ionic liquids, solvate ionic liquids and super-concentrated electrolyte solutions. The term ‘eutectic’ was used for the first time by Guthrie in 1884.²⁵ However, it should be clear that a eutectic mixture is a homogeneous mixture that has a melting point lower than those of the constituents, and the lowest melting point over all the mixing ratios and the composition are called the eutectic temperature and the eutectic composition, respectively. As far as we are aware, the first deep eutectic solvents were mixtures of choline chloride and urea.⁷ Abbott et al. noted a behavior of a large depression in melting point in eutectic compositions of choline chloride and urea and named these compositions deep eutectic solvents. In fact, according to the phase diagram, the eutectic composition should be choline chloride:urea = 1:2. However, the physicochemical or thermodynamic definitions for the deep eutectic solvents are still unknown. Among them, some do not seem to indicate eutectic points at the eutectic composition. AlCl₃-amide and AlCl₃-carbamide systems also yield molten salt mixtures at ambient temperature. It was considered that AlCl₃ dissolves in the solvents and a disproportionation reaction of 2AlCl₃ ↔ AlCl₂⁺ + AlCl₄⁻ occurs, so these mixtures were first proposed as ionic liquids composed of a dichloro-aluminum complex cation [AlCl₂(solvent)_n]⁺ instead of an organic cation and a tetrachloroaluminate anion [AlCl₄]⁻.²⁶ After that, AlCl₃-amide and AlCl₃-carbamide systems were thought of as eutectic mixers or deep eutectic solvents,^26–29 though no phase diagrams have been reported for them as far as we know. On a phase diagram, the eutectic temperature is seen as the eutectic point. For these kinds of mixtures, the eutectic composition should be specific and fixed based on Gibbs’ phase rule. In this regard, it can be said that a choline chloride and urea mixture is a typical eutectic mixture. Here, let us go back to Wilkes’ chloroaluminate ionic liquids.^30,31 In the phase diagrams of AlCl₃–1-ethyl-3-methylimidazolium chloride (C₂mimCl) and AlCl₃–1-ethylpyridinium chloride mixtures, specific and fixed eutectic points can be recognized at a glance.³² Furthermore, these melting points at the eutectic compositions are extremely low relative to other non-eutectic compositions and yield narrow deep valleys.

Based on the above discussion, we again recognize the importance of phase diagrams for solvent mixtures with a large and/or excess amount of salt. Fig. 3a and b show roughly drawn phase diagrams for LiTFSA (TFSA: N(SO₂CF₃)₂⁻)–PS and LiTFSA–SL systems. Differential scanning calorimetry (DSC) and thermogravimetry (TG) thermograms are also shown in Fig. S2 and S3,† respectively. As clearly shown in the DSC thermograms for neat PS and SL, these solvents show a solid/solid phase transition, probably due to the plastic crystal phase. For both systems, upon adding LiTFSA salt, the peaks due to melting/freezing disappeared, and subsequently the peaks due to a solid/solid phase transition also disappeared. Thermal anomalies were only found for the glass transition in the range of 0.1 < x_Li < 0.8 for PS and 0.1 < x_Li < 0.4 for SL, where x_Li stands for the mole fraction of LiTFSA with T_g (glass transition temperature) increasing up to around −13 °C for PS and −38 °C for SL with increasing amount of LiTFSA. It should be noted that hot crystallization was observed at x_Li = 0.1 for both systems, though both molecules were of small molecular weight. These facts suggest that both mixtures of a specific composition range behave like a polymer or a gel, hence it could be said that these mixtures of the respective specific composition range should be a new class of liquids called glass-forming liquid electrolytes, not super-concentrated solutions, solvate ionic liquids or deep eutectic solvents. One of the most important features of the glass-forming liquid electrolyte is the melting/freezing temperature lowering down to around room temperature. One of the most important characteristics of the glass-forming liquid electrolytes is that their melting/freezing temperatures drop to near room temperature. The temperature lowering, ΔT = T_ideal − T_m/f/g, is shown in Fig. 3c, where T_ideal and T_m/f/g represent the ideal melting/freezing point and the observed melting/freezing/glass transition point for the mixture. T_ideal should be given as T_ideal = T_{m/f, 1} + x₂ (T_{m/f, 2} − T_{m/f, 1}) where T_{m/f, 1}, T_{m/f, 2} and x₂ stand for the melting/freezing point for component 1 or 2, and the molar fraction of component 2, respectively. ΔT values for C₂mimCl–AlCl₃ mixtures are also shown in Fig. 3c for comparison. ΔT of both mixtures is about 150 °C, comparable to that of the C₂mimCl–AlCl₃ mixture.³² Here, it should be emphasized that the main motivation in this work is to clarify currently vague electrolyte solutions containing a solvent and a large or excess amount of salt. As is well known, numerous ionic liquids and super-concentrated electrolyte solutions exhibit a glass transition. We consider that the glass-forming liquid electrolytes should partly contain the characteristics of ionic liquids and super-concentrated electrolyte solutions.

Speciation and dipole reorientation dynamics

High ionic conductivity or high transference number of Li⁺ in solutions plays a key role for realization of high-energy-density rechargeable batteries. Such ion transport properties in solutions are governed by the intermolecular potentials. Structural knowledge of solutions, which provides insights into the intermolecular potentials, can be obtained as radial distribution functions or pair correlation functions using X-ray/neutron scattering techniques or molecular dynamics and Monte Carlo simulations. On the other hand, vibrational spectroscopies such as Raman/IR spectroscopy are frequently used for electrolyte solutions consisting of the anion/solvent of polyatomic molecules. Formation distribution functions determined by analyzing Raman/IR spectra can give a mean solvation number or a mean anion binding number, which are just the stoichiometric number of the ionic species existing in the solution. For this purpose, the MCR-ALS (multivariate curve resolution-alternating least squares) technique^33,34 seems to be useful. We recently developed the CLSA (complementary least-squares analysis) technique¹¹ that similarly gives the formation distribution functions of the anion/solvent in solution.

CLSA was applied to LiTFSA–PS and LiTFSA–SL systems to evaluate mean solvation numbers or mean anion binding numbers to Li⁺ for LiFSA–PS, as shown in Fig. 4. As clearly exhibited in this figure, on average two solvent molecules and one anion are directly coordinated to the Li⁺-aggregated species of [Li(solvent)₂(TFSA)] as a constituent unit, which mainly exists in both mixtures at a high LiTFSA concentration of 3 mol dm⁻³. As expected from the weaker solvation ability of PS relative to SL, the contact ion-pair (CIP) and/or aggregate (AGG) formation occurs at a lower LiTFSA concentration in PS than in SL. Raman scattering factors and formation distribution functions are also shown in Fig. S5 and S6.† For both mixture systems, the respective solvent and anion exist as 3 species: free, bound1 and bound2 for the solvent; and free, CIP/AGG and AGG for the anion. Free can be ascribed to the bulk solvent and those in the Li⁺ solvation shell higher than the second one of the solvent, and the free anion and those in the Li⁺ coordination shell higher than the first one for the anion. Solvent bound1 can be assigned to directly coordinated solvent molecules in the solvated Li⁺ species and in the CIP and/or the AGG and anion CIP/AGG also to the latter. Solvent bound2 and anion AGG can be attributed to the bridging ones in the Li⁺–TFSA–solvent aggregates. DFT calculations were performed for some typical models for the CIP and AGG at the B3LYP/cc-pVDZ level of theory to obtain the optimized geometries and theoretical Raman spectra as shown in Fig. S7–S10.† The Raman scattering factor for the CIP/AGG is larger than those for the free species, which suggests that the CIP/AGG contains a few anions. Similarly, a Raman scattering factor for the AGG is the most intense, which indicates that the AGG should include several anions. It is considered that the formation of AGGs in highly concentrated lithium salt solutions plays a key role with respect to the larger self-diffusion coefficient for Li⁺ in the super-concentrated lithium salt SL solutions.²⁰ Here, the AGG formation is directly evidenced in LiTFSA–PS and LiTFSA–SL mixture systems.

Knowledge on the rotational motion in solutions can be obtained using some experimental techniques. By means of DRS, the dipole reorientation relaxation in a liquid can be observed as the complex permittivity ε^* = ε′ + iε′′, where ε′, ε′′ and i represent the relative dielectric constant, dielectric loss and an imaginary unit, respectively. DRS has some advantages against NMR: (1) DRS can determine both the relaxation time and strength for the respective dipole of polar species in solution, while NMR gives only a time-averaged relaxation time due to the chemical exchange; and (2) the relaxation strength is related to the effective dipole moment of the polar species in solution, so that a space scale for the dipole can be obtained in some cases. On the other hand, DRS also has a weak point; it is difficult to the deconvolute and assign the observed relaxation when some relaxation times are close among them. In addition, it may be hard to observe the dipole reorientation relaxation of a short-lifetime polar species, such as the solvent-separated and solvent-shared ion-pairs. DRS is now used as a new technique for speciation analysis and/or dipole reorientation dynamics of Li and Mg salt solutions.^35–41 Among them, the attribution of SIP and CIP is not always correct, considering the lifetime of the labile/inert ion-pairs.

The observed complex permittivities and typical deconvolution results using superposition of the Debye relaxation model for both systems are shown in Fig. S11.† The fastest relaxation composed of two Debye functions can be ascribed to the bulk solvent molecules. However, the assignment is difficult for two of the newly appeared relaxation peaks upon addition of LiTFSA. Here, we attempted an entirely different approach from the conventional DRS analysis. The relaxation strength for the ith polar species, S_i, can be expressed using the Cavell equation:^42,43

where ε_s, ε₀, k_B, N_A, T and μ_eff are the relative permittivity, permittivity in vacuo, the Boltzmann constant, the Avogadro number, the absolute temperature and the effective dipole moment, respectively. The effective dipole moment represents the result in solution, not the gas phase, so that it would be more useful to discuss polar species in solution. Discussions are given later in detail on the evaluation of the effective dipole moment based on the Cavell equation. On the other hand, the Cavell equation shows that the relaxation strength, and hence the intensity of the DRS as well as the Raman spectra, is directly proportional to the concentration of the polar species. This suggests that the DRS intensity, particularly that in the imaginary part, should directly correlate with the Raman intensity. Consequently, as the first step, the MCR-ALS analysis was applied to the DRS data as exhibited in Fig. S12.† For both systems, the relaxations of the three species were necessary to describe the observed dielectric loss spectra based on the eigenvalues. The observed dielectric spectra were reproduced with the Hamilton R factors of 2.2 and 2.3% for the PS and SL system, respectively. This highly suggests that the observed relaxation is ascribed to the chemical species in solution. However, the evaluated resolved relaxations and resolved concentration seem to lose physical meaning; the resolved relaxations for the species2 and species3 cannot be described by using a sole model function such as Debye, Cole–Cole and Cole–Davidson.

As for the second step, we applied two types of Raman/DRS two-dimensional (2D) correlation analysis to both mixtures. Here, 2D refers to the apparent Raman spectra, which is the Raman spectra divided by the scattering species concentrations, and the dielectric loss (the imaginary part of the complex permittivity). The first one is a simple Pearson correlation coefficient⁴⁴ and the second one is Noda–Ozaki 2D hetero-correlation spectroscopy,^45,46 as shown in Fig. 5 and 6 for PS and SL systems, respectively. The Pearson correlation coefficient is defined as the ratio between the covariance of two variables and the product of their standard deviations. Details of the Noda–Ozaki 2D hetero-correlation spectroscopy are described in the ESI.† One of the characteristics of the Noda–Ozaki 2D hetero-correlation spectroscopy is that it can express the correlation between two different spectra in terms of two kinds of correlations called synchronous and asynchronous. Synchronous correlation is the relationship between two spectral intensities increasing or decreasing in the same or opposite direction, similar to the Pearson correlation coefficient. Asynchronous correlations are significant when the apparent Raman intensity at a given wavenumber and the dielectric loss intensity at a given frequency increase or decrease with increasing lithium salt concentration, respectively, and when there is a divergence between these changes.

As clearly indicated by the Pearson correlation coefficients and the Noda–Ozaki synchronous correlation shown in Fig. 5 and 6, the relaxation of the highest frequency can be attributed to the free solvent for both the PS and SL systems. With regard to the free solvent, the effective dipole moment can be successfully evaluated to be 8.5 and 6.0 for PS and SL, respectively, as shown in Fig. 7, which are consistent with theoretical values in the isolated gas phase. Additionally, the relaxation of the lowest frequency can also be ascribed to the AGG. Though the assignment is clear in the 2D correlation, it is difficult to estimate accurately and precisely the effective dipole moment for the AGG due to the lack of observed frequency and concentration ranges for the PS and SL systems, respectively. However, it is worth noting that the obtained values are rather small; 7.7 and 7 for the PS and SL systems, respectively, as shown in Fig. 7. Such a small dipole moment for the AGG is discussed later. On the other hand, the assignment is unclear for the relaxation of the intermediate frequency. Therefore, we attempted to assess the 2D correlation between the respective calculated spectra to obtain good correlation between the anion bound1 in the Raman results and the intermediate relaxation in the DRS results as shown in Fig. S13.† However, it should be noted that the synchronous and the asynchronous correlations in the Noda–Ozaki 2D hetero-correlation spectroscopy are very small values and remarkably significant relative to those shown in Fig. 5, so that it is difficult to attribute the relaxation of the intermediate frequency to Raman-visible species. However, it is worth evaluating the effective dipole moment based on the anion bound1 concentration, although the estimated effective dipole moment is less accurate and has a larger uncertainty compared to the free solvent, and has values 13 and 11 for PS and SL, respectively, as shown in Fig. 7. This implies that the intermediate relaxation observed in the DRS results should consist of several polar species; the solvated solvent molecules, the CIP and probably small-size AGGs. Moreover, we should emphasize that while the solvent-separated/shared ion-pairs only yield a silent response in the Raman spectra, they could yield a dipole response in the DRS spectrum. In contrast, the ion-pairs with symmetric charges, like a double dimer, should be silent in DRS even though it involves both solvent and anion Raman bands as the bound species.

Finally, we discuss a time scale of the dipole reorientation observed in the DRS results. Fig. S14† exhibits the relationship between the solution viscosity η and the relaxation time determined using DRS. All of the indexes evaluated from logη vs. logτ were significantly greater than unity, which indicates that the Stokes–Einstein–Debye relationship breaks down for the electrolyte solutions, as easily expected by taking into account that the Stoke–Einstein relationship breaks down for the electrolyte solutions. However, it worth mentioning that the relaxation times for the AGGs in both mixtures are rather slow. As aforementioned, with regard to the glyme-based solvent ionic liquids, an extremely slow relaxation was found and is associated with specific Li⁺ conduction in which Li⁺ is transported by solvent/anion exchange.²⁰ The relaxation time for the AGGs in the PS and SL systems is also significantly slow, which suggests that the space scale of the AGGs should be large. In addition, the estimated effective dipole moments for the AGGs in the PS and SL systems are rather small, though with less accuracy and large uncertainties, which can be attributed to the formation of large-space-scale AGGs with symmetric charges. However, taking into consideration the fact that Li⁺ can be transported very efficiently by AGG formation in the PS and SL systems, much more effort should be made to clarify the extremely slow relaxation found in the new classes of liquid electrolyte.

Conclusions

LiTFSA–PS and LiTFSA–SL binary-mixture systems were studied by means of DSC, Raman spectroscopy and DRS techniques, and DFT calculations. Only the glass transition was observed in the range of 0.1 < x_Li < 0.8 for PS and 0.1 < x_Li < 0.4 for SL, which proves that these ionic fluids should be liquid within wide ranges of lithium composition and temperature. Hence, we believe this class of ionic fluids should be called ”glass-forming liquid electrolytes”. Mean solvation numbers and mean anion binding numbers were evaluated for these binary mixtures, which suggested that free, CIP/AGG and AGG species of both the solvent and the anion exist in these binary mixtures. Furthermore, the dipole reorientation dynamics of the bulk solvent and the AGGs were elucidated, and these suggested that the AGGs in particular (in which the charges are symmetric, yielding a relatively small dipole moment) have a large space scale. The AGGs could play a key role in the unique Li⁺ conduction found in PS- and SL-based glass-forming liquid electrolytes.

Author contributions

Y. U.: supervision and conceptualization of the research, writing a draft and the final manuscript. E. O.: measurement and analysis of Raman spectra and dielectric relaxation spectra. H. W. and J. H.: conceptualization and discussion of the results.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

The authors express their deep thanks to Prof. Shiro Seki (Department of Environmental Chemistry and Chemical Engineering, School of Advanced Engineering) for technical assistance with the diffusion coefficients using NMR and Prof. Shigeaki Morita (Department of Engineering Science, Osaka Electro-Communication University) for his valuable discussions. This work was partially supported by Grant-in-Aid for Scientific Research (No. 18H01994, 18H03926, 20H05663) from the Japan Society for the Promotion of Science (JSPS) and Ministry of Education, Culture, Sports, Science and Technology (MEXT), Research Fellowships (No. 17J02361 and 20J14822) from the JSPS and the Japan Science and Technology Agency Advanced Low Carbon Technology Research and Development Program Specially Promoted Research for Innovative Next Generation Batteries (JST ALCA-SPRING) Grant Number JPMJAL1301, Japan and JST Green Technologies of Excellence (Gtex) Grant Number JPMJGX23S0, Japan.

Notes and references

1. Y. Yamada, K. Furukawa, K. Sodeyama, K. Kikuchi, M. Yaegashi, Y. Tateyama and A. Yamada, J. Am. Chem. Soc., 2014, 136(13), 5039–5046 CrossRef CAS PubMed .
2. Y. Yamada and A. Yamada, J. Electrochem. Soc., 2015, 162, A2406 CrossRef CAS .
3. K. Yoshida, M. Tsuchiya, N. Tachikawa, K. Dokko and M. Watanabe, MRS Online Proc. Libr., 2012, 1473, 20–25 Search PubMed .
4. K. Ueno, K. Yoshida, M. Tsuchiya, N. Tachikawa, K. Dokko and M. Watanabe, J. Phys. Chem. B, 2012, 116(36), 11323–11331 CrossRef CAS PubMed .
5. M. Watanabe, K. Dokko, K. Ueno and M. L. Thomas, Bull. Chem. Soc. Jpn., 2018, 91, 1660–1682 CrossRef CAS .
6. D. J. Eyckens and L. C. Henderson, Front. Chem., 2019, 7, 263 CrossRef CAS PubMed .
7. A. P. Abbott, G. Capper, D. L. Davies, R. K. Rasheed and V. Tambyrajah, Chem. Commun., 2003, 70–71 RSC .
8. C. A. Angell, J. Electrochem. Soc., 1965, 112, 1224 CrossRef .
9. C. A. Angell and D. M. Gruen, J. Am. Chem. Soc., 1966, 88(22), 5192–5198 CrossRef CAS .
10. Y. Yamada, K. Usui, K. Sodeyama, S. J. Ko, Y. Tateyama and A. Yamada, Nat. Energy, 2016, 1, 16129 CrossRef CAS .
11. N. Arai, H. Watanabe, E. Nozaki, S. Seki, S. Tsuzuki, K. Ueno, K. Dokko, M. Watanabe, Y. Kameda and Y. Umebayashi, J. Phys. Chem. Lett., 2020, 11(11), 4517–4523 CrossRef CAS PubMed .
12. A. Angell, N. Byrne and J. P. Belieres, Acc. Chem. Res., 2007, 40(11), 1228–1236 CrossRef PubMed .
13. R. Buchner, in Novel Approaches to the Structure and Dynamics of Liquids: Experiments, Theories and Simulations, ed. J. Samios and V. A. Durov, NATO Science Series, 2004, pp. 265–288 Search PubMed .
14. H. Watanabe, T. Umecky, N. Arai, A. Nazet, T. Takamuku, K. R. Harris, Y. Kameda, R. Buchner and Y. Umebayashi, J. Phys. Chem. B, 2019, 123(29), 6244–6252 CrossRef CAS PubMed .
15. N. Arai, H. Watanabe, T. Yamaguchi, S. Seki, K. Ueno, K. Dokko, M. Watanabe, Y. Kameda, R. Buchner and Y. Umebayashi, J. Phys. Chem. C, 2019, 123(50), 30228–30233 CrossRef CAS .
16. J. Alvarado, M. A. Schroeder, M. Zhang, O. Borodin, E. Gobrogge, M. Olguin, M. S. Ding, M. Gobet, S. Greenbaum, Y. S. Meng and K. Xu, Mater. Today, 2018, 21(4), 341 CrossRef CAS .
17. K. Dokko, D. Watanabe, Y. Ugata, M. L. Thomas, S. Tsuzuki, W. Shinoda, K. Hashimoto, K. Ueno, Y. Umebayashi and M. Watanabe, J. Phys. Chem. B, 2018, 122(47), 10736–10745 CrossRef CAS PubMed .
18. A. Nakanishi, K. Ueno, D. Watanabe, Y. Ugata, Y. Matsumae, J. Liu, M. L. Thomas, K. Dokko and M. Watanabe, J. Phys. Chem. C, 2019, 123(23), 14229–14238 CrossRef CAS .
19. T. Sudoh, S. Ikeda, K. Shigenobu, S. Tsuzuki, K. Dokko, M. Watanabe, W. Shinoda and K. Ueno, J. Phys. Chem., 2023, 127, 12295–12303 CrossRef CAS .
20. J. S. Wilkes, Green Chem., 2002, 4(2), 73–80 RSC .
21. T. Welton, Chem. Rev., 1999, 99(8), 2071–2084 CrossRef CAS PubMed .
22. D. MacFarlane, M. Kar and J. M. Pringl, Fundamentals of Ionic Liquids: From Chemistry to Applications, Wiley-VCH., Weinheim, Germany, 2017 Search PubMed .
23. M. Freemantle, An Introduction to Ionic Liquids, Royal Society of Chemistry, 2009 Search PubMed .
24. C. A. Angell, Y. Ansari and Z. Zhao, Faraday Discuss., 2012, 154, 9–27 RSC .
25. F. Guthrie, Philos. Mag., Ser. 5, 1884, 17(108), 462–482 CrossRef .
26. H. M. A. Abood, A. P. Abbott, A. D. Ballantyne and K. S. Ryder, Chem. Commun., 2011, 47, 3523–3525 RSC .
27. M. Li, B. Gao, C. Liu, W. Chena, Z. Shi, X. Hu and Z. Wang, Electrochim. Acta, 2015, 180, 811–814 CrossRef CAS .
28. A. P. Abbott, R. C. Harris, Y.-T. Hsieh, K. S. Rydera and I.-W. Sunb, Phys. Chem. Chem. Phys., 2014, 16, 14675 RSC .
29. V. S. Cvetković, N. M. Vukićević, N. Jovićević, J. S. Stevanović and J. N. Jovićević, Trans. Nonferrous Met. Soc. China, 2020, 30, 823–834 CrossRef .
30. J. S. Wilkes, J. A. Levisky, R. A. Wilson and C. L. Hussey, Inorg. Chem., 1982, 21(3), 1263–1264 CrossRef CAS .
31. J. S. Wilkes, J. A. Levisky, J. L. Pflug, C. L. Hussey and T. B. Scheffler, Anal. Chem., 1982, 54(13), 2378–2379 CrossRef CAS .
32. A. A. Fannin Jr., D. A. Floreani, L. A. King, J. S. Landers, B. J. Piersma, D. J. Stech, R. L. Vaughn, J. S. Wilkes and J. L. Williams, J. Phys. Chem., 1984, 88, 2614–2621 CrossRef .
33. M. Esteban, C. Ariño, J. M. Díaz-Cruz, M. S. Díaz-Cruz and R. Tauler, TrAC, Trends Anal. Chem., 2000, 19, 49–61 CrossRef CAS .
34. J. Jaumot, R. Gargallo, A. de Juan and R. Tauler, Chemom. Intell. Lab. Syst., 2005, 76, 101–110 CrossRef CAS .
35. S. Hwang, D.-H. Kim, J. H. Shin, J. E. Jang, K. H. Ahn, C. Lee and H. Lee, J. Phys. Chem. C, 2018, 122(34), 19438–19446 CrossRef CAS .
36. H. Lee, S. Hwang, M. Kim, K. Kwak, J. Lee, Y.-K. Han and H. Lee, J. Phys. Chem. Lett., 2020, 11(24), 10382–10387 CrossRef CAS PubMed .
37. J. Self, N. T. Hahn, K. D. Fong, S. A. McClary, K. R. Zavadil and K. A. Persson, J. Phys. Chem. Lett., 2020, 11(6), 2046–2052 CrossRef CAS PubMed .
38. J. Kim, B. Koo, J. Lim, J. Jeon, C. Lim, H. Lee, K. Kwak and M. Cho, ACS Energy Lett., 2022, 7(1), 189–196 CrossRef CAS .
39. B. Koo, H. Lee, S. Hwang and H. Lee, J. Phys. Chem. C, 2023, 127(12), 5676–5682 CrossRef CAS .
40. B. Koo, H. Lee, S. Hwang, J. Lee, Y.-K. Han, K. Ahn, C. Lee and H. Lee, J. Phys. Chem. C, 2023, 127, 18271–18278 CrossRef CAS .
41. J. Self, N. T. Hahn and K. A. Persson, Energy Environ. Mater., 2024, 7, e12494 CrossRef CAS .
42. E. A. S. Cavell, P. C. Knight and M. A. Sheikh, Trans. Faraday Soc., 1971, 67, 2225–2233 RSC .
43. J. Barthel, H. Hetzenauer and R. Buchner, Ber. Bunsenges. Phys. Chem., 1992, 96(10), 1424–1432 CrossRef CAS .
44. K. Pearson and A. Lee, Biometrika, 1903, 2(4), 357–462 CrossRef .
45. I. Noda and Y. Ozaki, Two-Dimensional Correlation Spectroscopy, John Wiley & Sons, 2005 Search PubMed .
46. Y. Ozaki, S. Šašic, T. Tanaka and I. Noda, Bull. Chem. Soc. Jpn., 2001, 74, 1–17 CrossRef CAS .

---

Footnote

† Electronic supplementary information (ESI) available: The details of the experimental methods, data analysis, Raman and DRS spectra, 2D correlation analysis between the Raman spectra and the dielectric loss are described in the ESI. See DOI: https://doi.org/10.1039/d4fd00050a

This journal is © The Royal Society of Chemistry 2024