Page 121 - Read Online
P. 121

Ahmed et al. Energy Mater. 2025, 5, 500079  https://dx.doi.org/10.20517/energymater.2024.209   Page 7 of 13






               where S  and S  are the scale factors for the crystalline and liquid phases, and Z, M, and V are the number of
                      c
                           L
               chemical formulas per unit cell, the molecular weights of these unit cells, and their respective volumes,
               respectively [37,38] . For simplicity, we consider that (ZMV)  = (ZMV) . In the fresh sample, W  ≈ 19%. Next,
                                                               C
                                                                        L
                                                                                              D
               the previously obtained standards for the liquid and crystalline phases were used to determine the weight
               fractions of the phases in the aged sample [Figure 4B]. In this case, W  ≈ 50%. One should notice that, with
                                                                          D
               the considerations made in this work, these weight fractions are not absolute values and must be analyzed as
               a qualitative comparison between the samples, suggesting that the aged sample has a ~2.5 times higher
               fraction of the disordered state.

               DISCUSSION
               Based on the presented experimental data of [P ][TFSI] and our previously published results obtained for
                                                        12
               [P ][PF ], where we performed a similar study , we can derive the common pattern about ion dynamics
                                                        [21]
                 1224
                       6
               in these OIPCs. Figure 5 presents the conductivity and LS spectra of the melted and solid phases. As
               mentioned above, the AC-DC crossover or τ  defines the timescale where ions escape from the cage of
                                                      σ
               surrounding ions and start the normal diffusion regime. In the conductivity spectra, it results in formation
               of the σ  plateau, and the melt phase demonstrates this typical behavior (upper panels in Figure 5).
                      DC
               However, in the solid phase, the initial high-frequency plateau is terminated by Process I, and the second
               low frequency plateau is formed with lower σ . The relaxation process observed in the light scattering
                                                       DC
               spectra corresponds to the structural relaxation and can be associated with the same time of local ion
               rearrangements (escape from the cage), τ . The light scattering relaxation processes for solid and melt
                                                   LS
               phases have almost the same positions and coincide with the high-frequency AC-DC crossover in both
               OIPCs, which indicates that local ionic mobility is comparable in liquid and solid phases. Howver, unlike
               the melt state, there is a mechanism of charge trapping in the solid phases, appearing as Process I. At the
               same time, there is no signature of this process in LS spectra. It might be explained by charge trapping
               leading to formation of large dipole moment. The dielectric response is proportional to the square of the
               dipole moment fluctuation, and therefore, this charge trapping appears as strong Process I. However, this
               process involves many individual ion jumps. In contrast, light scattering is sensitive to fluctuations in
               polarizability and is insensitive to dipole reorientation. As a result, LS detects individual ion jumps, but not
               the charge trapping process.


               The high ionic mobility indicated by the BDS and LS spectra is validated by direct measurements of the
               cation and anion diffusion coefficients in the liquid and solid phases by PFG-NMR. Both cation and anion
               self-diffusion coefficients show no sharp changes at the melt-solid phase transition in both OIPCs, while
               σ , defined by the low-frequency DC plateau (after Process I), drops significantly [Figure 2]. However,
                DC
               there are important differences between [P ][TFSI] and [P ][PF ] data. First, unlike the [P ][PF ], where
                                                                                             1224
                                                                       6
                                                                 1224
                                                   12
                                                                                                   6
               all  ions  are  mobile  in  Phase  I [11,21] , only  around  20%  of  cations  and  anions  are  mobile
               [Supplementary Figure 2B] in Phase I of [P ][TFSI]. Second, there are much stronger drops of high-
                                                     12
               frequency DC and AC conductivity and amplitude of structural relaxation peak in LS spectra in the solid
               phase of [P ][TFSI] in comparison with the same for [P ][PF ] [Figure 5]. The reduction in the fraction of
                                                                    6
                         12
                                                              1224
               mobile ions can only explain part of the drop in conductivity, indicating that there are additional
               mechanisms of conductivity suppression in the [P ][TFSI] system. To understand these mechanisms, we
                                                          12
               should discuss the key difference between ion diffusion and conductivity. Ion self-diffusion, measured by
               PFG-NMR, presents only self-part of velocity correlation functions, as given in
   116   117   118   119   120   121   122   123   124   125   126