Ammonium-Based Plastic Crystals as Solid-State Electrolytes for Lithium and Sodium Batteries

Ammonium-Based Plastic Crystals as Solid-State Electrolytes for Lithium and Sodium Batteries Manuel Salado, Thomas H. Smith, Nanditha Sirigiri, Fangfang Chen, Luke A. O’Dell, Jennifer M. Pringle, and Maria Forsyth* Cite This: JACS Au 2025, 5, 1663−1676 Read Online ACCESS Metrics & More Article Recommendations * sı Supporting Information ABSTRACT: Organic ionic plastic crystals (OIPCs) are a promising class of solid materials composed of organic cations and inorganic anions, increasingly explored for use as solid-state electrolytes (SSEs). These materials offer a safer alternative to conventional carbonate-based electrolytes in lithium and sodium ion batteries. In this study, lithium and sodium salts were incorporated into tetramethylammonium bis(fluorosulfonyl)imide ([N1111][FSI]), yielding solid state electrolytes with notable properties, including high ionic conductivities (1.79 mS·cm−1for LiFSI doped and 3.2 mS·cm−1NaFSI doped, both at 80 °C), elevated diffusion coefficients (up to 3.83 ×10−11 m2·s−1for Li+at 80 °C), and high transference numbers (0.8 and 0.4, for Li and Na, respectively). To date, except for ceramic and glassy ion conductors, there has been no significant research demonstrating true solid-state behavior with Li+or Na+ion transport fully decoupled from the motion of the host structure. Furthermore, these electrolytes have exhibited impressive current densities up to 3.5 mA·cm−2during Li|Li and 2.9 mA·cm−2for Na|Na symmetric cell cycling at room temperature. As a result, these materials hold considerable potential for enhancing both Li and Na electrochemical energy storage technologies, combining both improved efficiency and safety features. KEYWORDS: OIPC, single-ion-conductor, solid-electrolyte, energy storage ■INTRODUCTION Lithium-ion batteries (LIBs) play a pivotal role in today’s world, largely due to their widespread use in portable electronics. Their importance is expected to grow further, particularly in the fast-evolving sectors of electric vehicles (EVs) and grid storage for renewable energy. 1 Additionally, sodium-ion batteries (NIBs) are emerging as a promising alternative, particularly for large-scale applications where cost is a more critical factor than size, such as in grid energy storage. 2 Both LIBs and NIBs commonly utilize liquid carbonate-based electrolytes due to their low cost, high polarity, and broad electrochemical stability windows. 3 However, these electrolytes pose significant safety risks due to their high volatility and flammability. 4 As a result, there is a growing interest in exploring alternative electrolyte options, with solid-state electrolytes (SSEs) emerging as a promising solution. SSEs offer enhanced safety, as their lack of organic solvents makes them nonflammable. Moreover, their ability to suppress dendrite formation allows for the use of lithium and sodium metal anodes. 5 Despite these advantages, a key challenge with SSEs is their relatively low ionic conductivity, which can be several orders of magnitude lower than that of liquid electrolytes. 6 Organic ionic plastic crystals (OIPCs) have recently garnered attention as promising SSEs. OIPCs are composed of organic cations paired with inorganic anions, and while they share chemical similarities with ionic liquids (ILs), a key distinction is that OIPCs remain solid at room temperature. This solid state is due to the typically smaller cations in OIPCs, which enable tighter molecular packing and consequently higher melting points compared to ILs. 7 Within an OIPC, there is long-range crystalline order, but the individual cations and anions can present orientational disorder or indeed retain some degree of rotational and/or translational motions which contributes to the material’s plasticity. OIPCs are also characterized by their ability to undergo solid−solid phase transitions at temperatures below their melting point. These transitions can represent a rotator phase change with an increase in ion mobility, corresponding to progressively more disordered structures. 8 By convention, the phase immediately below the melting point is labeled phase I, with subsequent phases at lower temperatures designated as phase II, III, and so on. Among these, phase I generally exhibits the highest ionic conductivity. 9 OIPCs have several remarkable features that make them promising candidates for SSEs including nonflammability and nonvolatility, 10 as well as high thermal stability 11 and wide Received: November 14, 2024 Revised: March 5, 2025 Accepted: March 6, 2025 Published: March 26, 2025 Articlepubs.acs.org/jacsau © 2025 The Authors. Published by American Chemical Society 1663 https://doi.org/10.1021/jacsau.4c01086 JACS Au 2025, 5, 1663−1676 This article is licensed under CC-BY-NC-ND 4.0 Downloaded via UNIV DEL PAIS VASCO on May 8, 2025 at 13:35:07 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles. electrochemical stability windows. 12 Further, their plasticity means that they are more malleable which allows for good contact with electrodes, 13 especially when compared to other SSEs such as ceramics, that are more rigid and brittle. They have therefore, been proposed for various applications including fuel cells, 14,15 capacitors, 16,17 solar cells 18,19 as well as both lithium 20 and sodium 21 batteries. To be used as SSEs in lithium or sodium batteries, the OIPC must be combined with salts of the target ion (i.e., Li+or Na+), this typically increases the conductivity of the OIPC. 22 Like ILs, the chemistry of OIPCs can be easily varied, both the cations and anions can be easily modified, and this can have dramatic effects on the properties of the material. For example, OIPCs have been reported with cations including, but not limited to, ammonium, 23 pyrrolidinium, 24 imidazolium 25 and phosphonium. 21 Likewise, several different anions have been reported including bis(fluorosulfonyl)imide ([FSI]−), 9 bis(trifluoromethanesulfonyl)imide ([TFSI]−), 26 hexafluorophosphate ([PF6]−) 27 and tetrafluoroborate ([BF4]−). 28 Further changes such as alkyl chain length have been shown to have a large influence on properties such as thermal behavior and conductivity. 29 Quaternary ammonium cation based OIPCs have previously been studied with small cations such as tetramethylammonium ([N1111]+), methyl(triethyl)ammonium ([N1222]+) and tetraethylammonium ([N2222]+), 30 as well as bis-quaternary ammonium cations. 31 To make OIPCS, these cations have been combined with [FSI]−and [TFSI]−anions. The resulting OIPCs showed good performance in terms of mechanical properties, ionic conductivity, thermal and electrochemical stabilities. Further, when [N1222][FSI] was combined with Li salts (either LiFSI or LiTFSI) to make SSEs, the resulting materials were found to have impressive transference number and good interface stabilities against lithium metal. 32,33 This work focuses on the [N1111]+cation. The thermal properties of the materials formed with this cation were first reported Yunis et al. 30 When the [FSI]−anion was used, the resulting material was found to have only a single-phase transition before the melting point. Whereas, when [TFSI]− was used, the resulting material was found to be a crystalline solid without any phase transitions below the melting point. The thermal behavior of [N1111][FSI] was then further studied through a combination of experimental techniques and molecular dynamics simulations by Sirigiri et al. 34 It was found that strong interionic forces accounted for restricted ion motion, resulting in low plasticity and high structural order. Further, the relatively simple motion of the cations and anions, with cations undergoing rotation and the cis−trans transition of the anion, led to the simple phase behavior of the material. Despite showing promising characteristics, [N1111][FSI] has never been tested as a SSE in lithium or sodium devices. Herein, we investigate the incorporation of Li+or Na+ions, using lithium bis(fluorosulfonyl)imide (LiFSI) or sodium bis(fluorosulfonyl)imide (NaFSI), into [N1111][FSI] with a variety of analytical techniques. The phase behavior of the resulting materials was studied with DSC and XRD. All materials showed two phases, with a phase change at approximately 78 °C. NMR was used to study the ion mobility in the materials. The LiFSI systems showed only Li+ions were mobile in the low temperature phase (II) with the [N1111]+ cations and [FSI]−anions only becoming mobile in phase I. The NaFSI systems showed similar behavior for the [N1111]+ cations and [FSI]−anions, however the Na+ions had a more restricted mobility, with a noticeable amount of Na+ions remaining immobile above the phase transition. Initial electrochemical experiments were performed for the most promising candidate for each salt (10% LiFSI and 25% NaFSI, as determined by diffusion coefficients and ionic conductivities). It was found that impressive transference numbers could be achieved as well and good compatibilities with Li and Na metals. We therefore present [N1111][FSI] doped with LiFSI or NaFSI as highly promising candidates for SSEs in both Li and Na metal batteries. ■RESULTS AND DISCUSSION LiFSI Doped Samples The thermal behaviors and phase transitions of neat [N1111][FSI] and [N1111][FSI] doped with different amounts of LiFSI (5 mol−25 mol %) were investigated by DSC measurements and the results are shown in Figure 1. Previous studies 30 have shown that neat [N1111][FSI] has a simple phase behavior compared to other OIPC systems, 35 with only one solid−solid transition around 78 °C and a melting temperature at 300 °C. This phase transition at 78.3 °C remains constant for the samples doped with LiFSI. Nevertheless, at higher doping levels (15 and 25% LiFSI), samples exhibit notable glass transitions upon heating, observed at −57.9 and −59.7 °C, respectively (Figure S2). This glass transition can be attributed either to a liquid phase that freezes into a glass below Tgor to a freezing in of orientational disorder below Tg (essentially a material where the short-range disorder is “frozen” and which gains rotational degrees of freedom at Tg). This is in contrast to the solid−solid first order phase transitions observed in going from an ordered crystal to a disordered rotator phase. 36−38 Interestingly, an exothermic devitrification event is observed after the solid−solid transition, which shifts to lower temperatures (from 141.7 to 122.4 °C) as the proportion of LiFSI doped into the system increases. This exothermic event can be attributed to the existence of a small fraction of amorphous phase within an otherwise mostly crystalline OIPC. Again, this amorphous phase could be a small fraction of liquid phase which must have the coexistence of all ions in the materials, or amorphous regions that Figure 1. DSC of first heating traces of the neat [N1111][FSI] and with different molar doping with LiFSI. JACS Au pubs.acs.org/jacsau Article https://doi.org/10.1021/jacsau.4c01086 JACS Au 2025, 5, 1663−1676 1664 recrystallize into a more ordered structure. We will observe later that the XRD and NMR data do not support the presence of a liquid phase. Consequently, the exothermic peak increases accordingly with the amount of the LiFSI in the sample (from 0.12 kJ·mol−1of 5% LiFSI to 0.79 kJ·mol−1of 25% LiFSI). During the second heating, the composites showed slightly higher OIPC enthalpies (per mol of OIPC) of transitions (Table 1) possibly due to the recrystallization of the OIPC after the solid−solid transition. The exothermic peak also disappears on subsequent heating, indicating a stabilization of the structure after the first heating. As previous studies have reported, 35 mixing of Li salts with plastic crystals often lowers the solid−solid transition temperature. This is consistent with the DSC trace in the second heating as depicted in Figure S3 as well as Table 1. The entropies of solid−solid transition (ΔSS−S) were calculated from the melting endotherm area (ΔHS−S) using the relationship ΔSS−S=ΔHS−S/TS−S. Crystalline solid−solid transitions are first-order phase transitions characterized by discontinuous changes in volume, enthalpy, and entropy, resulting from alterations in crystal packing (Table 1). However, the magnitude of these changes is typically smaller compared to those seen in crystalline solid−liquid transitions. While the transition involves a symmetry break in the crystal structure, it requires cooperative molecular rearrangement between the two phases, with the positional shifts of the molecules remaining relatively small to ensure a coherent transformation. 39,40 The extracted values of ΔSS−Sdecrease as the proportion of LiFSI doped into the systems is increased compared to neat [N1111][FSI]. This indicates that the increased additional LiFSI leads to the plastic crystal mixtures becoming increasingly more disordered. 41 We note that we do not heat the samples to the melting transition to avoid decomposition of the material as the melting point of the neat [N1111][FSI] is greater than 300 °C. Synchrotron X-ray diffraction was employed to investigate the crystal structures and provide deeper insight into the phase diagram of [N1111][FSI]/[LiFSI] OIPC mixtures. Figure 2 presents the powder diffraction patterns for pure [N1111][FSI], as well as for samples doped with 2 and 10 mol % LiFSI at 40 and 110 °C. These measurement temperatures were selected based on DSC analysis, corresponding to conditions before and after the phase II−I transitions. The crystal structure of [N1111][FSI] was previously studied by Sirigiri et al. 34 using powder X-ray diffraction patterns. As in the aforementioned work, the diffraction pattern at phase II temperatures displays multiple peaks related to a monoclinic crystal structure type with P21/m space group. After the analysis of the diffractograms, the parameters of this specific unit cell correspond to a = 10.26 Å, b= 13.48 Å, c= 7.09 Å and three angles of α= 90°, ß = 90.69°,γ= 90°, with a unit cell volume of 981.09 Å3. The difference in the pattern of 10% LiFSI, (e.g., change in the intensity at high 2θand the disappearance of peaks) suggest the formation of a different crystal structure due to the inclusion of orthorhombic LiFSI. Above the solid−solid transition temperature, the XRD pattern shows the abrupt disappearance of a distinct set of peaks after the transition from phase II to phase I (Figure S4). Only two distinctive XRD peaks are found for phase I, indicating that the monoclinic crystal structure at low temperatures (phase II) has transformed into a simple structure with higher symmetry, such a body-centered cubic structure. In addition, the disappearance of the higher-angle Bragg peaks is due to a loss of short-range molecular order in the system, caused by local molecular motion. Extracting the unit cell parameters a=b=c= 9.51 Å and α= ß = γ= 90°, the volume decreases to 860 Å3in the case of pure [N1111][FSI]. In previous research on various OIPC systems, similar alterations in XRD patterns following phase transitions were noted. 10,28,42 This phenomenon is attributed to the initiation of ion rotations at fixed lattice sites, resulting in crystal structures with higher symmetry. Interestingly, among the various solid-state OIPC systems studied, [C2mpyr][BF4] exhibited a similar crystal structure change after the addition of a lithium salt. Specifically, upon the addition of LiBF4, the system adopts three distinct crystal structures: a dominant monoclinic phase (Phase III), a secondary trigonal phase, and a cubic phase (Phase I). Notably, the ionic conductivity of these compounds increases significantly at higher temperatures (Phase I) compared to lower temperature phases, as occurs in the [N1111][FSI] system as presented below. 43,44 Table 1. Phase Transition Temperatures, Enthalpy and Final Entropy of Fusion for [N1111][FSI] and Their LiFSI-Doped Composites [N1111][FSI] doped with LiFSI 1st heating 2nd heating TS−S(°C) ΔHS−S(kJ·mol−1)ΔHS−S(kJ·mol−1)TS−S(°C) ΔHS−S(kJ·mol−1)ΔHS−S(kJ·mol−1) [N1111][FSI] 78.3 23.4 0.066 5% LiFSI 78.3 11.64 0.034 78.3 13.0 0.038 10% LiFSI 78.3 8.7 0.026 74.2 8.44 0.026 15% LiFSI 78.3 6.2 0.019 74.2 6.22 0.020 25% LiFSI 78.3 4.41 0.015 74.2 4.60 0.016 Figure 2. Synchrotron powder X-ray diffraction patterns of [N1111][FSI] and with 2 and 10% LiFSI doping. The left part corresponds to Phase II and right to phase I (above 78.3 °C). JACS Au pubs.acs.org/jacsau Article https://doi.org/10.1021/jacsau.4c01086 JACS Au 2025, 5, 1663−1676 1665 Static solid-state NMR was used to study the temperaturedependent ion dynamics of the three components of the OIPCs. This is possible as 1H is only present in the [N1111]+ cation, 7Li is only present as the Li+cation and 19F is only present in the [FSI]−anion. Specifically, static measurements were performed as the line widths and relative intensities can be used to provide information about the dynamics and relative populations of the ions present. Broad signals correspond to ions having limited dynamics and/or motion happening below the NMR time scale. Conversely, when motion does occur, NMR interactions are averaged out to give narrower signals. Typically, as the temperature is increased narrower signals grow relative to the broad signals, as motion increases. However, more dramatic changes often occur due to phase changes. 45 The static 1H, 7Li and 19F NMR spectra for [N1111][FSI] with 10% and 25% LiFSI at temperatures ranging from −20 to 140 °C are given in Figure 3. For both samples, the 7Li spectra show relatively narrow peaks that do not change much on heating, in particular above room temperature. However, the spectra for the 10% LiFSI sample show a noticeably broader component whereas the 25% LiFSI sample appears as a single peak. Therefore, the spectra for [N1111][FSI] with 10% LiFSI were fitted with broad and narrow peaks, with full width at half maxima (fwhm) values of approximately 7 kHz and 1 kHz, respectively. An example of this fitting is shown in Figure S5. Interestingly, the relative area of the narrow peak (Figure S6b) is roughly constant between 30 and 35% and only increases to 46% at 140 °C. Previous work in polymer electrolyte materials have also observed broad components at the base of a narrower 7Li signal and have assigned this to quadrupolar features as 7Li has a spin of 3/2. 46 The relative ratio of the broad and narrow lines is consistent with this explanation. The FWHM of the narrow peaks can then be compared to the peak widths of the 25% LiFSI spectra without fitting, as shown in Figure S6a. For 10% LiFSI there is an initial step down in FWHM values from 1280 to 954 Hz as the temperature is increased from −20 to 0 °C, the FWHM values for both samples then steadily become narrower as the temperature is increased to 60 °C (890 Hz for 10% LiFSI and 780 Hz for 25% LiFSI). The total narrowing of the 7Li signal for both samples from 60 to 100 °C is relatively insignificant (890 to 700 Hz for 10% LiFSI and 780 to 660 Hz for 25% LiFSI), notwithstanding these line widths are narrow enough to suggest significant Li+ion diffusion in both cases and at all temperatures as discussed below. Given the phase change (from II to I), was previously observed in DSC experiments at 78.3 °C, it is interesting to note that there is no significant change in lithium mobility as we move through this transition. The 19F spectra however show a more noticeable change on heating. The low temperature spectra are dominated by broad chemical shift anisotropy (CSA) powder patterns. These CSA patterns arise from the superposition of multiple peaks representing all orientations of the anions relative to the magnetic field and are sensitive to molecular motions and alignments. 47 The CSA patterns can be described by the Haeberlen convention, 48 where δis the reduced anisotropy (δ =δzz−δiso) and ηis the asymmetry parameter (η= (δxx−δyy)/ δ) which can take a value between 0 and 1. As the temperature is increased, narrow symmetric peaks with fwhm values of approximately 2 kHz are also observed, which correspond to a very small fraction of anions that are rotating isotropically (and potentially diffusing) as previously observed even in pure OIPCs in the presence of mobile defects. 27,28 To quantify this effect, the 19F spectra were fitted with two components, a broad CSA pattern and a narrow symmetric peak (an example of this fitting is given in Figure S7). For both samples, ηfor the CSA patterns is approximately constant at 0.45. The FWHM values of the narrow peaks and δof the CSA patterns, as well as the relative areas of narrow peaks at all temperatures are given in Figure S8a,b, respectively. The spectra collected at 40 °C have a nonideal CSA line shape most likely due to intermediate time scale dynamics causing partial averaging of the CSA. It was therefore excluded from further analysis. For phase II (up to 60 °C), the CSA patterns correspond to the majority of the signals, indicating most of the [FSI]−is immobile. However, the CSA patterns narrow as the temperature is increased, with δdecreasing from about 69 kHz at −20 °C to 61 kHz at 60 °C, indicating the onset of some nonisotropic motion. The FWHM of the narrow peaks stay roughly constant at 2 kHz. It is noticeable that the narrow signals have a slightly higher relative intensity for 25% LiFSI (6% at 60 °C compared to 2% for 10% LiFSI) consistent with a greater degree of disorder observed from the DSC. After the transition to phase I at 80 °C, the narrow peaks increase in relative intensity, such that they correspond to 44 and 84% of the signals for 10% and 25% LiFSI, respectively, and the CSA patterns narrow further. For the spectra corresponding to 100 °C and above, only the narrow peaks are present, and they continue to narrow as the temperature is increased. This suggests that for phase II the [FSI]−anions are almost entirely immobile and for phase I the [FSI]−anions are almost all undergoing isotropic rotations. This is in sharp contrast to the 7Li behavior and therefore suggests that there is no significant fraction of liquid phase present in the OIPC upon doping, in contrast to previously studied systems where a noticeable liquid phase coexists with the OIPC phase 49,50 ; if a liquid phase coexisted, we would expect to see at least as many mobile [FSI]−anions as Li+ cations across all temperatures. The 1H spectra show similar behavior to the 19F spectra. The low temperature spectra consist of relatively broad peaks, with FWHM values of approximately 10 kHz, indicating that Figure 3. Static solid-state (a) 7Li, (b) 19F and (c) 1H NMR spectra of [N1111][FSI] with 10% LiFSI, and (d) 7Li, (e) 19F and (f) 1H spectra of [N1111][FSI] with 25% LiFSI collected at temperatures ranging from −20 to 140 °C, maximum intensities normalized. JACS Au pubs.acs.org/jacsau Article https://doi.org/10.1021/jacsau.4c01086 JACS Au 2025, 5, 1663−1676 1666 the [N1111]+cations have limited mobility in phase II. As with the 19F spectra, low intensity narrow peaks are also observed, which grow in intensity, particular after the phase II−I transition. The broad peaks are not purely Gaussian/ Lorentzian which suggests that there may be some distribution in 1H environments or dynamics present, however a reasonably good fit was achieved with two components (broad and narrow), an example of this fitting is given in Figure S9. The FWHM values of the components and the relative area of the peaks are given in Figure S10a,b, respectively. For phase II (60 °C and below) there are only broad peaks, which narrow slightly as the temperature is increased (from about 9.2 kHz at −20 °C to 8.6 kHz at 60 °C). This indicates that, in phase II, almost all of the [N1111]+cations are immobile, again consistent with a solid-state structure in phase II within which the Li+ions are able to diffuse. After the phase change, the broad peaks narrow at a faster rate and the narrow peak starts growing, such that by 140 °C the narrow peaks correspond to 59% of the signal for the 10% LiFSI sample and 72% for the 25% LiFSI sample. This is still significantly less than for the [FSI]−anions suggesting that the OIPC cations essentially hold together the phase I crystal structure. Pulsed Field Gradient (PFG) NMR was used to determine the diffusion coefficients of the individual components of the samples. Similarly, to the static NMR measurements, the 1H, 7Li and 19F data sets provide information about the motion of [N1111]+cations, Li+cations and [FSI]−anions, respectively. The full diffusion coefficient data sets for [N1111][FSI] doped with 5, 10, 15 and 25% LiFSI, example PFG NMR spectra collected at 60 °C without the gradient field applied and attenuation curves for all nuclei (for the 10% LiFSI sample) are given in Figures S11−S13, respectively, and the accompanying activation energies are given in Table S1. All diffusion coefficients increase with temperature, as expected. For all samples, there is a clear trend where the highest diffusion coefficients observed for 7Li. Figure 4a shows the 7Li diffusion coefficients for all samples at 20, 60 and 80 °C, at each temperature the highest diffusion coefficients are observed for 10% LiFSI. Together with the static NMR results, this indicates that while a greater proportion of Li+cations are mobile in 25% LiFSI, those ions that are mobile in 10% LiFSI have higher mobilities. It is worth noting that diffusion coefficients only apply to the mobile components of the sample, which static solid-state NMR experiments showed to be a small minority of the [N1111]+cations and [FSI]−anions in phase II. To demonstrate this, Figure 4b−e show 19F and 1H diffusion coefficients at 20, 60 and 80 °C from samples with 10 and 25% LiFSI with the proportion of mobile species. Again, the diffusion coefficients are greater for 10% LiFSI. But in phase II, this only corresponds to about 5% of the [FSI]−anions and a trace amount of the [N1111]+cations. After the phase transition at 80 °C, there is substantial increase in the amount of [FSI]−anions mobile, but not for the [N1111]+cations which static NMR showed only become mobile as the temperature is increased further. Thus, while both the OIPC anion and cation also have significant diffusion values as measured by 19F and 1H PFG NMR, only a small fraction of these ions are actually diffusing, until the phase II to I transition occurs, as indicated by the data in Figure 4b−e. Unfortunately, diffusion coefficients could not be measured at temperatures above 80 °C due to probe limitations. It has been already demonstrated that mixing OIPCs with alkali halide salts can enhance the transport of specific “target” ions. 44 For instance, MacFarlane et al. 51 found that introducing less than 1% Li+ions to [C1mpyr][TFSI] and [C2mpyr]- [TFSI] compounds increased their ionic conductivity by a factor of 20 at 25 °C. However, in most OIPCs it is now accepted that this is likely due to the presence of a liquid phase at the grain boundaries within the OIPC. 49,52 On the other hand [C2mpyr][BF4] is a material that has garnered significant attention due to its promising electrochemical properties that appear to exist in the true solid-state phase, leading to various studies exploring its potential. For example, Shekibi et al. 43 demonstrated that doping with 10 mol % LiBF4resulted in optimal performance, maintaining a consistent conductivity of 10−3S·cm−1over a broad temperature range in phase I. Later, Iranipour et al. 11 conducted a comprehensive study of the material’s structural changes and ionic diffusion using a variety of characterization techniques. Their findings revealed the presence of a secondary phase within the microstructure of the lithium-doped OIPC, which increased in proportion relative to the matrix phase as lithium concentration rose. This mixedphase microstructure was found to limit ionic conductivity at lower temperatures (phase II) in samples with higher LiBF4 concentrations, in contrast to those with lower concentrations. Interestingly, the [C2mpyr][BF4] OIPC undergoes a crystal structure rearrangement during its solid−solid transition from a monoclinic to a cubic crystal structure, similar to the [N1111][FSI] system under consideration here and it is possible that this is favorable for hosting Li+cations and hence retaining a true solid-state material at lower temperatures. To evaluate the conductivity of [N1111][FSI] and LiFSIdoped systems as a function of temperature, impedance Figure 4. 7Li diffusion coefficients as determined by PFG NMR for all [N1111][FSI] doped with 5, 10, 15 and 25% LiFSI at 20, 60 and 80 °C (a). 19F (b, c) and 1H (d, e) diffusion coefficients, as well as the proportion of mobile species, as determined from static NMR measurements, at 20, 60 and 80 °C in [N1111][FSI] doped with 10 and 25% LiFSI. JACS Au pubs.acs.org/jacsau Article https://doi.org/10.1021/jacsau.4c01086 JACS Au 2025, 5, 1663−1676 1667 spectroscopy was performed on compositions containing 5−25 mol % LiFSI (Figure S14). Unlike pristine [N1111][FSI], the conductivity of the mixed systems at room temperature is 6 orders of magnitude higher (Figure 5). In contrast to the pure system, where a sharp one-order-of-magnitude increase in ionic conductivity is observed at the phase II−I transition temperature, the LiFSI-doped samples exhibit a continuous rise in conductivity with increasing temperature. Consistent with PFG NMR measurements, the optimal composition was found to be 10% LiFSI, yielding an ionic conductivity of 1.8 mS·cm−1at 80 °C. Further increases in LiFSI concentration beyond 15% led to a decline in conductivity, likely due to stronger ion−ion correlations at higher doping levels, which hinder ionic mobility and suppress conductivity. 53 The conductivity for 10% LiFSI at 25 °C was calculated using the Nernst−Einstein expression and the measured PFG 7Li diffusion coefficient obtaining a value of 5.2 ×10−4S cm−1compared with the value measured from EIS of 1.17 ×10−4S cm−1(see SI for calculation). These two values are very close considering the approximations involved in the calculation. It is tempting to suggest that the higher conductivity must arise from a liquid phase coexisting with the OIPC phase, however, neither the XRD nor the NMR data support this hypothesis. For example, the conductivity at 25 °C in phase II is greater than 10−4mS·cm−1but less than 0.1% of available [FSI]−or [N1111]+ions are diffusing, even for 25 mol % LiFSI, whereas all Li+ions are diffusing even at room temperature. We further calculate the activation energies from the [N1111][FSI] doped with LiFSI systems obtaining values of 17.06, 43.44, 11.13, and 16.8 kJ·mol−1for 5, 10, 15 and 25% LiFSI, respectively (Figure S15), and being 131.4 kJ·mol−1the activation energy of the neat [N1111][FSI]. It is generally believed that low activation energies (e.g., Ea< 38.6 kJ·mol−1) enable high ionic conductivity. Considering the Arrhenius equation, it is apparent that with a fixed σ0, a lower Eais expected for higher σat a fixed temperature. However, in some ionic conductor materials, according to Meyer−Neldel rule (MNR), 54 a linear relation is observed between ln(σ0) and Ea in the form of, = +Eln( ) 0 a Here, αand βare constants, where the reciprocal of αis termed as the Meyer−Neldel energy Δ0. Therefore, in solid conductors, the total ionic conductivity then depends on Arrhenius prefactor σ0and activation energy Ea. In this study, we utilized the Meyer-Neldel relationship to understand the correlation between conductivity or diffusivity and activation energy in amorphous solids and explore charge transfer in glassy materials, highlighting the critical role of molecular motions in facilitating electronic transport within these systems. According to previous studies, 55 the behavior of the material depends on the relative magnitude of Meyer−Neldel energy Δ0and the thermal energy kBT at a given application temperature. If we take the thermal energy at a temperature where the material is structurally stable (e.g., 25 °C) and compare it with Meyer−Neldel energy, we obtain that Δ0< kBT, and so a higher activation energy will lead to higher conductivity for this type of structured materials. According to Anderson and Straut, 56 the total activation energy for an ion diffusing in an amorphous solid has both contribution from strain energy as well as electrostatic interactions. The diffusion of smaller ions, such Li+in our system, increases the electrostatic energy which is likely the critical factor determining the activation energy. Despite the complex nature of these features, the relationship between ionic conductivity and these factors remains largely unexplored, with only limited insights available for specific cases. For example, it has been proposed that the orientational and/or rotational motions of OIPCs may promote faster Li+ion diffusion. Another study hypothesized 51 that salt doping could introduce additional defects or form new phases within grain boundary regions, thereby enhancing ion conduction in these areas. Gaining a deeper understanding of the role of grain boundaries could be key to explaining the observed increase in ionic conductivity following salt addition. NaFSI Doped Samples In their study of solid-state sodium OIPC-based electrolytes, Forsyth et al. 35 investigated the phase behavior of N-methyl-Nethyl-pyrrolidinium bis(trifluoromethanesulfonyl)imide ([C2mpyr][TFSI]) when mixed with the sodium salt Na- [TFSI]. As expected, the mixed electrolyte (e.g., 40 mol % Na[TFSI]) exhibited an ionic conductivity of 10−4S·cm−1at 60 °C, more than 3 orders of magnitude higher than the conductivity of the pure OIPC. Notably, the eutectic temperature of the sodium-based system (63 °C) is significantly higher than that of its lithium counterpart (30 °C), enabling solid-state conductivity across a broader temperature range. Another system, N-methyl-N-methylpyrrolidinium dicyanamide ([C1mpyr][N(CN)2]) mixed with Na[N(CN)2], displayed an even higher eutectic temperature (∼89 °C), while maintaining a stable ionic conductivity above 10−4S·cm−1at around 80 °C. 57 In a related study, Makhlooghiazad et al. 26 explored mixedphase electrolyte materials based on solid-state compositions of trimethylisobutylphosphonium bis- (trifluoromethanesulfonyl)imide ([P111i4][TFSI]) and high concentrations of Na[TFSI] (above 70 mol %). These materials remain solid at room temperature and retain a soft solid consistency at 50 °C. Remarkably, their ionic conductivity approaches that of the pure ionic liquid at 50 °C, reaching values between 10−3and 10−2S·cm−1. This highlights the difference in material properties depending on the specific ions introduced into the electrolyte Figure 5. Ionic conductivity dependence with temperature of [N1111][FSI] and doped with LiFSI. JACS Au pubs.acs.org/jacsau Article https://doi.org/10.1021/jacsau.4c01086 JACS Au 2025, 5, 1663−1676 1668 system. In this study, systems doped with NaFSI instead of LiFSI also exhibit distinct behavior. A key reason for this difference could be that sodium systems tend to form new compounds even at low molar concentrations and exhibit higher eutectic temperatures. 58 The thermograms for neat [N1111][FSI] and those doped with varying molar percentage of NaFSI are shown in Figure 6. As happened with the previous lithium-doped samples, all exhibit an endothermic solid−solid transition followed by an exothermic process during the first heating cycle, with these transitions shifting to lower temperatures as the NaFSI concentration increases (from 166.2 °C at 5% NaFSI to 159.5 °C at 25% NaFSI). Unlike LiFSI, the solid−solid transition temperature for the highest NaFSI concentration (25%) shows a slight shift from 78.3 to 76.7 °C during the first heating cycle. Nevertheless, this shift is observed across all compositions in the second heating cycle (Figure S16 and Table 2), aligning with previous reports (e.g., [C2mpyr]- [TFSI], upon mixing with the sodium salt, Na[TFSI]). 35 Additionally, no evidence of glass transitions was detected within the investigated temperature range. However, mixing with the sodium salt (at higher concentrations (15−25%) also introduces an additional endothermic peak just before the solid−solid transition temperature (e.g., 50 °C) during the first heating cycle, which may suggest a eutectic melting and the presence of a new sodium-rich phase, although this is not consistent across the compositions. The ΔSS−Svalues per mole of OIPC for the solid−solid transitions (Table 2) show a similar trend to LiFSI, where they decrease with increased proportion of NaFSI, indicating an increased disordered system. The lower value of entropy observed in the second heating cycle is in contrast to the effect of LiFSI additions and again highlights the different behavior of introducing Na compared with Li cations into the OIPC. Figures 7 and S17 represent the crystal structure evolution with temperature of neat [N1111][FSI] and [N1111][FSI] doped with 2 and 10% NaFSI. After the analysis of the diffractograms, the parameters of this specific unit cell correspond to a= 10.26 Å, b= 13.48 Å, c= 7.09 Å and three angles of α= 90°, ß = 90.69°,γ= 90°, with a unit cell volume of 981.09 Å3. Although the XRD diffractograms show at first glance a solid−solid transition from monoclinic to cubic crystal phase (e.g., a=b= c= 9.51 Å and α= ß = γ= 90°, the volume decreases to 860 Å3) similar to lithium doped samples, the diagram observed at 10% NaFSI doping at 40 °C differs significantly from that at 2% due to the appearance of new diffraction peaks. The existence of these new peaks also indicates the possibility of forming complex defects because of coexisting different crystals domains, adding another level of complexity to the structure of [N1111][FSI] doped with NaFSI, as it was observed in the DSC measurements. After the solid−solid transition, variation of the lattice parameters and the unit cell volume as a function of temperature was found to be the same as for LiFSI doping. The observed decreased peak intensity, relative to the pure [N1111][FSI], suggests increased disorder within the OIPC lattice. This disorder likely arises from the introduction of structural defects or grain boundaries that can be attributed to the existence of mixed crystal phases, and which can cause Figure 6. DSC of first heating traces of the neat [N1111][FSI] and with different concentrations of NaFSI. Table 2. Phase Transition Temperatures, Enthalpy and Final Entropy of Fusion for [N1111][FSI] and their NaFSI-Doped Composites [N1111][FSI] doped with NaFSI 1st heating 2nd heating TS−S(°C) ΔHS−S(kJ·mol−1)ΔHS−S(kJ·mol−1)TS−S(°C) ΔHS−S(kJ·mol−1)ΔHS−S(kJ·mol−1) [N1111][FSI] 78.3 23.4 0.066 5% NaFSI 78.3 11.7 0.034 78.3 12.97 0.038 10% NaFSI 78.3 10.45 0.032 75.2 9.77 0.030 15% NaFSI 78.3 9.22 0.029 75.2 8.54 0.027 25% NaFSI 76.7 5.88 0.020 71.4 2.73 0.009 Figure 7. Synchrotron powder X-ray diffraction patterns of [N1111][FSI] and with 2 and 10% NaFSI doping. The left part corresponds to Phase II and right to phase I (above 78.3 °C). JACS Au pubs.acs.org/jacsau Article https://doi.org/10.1021/jacsau.4c01086 JACS Au 2025, 5, 1663−1676 1669 inhomogeneous strain within the crystals. These findings align with the increased lattice disorder inferred from the DSC. In order to have more insight into the formation of new crystal phases, we investigated a higher NaFSI doping. According to the diffractogram, the sample displays at 40 °C multiple peaks related to a monoclinic crystal structure. After the analysis of the diffractograms, the parameters of this specific unit cell correspond to a= 8.50 Å, b= 11.17 Å, c= 5.86 Å and three angles of α= 90°, ß = 90.85°,γ= 90°, with a unit cell volume of 557.55 Å3. This indicates that at higher NaFSI doping, there is a change in the intensity of the maxima at higher angles as well as a contraction of the unit cell volume. Static 1H, 19F and 23Na NMR experiments were also carried out on [N1111][FSI] samples with 10% and 25% NaFSI (Figure 8). Similarly to LiFSI experiments, the three nuclei can provide information about the individual components of the system. The 23Na spectra show that Na+cations have a different behavior to Li+cations in the LiFSI samples, as both 10% and 25% NaFSI samples have broad and narrow components at all temperatures. A magic angle spinning (MAS) NMR spectrum was therefore collected for [N1111][FSI] doped with 25% NaFSI (Figure S18), and only one isotropic peak was observed, however the spinning sidebands have more complex line shapes which are due to satellite transitions. While satellite transitions can be affected by both the firstand second-order quadrupolar interactions, MAS experiments fully remove the first-order components, leaving only the second order interactions. These results indicate that there is only one Na chemical environment, but possibly two dynamic phases present, with one phase being immobile and the other being mobile. Therefore, the static spectra were fitted with broad and narrow components, to represent the immobile and mobile species. Example peak fittings are given in Figure S19 and the FWHM values of the peaks are given in Figure S20a. For phase II (60 °C and below), the FWHM values are approximately 11 and 2 kHz for the broad and narrow components, respectively. For phase I (80 °C and above), both components narrow to approximately 9 kHz, and less than 1 kHz, respectively. The narrow peaks then continue to narrow as the temperature is increased, indicating that the level of motion increases. The proportion of the signals corresponding to the narrow peaks are given in Figure S20b, it is notable that even up to 140 °C, the narrow peaks correspond to a minority of the signals (12% for 10% NaFSI and 28% for 25% NaFSI at 140 °C). This suggests that in phase I, the majority of the Na+ions are still immobile (like phase II), but what is mobile can move at a faster rate than the mobile species in phase II. The 19F spectra show a similar trend to the LiFSI samples, with broad CSA patterns (δof approximately 60 kHz) and narrow peaks (FWHM values close to 1 kHz), representing immobile and mobile species, respectively. Therefore, like the LiFSI data, the spectra were fitted with CSA patterns and narrow peaks, an example of this fitting is given in Figure S21. For both samples ηof the CSA patterns is constant at approximately 0.45. The FWHM values of the narrow peaks and δof the CSA patterns, as well as the proportion of the signals corresponding to the narrow peak are given in Figure S22a,b, respectively. Similarly to the LiFSI samples, the 40 °C spectra showed a nonideal CSA line shape, so they were excluded from further analysis. The broad CSA patterns dominate phase II (60 °C and below), and δdecreases slightly as the temperature is increased. The narrow peaks have more constant FWHM values close to 1 kHz, and they slowly grow in intensity as the temperature is increased, but only account for 2.4 and 3.9% of the 10% NaFSI and 25% NaFSI signals, respectively, at 60 °C. For phase I, the narrow peaks increase to 31% and 33% of the signals for 10% and 25% NaFSI, respectively, at 80 °C. Then at subsequent temperatures, only narrow signals are observed. This indicates that like the LiFSI samples, the [FSI]−anions are mostly immobile in phase II and mobile in phase I. The 1H spectra also show broad peaks (FWHM of 10 kHz) indicating that there might be multiple 1H environments present and narrow peaks grow in relative intensity as the temperature is increased. Therefore, the spectra were fitted with broad and narrow components, representing immobile and mobile environments; an example of this fitting is given in Figure S23, the FWHM values of the components and the proportion of the spectra corresponding to the narrow peak is given in Figure S24a,b, respectively. For phase II the narrow peak slowly grows in relative intensity up to 60 °C (2 and 6% for 10 and 25% NaFSI, respectively), and the broad peak narrows from 9 kHz to 8 kHz. After the phase transition, the narrow peak continues to increase in intensity and the wide peaks narrow more rapidly. Although at 140 °C the proportion of the signals corresponding to narrow peaks is 44 and 52%, for 10 and 25% NaFSI, respectively. This indicates that even in phase I, a significant proportion of [N1111]+remains immobile. Overall, the static NMR data shows a similar behavior for [N1111]+cations and [FSI]−anions when doped with NaFSI, where there is very limited mobility in phase II and increased mobility in phase I, although a significant amount of the [N1111]+cations remain immobile in phase I. For the Na+ cations, a significant proportion of the Na+cations also remain immobile in phase I, but those that are mobile can move a faster rate. Furthermore, all components show a slightly larger fraction of mobile ions when 25% NaFSI is added, compared to 10% NaFSI. Unfortunately, the fast relaxation of 23Na means PFG experiments could not be performed for the Na+cations. Therefore, PFG measurements were only carried out on for 1H and 19F for samples with 10 and 25% NaFSI, at temperatures Figure 8. Static solid-state (a) 23Na, (b) 19F and (c) 1H NMR spectra of [N1111][FSI] with 10% NaFSI, and (d) 23Na, (e) 19F and (f) 1H spectra of [N1111][FSI] with 25% NaFSI collected at temperatures ranging from −20 to 140 °C, maximum intensities normalized. JACS Au pubs.acs.org/jacsau Article https://doi.org/10.1021/jacsau.4c01086 JACS Au 2025, 5, 1663−1676 1670 ranging from 20 to 80 °C. The full diffusion coefficient data sets are given in Figure S25. Example PFG NMR spectra collected at 60 °C without the gradient field applied and attenuation curves are given in Figures S26 and S27, respectively, and the activation energies associated with the diffusion coefficients are given in Table S2. A similar trend was observed to the LiFSI systems, where 19F diffusion coefficients were higher than 1H, indicating that like the LiFSI systems, [FSI]−anions are more mobile than the [N1111]+cations. However, the diffusion coefficients decrease much more significantly for the NaFSI samples at lower temperatures, compared to the LiFSI samples, which is represented by their higher activation energies. It is worth noting that these diffusion coefficients only correspond to the small proportion of [N1111]+cations and [FSI]−anions that are mobile in the phase and the differences between phases I and II are demonstrated in Figure 9. When the diffusion coefficients are compared at specific temperatures, the values are slightly larger for 25% LiFSI. This is different to the LiFSI systems, where 10% LiFSI had the highest diffusion coefficients. The ionic conductivity of [N1111][FSI] and NaFSI mixtures was measured to determine how the ion dynamics are affected by NaFSI content. Figure 10a shows the ionic conductivity of [N1111][FSI] doped with NaFSI (5−25 mol %), from 25 to 80 °C extracted from the Nyquist plot (Figure S28). As occurred with LiFSI doping, the ionic conductivity increases 5 orders of magnitude at room temperature compared to pure [N1111][FSI]. Unlike lithium doped systems where an optimum concentration was found (e.g., 10 mol % LiFSI), in the case of Na doped samples, 25 mol % NaFSI delivered an ionic conductivity 3.2 mS·cm−1with values very similar to 15 mol % NaFSI (1.87 mS·cm−1) samples at 80 °C. When we represent lnσvs 1/T(Figure S29), a change in the slope after the solid−solid transition is clearly observed, indicating a change in the slope of the Ea. We suggest that this change in the slope can be attributed to a lattice restructuring or a change in the tendency of the ionic mobility for the appearance of new phases as suggested by DSC and XRD data. Accordingly, activation energies were obtained for [N1111][FSI] doped with NaFSI systems before the phase transition obtaining values of 74.36, 66.89, 59.57, and 49.04 kJ·mol−1for 5, 10, 15 and 25% NaFSI, respectively. These values are greater than the values for equivalent LiFSI systems, that it is in accordance with the slightly lower ionic conductivities calculated (Figure 10b). Electrochemical Measurements: Proof of Concept (LiFSI and NaFSI Doped Samples) The lithium or sodium transference numbers (tLi+ and tNa+) are important properties that show the proportion of lithium or sodium that is conducted. Ideally, an electrolyte would have a Li transference number of unity so that the total charge is only carried by the (Li or Na) ions. Since 10% LiFSI and 25% NaFSI exhibited the highest ionic conductivity values among the different compositions prepared in this study, the Li or Na transference numbers were investigated by impedance spectroscopy and chronoamperometry, using symmetrical cells (e.g., Li|Li and Na|Na) at room temperature (Figure 11). The measured tLi+ and tNa+ values of 10% LiFSI and 25% NaFSI was 0.77 and 0.41, respectively. For the 10% LiFSI sample, the high diffusion coefficients observed in PFG NMR measurements are consistent with the elevated ionic mobility prior to the solid− solid transition compared with the rest of the composition studied in this paper. This indicates that the material’s behavior as a solid-state conductor not only promotes fast lithium-ion hopping, but also it would entail a lower concentration polarization at the electrode interfaces. 59 In contrast, the 25% NaFSI sample demonstrates a lower transference number compared to the lithium-doped system, which may suggest that the structural diffusion mechanism is influenced differently under an electric field. This could be attributed to factors such as a different ion transport mechanism across the bulk material and grain boundaries, as discussed earlier. We note that these measurements were undertaken on relatively thick solid-state Figure 9. Selected 19F and 1H PFG diffusion coefficients and proportion of mobile species as determined by static NMR experiments for [N1111][FSI] doped with 10 and 25% NaFSI. Figure 10. (a) Ionic conductivity dependence of [N1111][FSI] and with NaFSI doping with temperature. (b) Comparison of the ionic conductivity values of [N1111][FSI] doped with the same molar ratio of LiFSI and NaFSI. Figure 11. Steady-state polarization curve of [N1111][FSI] doped with (a) 10% LiFSI and (b) 25% NaFSI assembled into symmetric cell at 10 mV at room temperature. The internal illustration shows the electrochemical impedance spectra before and after electrolyte polarization. JACS Au pubs.acs.org/jacsau Article https://doi.org/10.1021/jacsau.4c01086 JACS Au 2025, 5, 1663−1676 1671