Plastic Crystals Utilising Small Ammonium Cations and Sulfonylimide Anions as Electrolytes for Lithium Batteries

Organic ionic plastic crystals (OIPCs) are increasingly promising as a class of solid-state electrolyte for developing safer lithium batteries. However, their advancement relies on expanding the range of well-characterised cation/anion combinations. Here, we report the synthesis and characterization of OIPCs utilising small ammonium cations tetramethylammonium ([N 1111 ] + ), triethylmethylammonium ([N 1222 ] + ) and tetraethylammonium ([N 2222 ] + ), chosen to encourage signi ﬁ cant rotational and translational motion, with the charge-diffuse and electrochemically stable bis( ﬂ uorosulfonyl)imide ([FSI] ˉ ) and bis(tri ﬂ uoro-methanesulfonyl)imide ([NTf 2 ] ˉ ) anions. To investigate the physico-chemical properties of the OIPCs, the free volume was measured by positron annihilation spectroscopy (PALS) and correlated with the ionic conductivity and thermal analysis (DSC). Solid-state NMR analysis of the salts, is also reported. The salts with the less symmetric cation, [N 1222 ][FSI] and [N 1222 ][NTf 2 ], were identi ﬁ ed as the most promising electrolyte materials, and thus the electrochemical properties after mixing with 10 and 90 mol% lithium bis( ﬂ uorosulfonyl)imide (LiFSI) or lithium bis(tri ﬂ uoromethanesulfonyl)imide (LiNTf 2 ), respectively, were investigated. This study demonstrates the ef ﬁ cacy of these OIPC materials as new quasi-solid state electrolytes with advantageous properties such as high conductivity, good thermal and electrochemical properties, the ability to incorporate high lithium salt concentrations and support ef ﬁ cient lithium electrochemistry.

The move towards more sustainable energy such as solar or wind, to mitigate the effects of climate change and decreasing fossil fuel resources, requires the support of safer, more efficient energy storage devices. Similarly, increased use of small consumer electronic devices and electric transportation demands significant improvements in device safety-most critically, elimination of flammable and toxic liquid electrolytes. Thus, the development of solid-state electrolytes is attracting global attention.
Plastic crystals represent a unique and increasingly promising class of material that can be utilised as solid-state electrolytes. These fall into two classes-molecular plastic crystals, such as succinonitrile, that have been known since the 1960 s, 1 and the more recently discovered organic ionic plastic crystals (OIPCs). 2 Plastic crystals have a long-range crystalline lattice but short range rotational and/or translational motion of the molecules/ions, soft mechanical properties, and display one or more solid-solid phase transition before melting. The intrinsic disorder in OIPCs can result in significant ionic conductivity, support the transport of target ions such as lithium when a lithium salt is added and, thus, allows their use as solid-state electrolytes. However, while the development and application of OIPCs has advanced significantly in recent years, 3 fundamental understanding is still lacking as to how different combinations of cation and anions impact plasticity, lattice structure, the extent of disorder and/or vacancies, and the resultant conductivity. 4 Key to addressing this dearth of understanding is increasing the range of well-characterised OIPC families.
For the fundamental OIPC investigations and subsequent electrolyte development reported here, tetraalkylammonium cations with short alkyl chains were identified as particularly advantageous for OIPC formation as their high symmetry and small size is likely to promote significant rotational disorder. Thus, the tetramethylammonium (N 1111 ), methyl(triethyl)ammonium (N 1222 ) and tetraethylammonium (N 2222 ) cations were chosen for investigation. Combining these ammonium cations with anions that exhibit conformational changes-i.e. cis/trans-was pursued in order to engineer a combination of disordering mechanisms anticipated to result in highly disordered and conductive materials. Further, the asymmetry and larger size of the [FSI] and [NTf 2 ] anions was predicted to hinder the packing and thus reduce the lattice energy (and thus melting point) compared to analogous ammonium [BF 4 ] or [PF 6 ] salts. 17 This is advantageous for achieving the optimum disorder in phase I within the temperature range most relevant for practical applications.
A range of ammonium salts with small alkyl chains have been previously synthesized. However, aside from the detailed study by Henderson et al. 18 12 Further, to enable the use of OIPCs as electrolytes for lithium or sodium batteries, they must be combined with salts of the target ion (i.e. Li + or Na + ). The addition of alkali metal salts can significantly increase the ion mobility in the electrolyte, as demonstrated by Henderson et al., 18 where the addition of 1 to 70 mol% lithium bis (trifluoromethanesulfonyl)imide (LiNTf 2 ) into [N 2222 ][NTf 2 ] resulted in a several fold increase in the ionic conductivity compared to the neat OIPC. Moreover, the utilization of high lithium salt concentration has been associated with reduced dendrite formation, higher lithium transference numbers and reduced degradation of sulfur cathodes in lithium-sulfur cells. [26][27][28] The electrochemical performance of N,N-diethyl-N-methyl-N-(npropyl)ammonium trifluoromethyltrifluoroborate ([N 1223 ][CF 3 BF 3 ]) has been reported for lithium metal battery applications. 9 9 The ability of the electrolyte to stabilise and cycle lithium metal was demonstrated by symmetrical Li | Li cell cycling for ∼800 h at 0.2 mA cm −2 current density. 9 An alternative polymer electrolyte was developed by mixing [N 1222 ][FSI] and poly(diallyldimethylammonium) bis(trifluoromethanesulfonyl) imide (PDADMA NTf 2 )/LiNTf 2 , which achieved good mechanical properties, high conductivity (2 × 10 −4 S cm −1 at 30°C) and good electrochemical stability. 10 However, analysis of the electrolyte mixture of [N 1222 ][FSI] and lithium bis(fluorosulfonyl)imide (LiFSI) without polymer was not reported. Moreover, in the body of recent studies, none has tried to delineate the key factors that determine electrolyte performance by systematically making stepwise variations in cation-anion composition.
To address this situation, and in the process pursue a deeper understanding of this promising family of OIPCs, we herein report the synthesis of six different ammonium salts ( Fig. 1) [FSI]. PALS analysis has been performed on these six salts for the first time, to determine the size and concentration of vacancies and compare these to the free volume within other OIPCs.
Solid-state NMR was used to investigate the effect of anion and cation alkyl chain length on the ion dynamics. Finally, the efficacy of these materials as quasi-solid-state electrolytes has been demonstrated by addition of lithium salts to [N 1222 ] [FSI] and [N 1222 ][NTf 2 ], utilising a common anion, to make electrolytes with either 10 mol% lithium salt as "lithium salt in OIPC" or 90 mol% lithium salt as "OIPC in lithium salt." The thermal, transport and electrochemical properties of these mixtures demonstrate their promise as electrolytes for next generation solid-state lithium batteries. Ion selective electrodes (ISE).-The residual chloride and bromide content were determined by means of an Ionode IJ-Cl or IJ-Br ion-selective electrode, after calibration with 10 and 100 ppm solutions of the chloride (potassium chloride) or bromide (sodium bromide) solutions. Potassium nitrate was used as an ion strength adjuster (ISA) and was added to both standards and samples. A known dried amount of OIPC (50-100 mg) was added to known amount of distilled water and the same ratio of ISA was used in the calibration step. Inductively coupled plasma-mass spectroscopy (ICP-MS).-Quantification of lithium and potassium was carried out using an inductively coupled plasma-mass spectrometer (ICP-MS; NexION 350X, PerkinElmer, USA). The internal standards Sc (200 ppb) and Rh (20 ppb) in 1% aqua regia were used for correction of matrix effects. The internal standard solution was mixed prior to the nebulizer using a T-piece in a 1 : 1 ratio. Calibration standards for lithium and potassium (Perkin Elmer, Lithium or Potassium standard 1000 ppm in 2% HNO 3 ) were prepared at 0.1, 1, 10, 50, 100 and 500 ppb with 2% suprapur nitric acid (HNO 3 ) in each. The mass spectrometer was operated in kinetic energy discrimination mode (KED) with 50 ms dwell times, 20 sweeps, one reading and three replicates. The plasma source conditions were: nebulizer gas flow 1.02 l min −1 , auxiliary gas flow 1.2 l min −1 , plasma gas flow 15 l min −1 , ICP RF power 1500 W. Data analysis was carried out using Syngistix (PerkinElmer) software. Signal responses were normalized to the scandium internal standard.
Thermal gravimetric analysis (TGA).-Thermogravimetric analysis (TGA) was performed on a Mettler Toledo TGA/DSC 1 STARe System over a range of 25°C to 550°C under N 2 (30 ml min −1 ), at a heating rate of 10°C min −1 . The samples (5 to 10 mg) were held in aluminium pans. Heating rate was 10°C min −1 .
Differential scanning calorimetry (DSC).-Before performing DSC, the melting points were determined visually using a Gallenkamp melting point apparatus. Before measurement the heat flow and temperature for DSC was calibrated using cyclohexane. DSC was measured on a Mettler Toledo DSC STARe with a scan rate of 10°C min −1 for both the heating and cooling cycles (except for 10  ], the conductivity was measured from 30°C to 90°C after pressing the OIPC into a pellet. The conductivities of the samples were determined from the real axis intercept in the Nyquist plot of the impedance data. NMR.-Static and magic angle spinning (MAS) NMR was measured on a Bruker Avance III 500 MHz wide-bore NMR spectrometer equipped with a 4 mm H/F-X double resonance MAS probe. The samples were packed in 4 mm NMR MAS rotors in argon glove box and sealed with Vespel caps. The static and MAS spectra were recorded with 4 μs RF pulse length, 10 s recycle delay and 10 kHz spinning rate (MAS spectra). Spectra were recorded from 20°C to 70°C.
Positron annihilation lifetime spectroscopy.-Positron Annihilation Lifetime spectroscopy (PALS) was performed on EG&G Ortec spectrometers using nitrogen airflow in the sample cell during analysis. The 3.5 MBq positron source ( 22 NaCl) was heat sealed in a Mylar envelope and placed in the middle of two OIPC pellets (1 mm thickness each) and placed between the spectrometers. A minimum of 1 × 10 6 integrated counts was collected for each file with five or more files collected for each sample. The spectra was fitted using LT-v9 software, 30 using a 3 component fit and a source correction (1.501 ns and 3.48%). The first component was attributed to para-positronium (p-Ps) annihilation with the bound state of the positron and electron with opposite spins and fixed to 0.125 ns. The second component (∼0.36 ns) was due to the annihilation of the positrons with free electrons within the sample. The final component was due to ortho-positronium (o-Ps) annihilation where the positron forms a bound state with an electron of the same spin. This longer lifetime for τ 3 was used to calculate the average pore radius (R) within the OIPC sample using the Tao-Eldrup quantum based empirical equation 31,32 where R 0 is the thickness of the electron layer within the potential well; Electrochemical measurements.-A Biologic SP-200 potentiostat was used to perform the electrochemical plating and stripping of lithium. Cyclic voltammetry was performed in a three-electrode cell at 50°C in an Argon filled glove box. The cell contained a platinum disk working electrode with surface area of 2.0 mm 2 , a lithium metal coil and a lithium metal strip as counter and reference electrodes respectively. Lithium metal was brushed with cyclohexane prior to use. 20 mV s −1 scan rate was used for all measurements. -Tetramethylammonium bromide (6.0767g, 39 mmol) and potassium bis(fluorosulfonyl)imide (8.735 g, 40 mmol) were dissolved separately in 50 ml of water each to get clear solutions. Upon mixing the above solutions, a white precipitate formed instantly, and the solution was left to stir for an hour at room temperature. The solution was filtered, and the filtrate was washed with water (7 × 50 ml). The white solid was dissolved in hot water (50 ml) and cooled down slowly at room temperature to get colourless needle-like crystals. These crystals were filtered and dried in vacuo at 50°C for 72 h to get [ ].-Triethylmethylammonium chloride (5.060g, 33 mmol) and lithium bis(trifluoromethanesulfonyl)imide (9.6663 g, 34 mmol) were dissolved separately in 50 ml of water each to get clear solutions. Upon mixing the above solutions, a white precipitate formed instantly, and the solution was left to stir for one hour at room temperature after addition of CH 2 Cl 2 (100 ml). The organic layer was washed with water (7 × 50 ml). The organic layer was separated and dried in vacuo for 72 h at 50°C to get [N 1222 ][NTf 2 ] (8 g, 61%). This sample was crystallized from hot water to get colourless single crystals. 1  ].-Tetraethylammonium bromide (5.1337 g, 24 mmol) and lithium bis(trifluoromethanesulfonyl)imide (7.0692 g, 25 mmol) were dissolved separately in 50 ml of water each to get clear solutions. Upon mixing the above solutions, a white precipitate formed instantly, and the solution was left to stir for an hour at room temperature after the addition of CH 2 Cl 2 (100 ml). The organic layer was washed with water (7 × 50 ml). The organic layer was separated, removed in vacuo and dried for 72 h to get a white solid of [N 2222 ][NTf 2 ] (8 g, 78%). The crystals of this sample were obtained by cooling down a hot solution in water. 1

Results and Discussion
Thermal properties.-The ammonium salts were synthesized by anion metathesis between commercially available ammonium halides and lithium bis(trifluoromethanesulfonyl)imide (LiNTf 2 ) or potassium bis(fluorosulfonyl)imide (KFSI). The synthesis of all six ammonium salts are very amenable to scale-up as they are synthesised in one step from commercially available, relatively low-cost starting materials, which is a significant benefit to their widespread utility. The purity of the salts was confirmed by 1 H, 13 C, 19 F NMR, mass spectroscopy, residual halide content by ion selective electrode, residual Li + or K + content by ICP-MS and microanalysis. The onset decomposition temperature (T d(onset) ) was determined by rising temperature TGA (Fig. 2). This technique can result in an overestimate of long-term thermal stability compared to an isothermal measurement 33  [FSI]) (T d(onset) = 220°C)) at a similar heating rate (10°C min −1 ). 6 The NTf 2 salts have onset decomposition temperatures approximately 100°C higher than the corresponding FSI salts. This is consistent with previously reported thermal decomposition behaviour of other FSI and NTf 2 salts. 34 The impact of the nature of the anion on the decomposition temperature is also much more significant than the effect of changing the length of the alkyl chain on the cation.   6 In-depth characterisation by, for example, TGA-MS would be required to elucidate fully the decomposition mechanisms. Nevertheless, and despite the lower onset decomposition temperature for FSI salts, the thermal stability of all of the reported ammonium salts is high enough for most electrochemical applications.
Of the six ammonium salts synthesized, only five display one or more solid-solid phase transitions in the temperature range from −120°C to melting (Fig. 3, Table I), which is a key indicator of plasticity. By convention, the highest temperature solid phase is termed phase I, while the lower temperature phases are termed phase II, III and so on. There is a clear dependence of both cation and anion on thermal behaviour. Generally, FSI-based OIPCs display more disorder than those with NTf 2 anions, 3,6,16 as evidenced by smaller entropies of fusion, and this is also observed in the current study.  6 cations. However, here the opposite behaviour is displayed in salts with the larger ammonium cation ([N 2222 ] + ), highlighting the fact that the nature of both the cation and anion can be important in determining the crystal packing and lattice energy.
[N 1111 ][FSI] displays one solid-solid phase transition before melting and the onset of decomposition above 300°C ( Fig. 1 and  Fig. S1). This high melting point was confirmed visually, and is consistent with the previously reported solid-solid phase transition at 71°C and a melting at 294°C, 21 (the previously reported 20 T m of 70°C is, we suggest, a solid-solid transition ). The entropy of melt agrees with Timmerman's suggested criterion for plasticity of ΔS f ⩽ 20 J K −1 mol −1 , 1 indicating significant disorder in phase I.
[N 1222 ][NTf 2 ] melts at 98°C with an entropy of fusion of 25 J K −1 mol −1 . This entropy slightly exceeds Timmerman's criterion for plasticity, but the material nonetheless has a very advantageous temperature range for phase I, starting from −4°C and spanning the temperature range of relevance for most electrochemical device applications.
[  3 However, as noted above, it is more common for [FSI] − salts to have higher melting points than their NTf 2 analogues, as seen in [C 2 mpyr] + and [C 2 epyr] + and the other ammonium salts investigated here. 35 Transport properties.-The ionic conductivity of OIPCs is one of the most important parameters that determine their suitability for electrochemical applications, while a wide phase I temperature range is also very valuable ( Fig. 4 and Table II). Overall, the FSI-based ammonium salts display higher conductivity in phase I than the NTf 2 -based ammonium salts. Of the different ammonium-based OIPCs studied here, [N 1222 ][FSI] displays the highest conductivity of 9.9 × 10 −6 S cm −1 at 30°C (consistent with the previous report), 9 indicating an advantageous increase in disorder as a result of use of the less symmetrical cation.  16 (Fig. 4b) are present in phase II at 30°C and an increase in conductivity is seen as they transition from phase II to phase I with heating. The thermal properties of the salts, determined by DSC (Fig. 3), suggest that [N 2222 ][FSI] is the most disordered of the materials in phase I-reflected in the lowest entropy of fusionand the material has the highest conductivity (5 × 10 −5 S cm −1 at 50°C) of all the salts studied here. [N 2222 ][FSI] is also the most conductive of all the salts at 60°C, when all five ammonium OIPCs are in phase I. However, the narrow temperature domain for phase I is a disadvantage. The conductivity measured for [N 2222 ][NTf 2 ] here is lower than in the prior report, 18 which is attributed to differences in experimental setup and material preparation. The lowest conductivity is displayed by [N 1111 ][NTf 2 ] because it is a crystalline solid with no solid-solid phase transitions and no evidence of plasticity.
Positron annihilation lifetime spectroscopy (PALS) of the neat OIPCs.-The six ammonium salts were characterized by PALS analysis at room temperature, to study the impact of different anions and the cation chain length. This technique allows investigation of the defects / free volume size (τ 3 ) and relative abundance of the vacancies in a material. The investigation of vacancy concentration in OIPCs is highly valuable as they are intrinsically involved in the mechanism of ionic conductivity. 36 The measured positronium lifetimes (which are correlated to defect size), intensities (which give a measure of relative defect concentration), the calculated pore size and volume for the six ammonium salts at room temperature are shown in Table II.
An important point to note when comparing the salts is the influence of the thermal phase-at room temperature, the [N 1222 ] salts are in phase I, whereas the other salts are in phase II or are crystalline solids. The phase I of OIPCs is normally the most conductive and disordered, but when comparing different OIPCs, the size and shape of the ions will also impact defect number and volume.    Fig. S1.
Overall, larger PALS volumes are consistently evident for NTf 2 -based salts than the FSI ones, which reflects the influence of the larger anion on the crystal packing. There appears to be a general correlation between the fractional free volume (FFV = Average Pore Volume × * Intensity) and conductivity across the NTf 2 salt series (Fig. 5a) 3 had comparable free volumes but conductivities which varied by orders of magnitude (note the log scale in Fig. 5b). It therefore appears that the free volume is not the controlling variable for conductivity of the FSI salts.
Solid-state NMR analysis of the neat plastic crystals.-Solid state NMR can be highly insightful for elucidating the chemical environment and molecular motions of the materials at a molecular level. While NMR spectra of solutions normally have very sharp peaks, and a static NMR spectra of a solid powder sample would be relatively broad, OIPCs are often sufficiently disordered to give either relatively narrow lines or a combination of broad and narrow lines that reflect the presence of both mobile and less mobile components. 38 Thus, in static NMR spectra, the linewidths may provide information about the relative mobility of the nuclei; in this case, the cation and anion by 1 H and 19 F respectively. Further, the spectra can indicate the presence of residual dipole-dipole couplings and/or chemical shift anisotropy (CSA), which relates to the orientation dependent electron shielding environment of the nuclei. 39 For example, for the NTf 2 anion the 19 F chemical shift depends on how the C-F bond is oriented with respect to the magnetic field. As the OIPCs are powders rather than single crystals, the crystallites will be orientated in many different directions to the magnetic field, thus the orientation of the 19 F shielding tensor will not be the same, and a sum of all possible orientations in space would result in an asymmetric peak -the so-called chemical shift anisotropy (CSA) powder pattern.
Here, solid-state static and magic angle spinning (MAS) NMR was used to investigate the structure and dynamics within the four most conductive salts ( 19 F CSA was observed but with significant distortion in the broad component of the spectra. This distortion is often observed for broad CSA patterns and is caused by the extremely fast T 2 relaxation, and as a consequence the loss of a few initial data points in the time domain. Therefore, in this study, only MAS NMR data is shown for 19 F NMR spectra. For the [N 1222 ][NTf 2 ] (top spectra, green), the static 19 F spectrum (Figs. 6e and 6f) is broad but almost symmetrical, indicating very little chemical shift anisotropy (CSA). This is relatively unusual for NTf 2 salts and suggests tumbling of the anion, but not significant translational motion as the conductivity is relatively low. The static 1 H NMR (Figs. 6b and 6c) (Fig. 2b, pink) (Fig. 6). The spectra recorded with increasing temperature from 30°C to 60°C are shown in Fig. S5.
The [N 1222 ][FSI] (purple spectra) is a much more disordered OIPC and thus no CSA is observed. This OIPC gives significantly narrower 1 H (Fig. 6b) and 19 F (Fig. 6e) static NMR line widths than the other OIPCs studied, suggesting faster dynamics of the cation and anion. In particular, the 19 F shows an extremely narrow (liquidlike) line, suggesting fast isotropic rotation/reorientation motion of the FSI anion. This almost liquid-like behaviour is consistent with the high conductivity at room temperature.
Finally, the [N 1111 ][FSI] (blue spectra) is the least conductive at room temperature. Chemical shift anisotropy is evident in the MAS NMR (Fig. 6d), which shows an asymmetric signal even with spinning. This suggests that the electron shielding of the 19   were selected for further investigation as electrolytes for lithium batteries. This choice was based on the high conductivity of these two salts and a strategy of investigating the influence of the anion while keeping the cation constant. To develop these salts as electrolytes, they were mixed with LiFSI and LiNTf 2 respectively, thereby introducing the target ion (Li + ) while keeping the anion consistent.
Thermal analysis of the neat salts and the 10 and 90 mol% lithium salt compositions are shown in Fig. 7 and Table SI. Addition of 10 mol% LiFSI to [N 1222 ][FSI] forms a composition described as "lithium salt in OIPC." Compared to the neat OIPC (T m = 133°C), the melting transition of this electrolyte is broader and decreased, and a very small and broad peak appears around 66°C (more visible in Figs. S2 and S3). A broadening and melting point decrease is a common observation upon introduction of a second component to an OIPC. 18 Visually, this material appears as a quasi-solid state electrolyte that does not flow under gravity, but with the presence of a small amount of liquid phase. Thus, as a result of the  19 F MAS NMR at room temperature at 8 kHz, (e) 19 F static NMR at 20°C and (f) 19 F static NMR at 70°C. heterogeneous nature, the melting peak is very broad. However, the electrolyte mixture still displays three solid-solid transitions consistent with the presence of a plastic phase.
Addition of 90 mol% LiFSI to [N 1222 ] [FSI], to form a composition described as an "OIPC in lithium salt," produces a high lithium concentration quasi-solid state electrolyte as a soft solid. This mixture displays a glass transition (−76°C), a solid-solid transition (−53°C) and a broad melting (102°C). The solid-solid transition in the 90 mol% LiFSI mixture indicates the presence of a LiFSI solid phase in the electrolyte, as neat LiFSI also undergoes a solid-solid phase transition at this temperature, before melting at 140°C (Fig. 7a).
Similar electrolyte compositions were investigated with the NTf 2 OIPC and Li salt analogues, to assess the influence of the anion (Fig. 7b and Table SI). The NTf 2 anion has been shown to have higher electrochemical and thermal stability compared to FSI and other anions. [40][41][42] In addition, NTf 2 is known to form a stable SEI layer in contact with lithium metal, 43,44 and therefore there are significant possible benefits of a high LiNTf 2 salt content electrolyte for prolonged battery cycling. 40 (Fig. 7a). However, the appearance of new phases were observed in this system, evidenced by a new crystallization and solid-solid phase transition (25°C). When this sample was held for an hour at −120°C (Fig. S4) and heated at 2°C min −1 , this crystallization and T s-s did not appear during the first heating cycle, suggesting that the sample had fully crystallized during the isothermal sequence. However, during the second and third heating run, this solid-solid transition re-appeared even when the sample was kept at −120°C for an hour. Thus, there is evidence of meta-stability in this particular system, with an impact of thermal history on the phase that is present at around room temperature.
The 90 mol% LiNTf 2 in [N 1222 ][NTf 2 ] (Fig. 7b) has a higher glass transition temperature (T g ) than the 90 mol% FSI-based electrolyte (Fig. 7a). The broad peak at around 149°C in the NTf 2 system (Fig. 7b) is attributed to a solid-solid phase transition, consistent with the presence of a LiNTf 2 rich phase, and the electrolyte melts at around 200°C (confirmed visually).
Ionic conductivity.-The ionic conductivities of the four quasisolid state electrolytes with lithium salt are shown in Fig. 7c , which is consistent with the presence of a small amount of liquid phase. It is important to note, however, that the overall physical nature of these materials are quasi-solid state and thus they will significantly reduce electrolyte leakage from a device compared to a liquid electrolyte.
Electrochemistry.-For further assessment of the applicability of the new materials for device applications, cyclic voltammetry (CV) was used to investigate the electrochemical stability of the 10 mol% and 90 mol% LiNTf 2 in [N 1222 ][NTf 2 ] electrolytes as well as the lithium plating/stripping kinetics. Additionally, a preliminary quantification of cycling efficiency was obtained by estimation of the Coulombic efficiency (CE), the percentage of lithium stripped to lithium plated during each cycle. The NTf 2 systems were chosen for electrochemical studies as this anion has higher thermal stability than the analogous FSI systems, as shown in the TGA plots in Fig. 2. Use of the NTf 2 anion is also believed to be advantageous for long device life. The CV of 10 mol% NTf 2 electrolyte (Fig. 8a) shows distinct lithium plating peak at −0.54 V (vs Li/Li + ) and stripping peak at 0.95 V with stable current density over six cycles. The second stripping peak at 1.5 V is attributed to lithium alloying with platinum. 45 The Coulombic efficiency (CE) of the first cycle is a moderate 75%, most likely as a result of solid electrolyte interface (SEI) formation, which typically forms due to degradation of the electrolyte upon first contact with lithium metal and during the first cycle. 46 However the CE dramatically increases and stabilises at ∼90% after the third cycle, indicating the formation of a stable SEI.
Similarly, the 90 mol% LiNTf 2 (Fig. 8b) electrolyte shows good cycling behaviour with stable current density. This achievement of consistent plating currents at potentials higher than -0.5 V is commonly seen in highly concentrated systems; the large amounts of lithium readily available at the electrode surface mitigates mass transport limitations. 28,47 The lithium stripping peak occurs at 1.5 V, which is a higher peak potential than in the 10 mol% system. This is indicative of higher resistance in the 90 mol% electrolyte, consistent with lower ionic conductivity compared to the 10 mol% system (Fig. 7c). Despite this, the highly concentrated 90 mol% system could prove beneficial for lithium cycling due to the large availability of lithium in the vicinity of the electrodes. Furthermore, the Coulombic efficiency is consistently high at ∼90% after the fourth cycle. Given that CE values from cyclic voltammetry typically underestimate significantly the values obtained with more accurate methods, these two electrolytes are promising candidates for battery applications. Future work will investigate symmetrical and full cell cycling of these electrolytes.

Conclusions
In summary, we report here the synthesis of six salts utilising small ammonium-based cations with sulfonylimide anions. were mixed with lithium FSI or NTf 2 salts, respectively, in two different ratios (10 and 90 mol%) to get quasi-solid state electrolytes. The highest ionic conductivity (2.6 × 10 −3 S cm −1 at 30°C) was obtained for 10 mol% LiFSI in [N 1222 ][FSI]. The cyclic voltammetry was more promising for the 10 mol% LiNTf 2 in [N 1222 ][NTf 2 ] than for the 90 mol% analogue, but the latter achieved higher Coulombic efficiency as a result of the higher availability of Li + . Thus, this study demonstrates the impact of cation and anion chemistry, size and symmetry on the thermal, physical and transport properties, and provides an initial demonstration of their promise as quasi-solid state electrolytes ahead of full electrochemical device development and testing.