Ionic Liquid-Based Electrolyte Membranes for Medium-High Temperature Lithium Polymer Batteries

Li+-conducting polyethylene oxide-based membranes incorporating N-butyl-N-methylpyrrolidinium bis(trifluoromethanesulfonyl)imide are used as electrolyte separators for all-solid-state lithium polymer batteries operating at medium-high temperatures. The incorporation of the ionic liquid remarkably improves the thermal, ion-transport and interfacial properties of the polymer electrolyte, which, in combination with the wide electrochemical stability even at medium-high temperatures, allows high current rates without any appreciable lithium anode degradation. Battery tests carried out at 80 °C have shown excellent cycling performance and capacity retention, even at high rates, which are never tackled by ionic liquid-free polymer electrolytes. No dendrite growth onto the lithium metal anode was observed.


Introduction
Large-scale applications such as automotive, stationary, deep-sea drilling devices need batteries to be capable of operating safely at medium-high temperatures with very good performance and cycle life; i.e., without appreciable degradation phenomena. In addition, even devices generally operating around room temperature could be accidentally subjected to prolonged overheating, thus requiring high thermal stability. In this scenario, electrolytes play a key role.
Rechargeable lithium batteries are an excellent choice as advanced electrochemical energy storage systems due to their high energy density and cycle life [1,2]. Recently published manuscripts report that ion conducting polymer membranes, realized through common materials and up-scalable processes, can act as electrolyte separators for rechargeable lithium battery systems. J.R. Nair et al. [3] have prepared methacrylic-based PEs, reinforced with both cellulose hand-sheets and nanosize cellulose fibers, by UV-induced free radical photo-polymerization. Similarly, rigid-flexible composite electrolyte membranes, based on poly(ethyl α-cyanoacrylate) and cellulose backbone, have been prepared through an in-situ polymerization process by P. Hu et al. [4]. These cross-linking techniques, also successfully proposed for Na + conducting PEs [5], have shown short processing times, easy up-scalability and eco-compatibility, and have enabled gel polymer electrolytes (GPEs) with wide electrochemical stability windows and high room temperature ionic conductivity in combination with good mechanical

Synthesis of the Ionic Liquid
The PYR 14 TFSI ionic liquid was synthesized through an eco-friend route, reported in detail elsewhere [35,36].

Preparation of the Polymer Electrolyte and the Composite Cathode
The ionic liquid-based polymer electrolyte and composite cathode were prepared through a solvent-free process [33] carried out in a very low relative humidity dry-room (R.H. < 0.1% at 20 • C). The material components, i.e., PEO (Dow Chemical, Midland, MI, USA, WSR 301, M W = 4,000,000 a.u.), lithium bis(trifluoromethanesulfonyl)imide (LiTFSI, 3M, battery grade) and PYR 14 TFSI, were vacuum dried at 50 • C for 48 h (PEO) and at 120 • C for 24 h (lithium salt and ionic liquid). PEO and LiTFSI (EO:Li mole ratio = 1:0.1) were intimately mixed in a mortar, and then PYR 14 TFSI was added to achieve a (PYR 14 ) + /Li + mole ratio equal to 1:1. In previous papers [24,33], we have shown that this ratio represents a good compromise between ion transport properties and interfacial stability. The P(EO) 1 (LiTFSI) 0.1 (PYR 14 TFSI) 0.1 past-like electrolyte blend was annealed under vacuum at 100 • C overnight in order to allow the full diffusion of the lithium salt and ionic liquid through the PEO host, therefore obtaining a homogeneous mixture. Finally, the so-obtained rubber-like material was hot-pressed at 100 • C for 2 min to form 70-80 µm thick films. Ionic liquid-free, P(EO) 1 (LiTFSI) 0.1 binary polymer electrolytes were prepared for comparison purposes.
The cathode tape was prepared by intimately blending LiFePO 4 active material (Sud Chemie, Munich, Germany) and KJB carbon (electronic conductor, Akzo Nobel, Amsterdam, The Netherlands). LiFePO 4 and KJB were previously vacuum dried at 120 • C for at least 24 h. Separately, PEO, LiTFSI and PYR 14 TFSI were roughly mixed (to obtain a paste-like mixture) and then added to the LiFePO 4 -KJB blend. The resulting cathodic mixture was firstly annealed at 100 • C overnight and then hot-pressed to form preliminary films (200-300 µm thick) which were cold-rolled to obtain the final cathode tape (<50 µm) and to remove any porosity within the composite cathode [37]. Finally, 12 mm diameter cathode discs (active area equal to 1.13 cm 2 ) were punched for the battery tests. The active material mass loading ranged from 4 to 5 mg·cm −2 , corresponding (accounting for a theoretical capacity of LiFePO 4 equal to 170 mA·h·g −1 ) to a capacity from 0.7 to 0.8 mA·h·cm −2 .

Thermal Analysis
DSC measurements were run using a differential scanning calorimeter (TA Instruments, model Q100, New Castle, DE, USA). The samples, upon housing (within the dry room) in sealed Al pans, were cooled (10 • C·min −1 ) from room temperature down to −140 • C and then heated (10 • C·min −1 ) up to 150 • C.
The thermal stability was verified in a nitrogen atmosphere through TG analysis carried out by a SDT 2960 equipment, simultaneous TG-DTA (TA Instruments, New Castle, DE, USA) with Thermal Solution Software (version 1.4, Thermal Solutions Inc, Ann Arbor, MI, USA). During the experiments, the atmosphere above the samples was fixed by flowing high purity nitrogen atmosphere at a flow rate of 100 mL·min −1 . The experiments were performed on 5-10 mg samples (handled in the dry room), which were housed in platinum crucibles. The thermal stability was initially investigated by running a heating scan from room temperature up to 500 • C at a scan rate of 10 • C·min −1 .

Cell Assembly
The electrochemical measurements on the polymer electrolyte samples were carried out on two-electrode cells fabricated in the dry room. Two different cell types (active area equal from 2 to 3 cm 2 ) were assembled by sandwiching a polymer electrolyte separator between (i) two Li foil electrodes (50 µm thick, supported onto Cu grids as the current collectors) for determining, respectively, the resistance at the interface with the lithium anode and the limiting diffusion current density; (ii) a nickel foil (working electrode, 100 µm thick, used also as the current collector) and a lithium foil (counter electrode, 50 µm thick, supported onto a Cu grid as the current collector) for the linear sweep voltammetry tests. In the latter kind of cell, a tiny lithium strip (50 µm thick, supported onto a Ni grid as the current collector) was used as the reference electrode.
The electronic conductivity of the ionic liquid-containing LiFePO 4 composite cathode was investigated as a function of the carbon content by carrying out impedance measurements on symmetrical Al/cathode/Al cells. The composite cathode tape was interlayered between two Al foils (20 µm thick), which were also used as the current collectors.
The solid-state Li/LiFePO 4 batteries (cathode limited) were fabricated (inside the dry room) by laminating a lithium foil (50 µm thick), a P(EO) 10  grids were used as the cathodic and anodic current collector, respectively. The electrochemically active area of the Li/LiFePO 4 cells was 1.13 cm 2 .
All assembled cells were housed in soft envelopes, evacuated for at least 1 h (10 −2 mbar) and then vacuum-sealed. Finally, the cells were laminated twice by hot-rolling at 100 • C to improve the electrolyte/electrode interfacial contact.

Electrochemical Tests
Impedance measurements were performed on symmetrical Li/polymer electrolyte/Li (frequency range: 65 kHz-10 mHz; temperature range: 20-80 • C) and Al/composite cathode/Al (10 kHz-1 Hz, 20 • C) cells by a Frequency Response Analyzer, F.R.A. (Schlumberger Solartron, mod. 1260, Leicester, UK). The analysis of the AC responses was carried out by an equivalent circuit model taking into account all possible contributes to the impedance of the cell under test [38]. The validity of the selected circuit was confirmed by fitting the AC responses using a non-linear least-square (NLLSQ) software developed by Boukamp [39,40] (only fits characterized by a χ 2 factor lower than 10 −4 were considerable acceptable [39,40]).
The electrochemical stability window (ESW) of the P(EO) 1 (LiTFSI) 0.1 (PYR 14 TFSI) 0.1 polymer electrolyte was evaluated by linear sweep voltammetries (LSVs) run at 0.5 mV·s −1 in the 20-80 • C temperature range. The measurements were performed by scanning the cell potential from the open circuit value (OCV) towards more negative or positive potentials to determine the cathodic and anodic electrochemical stability limits, respectively. The LSVs were performed at least twice on each electrolyte to confirm the results obtained, using fresh samples and clean electrodes for each test. The measurements were performed at 20 • C using an Electrochemical Interface (Schlumberger Solartron, mod. 1287, Leicester, UK).
The limiting diffusion current density of the P(EO) 1 (LiTFSI) 0.1 and P(EO) 1 (LiTFSI) 0.1 (PYR 14 TFSI) 0.1 polymer electrolytes was determined by potentiodynamic measurements on symmetrical Li/electrolyte/Li cells, i.e., the cell voltage was linearly increased from the OCV value (a few mV) at a scan rate of 0.01 mV·s −1 until the current response achieves a steady state. The measurements were performed at temperatures ranging from 40 to 80 • C by a potentiostat/galvanostat (MACCOR, mod. 4000, Tulsa, OK, USA).
The cycling performance of the Li/LiFePO 4 polymer cells was evaluated under charge/discharge rates ranging from 0.1C (j = 0.07-0.08 mA·cm −2 ) to 1C (j = 0.7-0.8 mA·cm −2 ) at 80 • C. The battery tests were performed using a multiple battery tester (MACCOR, mod. S4000, Tulsa, OK, USA). The voltage cut-offs were fixed at 4.0 V (charge step) and 2.0 V (discharge step), respectively. During the experiments, the cells were held in a climatic chamber (Binder GmbH, mod. MK53, Tuttlingen, Germany) with a temperature control of ±0.1 • C.

Ionic Liquid-Based Polymer Electrolytes
The solvent-free procedure allowed homogeneous, freestanding, polymer electrolyte membranes with good mechanical properties to be obtained. In addition, the ionic liquid-containing P(EO) 1 (LiTFSI) 0.1 (PYR 14 TFSI) 0.1 sample looks rather sticky, thus resulting (even if not easily handled) in improved contact at the interface with electrodes.
The results of the DSC investigation are illustrated in Figure 1a. The P(EO) 1 (LiTFSI) 0.1 electrolyte shows a broad endothermic melting peak centered around 60 • C [21,41] and a weak glass transition (T g ) feature located at −39 • C. The pure PYR 14 TFSI ionic liquid, reported for comparison purposes, exhibits only a melting peak around −7 • C [42]; i.e., the absence of glass transition and exothermal "cold crystallization" features suggest that the IL sample was fully crystallized prior to running the DSC measurements [43]. The incorporation of PYR 14 TFSI into the P(EO) 1 (LiTFSI) 0.1 electrolyte results in almost complete disappearance of the melting peak in the DSC trace, which displays only the T g feature around −55 • C, clearly indicating that the P(EO) 1 (LiTFSI) 0.1 (PYR 14 TFSI) 0.1 electrolyte is amorphous even at room temperature.
The thermal stability is a mandatory requirement for electrolytes to be addressed to battery systems for medium-high temperature applications. Figure 1b compares the TGA trace (in nitrogen atmosphere) of the P(EO) 1 (LiTFSI) 0.1 and P(EO) 1 (LiTFSI) 0.1 (PYR 14 TFSI) 0.1 electrolyte membranes. The IL-free sample exhibits a weight loss above 180 • C, whereas the addition of the ionic liquid component results in thermal stability increase up to 220 • C. It should be noted that PYR 14 TFSI is seen to be thermally stable up 290 • C. Therefore, we can reasonably hypothesize that the ionic liquid, properly incorporated within the polymer host, is able to protect the PEO chains by thermal degradation. Something similar was previously observed in other PEO electrolytes [41], in which the IL agent, suitably dispersed through the polymeric matrix, was seen to prevent the oxidation of the polymer host above 4 V (vs. Li + /Li • ). The IL-free sample exhibits a weight loss above 180 °C, whereas the addition of the ionic liquid component results in thermal stability increase up to 220 °C. It should be noted that PYR14TFSI is seen to be thermally stable up 290 °C. Therefore, we can reasonably hypothesize that the ionic liquid, properly incorporated within the polymer host, is able to protect the PEO chains by thermal degradation. Something similar was previously observed in other PEO electrolytes [41], in which the IL agent, suitably dispersed through the polymeric matrix, was seen to prevent the oxidation of the polymer host above 4 V (vs. Li + /Li°).  The effect of the incorporation of the PYR14TFSI ionic liquid on the ion transport properties of the polymer electrolyte is summarized in Table 1. A remarkable conductivity increase is observed, especially at ambient temperature and below. For instance, the P(EO)1(LiTFSI)0.1(PYR14TFSI)0.1 sample shows ion conduction values three and two orders of magnitude higher than that of the IL-free sample at −20 °C and 20 °C [31,33], respectively. More than 10 −4 S·cm −1 are exhibited at 20 °C, this is of interest for applications in practical devices and commonly not approached in polymer electrolyte membranes. These results support faster ion conduction through the PEO electrolyte due both to a much larger content of the amorphous phase, in agreement with the DSC data of Figure 1a, and to the enhanced mobility of the Li + cations resulting from the presence of PYR14TFSI; i.e., the addition of ionic liquid results in large anion excess with respect to the lithium cations. Therefore, the strength of the Li + ⋯Anion − interaction reduces the role of the PEO chains in the coordination of the lithium cations, e.g., as a result from the competition with the PEO⋯Li + interactions [24]. At medium-high temperatures, the conductivity of the P(EO)1(LiTFSI)0.1(PYR14TFSI)0.1 electrolyte is seen to approach or exceed 10 −3 S·cm −1 , still displaying a substantial rise with respect to that of the binary IL-free P(EO)1(LiTFSI)0.1 [31,33]. Table 1. Ionic conductivity and Li anode/polymer electrolyte interface resistance of the poly(ethyleneoxide) (P(EO))1(LiTFSI)0.1 and P(EO)1(LiTFSI)0.1(PYR14TFSI)1 polymer electrolytes at different temperatures. (*) from ref. [31].

Polymer Electrolyte Sample
Ionic The effect of the incorporation of the PYR 14 TFSI ionic liquid on the ion transport properties of the polymer electrolyte is summarized in Table 1. A remarkable conductivity increase is observed, especially at ambient temperature and below. For instance, the P(EO) 1 (LiTFSI) 0.1 (PYR 14 TFSI) 0.1 sample shows ion conduction values three and two orders of magnitude higher than that of the IL-free sample at −20 • C and 20 • C [31,33], respectively. More than 10 −4 S·cm −1 are exhibited at 20 • C, this is of interest for applications in practical devices and commonly not approached in polymer electrolyte membranes. These results support faster ion conduction through the PEO electrolyte due both to a much larger content of the amorphous phase, in agreement with the DSC data of Figure 1a, and to the enhanced mobility of the Li + cations resulting from the presence of PYR 14 TFSI; i.e., the addition of ionic liquid results in large anion excess with respect to the lithium cations. Therefore, the strength of the Li + · · · Anion − interaction reduces the role of the PEO chains in the coordination of the lithium cations, e.g., as a result from the competition with the PEO· · · Li + interactions [24]. At medium-high temperatures, the conductivity of the P(EO) 1 (LiTFSI) 0.1 (PYR 14 TFSI) 0.1 electrolyte is seen to approach or exceed 10 −3 S·cm −1 , still displaying a substantial rise with respect to that of the binary IL-free P(EO) 1 (LiTFSI) 0.1 [31,33]. Table 1. Ionic conductivity and Li anode/polymer electrolyte interface resistance of the poly(ethyleneoxide) (P(EO)) 1 (LiTFSI) 0.1 and P(EO) 1 (LiTFSI) 0.1 (PYR 14 TFSI) 1 polymer electrolytes at different temperatures. (*) from ref. [31].
An important requirement for any electrolyte is its capacity to successfully and efficiently allow electrode reactions, at the operating temperature of the device, without appreciable electrochemical degradation (oxidation/reduction) phenomena. Therefore, the electrochemical stability window (ESW) of the P(EO) 1 (LiTFSI) 0.1 (PYR 14 TFSI) 0.1 electrolyte system was investigated as a function of the temperature. The results, reported in Figure 2 as linear sweep voltammetry curves, evince only a moderate, even if progressive, reduction of the ESW on passing from 20 to 80 • C. In particular, the anodic stability (related to oxidation processes of the electrolyte) detected at 80 • C differs by just 200 mV with respect to that recorded at 20 • C. Conversely, no practical variation is observed on the cathodic side with the temperature increase, displaying massive electrolyte reduction well below 0 V vs. Li + /Li • , which allows lithium plating also at 80 • C. A very low current flow (<25 µA·cm −2 ) is observed up to the anodic breakdown voltage, thus supporting the high purity of the P(EO) 1 (LiTFSI) 0.1 (PYR 14 TFSI) 0.1 sample. On the cathodic verse, three weak (≤20 µA·cm −2 ) features, progressively evinced with the temperature increase, are observed around 1.5 V, 0.9 V and 0.5 V vs. Li + /Li • , respectively. Results previously reported in the literature [44] suggest that the peaks located at 1.5 V and 0.5 V vs. Li + /Li • are ascribable to the Li + cation intercalation process into the native Ni x O film onto the nickel working electrode surface, whereas the feature at 0.9 V is likely due to impurities, i.e., probably water [45]. To summarize, the P(EO) 1 (LiTFSI) 0.1 (PYR 14 TFSI) 0.1 electrolyte is allowed to successfully operate at medium-high temperatures. An important requirement for any electrolyte is its capacity to successfully and efficiently allow electrode reactions, at the operating temperature of the device, without appreciable electrochemical degradation (oxidation/reduction) phenomena. Therefore, the electrochemical stability window (ESW) of the P(EO)1(LiTFSI)0.1(PYR14TFSI)0.1 electrolyte system was investigated as a function of the temperature. The results, reported in Figure 2 as linear sweep voltammetry curves, evince only a moderate, even if progressive, reduction of the ESW on passing from 20 to 80 °C. In particular, the anodic stability (related to oxidation processes of the electrolyte) detected at 80 °C differs by just 200 mV with respect to that recorded at 20 °C. Conversely, no practical variation is observed on the cathodic side with the temperature increase, displaying massive electrolyte reduction well below 0 V vs. Li + /Li°, which allows lithium plating also at 80 °C. A very low current flow (<25 μA·cm −2 ) is observed up to the anodic breakdown voltage, thus supporting the high purity of the P(EO)1(LiTFSI)0.1(PYR14TFSI)0.1 sample. On the cathodic verse, three weak (≤20 μA·cm −2 ) features, progressively evinced with the temperature increase, are observed around 1.5 V, 0.9 V and 0.5 V vs. Li + /Li°, respectively. Results previously reported in the literature [44] suggest that the peaks located at 1.5 V and 0.5 V vs. Li + /Li° are ascribable to the Li + cation intercalation process into the native NixO film onto the nickel working electrode surface, whereas the feature at 0.9 V is likely due to impurities, i.e., probably water [45]. To summarize, the P(EO)1(LiTFSI)0.1(PYR14TFSI)0.1 electrolyte is allowed to successfully operate at medium-high temperatures. The compatibility with the lithium anode is a key parameter for applications as electrolyte separators in Li metal polymer batteries. Figure 3 compares the impedance plots of Li/P(EO)1(LiTFSI)0.1/Li and Li/P(EO)1(LiTFSI)0.1(PYR14TFSI)0.1/Li cells obtained at different temperatures. The AC responses are constituted by a semicircle, taking into account the overall Li/polymer electrolyte interfacial resistance (i.e., charge transfer + passive layer) [38], whereas the high frequency intercept with the real axis is associated with that of the electrolyte bulk [38]. It should be noted that, at 20 °C (panel a), the IL-free electrolyte shows a partial semicircle at high-medium frequencies, due to the relatively low conductivity of the sample P(EO)1(LiTFSI)0.1 [31]. Finally, the inclined straight-line, observed at low frequencies, is attributed to diffusive phenomena through the electrolyte (Warburg contribution) [38]. The impedance plots of Figure 3 clearly confirm how the incorporation of ionic liquid results in a significant decrease of the electrolyte resistance, especially The compatibility with the lithium anode is a key parameter for applications as electrolyte separators in Li metal polymer batteries. Figure 3 compares the impedance plots of Li/P(EO) 1 (LiTFSI) 0.1 /Li and Li/P(EO) 1 (LiTFSI) 0.1 (PYR 14 TFSI) 0.1 /Li cells obtained at different temperatures. The AC responses are constituted by a semicircle, taking into account the overall Li/polymer electrolyte interfacial resistance (i.e., charge transfer + passive layer) [38], whereas the high frequency intercept with the real axis is associated with that of the electrolyte bulk [38]. It should be noted that, at 20 • C (panel a), the IL-free electrolyte shows a partial semicircle at high-medium frequencies, due to the relatively low conductivity of the sample P(EO) 1 (LiTFSI) 0.1 [31]. Finally, the inclined straight-line, observed at low frequencies, is attributed to diffusive phenomena through the electrolyte (Warburg contribution) [38]. The impedance plots of Figure 3 clearly confirm how the incorporation of ionic liquid results in a significant decrease of the electrolyte resistance, especially from room to medium temperature, in agreement with the conductivity data reported in Table 1. However, a gain, even if moderate, in interface resistance is also detected. For instance, the P(EO) 1 (LiTFSI) 0.1 (PYR 14 TFSI) 0.1 sample shows, at the interface with Li metal, a resistance of 10-11% lower (i.e., from 830 to 750 cm 2 at 20 • C and from 7.0 to 6.3 cm 2 at 80 • C) than that of the IL-free electrolyte (Table 1), in the whole investigated temperature range (20-80 • C). We can hypothesize that the ionic liquid improves the Li + cation mobility at the electrolyte/lithium interface.  Applications such as in automotives, smart grids, etc. require high power and for energy to be readily available; this means that this requires the battery system to be feasibly discharged and charged at high current rates without significantly depleting its performance. For instance, the increase of the current rate promotes the diffusive phenomena within the battery, thus lowering the content of the stored/delivered energy. In electrochemical cells, the redox process kinetics are generally much faster than the active species diffusion through the electrolyte separator. By increasing the current value, the matter transferring process becomes more and more predominant with respect to those at the interfaces with the electrodes. When the current flow through the cell achieves a limiting value, JL (diffusion limiting current), the electrochemical processes are fully governed by the ion diffusion from the electrolyte bulk to the electrode surface and vice versa. Therefore, JL is a key parameter for evaluating the feasibility of an electrolyte at high current rates. The limiting current value was determined as reported in Materials and Methods. For instance, linear sweep voltammetry tests were run (at 0.01 mV·s −1 ) on symmetrical Li/P(EO)1(LiTFSI)0.1/Li and Li/P(EO)1(LiTFSI)0.1(PYR14TFSI)0.1/Li cells at temperatures ranging from 40 to 80 °C. Figure 4 plots the current density values, recorded during the potentiodynamic measurements, as a function of the cell overvoltage. After an initial step increase, in which the electrolyte membrane shows a quasi-ohmic behavior, the current density is seen to progressively level off, likely associated with the establishment of a concentration gradient within the electrolyte membrane [46], around a time-stable value. Such a behavior indicates that the current density through the cell has reached the limiting value (JL), e.g., the ion conduction processes inside the electrolyte membrane are governed by diffusion phenomena (the concentration gradient extends through the overall electrolyte thickness). In Figure 4, it is shown how the JL value remarkably increases with the operating cell temperature but is not affected by the presence of PYR14TFSI, i.e., from 0.13-017 to 1.2-2.0 mA·cm −2 (about one order of magnitude) in passing from 40 to 80 °C for both the IL-free (panel a) and the IL-containing (panel b) electrolyte. Therefore, the ionic liquid does not seem to reduce the diffusive phenomena through the PEO electrolyte. However, the current density of the P(EO)1(LiTFSI)0.1 sample, upon achieving the limiting value, quickly shows an abrupt feature during the potentiodynamic measurements at 60 °C and 80 °C (Figure 4a). This behavior, repeatedly confirmed by several Applications such as in automotives, smart grids, etc. require high power and for energy to be readily available; this means that this requires the battery system to be feasibly discharged and charged at high current rates without significantly depleting its performance. For instance, the increase of the current rate promotes the diffusive phenomena within the battery, thus lowering the content of the stored/delivered energy. In electrochemical cells, the redox process kinetics are generally much faster than the active species diffusion through the electrolyte separator. By increasing the current value, the matter transferring process becomes more and more predominant with respect to those at the interfaces with the electrodes. When the current flow through the cell achieves a limiting value, J L (diffusion limiting current), the electrochemical processes are fully governed by the ion diffusion from the electrolyte bulk to the electrode surface and vice versa. Therefore, J L is a key parameter for evaluating the feasibility of an electrolyte at high current rates. The limiting current value was determined as reported in Materials and Methods. For instance, linear sweep voltammetry tests were run (at 0.01 mV·s −1 ) on symmetrical Li/P(EO) 1 (LiTFSI) 0.1 /Li and Li/P(EO) 1 (LiTFSI) 0.1 (PYR 14 TFSI) 0.1 /Li cells at temperatures ranging from 40 to 80 • C. Figure 4 plots the current density values, recorded during the potentiodynamic measurements, as a function of the cell overvoltage. After an initial step increase, in which the electrolyte membrane shows a quasi-ohmic behavior, the current density is seen to progressively level off, likely associated with the establishment of a concentration gradient within the electrolyte membrane [46], around a time-stable value. Such a behavior indicates that the current density through the cell has reached the limiting value (J L ), e.g., the ion conduction processes inside the electrolyte membrane are governed by diffusion phenomena (the concentration gradient extends through the overall electrolyte thickness). In Figure 4, it is shown how the J L value remarkably increases with the operating cell temperature but is not affected by the presence of PYR 14 TFSI, i.e., from 0.13-017 to 1.2-2.0 mA·cm −2 (about one order of magnitude) in passing from 40 to 80 • C for both the IL-free (panel a) and the IL-containing (panel b) electrolyte. Therefore, the ionic liquid does not seem to reduce the diffusive phenomena through the PEO electrolyte. However, the current density of the P(EO) 1 (LiTFSI) 0.1 sample, upon achieving the limiting value, quickly shows an abrupt feature during the potentiodynamic measurements at 60 • C and 80 • C (Figure 4a). This behavior, repeatedly confirmed by several (potentiodynamic) tests carried out on different Li/P(EO) 1 (LiTFSI) 0.1 /Li cells and never observed in the P(EO) 1 (LiTFSI) 0.1 (PYR 14 TFSI) 0.1 sample, is ascribable to dendrite growth onto the Li electrode at current rates above 1 mA·cm −2 . The results reported in Figure 4a suggest that the IL-free electrolyte is not able to sustain high current rates. Conversely, the ionic liquid plays a key role in improving the compatibility at the interface with the lithium anode, in particular when the cell is subjected to high current rates instead of in an open circuit condition as plotted in Figure 3. It is a plausible hypothesis that PYR 14 TFSI behaves as a protective agent towards the Li metal electrode, allowing the running of charge/discharge cycling tests at a high current density without appreciable degradation phenomena of the lithium anode. Once more, this confirms the beneficial effect resulting from ionic liquid incorporation on battery performance.

Composite Electrodes
The LiFePO4 electrode formulation was optimized in terms of carbon content in order to reach a good compromise between electronic conductor content and cathode performance. Therefore, electrode samples containing different carbon contents were prepared and investigated in terms of their electronic conductivity by impedance spectroscopy. The results are reported in Figure 5 as AC responses (panel a) and electronic conductivity vs. carbon content dependence (panel b). The impedance plot of the carbon-free sample (Figure 5a) is constituted by a semicircle (not starting from the axis origin) which does not display any capacitive contribution, indicating charge transfer at the interfaces with the Al° collectors [38]. This behavior-i.e., supporting electron conduction through the composite electrode-suggests the establishment of a three-dimensional network (percolation) formed by LiFePO4 particles and, therefore, electronic continuous pathways through the composite cathode [37]. It should be noted that the as-received active material is provided as superficially carbon-coated; this supports the not-very-low electronic resistance (given by the AC plot intercept with the real axis at low frequencies [38]) of the composite cathode (i.e., pure LiFePO4 material exhibits very low electronic conductivity [47]). The addition of KJB carbon around 3-4 wt. % results in a remarkable reduction of the semicircle diameter and a shifting of the low frequency intercept with the real axis towards smaller impedance values, highlighting a decrease of the electronic resistance of the cathode. At a KJB content equal to 6 wt. %, the semicircle practically reduces to a quasi-single point on the real axis, indicating that the electronic conductivity is largely overcome with respect to the ionic one (the electron and ion conductions through the polymer electrolyte are in parallel) of the polymer electrolyte incorporated within the electrode. In such a condition, the electronic resistance of the composite cathode is given by the distance of the "spot" response intercept

Composite Electrodes
The LiFePO 4 electrode formulation was optimized in terms of carbon content in order to reach a good compromise between electronic conductor content and cathode performance. Therefore, electrode samples containing different carbon contents were prepared and investigated in terms of their electronic conductivity by impedance spectroscopy. The results are reported in Figure 5 as AC responses (panel a) and electronic conductivity vs. carbon content dependence (panel b). The impedance plot of the carbon-free sample (Figure 5a) is constituted by a semicircle (not starting from the axis origin) which does not display any capacitive contribution, indicating charge transfer at the interfaces with the Al • collectors [38]. This behavior-i.e., supporting electron conduction through the composite electrode-suggests the establishment of a three-dimensional network (percolation) formed by LiFePO 4 particles and, therefore, electronic continuous pathways through the composite cathode [37]. It should be noted that the as-received active material is provided as superficially carbon-coated; this supports the not-very-low electronic resistance (given by the AC plot intercept with the real axis at low frequencies [38]) of the composite cathode (i.e., pure LiFePO 4 material exhibits very low electronic conductivity [47]). The addition of KJB carbon around 3-4 wt. % results in a remarkable reduction of the semicircle diameter and a shifting of the low frequency intercept with the real axis towards smaller impedance values, highlighting a decrease of the electronic resistance of the cathode. At a KJB content equal to 6 wt. %, the semicircle practically reduces to a quasi-single point on the real axis, indicating that the electronic conductivity is largely overcome with respect to the ionic one (the electron and ion conductions through the polymer electrolyte are in parallel) of the polymer electrolyte incorporated within the electrode. In such a condition, the electronic resistance of the composite cathode is given by the distance of the "spot" response intercept with the real axis from the origin of the axes [38]. Figure 5b illustrates the electronic conductivity of the composite LiFePO 4 cathode as a function of the carbon content. As evinced in Figure 5a, the electron conduction raises up to 7 wt. % of KJB with a gain of about 1.5 orders of magnitude. The further addition of carbon does not lead to any improvement of the electron transport properties, whereas it depletes the active material content and, therefore, the energy density of the composite cathode. Therefore, the KJB content in the LiFePO 4 electrode was fixed to 7 wt. %.

Battery Tests at 80 °C
Upon investigation of the electrochemical performance, the P(EO)1(LiTFSI)0.1(PYR14TFSI)0.1 ionic liquid-based, polymer electrolyte was subjected to tests in Li/LiFePO4 cells at 80 °C. Figure 6a compares the voltage vs. capacity profile referring to the 1st charge-discharge cycle run at different current rates. A flat plateau, typical of the Li + insertion/de-insertion process into the LiFePO4 active material [24,33,34], is observed (in the 3.3-3.6 V range) even at higher rates, with a coulombic efficiency close to 99%. This highlights that IL-incorporating Li/LiFePO4 cells are capable of maintaining the same voltage during almost the entire charge/discharge step. Only a 100 mV increase in ohmic drop is observed on passing from 0.1C through 1C. An initial capacity corresponding to the theoretical value (170 mA·h·g −1 ) is delivered up to the medium rate (0.33C) with just a moderate decrease at high current rates, i.e., more than 160 mA·h·g −1 (>94.1% of the theoretical capacity) are discharged at 1C. Figure 6b,c compares the voltage profiles of the selected charge/discharge cycles at 0.1C and 1C, respectively. It is worth noting that the excellent reproducibility of the battery performance, i.e., the profile feature and the delivered capacity, are practically unchanged after 100 consecutive cycles run (at 100% of deep of discharge, DOD) even at high current rates, which is not often reported for lab-scale, lithium metal polymer cells [24]. These results clearly show the very good reversibility of the Li + intercalation process even under hard operating conditions in combination with an excellent compatibility at the electrolyte/electrode interface and negligible degradation phenomena occurring within the cell components. Such a performance score, however, can be achieved only through good manufacturing of the electrolyte/electrode components, i.e., high purity levels and careful optimization of the formulation, and of the full cells.

Battery Tests at 80 • C
Upon investigation of the electrochemical performance, the P(EO) 1 (LiTFSI) 0.1 (PYR 14 TFSI) 0.1 ionic liquid-based, polymer electrolyte was subjected to tests in Li/LiFePO 4 cells at 80 • C. Figure 6a compares the voltage vs. capacity profile referring to the 1st charge-discharge cycle run at different current rates. A flat plateau, typical of the Li + insertion/de-insertion process into the LiFePO 4 active material [24,33,34], is observed (in the 3.3-3.6 V range) even at higher rates, with a coulombic efficiency close to 99%. This highlights that IL-incorporating Li/LiFePO 4 cells are capable of maintaining the same voltage during almost the entire charge/discharge step. Only a 100 mV increase in ohmic drop is observed on passing from 0.1C through 1C. An initial capacity corresponding to the theoretical value (170 mA·h·g −1 ) is delivered up to the medium rate (0.33C) with just a moderate decrease at high current rates, i.e., more than 160 mA·h·g −1 (>94.1% of the theoretical capacity) are discharged at 1C. Figure 6b,c compares the voltage profiles of the selected charge/discharge cycles at 0.1C and 1C, respectively. It is worth noting that the excellent reproducibility of the battery performance, i.e., the profile feature and the delivered capacity, are practically unchanged after 100 consecutive cycles run (at 100% of deep of discharge, DOD) even at high current rates, which is not often reported for lab-scale, lithium metal polymer cells [24]. These results clearly show the very good reversibility of the Li + intercalation process even under hard operating conditions in combination with an excellent compatibility at the electrolyte/electrode interface and negligible degradation phenomena occurring within the cell components. Such a performance score, however, can be achieved only through good manufacturing of the electrolyte/electrode components, i.e., high purity levels and careful optimization of the formulation, and of the full cells.
0.1C and 1C, respectively. It is worth noting that the excellent reproducibility of the battery performance, i.e., the profile feature and the delivered capacity, are practically unchanged after 100 consecutive cycles run (at 100% of deep of discharge, DOD) even at high current rates, which is not often reported for lab-scale, lithium metal polymer cells [24]. These results clearly show the very good reversibility of the Li + intercalation process even under hard operating conditions in combination with an excellent compatibility at the electrolyte/electrode interface and negligible degradation phenomena occurring within the cell components. Such a performance score, however, can be achieved only through good manufacturing of the electrolyte/electrode components, i.e., high purity levels and careful optimization of the formulation, and of the full cells.
(a) (b) (c)  The cycling performance of the Li/P(EO) 1 (LiTFSI) 0.1 (PYR 14 TFSI) 0.1 /LiFePO 4 solid-state cells, tested at 80 • C and different current rates, is depicted in Figure 7a. An excellent capacity retention (as also evinced in Figure 6b,c) with a coulombic efficiency quickly leveling above 99.5% (100% at 0.1C) is recorded even at higher rates, i.e., more than 99.5% and 94% of theoretical capacity are initially delivered at 0.33C and 1C, respectively, with a very modest decay (>98% and 93.6%, respectively) after 100 consecutive cycles. This corresponds to a capacity fading around 0.005% per cycle and, in conjunction with the very good charge/discharge efficiency, once more highlights a highly reversible lithiation process in combination with the high purity level and high compatibility of the P(EO) 1 (LiTFSI) 0.1 (PYR 14 TFSI) 0.1 polymer electrolyte towards electrodes, in particular with the lithium metal anode, even at high current rates. Also, it should be noted that very clean lithium metal tapes were used for the cell manufacturing in order to obtain an optimal Li/electrolyte interface. Especially, we would like to point out the absence of dendrite growth on the Li electrode during prolonged cycling tests run also at 1C, i.e., very rarely encountered in lithium metal polymer batteries operating at medium-high temperatures under high rates [24]. These experimental data, in rather good agreement with the results derived from potentiodynamic measurements depicted in Figure 4, once more demonstrate that the incorporation of ionic liquids such as PYR 14 TFSI significantly improves the PEO electrolyte interface with the lithium anode, allowing high current rates to be sustained for prolonged cycling tests without appreciably depleting the cell performance. The cycling performance of the Li/P(EO)1(LiTFSI)0.1(PYR14TFSI)0.1/LiFePO4 solid-state cells, tested at 80 °C and different current rates, is depicted in Figure 7a. An excellent capacity retention (as also evinced in Figure 6b,c) with a coulombic efficiency quickly leveling above 99.5% (100% at 0.1C) is recorded even at higher rates, i.e., more than 99.5% and 94% of theoretical capacity are initially delivered at 0.33C and 1C, respectively, with a very modest decay (>98% and 93.6%, respectively) after 100 consecutive cycles. This corresponds to a capacity fading around 0.005% per cycle and, in conjunction with the very good charge/discharge efficiency, once more highlights a highly reversible lithiation process in combination with the high purity level and high compatibility of the P(EO)1(LiTFSI)0.1(PYR14TFSI)0.1 polymer electrolyte towards electrodes, in particular with the lithium metal anode, even at high current rates. Also, it should be noted that very clean lithium metal tapes were used for the cell manufacturing in order to obtain an optimal Li/electrolyte interface. Especially, we would like to point out the absence of dendrite growth on the Li electrode during prolonged cycling tests run also at 1C, i.e., very rarely encountered in lithium metal polymer batteries operating at medium-high temperatures under high rates [24]. These experimental data, in rather good agreement with the results derived from potentiodynamic measurements depicted in Figure 4, once more demonstrate that the incorporation of ionic liquids such as PYR14TFSI significantly improves the PEO electrolyte interface with the lithium anode, allowing high current rates to be sustained for prolonged cycling tests without appreciably depleting the cell performance. The capacity vs. current density dependence (80 °C) is plotted in Figure 7b, which evinces a very good rate capability. Above 94% of the theoretical value is still obtained at 1C, supporting an excellent rate capability up to 1C, i.e., corresponding to about 0.7 mA·cm −2 , which represents a very interesting current value for an all-solid-state polymer electrolyte. A further increase of the current rate up to 2C, The capacity vs. current density dependence (80 • C) is plotted in Figure 7b, which evinces a very good rate capability. Above 94% of the theoretical value is still obtained at 1C, supporting an excellent rate capability up to 1C, i.e., corresponding to about 0.7 mA·cm −2 , which represents a very interesting current value for an all-solid-state polymer electrolyte. A further increase of the current rate up to 2C, i.e., around 1.4 mA·cm −2 , leads to a reduction of the delivered capacity which levels off at 57% of the theoretical value. This behavior, ascribable to diffusive phenomena within the electrolyte separator, is in good agreement with the results obtained by the potentiodynamic measurements (Figure 3b), which indicates that above a current density of about 1.2 mA·cm −2 (determined as J L value), the electrochemical processes through the cell are controlled by the diffusive phenomena occurring within the polymer electrolyte. However, despite the capacity decay due to the operating current density exceeding the limiting value, the Li/P(EO) 1 (LiTFSI) 0.1 (PYR 14 TFSI) 0.1 /LiFePO 4 cells are still able to deliver about 100 mA·h·g −1 at a rate as high as 2C (about 1.4 mA·cm −2 ), i.e., representing a remarkable capacity value for an all-solid-state polymer electrolyte.
The battery performance of the P(EO) 1 (LiTFSI) 0.1 (PYR 14 TFSI) 0.1 electrolyte, detected at 80 • C in Li/LiFePO 4 cells, is compared with that of other lithium-conducting, ionic liquid-free, PEO membranes, recorded in Li/LiFePO 4 and Li/V 2 O 5 systems at temperatures from 90 • C to 100 • C [18,22,23]. The data, reported in Table 2, show how appreciable capacities, i.e., from 70 to 96% of the cell theoretical value, are delivered only at low-medium rates (0.2C-0.33C). However, a capacity decay down to 45-60% of the theoretical value is observed after 100 consecutive charge/discharge cycles, with a fading corresponding to 0.26-0.36% per cycle. Conversely, very modest capacities, i.e., from 8 to 14% of the theoretical value, are obtained when the current rate is increased up to 0.8C-1C. From the data illustrated in Figures 6 and 7 and Table 2, it is evident how, at medium-high temperatures, the PYR 14 TFSI-incorporating lithium polymer batteries behave much better in terms of their delivered capacity and cycling performance than the IL-free ones. For instance, the addition of suitable ionic liquid is able to largely improve the performance of the LPBs not only at ambient or near ambient conditions, as previously reported in the literature [18,20,22,23], but even at medium-high temperatures. Therefore, the PEO-LiTFSI-PYR 14 TFSI Li + -conducting membranes are very promising candidates as electrolyte separator systems for all-solid-state lithium polymer batteries operating around 100 • C. Table 2. Summary of the battery performance of the P(EO) 1 (LiTFSI) 0.1 (PYR 14 TFSI) 1 polymer electrolyte at 80 • C compared with that of lithium-conducting, ionic liquid-free, PEO membranes at medium-high temperatures. (a) From reference [22]; (b) from reference [23]; (c) from reference [18]; (d) this work.

Conclusions
PEO-LiTFSI Li + -conducting membranes, containing the PYR 14 TFSI ionic liquid, were prepared and studied to be addressed as electrolyte separators for all-solid-state lithium polymer batteries operating at medium-high temperatures. A solvent-free procedure was designed to prepare the PEO-LiTFSI-PYR 14 TFSI electrolytes. These ternary systems have shown remarkably improved thermal, ion transport and interfacial properties with respect to the ionic liquid-free electrolytes. Wide electrochemical stability was observed even at medium-high operating temperatures. In particular, the ionic liquid-based PEO electrolytes are able to sustain high current rates without any appreciable lithium anode degradation, which is not allowed in binary ionic liquid-free, PEO-LiTFSI systems, thus enabling their use in battery systems operating at 80 • C or above and high current rates. Battery tests carried out at 80 • C in Li/LiFePO 4 polymeric systems have shown excellent cycling behavior and capability retention at high current rates, e.g., more than 93.6% of the theoretical capacity (i.e., 99.5% of the initial value) is still delivered after 100 cycles run at 1C with a coulombic efficiency close 100%. This performance largely exceeds that of analogous, ionic liquid-free, polymer lithium batteries at the same operating conditions, nominating the PEO-LiTFSI-PYR 14 TFSI ternary system as an electrolyte separator for medium-high temperature lithium polymer batteries. It is worth highlighting the absence of dendrite growth on the Li anode during prolonged cycling tests even at high current rates, which is very often not observed in lithium metal polymer batteries.