Polymerized Ionic Liquid Block Copolymer Electrolytes for All-Solid-State Lithium-Metal Batteries

In this work, we present a polymerized ionic liquid block copolymer (PBCP) ﬁ lm where relevant properties such as ionic conductivity and electrochemical parameters are tailored by using a ternary system comprised of poly(styrene-b-1-((2-acryloyloxy) ethyl) − 3-butylimidazolium bis(tri- ﬂ uoromethanesulfonyl)imide), LiFSI salt and ethylene carbonate (EC) as a cosolvent. It was found that EC ef ﬁ ciently decreases the glass transition temperature of the ionic block, resulting in an improved ionic conductivity and ef ﬁ cient platting/stripping of lithium. By using an optimal ratio of EC/LiFSI at relatively high LiFSI amount, Li ∣ Li symmetrical cells at 50 °C show an overpotential as low as 70 mV at 0.1 mA.cm − 2 along with a high lithium transport number of 0.56 (t Li + ). All-solid-state full cells based on lithium iron phosphate cathode paired with a lithium metal anode reveal a rather stable cycling at both 50 °C and 70 °C. A negligible capacity fading is observed up to 30 cycles where a speci ﬁ c capacity as high as 161 mAh.g − 1 is achieved with a coulombic ef ﬁ ciency of 99.9%. Thus, this work demonstrates an important pathway for tailoring the properties of solid state polymer electrolytes for emerging and specially designed block copolymer architectures comprising domains that give both excellent ionic conduction along with desirable mechanical properties. ©

The introduction of Lithium-ion batteries (LIBs) in the market by Sony Corporation in 1991 was considered a breakthrough due to their ability to store substantial higher energy when compared to other available standard secondary batteries. 1,2 Interestingly, even after almost three decades of their announcement, LIBs are still dominant and can be found in several modern devices such as laptops, smartphones, portable gadgets and many other examples. Due to the intense and rapid evolution in the electronics technology and the additional features that continue to be developed in portable devices, LIBs have also evolved over time in order to catch up with higher energy requirements. [3][4][5] Nevertheless, the actual energy density of LIBs is approaching their limit (∼250-300 Wh kg −1 ) and there is a common agreement that in the near feature, LIBs will not be capable of addressing higher energy demands. 6 In addition, the growing interest in sustainable transportation, which have dramatically raised the commercialization of both plug-in hybridelectrical-vehicles (HEVs) and all-electric vehicles (EVs), further exposes the requirement for battery technologies that can store higher amounts of energy per mass and/or per volume. 7,8 Among several available and well-known batteries that present higher energy density than LIBs, lithium batteries based on lithium metal anodes are considered as the "holy grail" of batteries due to their superior energy densities. This is because the lithium metal anode shows 10 times higher theoretical specific capacity when compared to the state-of-art graphite anode materials that are now present in LIBs. 9 Moreover, the possible successful replacement of graphite by lithium metal would also allow the transition to the highest energy density batteries, such as Li-O 2 and Li-S. 10,11 However, both Li-O 2 and Li-S present severe issues that have yet to be overcome; drawbacks such as low coulombic efficiency, unsatisfactory active material utilization, reactive discharge products, and large volume expansion coupled with strong capacity fading hamper their prompt practical applications. 11,12 In this context, Li-metal batteries based on Li-intercalation type cathode materials are currently considered to be the most promising next generation batteries. However, the presence of lithium metal brings serious challenges that need to be overcome before practical implementation. During charge, lithium ions undergo an uneven electrodeposition of lithium which potentially results in dendrite formation; these dendrites will likely keep growing and will eventually pierce the separator and reach the cathode, causing a short circuit which would lead to catastrophic fire and explosion, especially when standard flammable liquid electrolytes are used. 13 In addition, lithium metal is a highly reactive metal and will potentially decompose the electrolyte; hence, the decomposition product of the electrolyte (solid electrolyte interphase, SEI) must be able to give a passivation layer that creates a barrier to electron transfer but at the same time provides a compact and efficient pathway for Li ions to move through (ionic conductor). 14 In order to address the aforementioned challenges that the lithium metal anode presents, solid polymer electrolytes (SPEs) have been considered as good candidates to achieve a safe, stable and long lifespan lithium metal batteries. 15 In general, SPEs are nonflammable and have the ability to form flexible membranes that can work as both electrolyte and separator, ensuring a high energy density battery. Furthermore, due to its solid state feature, it has been suggested that SPE can act as a physical barrier, creating a "shield" that may prevent further Li dendrite growth and short circuit; this characteristic relies on the toughness of the SPE which is dictated by its modulus. It is worth pointing out that such SPEs must be able to provide an intimate contact on the electrodes in order to minimize the risk of interfacial resistance. The inability on providing such characteristics will result in a cell with a short lifespan and a rapid capacity fading. 16,17 Indeed, the nature of the interfacial resistance between solid electrolytes and electrodes is one of the major problems encountered for the successful application of solid electrolytes, especially for highly conductive sulfide-and ceramicbased electrolytes. 18,19 When it comes to solid state ionic conductive polymers, Polyethylene oxide (PEO) and polymerized ionic liquids (PILs) stand out. 20,21 In both cases, a binary (polymer+salt) or ternary mixture (polymer+salt+solvent) is used to ensure ionic conductivity. Although PEO and its solvation properties when mixed with lithium salts and other compounds are relatively well-known, PEObased SPEs usually show low lithium transport number (t Li +), which can result in a poor electrochemical stability due to the concentration polarization; this behavior mainly relies on the strong coordination of Li ions with oxygen from PEO chains. 22,23 Moreover, the Li transport mechanism in PEO systems is believed to be effective mainly above PEO's melting point (∼65°C), imposing some limitations on its applicability. On the other hand, PILs' properties can be widely tuned based on a judicious choice of both cation and anion from their ionic liquid monomers. Furthermore, as ionic liquid-derived compounds, PILs show high chemical stability, wide electrochemical window and high thermal stability. Besides, PILs are able to solvate and interact with a wide range of lithium salts and solvents (ionic liquids or organic plasticizers), resulting in a very versatile material for solid or quasi solid state batteries. 24,25 Unfortunately, SPEs based on PIL homopolymers show unsatisfactory mechanical properties due to their low modulus value; such features would still not solve the cell failure problem based on the problematic dendrite growth and alternatives to reinforce the mechanical properties of these SPEs should be adopted. In this context, block copolymer electrolytes (BCPs) are considered as possibly the most appealing approach to overcome the low modulus of homopolymers. Block copolymers consist of dissimilar segments/ blocks that are covalently bound; and hence, due to the different nature of the blocks, the material microphase separates into periodic structures within nanoscopic domains. 26,27 For all-solid-state battery applications, an ideal low Tg ionic block should be responsible for lithium ion conduction whilst a non-ionic, high Tg block provides the mechanical stability required to provide overall robust film properties which may be important for lithium dendrite suppression. Polystyrene and poly (methyl methacrylate) are probably the most common investigated high Tg blocks found in BCPs whilst PEO dominates the literature regarding the low Tg block responsible for Li ion conductivity. [28][29][30][31] Due to the previously mentioned drawbacks of PEO, in this paper we report a polymerized ionic liquid as the ionic block coupled with polystyrene block to give a polymerized ionic liquid block copolymer (PBCP) as a solid polymer electrolyte for all-solid-state lithium metal batteries. We describe how physical, thermal and electrochemical properties of the PBCP is optimized by using a ternary system comprising the PBCP polymer, LiFSI salt and EC. The use of LiFSI salt is inspired by its ability to give higher conductivity and form a homogeneous SEI layer on lithium metal, as already demonstrated in many systems by our group. [32][33][34][35] In the PBCP used in this work we consider the polymer where the counterion in the PIL block is the TFSI anion. This leads to a mixed anion system which has previously been shown to have benefits for Li batteries with respect to increased stability and cycling performance, as well as enhanced Li ion transport. 36,37 The choice of ethylene carbonate solvent was recently demonstrated by Nguyen et al. in a single ion conductor-multi-block copolymer electrolyte as capable of solvating Li ions and leading to high conductivity membranes that supported cycling of an NMC cathode, although with significant capacity loss. 38 Furthermore, in that prior work, the EC contents considered were close to 90% by weight of the total electrolyte with the polymer providing the mechanical support. In the present work we consider a ternary system where a high LiFSI concentration is dissolved in a simple imidazolium based PIL block with polystyrene block providing the mechanical integrity, and the EC component being an efficient, easily applicable, low cost co-solvent that can effectively be added to the PBCP to help improve the mobility of Li ions through the PIL block. An all-solid-state Limetal battery based on Li|PBCP|LFP using this simple ternary system is successfully demonstrated at both 50°C and 70°C. Due to the low interfacial resistance in conjunction with high lithium transport number, full cells exhibit a high specific capacity which operates at 99.8%-99.9% coulombic efficiency with low capacity fading. Conductivity measurements.-The conductivities of the membranes were measured in a barrel cell equipped with a temperature controller (Eurotherm 2204 temperature controller) and electrochemical impedance spectroscopy (EIS) was applied to obtain the resistance of the membranes in a temperature range from 30°C-100°C. Polymer membranes were assembled in a sample holder and placed in between two circular polished stainless steel discs and subsequently sealed in a cylindrical barrel-type cell. The MTZ Impedance analyzer (MTZ-35, Biologic Science Instruments) was used to record the impedance spectrum at each temperature after equilibrium by using a frequency range from 1 × 10 7 -1 Hz at an amplitude of 10 mV. The spectra were fitted by using the MT-lab software (V. 1.00) and the conductivity values were obtained after using the Eq. 1 below: Where σ is the conductivity (S.cm −1 ), L is the thickness of the membrane (cm), R is the resistance obtained after fitting the spectra and A is the area of the membrane (cm 2 ).
Differential scanning calorimetry (DSC).-All DSC traces were obtained by using a Mettler Toledo DSC 1 Star e Instrument. Prior to measurements, the equipment was calibrated with cyclohexane. Polymer samples were prepared in an inert atmosphere (Ar glovebox, Kiyon) with extreme low level of oxygen and water (O 2 < 0.2 ppm and H 2 O ∼ 0 ppm). Samples were weighted to a minimum amount of material (>10 mg) and hermetically sealed in standard aluminum pans. The experiments were carried out by first cooling down to 173.15 K at a rate of 100 K.min −1 and maintained at 173.15 K for 15 min. The heating was recorded from 173.15 K to 393.15 K at a rate of 10 K.min −1 . This cycle was repeated three times and the third heating trace was used to report the results.
Small angle X-ray scattering (SAXS).-SAXS experiments involving all samples were performed at the Australian Synchrotron SAXS beamline. A beam energy of 12 keV and a 1338 mm camera were used. The SAXS patterns were recorded with a 1M Pilatus detector with 981 × 1043 pixel resolution.
Pulse field gradient diffusion NMR (PFG-NMR).-Lithium diffusion coefficients were determined by using a Bruker Avance III 300 MHz wide-bore NMR spectrometer. To record the data, a 5 mm Diff50 probe was used. According to the nuclei needed, different RF coils were used for 19 F and 7 Li NMR For the measurements, samples were firstly packed and sealed into rotors inside an Ar glovebox in order to avoid any moisture uptake which could interfere in the results. Subsequently, rotors were introduced in 5 mm glass tubes and finally taken into diffusion probe to start the measurement. Before starting the measurement, samples were allowed to reach equilibrium at each temperature (from 30°C-90°C) and then the pulse-field gradient stimulated echo (PFG-STE) pulse sequence was used to obtain the diffusion coefficients. A pulse of 4 ms duration was set for 7 Li and 10 ms for 19 F. The gradient was performed over 32 steps and the gradient strength varied between 0.01 and 25 T.m −1 in a log scale to obtain the NMR attenuation curve.
Electrochemical methods.-Li|Li symmetrical cells were assembled in an Ar glovebox by firstly cleaning Li ribbons with cyclohexane with the aid of a nylon brush. After cleaning, lithium was punched to give discs of 0.5 cm 2 . Symmetrical cells were fabricated by sandwiching the polymer membranes in between two lithium discs in a CR 2032 coin cell setup (Hohsen, Japan).
The fabrication of positive electrodes (LFP electrodes) was performed by using the slurry coating method and a detailed description of the process is provided in supplementary information.
Full cells were assembled in CR 2032 coin cells inside Ar glovebox by using the LFP electrodes as cathode, lithium metal foil (Gelon) as anode and solid polymer electrolyte membranes (PBCP's) as electrolyte. For cycling the batteries, a multi-channel VMP-3 potentiostat (Biologic) and the Neware battery cycler were used.
Electrochemical impedance spectroscopy (EIS) was used to characterize Li|Li symmetrical cells. To do so, a perturbation of 10 mV (amplitude) was applied at open circuit potential (OCP) at 50°C for all cells after 24 h rest at 50°C. The EIS spectra were obtained by using frequencies ranging from 100,000-0.1 Hz. The t Li + was determined by applying the Bruce-Vincent-Evans method as already described elsewhere. 39 Full cells were cycled by using a galvanostatic charge-discharge method (chronopotentiometry). The applied current (C-rate) was calculated based on the mass of the active material in the cathode (refer supplementary information) and taking into account the theoretical specific capacity of LFP (170 mAh.g −1 ). Based on the aforementioned details of our as-fabricated cathode, the C-rates reported throughout the manuscript, which are equal to C/20, C/10, C/5, C/4 and C/3 correspond to current densities of 0.023 mA.cm −2 , 0.046 mA.cm −2 , 0.092 mA.cm −2 , 0.115 mA.cm −2 and 0.153 mA.cm −2 , respectively.
A voltage range of 3.0-3.7 V vs Li/Li + was selected to cycle the Li-metal batteries.

Results and Discussion
Differential scanning calorimetry (DSC) and conductivity measurements.-To investigate the impact of the addition of both EC and LiFSI on the final conductivity of the solid polymer electrolytes, a matrix involving different combinations of the ternary system was formulated and is presented in Table I. It is worth mentioning that the amount of both lithium salt and EC are on a molar basis per mol of PIL unit. The reason behind this, as will be discussed later, is because both the Li salt and EC have been shown to be dissolved solely in the PIL block.
It is known that EC acts as a plasticizer when added to polymers, and this effect may result in a decrease of the polymer's glass transition (T g ). Based on that, it is helpful to analyze the effect of the additives on the conductivity of the membranes by concomitantly looking to the shift of the Tg of each sample. In Fig. 1a, the DSC traces of samples PBCP-215, PBCP-240 and PBCP-260 are presented. When 2 mols of LiFSI/PIL and 1.5 mols of EC/PIL are added to the polymer (PBCP-215), two glass transitions are observed at −6.8°C and 90.0°C. The first Tg at −6.8°C is ascribed to the PIL block and the second at 90.0°C corresponds to the mechanical block (PS-polystyrene block). The fact that two Tg's are observed reveal that PIL and PS blocks are microphase separated, which is beneficial to achieve higher conductivities if compared to random copolymers, not to mention superior mechanical properties. The conductivity of samples PBCP-215, PBCP-240 and PBCP-260 is provided in Fig. 1b. As shown, sample PBCP-215 shows a conductivity as low as 6.5 × 10 −8 S.cm −1 at 50°C and 2.3 × 10 −6 S.cm −1 at 100°C. Very interestingly, when the EC amount is increased from 1.5 to 4 mols/PIL (PBCP-240), the Tg of the PIL block drops from −6.8°C to −21.8°C, confirming that the EC is dissolved in the PIL block. Moreover, it is important to note that the Tg of the PS block is not affected, which suggests that the EC does not interact with the noncharged PS block. The conductivity of sample PBCP-240 dramatically raises when compared to PBCP-215 (from 6.5 × 10 −8 to 3.7 × 10 −6 S.cm −1 at 50°C); this jump can be ascribed to the additional EC in the PIL block which increases the free volume and decreases the cohesive forces between PIL chains, resulting in enhanced ion dynamics. By adding more EC and keeping the same ratio of LiFSI/ PIL (PBCP-260), it is still possible to decrease the Tg of the PIL block down to −30.8°C, reflecting an increase in conductivity from 3.7 × 10 −6 to 1.0 × 10 −5 S.cm −1 when compared to sample PBCP-240 at 50°C.
After assessing the role of EC on both Tg and conductivity, a second step based on the increase of LiFSI amount to the system was also investigated; this was an attempt to enhance the electrochemical performance of the membranes. In Fig. 1c, the DSC traces of samples PBCP-460 and PBCP-480 are presented. By increasing the ratio LiFSI/ PIL from 2 to 4 mols/PIL and keeping EC/PIL = 6 (PBCP-460), the Tg of the PIL slightly increases in comparison to sample PBCP-260. This suggests that the coordination environment has possibly changed due to the higher amount of lithium ions; higher concentration of lithium likely results in possible formation of aggregates which in turn contributes to the increase of the Tg. Although the measured Tg of sample 4 is slightly higher, the conductivity increases to reach a value of 1.34 × 10 −5 S.cm −1 at 50°C. By adding more EC (sample 5 -PBCP-480), the Tg of the PIL block further decreases and reaches the lowest value among all samples (−31.0°C). Due to the enhanced ion dynamics provided by extra EC molecules, the conductivity increases from 1.34 × 10 −5 S.cm -1 (PBCP-460) to 4.0 × 10 −5 S.cm -1 (PBCP-480) at 50°C. Higher amounts of added EC were also attempted in this system but this led to SPE membranes with poor mechanical integrity and were not pursued further.
Pulse field gradient nuclear magnetic resonance measurements (PFG NMR).-In order to investigate the diffusivity of both lithium ion and FSI/TFSI in PBCPs, NMR pulse field gradient self-diffusion was performed at temperatures ranging from 30°C-90°C. After fitting the spectra, 7 Li and 19 F diffusion coefficients were obtained and the results are depicted in Fig. 2. As a first screening of the results, it is interesting noting that the diffusion of Lithium species is usually faster than both FSI and TFSI species. This outcome reveals that the transport of lithium is facilitated in the ternary system and high lithium transport number is expected. As can be seen from Fig. 2a, the Li diffusion coefficient for PBCP-215 shows the smallest values for all temperatures, this may be due to the lower amount of added EC; such a low EC content allows an efficient interaction (coordination) of EC molecules with LiFSI salt and the ionic block of the polymer, resulting in a higher Tg of the PIL block and hence a more restricted ion mobility. Following this interpretation, as more EC is added to the system, the Li diffusivity tends to increase, mainly ascribed to the lower Tg of the PIL block as evidenced by DSC results, which enhances ion dynamics. The results obtained from PFG-NMR supports this trend and lithium moves faster when both temperature increases and the Tg of the PIL block decreases. Regarding the diffusion coefficients of fluorine-derived species (FSI/TFSI) presented in Fig. 2b, some diffusion values can be noted missing; this is due to the fast T2 relaxation time which does not allow the measurement of the diffusion coefficient using this technique. This is especially true at lower temperatures. Overall, the 19 F (FSI) follows the similar behavior as discussed for lithium species, but mostly showing smaller diffusion coefficients than Li species (when compared at same temperatures). Interestingly, for sample BPCP-215, 19 F (TFSI) diffusion shows higher mobility than 19 F (FSI) at 70°C, this suggests that for high concentration of LiFSI salt, the TFSI anions, originally from the polymer backbone, tend to move more freely and interact less with the imidazolium cation.
Small angle scattering X-ray scattering (SAXS).-The macrophase separation present in the PBCP solid electrolytes was analyzed by small-angle X-ray scattering at room temperature. As can be seen in Fig. 3, the neat PBCP exhibits two main reflections peaks at q = q * , 3q * , where q = (4π/λ), characteristic of a lamellar morphology. In all PBCP electrolytes, the addition of LiFSI and EC results in the formation of a less ordered lamellar morphology, with respect to the neat PBCP, as suggested by the broadening of the reflections peaks. Additionally, a reduction of the repeat distance, d, going from 15.8 to 13.6 nm, is observed, when comparing the neat PBCP to the PBCP-215. This reduction in d is likely due to the formation of dynamic ionic interchain crosslinks between the lithium ions and the PIL chains. This dynamic ionic crosslink phenomena was already observed in a ternary polymer electrolytes system containing the presently studied PBCP, an ionic liquid and the LiFSI salt. 33 Such a dynamic ionic crosslinking effect is again observed, with increasing of the LiFSI content of the PBCP electrolyte from 2 to 4 mol/PIL units (PBCP-260 vs PBC-P460), while keeping the EC content constant. This also coincides with an increase in the Tg associated with the PIL phase (See Fig. 1). In contrast, an increase of the repeat distance, d, going from 13.6 nm to 14.6 nm, is observed when increasing the EC content, from 1.5 to 6 mols/PIL units and keeping the LiFSI content fixed at 2 mols/PIL units (PBCP-215 vs PBCP-260). This increase in d suggests the disruption of the dynamic ionic crosslinks present in these system, likely due to the solvation of the Li ions by the EC molecules. Such disruption of the dynamic ionic crosslinks is likely to result in an increase in the segmental mobility of the PIL chain, which is supported by the DSC results, where a decrease of the glass transition temperature associated with the PIL phase is observed.
Electrochemical properties.-Symmetrical cells-Electrochemical Impedance Spectroscopy (EIS), galvanostatic charge-discharge and transport number.-The electrochemical performance of the solid polymer electrolytes was first assessed by assembling Li|PBCPs|Li symmetrical cells and analyzing the impedance of cells after 24 h rest. As shown in Figs. 4a and 4b, the Nyquist plot of all samples reveal the presence of two semi circles in the whole frequency range. These semicircles at high frequencies are ascribed to the bulk resistance of the PBCP whilst the smaller semi-circles at medium-low frequencies are related to the interfacial resistance (electrode-electrolyte interface). The Nyquist plot of all samples were fitted according to the equivalent circuit shown in Fig. 4c. It is clear that the addition of EC plays a role and effectively decreases the bulk resistance of the electrolyte as well as the interfacial resistance (refer Table SI (available online at stacks. iop.org/JES/167/070525/mmedia) for fitted parameters). When comparing samples PBCP-460 and PBCP-480 (Fig. 4b) with samples PBCP-215, PBCP-240 and PBCP-260 (Fig. 4a), the higher amount of Li salt (ratio LiFSI/PIL) was also shown to be important and lower resistances are observed. This first electrochemical assessment is in a good agreement with conductivity and DSC measurements shown in Fig. 1. In addition, the lithium transference number was also calculated by applying the Bruce-Evans-Vincent method. As shown in Fig. S1a, the Li transport numbers of all PBCPs are much higher when compared to the well-known PEO-based solid polymer electrolytes found in literature, except those based on single ion conductors. 25 Interestingly, sample 1 (PBCP-215) shows a quite high transport number (t Li + = 0.72); this is likely to the high concentration of LiFSI salt in the polymer (refer Table I), which is the highest among all samples. Indeed, the Li transport numbers of PBCPs are strongly dependent on the amount of LiFSI added, this behavior is clearly   shown in Fig. S1b. Despite the high t Li + of samples bringing 2 mols LiFSI/PIL (PBCPs-2XY); due to their low ionic conductivity and high interfacial resistance, they are obviously not interesting for applications in full cells. Overall, all PBCPs electrolytes showed high Li transport numbers (t Li + > 0.5) and solid polymer electrolytes that have 4 mols of LiFSI/PIL (PBCPs-4XY) are considered herein excellent solid polymer electrolytes and will be investigated with more detail.
The overpotential of lithium stripping/plating and the transport properties of lithium using PBCP as a solid polymer electrolyte material was evaluated by applying a constant current (1 h step) for both charge and discharge for 100 cycles (200 h polarization). As a standard, we concentrated on achieving a reliable cycling by using a current density of 0.1 mA.cm −2 at 50°C. Due to both high bulk and interfacial resistance of samples with 2 mols of LiFSI/PIL (PBCPs-2XY), some of them were limited to current densities lower than 0.1 mA.cm −2 (Fig. S2). As can be seen from Fig. S2, sample PBCP-215 can only be cycled at very low current density (0.02 mA.cm −2 ) and fails when the current is increased to only 0.04 mA.cm −2 . Sample PBCP-240 shows a stable cycling up to 0.08 mA.cm −2 , but shows instabilities when the current is raised to 0.1 mA.cm −2 . Conversely, sample BCP-260 show an improved cycling and is stable at 0.1 mA.cm −2 , however, the overpotential seems to be relatively high and reaches a value of 280 mV. In Fig. 2d, the long term stripping and platting of lithium reveals a rather stable behavior of the polymer electrolytes (PBCP-460 and PBCP-480) with a smooth cycling. An overpotential of 190 mV at 0.1 mA.cm −2 is observed for PBCP-460 at 50°C and this value negligibly changes until the end of the cycling. Due to a lower impedance of PBCP-480, the cycling evidences an overpotential as low as 74.5 mV for this sample. In Fig. 2e, a maximization of the cycling further compares both samples; it is interesting to see that in the case of sample PBCP-480, apart from having a lower overpotential, it also shows a superior profile; this behavior is clearly seen during plating and stripping. During cycling, PBCP-460 shows a linear increase in its overpotential (during each step), likely due to some mass transfer limitations, reflecting the pronouncement of concentration polarization. On the other hand, sample 5 (PBCP-480) tends to form a plateau around 74 mV and its potential negligibly changes during the 1 h steps; this behavior evidences a superior electrochemical performance of PBCP-480 membrane. For this reason, the membrane PBCP-480 was selected as a solid polymer electrolyte to further assemble a lithium metal battery by using LiFePO 4 as cathode material.
Full all-solid-state Lithium-metal battery tests.-To evaluate the performance of the polymer electrolyte (PBCP-480), full cells were assembled comprising lithium metal as anode, LiFePO 4 (LFP) as cathode and PBCP-480 as both solid polymer electrolyte and separator. These all-solid state full cells were cycled at both 50°C and 70°C and the results are exhibited in Fig. 5. At 50°C (Fig. 5a), the voltage profiles show a flat charge-discharge curve and the first cycle reaches a specific capacity of 154 mAh.g −1 at C/20 C-rate. As the cycling goes on, the cell shows a very stable behavior and only a very small capacity fading is observed in the first 25 cycles; the coulombic efficiency, as shown in Fig. 5b remains at 99.8%. After that first stage of cycling, the C-rate was increased to C/10 for further 100 cycles in order to assess the long-term stability of the battery. By analyzing the voltage profiles in Fig. 5a, a small polarization effect can be seen from the 25th to 125th cycle, however, the specific capacity at the end of the 125th cycle is 150 mAh.g −1 , this value represents a capacity retention as high as 97.4% and a capacity fading of only 0.02% per cycle is calculated. Figure 5b further highlights the efficient cycling of the battery with a coulombic efficiency of 99.8%-99.9%; this successful long-term cycling reinforces the suitability of EC in this particular ternary system and can be potentially used in other ternary systems. Moreover, the ternary system devised in this work clearly shows that a good SEI layer has been formed; such effective and homogeneous SEI layer on lithium metal prevents continuous electrolyte degradation and low capacity fading is observed in a long-term. It is worth mentioning that the performance of the solid electrolyte may be enhanced by using, for example, a different polymerized ionic liquid in the block copolymer, this will be the topic of a future work from our group.
When the battery is cycled at 70°C using C/20 C-rate, a specific capacity of 161 mAh.g −1 is obtained at the first cycle (Fig. 5c). Interestingly, this value remains stable with no apparent capacity fading up to 10 cycles. This high capacity close to the theoretical value is comparable or superior to results reported in literature for polymer electrolyte-based Lithium-metal batteries at T ⩾ 50°C. [40][41][42][43][44] It is worth noticing that the overpotential at 70°C is smaller when compared to the cell at 50°C; this is likely due to the higher conductivity of the solid electrolyte at higher temperature and therefore a more efficient cycling is expected. As shown in Fig. 5d, after 10 cycles, the rate is increased to C/10 and the cycling is performed for additional 20 cycles. Interestingly, the specific capacity does not decrease when compared to the cycling at C/20 Crate and a small polarization effect in the voltage profile (Fig. 5c) is observed when compared to C/20 C-rate. This behavior again confirms the efficient Li transport observed in Li|Li symmetrical cells presented in Fig. 4e.
The cycling rate is an important parameter in batteries since high current densities are sometimes crucial for particular applications, and rates of C/20 and C/10 may be only suitable for applications where fast charge-discharge is not paramount. In order to assess the rate capability of our all-solid-state batteries, cells at both 50°C and 70°C were cycled at higher rates (up to C/3). In Fig. S3a, the voltage profiles of a battery operating at 50°C are presented. As can be seen, at C/5 C-rate, the battery is still able to deliver a capacity of 100 mAh.g −1 but a stronger polarization effect is observed, which leads to a sloppy charge discharge curve; at C/4, the capacity drops to 60 mAh.g −1 , suggesting that at 50°C the battery is limited to operate at C-rates < C/4. On the other hand, at 70°C (Fig. S3b), the battery cycles very well at C/5 and brings a flat charge-discharge curve, delivering a capacity as high as 150 mAh.g −1 . At C/4, a small polarization effect is observed due to the higher applied current, but surprisingly, the capacity is still maintained at 150 mAh.g −1 . At C/3, the battery maintains a flat and smooth charge-discharge and a specific capacity of 138 mAh.g −1 is obtained. The superior performance at elevated temperature is basically due to the enhanced conductivity of the polymer electrolyte, as already discussed previously. Overall, our solid polymer electrolyte PBCP-480 has been shown to be very stable in LFP cells and the employment of EC as a cheap and sustainable additive was successfully demonstrated.

Conclusions
In summary, we have investigated the application of ethylene carbonate (EC) as a cheap, widely available and benign material to tune the properties of a polymerized ionic liquid block copolymer by developing a ternary system where LiFSI is used as lithium salt. After studying and optimizing the system by using several compositions involving different amounts of both EC and LiFSI, it was found that EC effectively decreases the Tg of the PIL block and therefore superior ion dynamics are achieved. The best solid electrolytes were obtained when the ratio EC/LiFSI~2 and the properties are enhanced as the ratio LiFSI/PIL increases until it reaches the limit of the mechanical stability of the flexible solid polymer membrane electrolyte. The best solid electrolyte was characterized and its suitability as a solid electrolyte membrane for lithium metal batteries was demonstrated. Particularly in full cells using LFP as cathode and lithium metal as anode, the batteries showed excellent cyclability and stability with a negligible capacity fading at both 50°C and 70°C, meaning that EC is an excellent alternative to expensive and harmful solvents.