(Invited) Progress in Polymer Electrolytes for Li° Electrode

Abstract

The incentives for the use of lithium metal negative electrode are easily justified considering the gain in energy density. Polymer electrolytes disclosed almost fifty years ago are still considered as an option for the most thought-for all solid-state systems. In fact, the only solid-state batteries commercialised are those from Blue Solutions® powering busses in Europe, using a Li° negative, a PEO [poly(ethylene oxide)] matrix with a LiTFSI solute and an LiFePO4 cathode materials and operating at 70°C. 1500 -2000 cycles are routinely obtained. The main drawbacks of PEO, which is stable to lithium metal is a low conductivity at room temperature (10-5 - 10-6 Ohm-1.cm-1 at 25°C) while it reaches easily 10-3 - 10-4 Ohm-1.cm-1 at 70°C. The other impedimenta of PEO is the relatively low transference/transport number of the Li+ cation (T+ ≈ 0.2), resulting in the formation of a salt concentration gradient in the electrolyte upon operation, with concomitant polarisation and the triggering of dendrites when the Sand time is reached, shortening the cycle life-time. Other polymer matrices like poly(ε-caprolactone) PCL and poly(propylene carbonate) PPC show slightly better conductivity and higher transference number (T+ ≈ 0.5), but are not stable in contact with lithium metal.Our strategy to increase the conductivity at lower temperature consisted in modifying the polymer architecture, from linear (PEO) to comb polymer in which medium length PEG segments (Mw 1000 - 2000) are attached to the main backbone via a flexible non-solvating tether (PPO). This optimises the speed of re-orientation of the solvating units, speed needed to favour the ionic motion. With these polymers and delocalised salts like LiTFSI or LiFSI, conductivities close to 10-4 Ohm-1.cm-1 at 25°C are obtained though with the same transference number as PEO.In order to improve the selectivity of the cation transport in PEs, it is necessary to immobilise the negative charges of the salt by attaching them to a polymer or nano-particles. The other option consists in modifying the salt architecture to slow down the diffusion/migration of the anions by increasing the interactions of the anions with themselves or with the polymer backbone.We have found that, starting for a simple cellulose derivative, the action of FSO2N=C=O results in the attachment of anions to the backbone and that the poly(salts) are now able to form an alloy with PEO or other poly(ethers), with excellent mechanical properties imparted by the rigid cellulose matrix, yet giving Li-only conductivities allowing battery operation. Alternatively, SiO2 or Al2O3 nano-particles can be grafted with delocalised anionic moieties, which result, when dispersed into PEO, in conducting composites with again excellent electrochemical and mechanical properties.Modification of the well-known anion TFSI to increase its interactions and slow it down have been undertaken: either hydrogen bonds, dipole interactions, chain entanglement, or π - π interactions have been introduced in the RSO2N(-)SO2CF3 anion with R equal respectively CF2H—, (CH3)2N—, (CH3OC2H4)2N—, C6H5—. In all cases, higher T+ were obtained as compared with TFSI-, though the total conductivity was lower, but resulting still in σLi+ that translates into better lithium plating efficiencies as seen in the classical Li°/PE/Li° cycling tests and in batteries lifecycle.A final drawback of PEO is its inability to withstand voltages above 3.9 V vs. Li+:Li° while PPC for instance, stable to 4.5V, does decompose rapidly on the negative Li°. The obvious idea to combine a PEO based anolyte and a PPC catholyte stumbles on the consideration that PEO is far more solvating than PPC and deplete totally the polyester compartment, with no conductivity remaining. We solved this problem with the use of an immobilised polyanion in both the anolyte and catholyte, resulting in the stable operation of a Li°/NMC622 battery.All these research results will be discussed in details.