Stefan Adams1,Nahong Zhao1,Aniruddh Ramesh1,Rayavarapu Prasada Rao1
National University of Singapore1
Stefan Adams1,Nahong Zhao1,Aniruddh Ramesh1,Rayavarapu Prasada Rao1
National University of Singapore1
In order to realize the promise of superior energy density, the solid electrolytes in solid state batteries have to combine fast ionic conductivity with a low contribution to the overall cell mass. While the dead mass of the “separator” solid electrolyte layer can be minimized by reducing its thickness to the extent limited by operational safety and processability, in high energy density batteries 80-90% of the electrolyte in the cell will be the catholyte material within the inevitably thick composite cathodes. Thus, in the absence of fast-ion conducting high energy density cathode materials, low density catholytes are of key importance for the achievable energy density of all-solid state batteries. Additional constraints for low-density fast-ion conducting solids include of course electrochemical and electromechanical compatibility, as well as interface wettability with the active materials.<br/>Here we discuss examples for two design strategies to realize such low density fast-ion conductors. The first strategy involves ultrathin ceramic-in-polymer Composite Solid State Electrolytes (CSSEs). We developed CSSEs combining Li-doped polyacrylonitrile with various ceramic electrolytes (Li<sub>1.5</sub>Al<sub>0.5</sub>Ge<sub>1.5</sub>(PO<sub>4</sub>)<sub>3</sub>, LiTa<sub>2</sub>PO<sub>8</sub> etc.) with and without plasticizers. Optimized composites reach conductivities of 0.1 – 1 mS cm<sup>-1</sup> at densities < 2 g cm<sup>-3</sup> and retain wide electrochemical windows. Temperature-dependent analysis of Li<sup>+</sup> paths shows that the ceramic filler induces a percolating network of fast ion-conductive interphases below the glass transition of the bulk polymer. Spin coating and in situ UV-curing the precursor slurry on Li anode or composite cathode drastically reduces interfacial resistance. The electrochemical and mechanical performance of CSSEs are investigated. Symmetric cells cycled with up to 2 mA cm<sup>-2</sup> over 500h retain a stable overpotential and no dendrite growth is observed. LFP/SSE/Li cells achieve a maximum specific discharge capacity of 148 mAh g<sup>-1</sup> and retain 89% of that after 200 cycles. Stable room temperature cycling is demonstrated for cells with a cathode mass loading of 6 mg cm<sup>-2</sup>. The strong binding and 3-dimensional architecture of the in situ-formed interfaces is crucial for the favourable cycling stability.[1]<br/>An alternative all-ceramic strategy is to systematically explore the correlation between fast ionic conductivity, density and glass forming ability in crystalline and glass-ceramic oxides and oxyhalides [2]. The higher number of monovalent light halides required to balance the charges of (typically heavier) high valent glass-former cations helps to reduce the overall density and turns the materials less brittle. Compared to often close-packed halides, oxyhalides typically have reduced coordination numbers, which helps to create free volume for ionic motion. In view of their favorable oxidation stability yet limited reduction stability, oxyhalides should in particular be suitable for use as catholytes. The recent report of fast room temperature ionic conductivity of 10 mS cm<sup>-3</sup> in LiMOCl<sub>4</sub> (M=Nb,Ta) [3] exemplifies the strong potential of the so far hardly explored class of oxyhalides. While the phases have originally been reported as orthorhombic and isostructural to moderate ion conductor LiVOF<sub>4</sub>, our redetermination of LiNbOCl<sub>4</sub> (density 2.7 g cm<sup>-3</sup>) finds a simpler, but Li-disordered, tetragonal crystal structure with smaller unit cell. The revised structure model yields more plausible bond valence sums and is consistent with DFT energy minimizations. Molecular dynamics simulations utilizing our bond-valence related embedded-atom method type forcefield softBV-EAM reveal the ion transport mechanism and its relation to the local rotational mobility of the (MOCl<sub>4</sub><sup>–</sup>)<sub>n</sub> chains [4].<br/><br/>References:<br/>[1] L. He et al. <i>Energy Stor. Mater.</i> 60 (2023) 102838.<br/>[2] R. Dai et al. <i>Chem. Mater.</i> 34 (2022) 10572.<br/>[3] Y. Tanaka et al. <i>Angew. Chem. Int. Ed.</i> 62 (2023) e202217581.<br/>[4] A. Ramesh et al. <i>in preparation</i>.