ИСТИНА

All-solid-state batteries (ASSBs) are the objects of increased research interest as a one of the most prospective electrochemical energy storage technologies. Among the materials of solid-state electrolytes, oxides of garnet-type are the objects of the high research interest due to the number of the attractive characteristics such as the competitive ionic transport characteristics, wide electrochemical window, stability with Li and its alloys, the transference number close to unity, and the negligible values of the electronic conductivity. In this study, the role of the composition, disorder and the processing of the compounds is considered. The objects of the given study, garnet-structured solid electrolytes, adopt the structure of the highest group of symmetry. We consider the group of symmetry of the solid electrolyte material as one of the most important criteria for assessing the potential for enhancing the ion transport properties. It is shown that, in a case of the high symmetry garnet-type oxides, the entropy factor resulted in the ion transport enhancement. Based on the experimental data available, one can infer that the dopants of Li sites A and of C cation octahedrally coordinated site of the garnet structure described by the formula AxB3C2O12 may accomplish the different functions and, hence, the co-doping strategy is required: the former ones are responsible for Li ion re-distribution between the sites forming the 3D Li-ion conductivity paths while the latter ones are responsible for the size of the diffusion paths. The joint analysis of the Li-ion conductivity, activation energy and lattice constant values showed the highly probable kinetic limitations of Li-ion transport exist that may be neglected through the cation substitution whereas the impact of the In3+, Te6+, Hf4+, W6+ and Mo6+ substituents are not completely understood due to the conflict of enhanced conductivity values with high activation energy values. The predictive Machine Learning (ML) models have been developed involving different types of descriptors and methods that include Bayesian optimization, regression and chemography techniques. The synthesis data have been used as the descriptor parameters, the impact of the synthesis route for the final materials functional properties is discussed.