■ 제 목: Machine-learning (ML) guided reaction prediction in energy storage devices
■ 연 사: 이병주 박사 (Lawrence Berkeley National Laboratory)
■ 일 시: 2020년 8월 4일(화) 오후 2시
■ Abstract : The stability and/or compatibility between chemical components are one of the essential parameters that need to be considered in the selection of functional materials in the system configuration. In configuring devices such as batteries or solar-cells, not only the functionality of individual constituting materials such as electrodes or electrolyte, but also a proper combination of materials which do not undergo unwanted side-reactions is critical in securing the reliable performance in the long-term operation. While the universal theory that can predict the general chemical reactivity between materials is long awaited and has been the subject of studies with a rich history, traditional ways proposed to date have been mostly based on simple electronic properties of the material such as the electronegativity, ionization energy, electron affinity and hardness/softness, and could be applied to only small group of materials. Moreover, the prediction has often far from the accuracy and failed to offer the general implications, thus they were practically inadequate as the selection criteria from a large material database, i.e. data-driven material discovery. Herein, we propose a new model for predicting a general reactivity and chemical compatibility among a large number of organic materials, realized by a machine-learning approach. As a showcase, we demonstrate that our new implemented model successfully reproduces the previous experimental results reported on side-reactions occurring in lithium-oxygen electrochemical cells. Furthermore, the map of the chemical stability among more than 90 available electrolyte solvents and the representative redox mediators is provided from this approach, presenting an important guideline in the development of stable electrolyte/redox mediator couples for lithium-oxygen batteries.