有机无机杂化钙钛矿体系中有机组分对材料性质影响机制的理论研究

快速开展高性能功能材料搜索和设计，有赖于对材料构效关系的深刻理解。本项目拟结合密度泛函理论和机器学习方法，以有机无机杂化钙钛矿体系为研究对象，发展一套能够有效剖析复杂材料体系构效关系且具有泛化能力的理论方案。虽然有机无机杂化钙钛矿材料在光伏洁净能源领域应用潜力巨大，但自身毒性和不稳定性却严重制约其广泛应用，且有机组分对材料性质及稳定性作用机制仍不明晰。针对原子信息描述符无法准确描述具有复杂结构的有机组分这一现实，本项目拟从有机组分角度出发，深入剖析有机组分内禀属性与材料性质及结构稳定性之间的深层次关联，进而构建准确表征材料性质的描述符和探寻有效界定材料稳定性的全新判据，以期加速材料搜索和设计。此外，本项目拟透过同分异构体系横向比较揭示有机组分对材料性质作用机理，以突破有机阳离子彼此巨大差异化所限制纵向比较分析的藩篱，并有助提升数据集质量和多样性，助力机器学习模型构建及深层次关联剖析。

Rapidly discovering and designing high-performance functional materials seriously relies on a deep understanding of structure-property relationships. In this project, combining density functional theory calculations with machine learning techniques, a strategy with high generalization ability intends to be proposed to effectively establish structure-property relationships of hybrid organic-inorganic perovskites. Hybrid organic-inorganic perovskites, as one of the most promising materials for photovoltaics, possess a broad application prospect in the field of clear energy, however two key challenges of intrinsic instability and toxicity issues restrain its large-scale commercial applications and the underlying physics of organic components on the material properties are still unclear. Moreover, due to its complex structure, it hard to rely on atomic descriptors to accurately define organic cations. From the perspective of organic components in hybrid perovskites, the proposed theoretical method here is introduced to reveal the correlations between the intrinsic properties of organic components and the electronic properties and structural stability, thus construct a set of valid descriptors for accurately characterizing electronic structural properties and look for effective and innovative criterions to evaluate the structure stability, in order to accelerate the discovery and design of promising hybrid perovskites for photovoltaics. In addition, the microscopic mechanism of organic cations on electronic structural properties and stabilities here intend to be investigated via introducing comparisons between isomeric perovskites, which is expected to overcome the traditional analysis difficulties induced by giant difference between organic cations. Moreover, the improvements of quality and diversity of dataset here also benefit to promote machine learning model construction and intrinsic correlation analysis.