数据科学范式下的钙钛矿结构铁电新材料研究

Design of new perovskite-type oxide ferroelectric materials within data science paradigm

材料基因组计划倡导预测式新材料研发理念,推进高通量数据生产和利用技术,关注材料全生命周期价值。因此,材料基因组计划的执行需要在材料科学系统工程的框架下,集成统一计算、实验和理论等研究方法,以数据科学新范式为牵引、协同运用实验观测、理论建模和计算仿真研究范式,最终建立相关材料体系的性能与材料基因(原子系统的组成与结构)、工艺参数与使役条件之间的量化关系和数据库,实现新材料的按需设计和应用。本文在简单探讨科学研究范式、材料基因组计划和材料科学系统工程基本概念和方法的基础上,以钙钛矿结构氧化物铁电压电材料研究为例,探讨了数据科学范式下的新材料研究实践。结果表明,数据挖掘驱动的新材料设计确实可以降低探索时间和实验任务,加快新材料的发现和应用进程。

The materials genome initiative(MGI) issued in 2011 convokes a new paradigm of data science complementing with empirical observation, theoretical model and computation simulation, and comprehensively integrates computing, experimental and theoretical methods to produce and deal with big-data under the framework of system engineering of material science. Thereafter, the relationships between properties and material genome(composition and structure of atom systems), processing parameters and service conditions are mined out of data for designing and deployment of new material in accordance with the desired goal. In this article, research paradigms, MGI and system engineering of material science are briefly introduced. Then how to design new materials within data science paradigm is presented in detail through an example in the field of perovskite-type oxide ferroelectric piezoceramics. Finally the result demonstrates that the method of data-mining driven designing within data science paradigm is able to reduce time-to-insight and human effort on synthesis, thus accelerating new materials discovery and deployment.