负热膨胀材料是体膨胀系数为负值的一类材料。根据产生的物理机制可分为声子驱动和电子结构相变驱动——两类负热膨胀材料。声子驱动负热膨胀来源于晶格中一些特殊的声子振动行为。过渡金属氧化物结构中普遍存在多面体结构单元，其中多面体间的共顶原子横向振动是声子驱动负热膨胀的主要来源。探索多面体结构单元种类、组合及堆垛方式与热膨胀之间的内在关联，是此类负热膨胀材料设计中亟待解决的关键问题。本项目拟以发现新型负热膨胀材料为目标，发展可以高效处理大量声子计算的负热膨胀材料设计方法；应用发展的方法在过渡金属氧化物体系中，开展负热膨胀材料高通量计算，获得具有大的负热膨胀系数和宽温区的新型负热膨胀材料。同时，采用机器学习算法，挖掘材料设计过程中产生的大量数据，获取过渡金属氧化物体系“组分-结构-热膨胀性质”对应关系背后的物理机制，以及建立负热膨胀性质调控的规律性认知，为负热膨胀材料的开发及应用提供知识储备。

Negative thermal expansion (NTE) materials are a class of materials with negative bulk expansion coefficients. Two fundamental mechanisms responsible for NTE are phonon-driven and electronic transition-driven. The phonon-driven NTE comes from some special vibration behaviors of phonons in the lattice. Polyhedral structural units are ubiquitous in transition metal oxides, in which the transverse vibration of corner-sharing atoms between polyhedrons is the main source of phonon-driven NTE. Exploring the inherent relationship between the types, combinations, and stacking of polyhedrons and the thermal expansion property is a key issue that needs to be resolved in the material-by-design study of the NTE materials. This proposal aims at discovering new NTE materials. We propose to develop a new material-by-design approach for NTE, in which the performance bottleneck problem caused by massive phonon calculations will be effectively solved. Then we will carry out material-by-design study (high-throughput calculation) of transition metal oxides by employing our developed method. We target at finding new NTE materials displaying large negative thermal expansion coefficient and wide temperature region. At the same time, by deploying machine learning algorithms to learn the data generated by the above calculations, we will unravel the physical mechanism underlying the component-structure-thermal expansion property relationship in the transition metal oxides. We also establish a uniform understanding of the physical principles determining how the NTE is deliberately modified in transition metal oxides. This is expected to provide useful knowledge accumulation for the application of NTE materials.