开发表面原子识别模型，对单个TiO2颗粒升温烧结过程中表面原子进行分类研究。模型对颗粒空间立方体网格化，利用标准球形颗粒体积积分法确定最佳网格尺寸为0.3 nm。通过表面网格识别实现表面原子分类，采用近邻网格中外部网格数量(Next)作为准则数判断目标网格是否为表面网格，确定最佳Next=9。基于LAMMPS软件模拟了半径为0.75 nm颗粒的升温过程，发现系统能量弛豫速度明显高于结构弛豫速度；利用表面识别模型分类分析原子特性，表面原子平均位移大于内部原子，且表面O原子迁移活性高于Ti原子；表面原子配位数低于内部原子，佐证表面结构规律性较差。研究结果为深入分析纳米材料活性位等结构分布奠定基础。
In this paper, we develop a surface atom identification model to perform atom identification and analysis of the surface atoms of a single TiO2 nanoparticle during the heating and sintering process. Cubic mesh was used to obtain the particle structure mesh, and the optimal mesh size of 0.3 nm was determined by volumetric integration of a spherical particle. Surface atoms were classified according to identification of surface meshes, and the number of external meshes in all of the neighbor meshes (Next) was set as a criterion to determine whether the target mesh was a surface mesh. The optimal value was Next=9. LAMMPS was used to simulate the heating process of a particle with a radius of 0.75 nm. The results show that energy relaxation is significantly faster than structure relaxation. The atom classification analysis with the developed surface atom identification model shows that displacement of the surface atoms is larger than the interior atoms, and surface O atoms are more active in migration than surface Ti atoms. The coordination number of surface atoms is lower than that of interior atoms. The present study provides fundamental information for analyzing the active structure distribution of nanoparticles