当前位置: X-MOL 学术Solid State Sci. › 论文详情
Our official English website, www.x-mol.net, welcomes your feedback! (Note: you will need to create a separate account there.)
Predicting lattice parameters for orthorhombic distorted-perovskite oxides via machine learning
Solid State Sciences ( IF 3.5 ) Pub Date : 2021-02-02 , DOI: 10.1016/j.solidstatesciences.2021.106541
Yun Zhang , Xiaojie Xu

Orthorhombic perovskite oxides, ABO3, have been studied extensively for application in advanced functional materials and devices, due to unique and versatile materials properties that arise from stoichiometry engineering. The lattice parameters, a, b, and c, have a critical impact on the structural stability, electronic structure, magnetic ordering, and thus materials performance. In this study, we develop the Gaussian process regression (GPR) model based on the Bayesian optimization to provide statistical relationships among the ionic radii, electronegativities, valence, and lattice constants for orthorhombic perovskite ABO3 compounds with 157 samples. The model shows a high degree of stability and accuracy. After being applied to predict the lattice parameters of doped- and non-synthesized oxides, the GPR model shows promising prediction power as well. The model may be used as an alternative method to obtain lattice parameters as compared to experimental approaches and other modeling methods.



中文翻译:

通过机器学习预测正交晶形畸变钙钛矿氧化物的晶格参数

斜方钙钛矿氧化物,ABO 3,由于化学计量工程学产生的独特和通用的材料特性,已经广泛研究了其在高级功能材料和设备中的应用。晶格参数a,b和c对结构稳定性,电子结构,磁有序性以及材料性能具有至关重要的影响。在这项研究中,我们开发基于贝叶斯优化的高斯过程回归(GPR)模型,以提供正交晶体钙钛矿ABO 3的离子半径,电负性,化合价和晶格常数之间的统计关系157个样品。该模型显示出高度的稳定性和准确性。在用于预测掺杂和非合成氧化物的晶格参数后,GPR模型也显示出有希望的预测能力。与实验方法和其他建模方法相比,该模型可以用作获得晶格参数的替代方法。

更新日期:2021-02-07
down
wechat
bug