Physically inspired atom-centered symmetry functions for the construction of high dimensional neural network potential energy surfaces

doi:10.1016/j.commatsci.2020.110071

IMR OpenIR

	Physically inspired atom-centered symmetry functions for the construction of high dimensional neural network potential energy surfaces
	Zhang, Kangyu 1,2; Yin, Lichang 1,2; Liu, Gang 1,2
通讯作者	Yin, Lichang(lcyin@imr.ac.cn)
	2021
发表期刊	COMPUTATIONAL MATERIALS SCIENCE
ISSN	0927-0256
卷号	186 页码:7
摘要	Among different atomistic neural network (AtNN) potential energy surfaces (PESs), the Behler-Parrinello neural network (BPNN) based on atom-centered symmetry functions (ACSFs) has been proved to be capable of constructing accurate PESs for various crystals. A judicious setting of the parameters of the ACSFs largely determines the accuracy of a BPNN PES. However, this is typically an ad hoc and tedious task requiring highly acute chemical intuition. To address this issue, we derived a set of physically inspired ACSFs from the effective densities of atoms, in which the radii of atoms are naturally incorporated. Therefore, the parameters of the physically inspired ACSFs can be directly chosen based on the types of chemical bonds within a target system. Compared with the original ones, the physically inspired ACSFs are more suitable for complex systems based on its better performance on predicting the formation enthalpies of molecules in QM9 database. Moreover, the physically inspired ACSFs can also effectively accelerate the convergence of the atomic forces during the training of an AtNN PES. With the physically inspired ACSFs, we constructed a highly accurate AtNN PES for a solid electrolyte Li10GeP2S12. Based on the AtNN PES, we studied the bulk Li ion diffusion within Li10GeP2S12 by molecular dynamics (MD) simulations. The MD results well reproduced the experimental results, indicating the high accuracy of the AtNN PES constructed with the physically inspired ACSFs.
关键词	Machine learning Potential energy surface Atom centered symmetry function Solid electrolyte Molecular dynamics simulation
资助者	National Natural Science Foundation of China (NSFC)
DOI	10.1016/j.commatsci.2020.110071
收录类别	SCI
语种	英语
资助项目	National Natural Science Foundation of China (NSFC)[51972312] ; National Natural Science Foundation of China (NSFC)[51472249]
WOS研究方向	Materials Science
WOS类目	Materials Science, Multidisciplinary
WOS记录号	WOS:000594489800005
出版者	ELSEVIER
引用统计	被引频次：10[WOS] [WOS记录] [WOS相关记录]
文献类型	期刊论文
条目标识符	http://ir.imr.ac.cn/handle/321006/158753
专题	中国科学院金属研究所
通讯作者	Yin, Lichang
作者单位	1.Chinese Acad Sci, Inst Met Res, Shenyang Natl Lab Mat Sci, 72 Wenhua Rd, Shenyang 110016, Peoples R China 2.Univ Sci & Technol China, Sch Mat Sci & Engn, 72 Wenhua Rd, Shenyang 110016, Peoples R China
推荐引用方式 GB/T 7714	Zhang, Kangyu,Yin, Lichang,Liu, Gang. Physically inspired atom-centered symmetry functions for the construction of high dimensional neural network potential energy surfaces[J]. COMPUTATIONAL MATERIALS SCIENCE,2021,186:7.
APA	Zhang, Kangyu,Yin, Lichang,&Liu, Gang.(2021).Physically inspired atom-centered symmetry functions for the construction of high dimensional neural network potential energy surfaces.COMPUTATIONAL MATERIALS SCIENCE,186,7.
MLA	Zhang, Kangyu,et al."Physically inspired atom-centered symmetry functions for the construction of high dimensional neural network potential energy surfaces".COMPUTATIONAL MATERIALS SCIENCE 186(2021):7.