Machine learning approach for the prediction of electron inelastic mean free paths

被引:9
作者
Liu, Xun [1 ,2 ,3 ]
Yang, Lihao [1 ,2 ,3 ]
Hou, Zhufeng [4 ]
Da, Bo [3 ,5 ]
Nagata, Kenji [3 ]
Yoshikawa, Hideki [3 ]
Tanuma, Shigeo [3 ]
Sun, Yang [6 ]
Ding, Zejun [1 ,2 ]
机构
[1] Univ Sci & Technol China, Hefei Natl Lab Phys Sci Microscale, Hefei 230026, Anhui, Peoples R China
[2] Univ Sci & Technol China, Dept Phys, Hefei 230026, Anhui, Peoples R China
[3] Natl Inst Mat Sci, Res & Serv Div Mat Data & Integrated Syst, 1-1 Namiki, Tsukuba, Ibaraki 3050044, Japan
[4] Chinese Acad Sci, Fujian Inst Res Struct Matter, State Key Lab Struct Chem, Fuzhou 350002, Peoples R China
[5] Natl Inst Mat Sci, Res Ctr Adv Measurement & Characterizat, 1-2-1 Sengen, Tsukuba, Ibaraki 3050047, Japan
[6] Columbia Univ, Dept Appl Phys & Appl Math, New York, NY 10027 USA
基金
美国国家科学基金会;
关键词
ENERGY-LOSS SPECTRA; SURFACE EXCITATION PARAMETERS; EFFECTIVE ATTENUATION LENGTHS; MONTE-CARLO-SIMULATION; QUANTITATIVE-ANALYSIS; DIELECTRIC FUNCTION; ELASTIC-SCATTERING; REFLECTION; SOLIDS; SENSITIVITY;
D O I
10.1103/PhysRevMaterials.5.033802
中图分类号
T [工业技术];
学科分类号
08 ;
摘要
The prediction of electron inelastic mean free paths (IMFPs) from simple material parameters is a challenging problem in studies using electron spectroscopy and microscopy. Herein, we propose a machine learning (ML) approach to predict IMFPs from some basic material property data. The ML model showed excellent performance based on the calculated IMFPs for a group of 41 elemental materials [Li, Be, C (graphite), C (diamond), C (glassy), Na, Mg, Al, Si, K, Sc, Ti, V, Cr, Fe, Co, Ni, Cu, Ge, Y, Nb, Mo, Ru, Rh, Pd, Ag, In, Sn, Cs, Gd, Tb, Dy, Hf, Ta, W, Re, Os, Ir, Pt, Au, and Bi] from a previous paper [Shinotsuka et al., Surf. Interface Anal. 47, 871 (2015); 47, 1132 (2015)]. which was comparable with that of the robust Tanuma-Powell-Penn (TPP-2M) formula. The developed ML model was then extended to materials that do not have reported IMFPs in the database. The IMFPs for 18 transition and lanthanide metals (Mn, Zn, Zr, Tc, Cd, La, Ce, Pr, Nd, Pm, Sm, Eu, Ho, Er, Tm, Yb, Lu, and Hg) were predicted by theML model. In the comparison with full-Penn algorithm-calculated IMFPs through two newly found experimental energy loss functions (ELFs), i.e., Mn and Zr, the Gaussian process regression-predicted IMFPs not only agreed well with those calculated using the TPP-2M formula in the energy range >50 eV but were also consistent with the trend of IMFPs calculated through experimental ELFs in the range of 2.7-50 eV, where the TPP-2M formula cannot be used. Our findings suggest that ML is very powerful and efficient and has great potential to complete a database of IMFPs for materials that can prove solutions closer to reality than empirical models on materials with similar physical and chemical properties and can be applied to other different situations for correlated information prediction.
引用
收藏
页数:23
相关论文
共 66 条
[1]   Dielectric description of wakes and stopping powers in solids [J].
Abril, I ;
Garcia-Molina, R ;
Denton, CD ;
Perez-Perez, FJ ;
Arista, NR .
PHYSICAL REVIEW A, 1998, 58 (01) :357-366
[2]  
Adachi S., 2012, The Handbook on Optical Constants of Metals: In Tables and Figures
[3]  
[Anonymous], 2010, ISO18115
[4]  
Bellaiche L, 2000, PHYS REV B, V61, P7877, DOI 10.1103/PhysRevB.61.7877
[5]  
Bethe H, 1930, ANN PHYS-BERLIN, V5, P325
[6]   Momentum-Dependent Lifetime Broadening of Electron Energy Loss Spectra: A Self-Consistent Coupled-Plasmon Model [J].
Bourke, J. D. ;
Chantler, C. T. .
JOURNAL OF PHYSICAL CHEMISTRY LETTERS, 2015, 6 (03) :314-319
[7]   Monte Carlo simulation of full energy spectrum of electrons emitted from silicon in Auger electron spectroscopy [J].
Cao, N. ;
Da, B. ;
Ming, Y. ;
Mao, S. F. ;
Goto, K. ;
Ding, Z. J. .
SURFACE AND INTERFACE ANALYSIS, 2015, 47 (01) :113-119
[8]   X-ray Spectroscopic Measurement of Photoelectron Inelastic Mean Free Paths in Molybdenum [J].
Chantler, Christopher T. ;
Bourke, Jay D. .
JOURNAL OF PHYSICAL CHEMISTRY LETTERS, 2010, 1 (15) :2422-2427
[9]  
CORTES C, 1995, MACH LEARN, V20, P273, DOI 10.1023/A:1022627411411
[10]  
Cullen D.E., 1997, EPDL97 The Evaluated Data Library, 97 Version