The Investigation of Lattice Properties for Group-IV Sigesn Ternary Alloy: By Using Machine Learning Method

Abstract:

Group-IV SiGeSn ternary alloy is a hot spot in the field of fabricating high-efficient Si-based light source due to its large lattice constant and bandgap variation range. However, due to the high cost and low speed of experimental and computational research, it is difficult to obtain their lattice constants comprehensively and quickly. Machine learning prediction based on statistics is an advanced method to solve this problem. In this paper, based on the existing data of group IV alloys, three machine learning methods such as Random Forest (RF), Support Vector Regression (SVR) and Gradient Boosting Decision Tree (GBDT) have been built to predict the lattice constants of SiGeSn. Firstly, the lattice constants of Group-IV alloys are collected to construct data set; Then, the data set are used to train the machine learning models which describe the quantitative relationship between concentrations and lattice constants; Finally, the prediction performance of these models are compared with each other, and the concentrations with appropriate lattice constants are predicted. The results show the comprehensive performance of SVR model is better than the other two, which means the SVR model can be used to directly predict the lattice constants of SiGeSn.

• Periodical: Nano Hybrids and Composites (Volume 34)
• Pages: 83-88
• DOI: https://doi.org/10.4028/p-o9669j
• Online since: February 2022
• Authors: Li Xin Zhang, Sheng Liu Sun, Wen Qi Huang*, Zhen Yu Chen, Hao Wang, Chun Qian Zhang
• Keywords: Lattice Constants, Machine Learning, SiGeSn Ternary Alloy

References

[1] Du, W., et al., Study of Si-Based GeSn Optically Pumped Lasers With Micro-Disk and Ridge Waveguide Structures. Frontiers in Physics, 2019. 7(147): p.7.

DOI: 10.3389/fphy.2019.00147

[2] Fang, C., et al., Germanium-tin alloys: applications for optoelectronics in mid-infrared spectra. Opto-Electronic Advances, 2018. 1(3): p.180004.

[3] WirthsS, et al., Lasing in direct-bandgap GeSn alloy grown on Si. Nat Photon, 2015. 9(2): pp.88-92.

[4] Homewood, K.P. and M.A. Lourenco, Optoelectronics: The rise of the GeSn laser. Nat Photon, 2015. 9(2): pp.78-79.

[5] Basu, R., et al., Predicted performance of Ge/GeSn hetero-phototransistors on Si substrate at 1.55\,\upmu \mathrm{m}. Optical and Quantum Electronics, 2015. 47(2): pp.387-399.

DOI: 10.1007/s11082-014-9921-3

[6] Coppinger, M., et al., Photoconductivity of germanium tin alloys grown by molecular beam epitaxy. Applied Physics Letters, 2013. 102(14): p.141101.

DOI: 10.1063/1.4800448

[7] Xiao, G., et al., Germanium–Tin (GeSn) p-Channel MOSFETs Fabricated on (100) and (111) Surface Orientations With Sub-400 oC Si2H6 Passivation. Electron Device Letters, IEEE, 2013. 34(3): pp.339-341.

DOI: 10.1109/led.2012.2236880

[8] Soref, R., Direct-bandgap compositions of the CSiGeSn group-IV alloy. Optical Materials Express, 2014. 4(4): pp.836-842.

DOI: 10.1364/ome.4.000836

[9] Wang, D., et al., Semiconducting SiGeSn High-Entropy Alloy: A Density Functional Theory Study. Journal of Applied Physics, (2019).

[10] Moontragoon, P., R.A. Soref, and Z. Ikonic, The direct and indirect bandgaps of unstrained SixGe1−x−ySny and their photonic device applications. Journal of Applied Physics, 2012. 112(7): p.073106.

DOI: 10.1063/1.4757414

[11] Yu, M.T., et al., Machine learning the Hubbard U parameter in DFT plus U using Bayesian optimization. Npj Computational Materials, 2020. 6(1): p.6.

DOI: 10.1038/s41524-020-00446-9

[12] Niu, Hu,et al.Application of Machine Learning in Material Informatics: A survey [J]. Materials Reports,2020,34(23):23100-23108.

[13] Tibshirani R. Regression shrinkage and selection via the lasso: A retrospective. J R Stat Soc Ser B Stat Methodol, 2011, 73: 273–282.

DOI: 10.1111/j.1467-9868.2011.00771.x

[14] Chun Song Jiang and Gui Qin Liang. Modeling shear strength of medium- to ultra-high-strength concrete beams with stirrups using SVR and genetic algorithm[J]. Soft Computing, 2021: 1-15.

DOI: 10.21203/rs.3.rs-338840/v1

[15] Zhang Jialin et al. FS-GBDT: identification multicancer-risk module via a feature selection algorithm by integrating Fisher score and GBDT. [J]. Briefings in bioinformatics, 2021, 22(3).

DOI: 10.1093/bib/bbaa189

[16] BREIMAN L. Random forests[J]. Machine Learning, 2001,45(1): 5-32.

[17] Zheng, Zhang, et al. different machine learning algorithms different machine learning algorithms. [J]. The Chinese Journal of Nonferrous Metals,2019,29(04):803-809.

[18] Liu Yaru and Wang Lei and Gu Kaixuan. A support vector regression (SVR)-based method for dynamic load identification using heterogeneous responses under interval uncertainties[J]. Applied Soft Computing Journal, 2021, 110.

DOI: 10.1016/j.asoc.2021.107599

[19] Zhang Yuefei et al. Development of an SVR model for microwave-assisted aqueous two-phase extraction of isoflavonoids from[J]. Chemical Engineering Communications, 2021, 208(7): 1005-1016.

DOI: 10.1080/00986445.2020.1734578

[20] Liu Yaru and Wang Lei and Gu Kaixuan. A support vector regression (SVR)-based method for dynamic load identification using heterogeneous responses under interval uncertainties[J]. Applied Soft Computing Journal, 2021, 110.

DOI: 10.1016/j.asoc.2021.107599

[21] Zhang Wanhu et al. Predictive model of cooling load for ice storage air-conditioning system by using GBDT[J]. Energy Reports, 2021, 7: 1588-1597.

DOI: 10.1016/j.egyr.2021.03.017