journal article·2025-03-13

Machine Learning-Assisted Dielectric Screening of Bismuth/Antimony-Based Compounds for Promising Optoelectronic Semiconductors

Guoliang Luo, Xiaoyu Yang, Yansong Zhou, Kun Zhou, Junjie Feng, Jiahao Xie, Xin He, Lijun Zhang

The Journal of Physical Chemistry C· Volume 129· Issue 10
Publisher page

Abstract

High-dielectric-constant materials facilitate carrier transport by reducing the scattering, trapping, and recombination of carriers through the shielding of charged defects and impurities, enabling their advantageous application in optoelectronics. Materials containing bismuth and antimony cations exhibit excellent dielectric properties due to their soft lattice and the presence of n s 2 lone pair electrons. In this study, we have combined first-principles calculations and machine learning methods to unravel the relationship between dielectric constant and physical features in 523 bismuth/antimony-based materials from the Inorganic Crystal Structure Database. We trained two Gradient Boosting Decision Tree regression models to predict the ionic and electronic dielectric constants and analyzed the critical features influencing these constants. Furthermore, utilizing Sure Independence Screening and Sparsifying Operator combined with these critical features, we fitted three dielectric descriptors, comprising four physical features─Polar center (PC), Packing fraction of Bi-/Sb-based polyhedron (PF Bi,Sb ), Energy of Highest Occupied Molecular Orbital (E HOMO ), and Fraction of p valence electron (VE p frac ). The descriptors indicate that the larger the PC and the smaller the PF Bi,Sb, the higher the ionic dielectric constants. Conversely, the larger the E HOMO and the smaller the VE p frac, the more conducive it is to high electronic dielectric constants. Finally, we trained a random forest model to identify materials with high dielectric constants (ε > 20). Using this model, we screened out 4933 experimentally unsynthesized, potentially high-dielectric-constant materials containing bismuth or antimony from Open Quantum Materials Database. Thirty-eight thermodynamically stable sulfur-containing compounds were selected for density functional perturbation theory calculations. The results show that 86.8% of these materials have a dielectric constant greater than 20. Finally, 12 materials were identified with potential optoelectronic applications. This study provides valuable insights and practical tools for the design and identification of materials with desirable dielectric properties for promising optoelectronic applications.
Source: openalex_inverted_index · Quality: baseline

Research context

OpenAlex topics: Semiconductor materials and interfaces · Machine Learning in Materials Science · Semiconductor materials and devices

SciVX discovery tags are terminology-based. OpenAlex topics are source-provided classifications.

Affiliations

State Key Laboratory of Integrated Optoelectronics, Key Laboratory of Automobile Materials of MOE and Key Laboratory of Material Simulation Methods & Software of MOE, School of Materials Science and Engineering

Jilin University

State Key Laboratory of Superhard Materials, International Center of Computational Method and Software, School of Physics

Funding

Department of Science and Technology of Jilin Province · National Natural Science Foundation of China

References (72)

Show source-provided references
1. 10.1557/mrc.2015.2610.1557/mrc.2015.26 ↗
2. 10.1021/ja412583t10.1021/ja412583t ↗
3. 10.1039/C4TA01198H10.1039/C4TA01198H ↗
4. 10.1126/science.124398210.1126/science.1243982 ↗
5. 10.1021/ic401215x10.1021/ic401215x ↗
6. 10.1103/PhysRevB.81.14411410.1103/PhysRevB.81.144114 ↗
7. 10.1126/science.aaa576010.1126/science.aaa5760 ↗
8. 10.1021/jacs.8b1068510.1021/jacs.8b10685 ↗
9. 10.1038/s41524-018-0071-110.1038/s41524-018-0071-1 ↗
10. 10.1021/acs.chemmater.7b0062810.1021/acs.chemmater.7b00628 ↗
11. 10.1063/1.486477810.1063/1.4864778 ↗
12. 10.1039/C4TA05033A10.1039/C4TA05033A ↗
13. 10.1021/jp512077m10.1021/jp512077m ↗
14. 10.1002/adma.20130628110.1002/adma.201306281 ↗
15. 10.1016/j.mssp.2014.08.02810.1016/j.mssp.2014.08.028 ↗
16. 10.1016/j.matlet.2008.04.00310.1016/j.matlet.2008.04.003 ↗
17. 10.1021/cm050293110.1021/cm0502931 ↗
18. 10.1039/C6TA04104C10.1039/C6TA04104C ↗
19. 10.1103/PhysRevB.87.04520510.1103/PhysRevB.87.045205 ↗
20. 10.1063/1.491753210.1063/1.4917532 ↗
21. 10.1038/nphys335710.1038/nphys3357 ↗
22. 10.1103/PhysRevB.88.16520310.1103/PhysRevB.88.165203 ↗
23. 10.1103/PhysRevB.86.14410810.1103/PhysRevB.86.144108 ↗
24. 10.1103/PhysRevB.82.04520310.1103/PhysRevB.82.045203 ↗
25. 10.1016/S0038-1098(03)00566-010.1016/S0038-1098(03)00566-0 ↗
26. 10.1002/adma.20200857410.1002/adma.202008574 ↗
27. 10.1103/PhysRevA.52.108610.1103/PhysRevA.52.1086 ↗
28. 10.1103/PhysRevB.55.1035510.1103/PhysRevB.55.10355 ↗
29. 10.1038/s43588-024-00632-510.1038/s43588-024-00632-5 ↗
30. 10.1021/jacs.3c0620710.1021/jacs.3c06207 ↗
31. 10.1021/jacs.4c0350710.1021/jacs.4c03507 ↗
32. 10.1021/jacs.3c1357410.1021/jacs.3c13574 ↗
33. 10.1038/s41563-023-01705-y10.1038/s41563-023-01705-y ↗
34. 10.1038/s43246-023-00373-410.1038/s43246-023-00373-4 ↗
35. 10.1021/acsami.3c0365710.1021/acsami.3c03657 ↗
36. 10.1039/D3TA04874H10.1039/D3TA04874H ↗
37. 10.1126/science.abn344510.1126/science.abn3445 ↗
38. 10.1039/D1CP04702G10.1039/D1CP04702G ↗
39. 10.1103/PhysRevMaterials.4.10380110.1103/PhysRevMaterials.4.103801 ↗
40. 10.1038/s41524-020-0337-210.1038/s41524-020-0337-2 ↗
41. 10.1103/PhysRevB.54.1116910.1103/PhysRevB.54.11169 ↗
42. 10.1103/PhysRev.140.A113310.1103/PhysRev.140.A1133 ↗
43. 10.1103/PhysRevB.50.1795310.1103/PhysRevB.50.17953 ↗
44. 10.1103/PhysRevLett.77.386510.1103/PhysRevLett.77.3865 ↗
45. 10.1063/1.156406010.1063/1.1564060 ↗
46. 10.1016/j.scib.2021.06.01110.1016/j.scib.2021.06.011 ↗
47. 10.1103/PhysRev.130.12910.1103/PhysRev.130.129 ↗
48. 10.1016/j.commatsci.2018.05.01810.1016/j.commatsci.2018.05.018 ↗
49. 10.1111/j.2517-6161.1996.tb02080.x10.1111/j.2517-6161.1996.tb02080.x ↗
50. Garnett, R. Bayesian Optimization; Cambridge University Press, 2023; pp 123–156.10.1017/9781108348973 ↗
51. 10.1038/sdata.2016.13410.1038/sdata.2016.134 ↗
52. 10.1038/s41598-018-33095-610.1038/s41598-018-33095-6 ↗
53. 10.1103/PhysRevMaterials.3.04460210.1103/PhysRevMaterials.3.044602 ↗
54. 10.1103/PhysRevApplied.7.02401510.1103/PhysRevApplied.7.024015 ↗
55. Drucker H. · 1996 · 155
56. 10.2307/253094610.2307/2530946 ↗
57. 10.1214/aos/101320345110.1214/aos/1013203451 ↗
58. Ho, T. K. In Random Decision Forests, Proceedings of 3rd International Conference on Document Analysis and Recognition, 1995; pp 278–282.
59. 10.1007/3-540-59119-2_16610.1007/3-540-59119-2_166 ↗
60. Chen, T.; Guestrin, C.In XGBoost: A Scalable Tree Boosting System, Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, 2016; pp 785–794.10.1145/2939672.2939785 ↗
61. 10.1038/s41598-023-49603-210.1038/s41598-023-49603-2 ↗
62. 10.1080/14686996.2020.172482410.1080/14686996.2020.1724824 ↗
63. 10.7567/1347-4065/ab34d610.7567/1347-4065/ab34d6 ↗
64. 10.1063/5.022389010.1063/5.0223890 ↗
65. 10.1103/PhysRevMaterials.2.08380210.1103/PhysRevMaterials.2.083802 ↗
66. Zou, X.; Hu, Y.; Tian, Z.; Shen, K.In Logistic Regression Model Optimization and Case Analysis, 2019 IEEE 7th International Conference on Computer Science and Network Technology, 2019; pp 135–139.10.1109/ICCSNT47585.2019.8962457 ↗
67. 10.1007/BF0099401810.1007/BF00994018 ↗
68. 10.1109/TIT.1967.105396410.1109/TIT.1967.1053964 ↗
69. 10.1103/PhysRevLett.108.06870110.1103/PhysRevLett.108.068701 ↗
70. 10.1021/acs.chemmater.7b0266410.1021/acs.chemmater.7b02664 ↗
71. 10.1021/acs.jpcc.5b0555110.1021/acs.jpcc.5b05551 ↗
72. 10.1007/s11082-023-05433-010.1007/s11082-023-05433-0 ↗