MELRSNet for accelerating the exploration of novel ultrawide bandgap semiconductors

Abstract

Ultrawide bandgap (UWBG) semiconductors, with bandgaps exceeding 3.4 eV of gallium nitride, offer the potential to overcome the limitations of conventional semiconductors and drive innovations in electronics and photovoltaics. However, discovering such materials remains a huge challenge due to the prohibitive cost of trial-and-error-based experiments and the complexity of cutting-edge quantum mechanical approaches. Here, we develop the Multistage Ensemble Learning Rapid Screening Network (MELRSNet), a data-driven hierarchical machine learning framework integrated with high-throughput first-principles calculations, designed for swift identification of UWBG semiconductors. Trained on the Materials Project dataset, MELRSNet utilizes elemental and structural features to classify, regress, and validate potential candidates. Its efficacy is underscored by the accurate prediction of bandgaps in UWBG oxides and the revelation of metric-bandgap relationships, aligning closely with first-principles calculations. Furthermore, MELRSNet's reliability is bolstered through the identification of eight novel ternary oxide compounds, derived from monoclinic hafnium oxide crystals, exhibiting high stability, desirable band gaps, and strong ultraviolet light absorption, marking them promising candidates for lab synthesis and subsequent applications. MELRSNet not only streamlines the discovery of UWBG semiconductors but also paves the way for high-throughput computational screening of other functional materials.

Keywords

Ultrawide bandgap semiconductor / machine learning / density functional theory / stacked generalization / LightGBM

References

[1] Zhou F,Xiao M.An avalanche-and-surge robust ultrawide-bandgap heterojunction for power electronics.Nat Commun2023;14:4459 PMCID:PMC10368629

[2] Kim H,Lien DH.Actively variable-spectrum optoelectronics with black phosphorus.Nature2021;596:232-7

[3] Luo L,Cheng K.MXene-GaN van der Waals metal-semiconductor junctions for high performance multiple quantum well photodetectors.Light Sci Appl2021;10:177 PMCID:PMC8410839

[4] Makita T,Kumamoto A.High-performance, semiconducting membrane composed of ultrathin, single-crystal organic semiconductors.Proc Natl Acad Sci USA2020;117:80-5 PMCID:PMC6955328

[5] Nguyen TK,Ashok A.Wide bandgap semiconductor nanomembranes as a long-term biointerface for flexible, implanted neuromodulator.Proc Natl Acad Sci USA2022;119:e2203287119 PMCID:PMC9388084

[6] Park SH,Chen D.Wide bandgap III-nitride nanomembranes for optoelectronic applications.Nano Lett2014;14:4293-8

[7] Shi J,Yang L,Qi DC.Wide bandgap oxide semiconductors: from materials physics to optoelectronic devices.Adv Mater2021;33:e2006230

[8] Wang Y,Fu L.Realization of robust and ambient-stable room-temperature ferromagnetism in wide bandgap semiconductor 2D carbon nitride sheets.ACS Appl Mater Interfaces2023;15:54797-807

[9] Tsao JY,Hollis MA.Ultrawide-bandgap semiconductors: research opportunities and challenges.Adv Elect Materials2018;4:1600501

[10] Zhang J,Dang K.Ultra-wide bandgap semiconductor Ga₂O₃ power diodes.Nat Commun2022;13:3900 PMCID:PMC9259626

[11] Xie C,Tong X.Recent progress in solar-blind deep-ultraviolet photodetectors based on inorganic ultrawide bandgap semiconductors.Adv Funct Materials2019;29:1806006

[12] Chen Y,Liu J.Growth of 2D GaN single crystals on liquid metals.J Am Chem Soc2018;140:16392-5

[13] Razeghi M.Short-wavelength solar-blind detectors-status, prospects, and markets.Proc IEEE2002;90:1006-14

[14] Wang J,Xu F.Group-III nitride heteroepitaxial films approaching bulk-class quality.Nat Mater2023;22:853-9

[15] Ohtomo A,Koida T.Mg x Zn_1−xO as a II–VI widegap semiconductor alloy.Appl Phys Lett1998;72:2466-8

[16] Han D,Yang J.Performance enhancement of a p-Si/n-ZnGa₂O₄ heterojunction solar-blind UV photodetector through interface engineering.J Mater Chem C2021;9:10013-9

[17] Yang Y,Wang X.Polarization-sensitive ultraviolet photodetection of anisotropic 2D GeS₂.Adv Funct Materials2019;29:1900411

[18] Zheng Y,Wang W,Li G.Large-size ultrathin α-Ga₂S₃ nanosheets toward high-performance photodetection.Adv Funct Materials2021;31:2008307

[19] Zhou N,Yang R.Nonlayered two-dimensional defective semiconductor γ-Ga₂S₃ toward broadband photodetection.ACS Nano2019;13:6297-307

[20] Yan Y,Du J.Cross-substitution promoted ultrawide bandgap up to 4.5 eV in a 2D semiconductor: gallium thiophosphate.Adv Mater2021;33:e2008761

[21] Maldovan M.Diamond-structured photonic crystals.Nat Mater2004;3:593-600

[22] Li L.Binary compound bilayer and multilayer with vertical polarizations: two-dimensional ferroelectrics, multiferroics, and nanogenerators.ACS Nano2017;11:6382-8

[23] Liu X,Liu H.Recent progress and applications of HfO₂-based ferroelectric memory.Tsinghua Sci Technol2023;28:221-9

[24] Schroeder U,Mikolajick T.The fundamentals and applications of ferroelectric HfO₂.Nat Rev Mater2022;7:653-69

[25] Garrity EM,Gorai P,Zakutayev A.Computational identification of ternary wide-band-gap oxides for high-power electronics.PRX Energy2022;1

[26] Perdew JP,Ernzerhof M.Generalized gradient approximation made simple.Phys Rev Lett1996;77:3865-8

[27] Gorai P,Grover S,Metzger WK.A search for new back contacts for CdTe solar cells.Sci Adv2023;9:eade3761 PMCID:PMC11804015

[28] Perdew JP.Physical content of the exact kohn-sham orbital energies: band gaps and derivative discontinuities.Phys Rev Lett1983;51:1884-7

[29] Heyd J,Scuseria GE.Energy band gaps and lattice parameters evaluated with the Heyd-Scuseria-Ernzerhof screened hybrid functional.J Chem Phys2005;123:174101

[30] Masood H,Toe CY.Enhancing prediction accuracy of physical band gaps in semiconductor materials.Cell Rep Phys Sci2023;4:101555

[31] Rosen AS,Ray D.Machine learning the quantum-chemical properties of metal–organic frameworks for accelerated materials discovery.Matter2021;4:1578-97

[32] Wang X,Xie X.Bayesian-optimization-assisted discovery of stereoselective aluminum complexes for ring-opening polymerization of racemic lactide.Nat Commun2023;14:3647 PMCID:PMC10282063

[33] Jain A,Hautier G.Commentary: the materials project: a materials genome approach to accelerating materials innovation.APL Materials2013;1:011002

[34] Saal JE,Aykol M,Wolverton C.Materials design and discovery with high-throughput density functional theory: The Open Quantum Materials Database (OQMD).JOM2013;65:1501-9

[35] Curtarolo S,Wang S.AFLOWLIB.ORG: A distributed materials properties repository from high-throughput ab initio calculations.Comput Mater Sci2012;58:227-35

[36] Sutton C,Ghiringhelli LM,Vreeken J.Identifying domains of applicability of machine learning models for materials science.Nat Commun2020;11:4428 PMCID:PMC7474068

[37] He L,Torrent D,Rabczuk T.Machine learning assisted intelligent design of meta structures: a review.Microstructures2023;3:2023037

[38] Li C,Xu B.A progressive learning method for predicting the band gap of ABO ₃ perovskites using an instrumental variable.J Mater Chem C2020;8:3127-36

[39] Lu S,Ma L,Wang J.Rapid discovery of ferroelectric photovoltaic perovskites and material descriptors via machine learning.Small Methods2019;3:1900360

[40] Lu S,Ouyang Y,Li Q.Accelerated discovery of stable lead-free hybrid organic-inorganic perovskites via machine learning.Nat Commun2018;9:3405 PMCID:PMC6109147

[41] Park H,Ali A,Bensmail H.Importance of structural deformation features in the prediction of hybrid perovskite bandgaps.Comput Mater Sci2020;184:109858

[42] Dan Y,Li X,Hu M.Generative adversarial networks (GAN) based efficient sampling of chemical composition space for inverse design of inorganic materials.npj Comput Mater2020;6:352

[43] Lee J,Shitara K,Tanaka I.Prediction model of band gap for inorganic compounds by combination of density functional theory calculations and machine learning techniques.Phys Rev B2016;93

[44] Shi Z,Dao M,Shapeev A.Deep elastic strain engineering of bandgap through machine learning.Proc Natl Acad Sci USA2019;116:4117-22 PMCID:PMC6410806

[45] Zhuo Y,Brgoch J.Predicting the band gaps of inorganic solids by machine learning.J Phys Chem Lett2018;9:1668-73

[46] Dau MT,Michon A,Vézian S.Descriptor engineering in machine learning regression of electronic structure properties for 2D materials.Sci Rep2023;13:5426 PMCID:PMC10070413

[47] Fung V,Hu G,Sumpter BG.Inverse design of two-dimensional materials with invertible neural networks.npj Comput Mater2021;7:670

[48] Lu S,Guo Y,Wu Y.Coupling a crystal graph multilayer descriptor to active learning for rapid discovery of 2D ferromagnetic semiconductors/half-metals/metals.Adv Mater2020;32:e2002658

[49] Zhuo Y,Oliynyk AO,Brgoch J.Identifying an efficient, thermally robust inorganic phosphor host via machine learning.Nat Commun2018;9:4377 PMCID:PMC6197245

[50] Wang Z,Xie X.Applications of machine learning in perovskite materials.Adv Compos Hybrid Mater2022;5:2700-20

[51] Shen H,Chen Z.First-principles study combined with interpretable machine-learning models of bayesian optimization for the design of ultrawide bandgap double perovskites.J Phys Chem C2023;127:21410-22

[52] Batista GEAPA,Monard MC.A study of the behavior of several methods for balancing machine learning training data.SIGKDD Explor Newsl2004;6:20-9

[53] Chawla NV.; Bowyer KW.; Hall LO.;Kegelmeyer WP. SMOTE: synthetic minority over-sampling technique.J Artif Intell Res2002;16:321-57

[54] Pedregosa F.; Varoquaux G.; Gramfort A, et al. Scikit-learn: machine learning in python.J Mach Learn Res2011;12:2825-30

[55] Wolpert DH.Stacked generalization.Neural Networks1992;5:241-59

[56] Guyon I,Barnhill S.Gene selection for cancer classification using support vector machines.Mach Learn46:389-422

[57] Ke G.; Meng Q.; Finley T, et al. LightGBM: a highly efficient gradient boosting decision tree. In: Proceedings of the 31st International Conference on Neural Information Processing Systems. Long Beach, California, USA: Curran Associates Inc.; 2017, p. 3149-57.

[58] Clark SJ,Pickard CJ.First principles methods using CASTEP.Z Kristallogr Cryst Mater2005;220:567-70

[59] Payne MC,Allan DC,Joannopoulos JD.Iterative minimization techniques for ab initio total-energy calculations: molecular dynamics and conjugate gradients.Rev Mod Phys1992;64:1045-97

[60] Kresse G.Efficient iterative schemes for ab initio total-energy calculations using a plane-wave basis set.Phys Rev B Condens Matter1996;54:11169-86

[61] Kresse G.From ultrasoft pseudopotentials to the projector augmented-wave method.Phys Rev B1999;59:1758-75

[62] Pfrommer BG,Louie SG.Relaxation of crystals with the quasi-newton method.J Comput Phys1997;131:233-40

[63] Monkhorst HJ.Special points for Brillouin-zone integrations.Phys Rev B1976;13:5188-92

[64] Heyd J,Ernzerhof M.Hybrid functionals based on a screened Coulomb potential.J Chem Phys2003;118:8207-15

[65] Krukau AV,Izmaylov AF.Influence of the exchange screening parameter on the performance of screened hybrid functionals.J Chem Phys2006;125:224106

[66] Refson K,Clark SJ.Variational density-functional perturbation theory for dielectrics and lattice dynamics.Phys Rev B2006;73

[67] Baroni S,Dal Corso A.Phonons and related crystal properties from density-functional perturbation theory.Rev Mod Phys2001;73:515-62

[68] Momma K.VESTA 3 for three-dimensional visualization of crystal, volumetric and morphology data.J Appl Crystallogr2011;44:1272-6

[69] Leeuwen R. Mapping from densities to potentials in time-dependent density-functional theory.Phys Rev Lett1999;82:3863-6

[70] Kühne TD,Del Ben M.CP2K: An electronic structure and molecular dynamics software package - Quickstep: Efficient and accurate electronic structure calculations.J Chem Phys2020;152:194103

[71] Lu T.Multiwfn: A multifunctional wavefunction analyzer.J Comput Chem2012;33:580-92

[72] Jain A,Moore CJ.A high-throughput infrastructure for density functional theory calculations.Comput Mater Sci2011;50:2295-310

[73] Borlido P,Huran AW,Marques MAL.Large-scale benchmark of exchange-correlation functionals for the determination of electronic band gaps of solids.J Chem Theory Comput2019;15:5069-79 PMCID:PMC6739738

[74] Yang J,Chen X.Recent advances in optoelectronic and microelectronic devices based on ultrawide-bandgap semiconductors.Prog Quantum2022;83:100397

[75] Yin Y,Sun Z,Jin G.Machine learning regression model for predicting the band gap of multi-elements nonlinear optical crystals.Comput Mater Sci2024;242:113109

[76] Chen X,Chen Q,Wang J.From bulk effective mass to 2D carrier mobility accurate prediction via adversarial transfer learning.Nat Commun2024;15:5391 PMCID:PMC11199574

[77] Biswas M.Thermodynamically metastable α-, ε- (or κ-), and γ-Ga₂O₃: From material growth to device applications.APL Materials2022;10:060701

[78] Gladkikh V,Hajibabaei A,Myung CW.Machine learning for predicting the band gaps of ABX₃ perovskites from elemental properties.J Phys Chem C2020;124:8905-18

[79] Fowler WB.Influence of electronic polarization on the optical properties of insulators.Phys Rev1966;151:657-67

[80] Curtarolo S,Hart GL.AFLOW: An automatic framework for high-throughput materials discovery.Comput Mater Sci2012;58:218-26

[81] Zagorac D,Ruehl S,Rehme S.Recent developments in the Inorganic Crystal Structure Database: theoretical crystal structure data and related features.J Appl Crystallogr2019;52:918-25 PMCID:PMC6782081

[82] Becke AD.A simple measure of electron localization in atomic and molecular systems.J Chem Phys1990;92:5397-403

[83] Su J,Lin Z.Unusual electronic and optical properties of two-dimensional Ga₂O₃ predicted by density functional theory.J Phys Chem C2018;122:24592-9

[84] Anam B.Structural, thermal, and electronic properties of two-dimensional gallium oxide (β-Ga₂O₃) from first-principles design.Chemphyschem2021;22:2362-70

[85] Kubota Y,Tsuda O.Deep ultraviolet light-emitting hexagonal boron nitride synthesized at atmospheric pressure.Science2007;317:932-4

[86] Wang Y,Tian Y.Deep ultraviolet photodetectors based on carbon-doped two-dimensional hexagonal boron nitride.ACS Appl Mater Interfaces2020;12:27361-7