Machine learning for optimized nonlinear optical rectification in core/shell nanodiscs under hydrostatic pressure

Ed-Dahmouny A., Zeiri N., Fakkahi A., BAŞER P., Sali A., Duque C.

Ceramics International, 2026 (SCI-Expanded, Scopus) identifier

  • Yayın Türü: Makale / Tam Makale
  • Basım Tarihi: 2026
  • Doi Numarası: 10.1016/j.ceramint.2026.01.045
  • Dergi Adı: Ceramics International
  • Derginin Tarandığı İndeksler: Science Citation Index Expanded (SCI-EXPANDED), Scopus, Chemical Abstracts Core, Compendex, INSPEC
  • Anahtar Kelimeler: Finite element method, Machine learning, Nonlinear optical rectification, Quantum disc
  • Sivas Cumhuriyet Üniversitesi Adresli: Evet

Özet

In this work, we investigate the nonlinear optical rectification (NOR) behavior of GaAs/AlGaAs core/shell nanodiscs (CSNDs) subjected to varying hydrostatic pressures and containing a central hydrogenic impurity. Using the finite element method (FEM), we numerically solve the Schrödinger equation under cylindrical symmetry to extract energy levels and wave functions. These results are employed to train several advanced machine learning (ML) algorithms, including artificial neural networks (ANN), convolutional neural networks (CNN), feed forward neural networks (FNN), gradient boosting regression (GBR), and support vector machines (SVM). The predictive performance of each model is evaluated by comparing predicted NOR coefficients with those obtained from FEM simulations across five different pressure values. The SVM and GBR models demonstrate near-perfect agreement with numerical data ((Formula presented) ), while the ANN and FNN models show strong generalization and learning capabilities. These findings highlight the potential of ML techniques to accurately and efficiently model complex optical phenomena in semiconductor nanostructures, offering promising pathways for the design of next-generation optoelectronic devices.